0% found this document useful (0 votes)

410 views

Physics 130 Notes

This document outlines the course contents for a quantum physics course. It covers 27 topics related to quantum mechanics including problems with classical physics, probability amplitudes, wave packets, operators, the Schrodinger equation, angular momentum, perturbation theory, and more. Each topic includes subsections on derivations, computations, examples, and homework problems.

Uploaded by

Porter Howland

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

410 views

Physics 130 Notes

Uploaded by

Porter Howland

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 714

Quantum Physics

(UCSD Physics 130)

April 2, 2015

Contents

TOC

Contents
1 Course Summary
1.1 Problems with Classical Physics . . . . . . . . . . . . . . . . . . . . . . .
1.2 Thought Experiments on Diffraction . . . . . . . . . . . . . . . . . . . .
1.3 Probability Amplitudes . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Wave Packets and Uncertainty . . . . . . . . . . . . . . . . . . . . . . .
1.5 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6 Expectation Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.7 Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.8 The Schr
odinger Equation . . . . . . . . . . . . . . . . . . . . . . . . .
1.9 Eigenfunctions, Eigenvalues and Vector Spaces . . . . . . . . . . . . . .
1.10 A Particle in a Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.11 Piecewise Constant Potentials in One Dimension . . . . . . . . . . . . .
1.12 The Harmonic Oscillator in One Dimension . . . . . . . . . . . . . . . .
1.13 Delta Function Potentials in One Dimension . . . . . . . . . . . . . . .
1.14 Harmonic Oscillator Solution with Operators . . . . . . . . . . . . . . .
1.15 More Fun with Operators . . . . . . . . . . . . . . . . . . . . . . . . . .
1.16 Two Particles in 3 Dimensions . . . . . . . . . . . . . . . . . . . . . . .
1.17 Identical Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.18 Some 3D Problems Separable in Cartesian Coordinates . . . . . . . . .
1.19 Angular Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.20 Solutions to the Radial Equation for Constant Potentials . . . . . . . .
1.21 Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.22 Solution of the 3D HO Problem in Spherical Coordinates . . . . . . . .
1.23 Matrix Representation of Operators and States . . . . . . . . . . . . . .
1.24 A Study of ` = 1 Operators and Eigenfunctions . . . . . . . . . . . . . .
1.25 Spin 1/2 and other 2 State Systems . . . . . . . . . . . . . . . . . . . .
1.26 Quantum Mechanics in an Electromagnetic Field . . . . . . . . . . . . .
1.27 Local Phase Symmetry in Quantum Mechanics and the Gauge Symmetry
1.28 Addition of Angular Momentum . . . . . . . . . . . . . . . . . . . . . .
1.29 Time Independent Perturbation Theory . . . . . . . . . . . . . . . . . .
1.30 The Fine Structure of Hydrogen . . . . . . . . . . . . . . . . . . . . . .
1.31 Hyperfine Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.32 The Helium Atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.33 Atomic Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.34 Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19
19
19
20
21
22
22
23
23
23
25
26
27
28
29
30
31
33
33
34
35
35
37
37
38
38
39
40
41
44
45
46
47
48
50
2

Contents

1.35
1.36
1.37
1.38
1.39
1.40
1.41
1.42
1.43
2 The
2.1
2.2
2.3
2.4
2.5

2.6

2.7
2.8

Time Dependent Perturbation

Radiation in Atoms . . . . . .
Classical Field Theory . . . .
The Classical Electromagnetic
Quantization of the EM Field
Scattering of Photons . . . .
Electron Self Energy . . . . .
The Dirac Equation . . . . .
The Dirac Equation . . . . .

TOC

Theory
. . . . .
. . . . .
Field .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .

.
.
.
.
.
.
.
.
.

Problems with Classical Physics

Black Body Radiation * . . . . . . . . . . . . . . . . . . . . . . .
The Photoelectric Effect . . . . . . . . . . . . . . . . . . . . . . .
The Rutherford Atom * . . . . . . . . . . . . . . . . . . . . . . .
Atomic Spectra * . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 The Bohr Atom * . . . . . . . . . . . . . . . . . . . . . .
Derivations and Computations . . . . . . . . . . . . . . . . . . .
2.5.1 Black Body Radiation Formulas * . . . . . . . . . . . . .
2.5.2 The Fine Structure Constant and the Coulomb Potential
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.1 The Solar Temperature * . . . . . . . . . . . . . . . . . .
2.6.2 Black Body Radiation from the Early Universe * . . . . .
2.6.3 Compton Scattering * . . . . . . . . . . . . . . . . . . . .
2.6.4 Rutherfords Nuclear Size * . . . . . . . . . . . . . . . . .
Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . .
Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . .

3 Diffraction
3.1 Diffraction from Two Slits . . . . . . . . . . . . . . .
3.2 Diffraction from Crystals . . . . . . . . . . . . . . .
3.3 The de Broglie Wavelength . . . . . . . . . . . . . .
3.3.1 Computing de Broglie Wavelengths . . . . . .
3.4 Single Slit Diffraction . . . . . . . . . . . . . . . . .
3.5 Wave Particle Duality (Thought Experiments) . . .
3.6 Doing the Critical (Diffraction) Thought Experiment
3.7 Examples . . . . . . . . . . . . . . . . . . . . . . . .
3.7.1 Intensity Distribution for Two Slit Diffraction
3.8 Homework . . . . . . . . . . . . . . . . . . . . . . . .
3.9 Sample Test Problems . . . . . . . . . . . . . . . . .

. .
. .
. .
. .
. .
. .
. .
. .
*
. .
. .

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.

50
51
54
55
57
59
61
62
71
75
77
82
84
87
90
92
92
92
93
93
94
95
97
98
100

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

101
101
104
105
106
107
108
112
113
113
113
114
3

Contents

TOC

4 The Solution: Probability Amplitudes

4.1

4.2

116

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 117

4.1.1

Review of Complex Numbers . . . . . . . . . . . . . . . . . . . . 117

4.1.2

Review of Traveling Waves . . . . . . . . . . . . . . . . . . . . . 118

Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

5 Wave Packets

120

5.1

Building a Localized Single-Particle Wave Packet . . . . . . . . . . . . . 120

5.2

Two Examples of Localized Wave Packets . . . . . . . . . . . . . . . . . 121

5.3

The Heisenberg Uncertainty Principle . . . . . . . . . . . . . . . . . . . 123

5.4

Position Space and Momentum Space . . . . . . . . . . . . . . . . . . . 124

5.5

Time Development of a Gaussian Wave Packet * . . . . . . . . . . . . . 126

5.6

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 128

5.7

5.6.1

Fourier Series * . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

5.6.2

Fourier Transform * . . . . . . . . . . . . . . . . . . . . . . . . . 129

5.6.3

Integral of Gaussian . . . . . . . . . . . . . . . . . . . . . . . . . 130

5.6.4

Fourier Transform of Gaussian * . . . . . . . . . . . . . . . . . . 131

5.6.5

Time Dependence of a Gaussian Wave Packet * . . . . . . . . . . 133

5.6.6

Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

5.6.7

The Dirac Delta Function . . . . . . . . . . . . . . . . . . . . . . 136

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.7.1

Can I See inside an Atom . . . . . . . . . . . . . . . . . . . . . 137

5.7.2

Can I See inside a Nucleus . . . . . . . . . . . . . . . . . . . . 138

5.7.3

Estimate the Hydrogen Ground State Energy . . . . . . . . . . . 138

5.8

Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

5.9

Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

6 Operators
6.1

145

Operators in Position Space . . . . . . . . . . . . . . . . . . . . . . . . . 145

6.1.1

The Momentum Operator . . . . . . . . . . . . . . . . . . . . . . 145

6.1.2

The Energy Operator . . . . . . . . . . . . . . . . . . . . . . . . 146

6.1.3

The Position Operator . . . . . . . . . . . . . . . . . . . . . . . . 146

6.1.4

The Hamiltonian Operator . . . . . . . . . . . . . . . . . . . . . 146

6.2

Operators in Momentum Space . . . . . . . . . . . . . . . . . . . . . . . 147

6.3

Expectation Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

6.4

Dirac Bra-ket Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.5

Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4

Contents

6.6

6.7

TOC

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 151

6.6.1

Verify Momentum Operator . . . . . . . . . . . . . . . . . . . . . 151

6.6.2

Verify Energy Operator . . . . . . . . . . . . . . . . . . . . . . . 151

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.7.1

Expectation Value of Momentum in a Given State . . . . . . . . 151

6.7.2
6.7.3

Commutator of E and t . . . . . . . . . . . . . . . . . . . . . . . 152

Commutator of E and x . . . . . . . . . . . . . . . . . . . . . . . 152

6.7.4

Commutator of p and xn

6.7.5

Commutator of Lx and Ly

. . . . . . . . . . . . . . . . . . . . . . 153
. . . . . . . . . . . . . . . . . . . . . 153

6.8

Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6.9

Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

7 The Schr
odinger Equation

157

7.1

Deriving the Equation from Operators . . . . . . . . . . . . . . . . . . . 157

7.2

Schr
odinger Gives Time Development of Wavefunction . . . . . . . . . . 159

7.3

The Flux of Probability * . . . . . . . . . . . . . . . . . . . . . . . . . . 159

7.4

The Schr
odinger Wave Equation . . . . . . . . . . . . . . . . . . . . . . 160

7.5

The Time Independent Schr

odinger Equation . . . . . . . . . . . . . . . 160

7.6

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 162

7.7

7.6.1

Linear Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

7.6.2

Probability Conservation Equation * . . . . . . . . . . . . . . . . 162

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.7.1

7.8

Solution to the Schr

odinger Equation in a Constant Potential . . 163

Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

8 Eigenfunctions, Eigenvalues and Vector Spaces

165

8.1

Eigenvalue Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

8.2

Measurement in Quantum Mechanics . . . . . . . . . . . . . . . . . . . . 167

8.3

Hermitian Conjugate of an Operator . . . . . . . . . . . . . . . . . . . . 170

8.4

Hermitian Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

8.5

Eigenfunctions and Vector Space . . . . . . . . . . . . . . . . . . . . . . 172

8.6

The Particle in a 1D Box . . . . . . . . . . . . . . . . . . . . . . . . . . 174

8.7

Momentum Eigenfunctions

8.8

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 180

8.6.1

The Same Problem with Parity Symmetry . . . . . . . . . . . . . 176

. . . . . . . . . . . . . . . . . . . . . . . . . 178

8.8.1

Eigenfunctions of Hermitian Operators are Orthogonal . . . . . . 180

8.8.2

Continuity of Wavefunctions and Derivatives . . . . . . . . . . . 181

Contents

8.8.3
8.9

TOC

Delta Function Implicit in Fourier Transforms . . . . . . . . . . . 182

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
8.9.1

Hermitian Conjugate of a Constant Operator . . . . . . . . . . . 182

8.9.2

Hermitian Conjugate of

. . . . . . . . . . . . . . . . . . . . . 183

8.10 Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

8.11 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
9 One Dimensional Potentials
9.1

187

Piecewise Constant Potentials in 1D . . . . . . . . . . . . . . . . . . . . 187

9.1.1

The General Solution for a Constant Potential . . . . . . . . . . 187

9.1.2

The Potential Step . . . . . . . . . . . . . . . . . . . . . . . . . . 188

9.1.3

The Potential Well with E > 0 * . . . . . . . . . . . . . . . . . . 190

9.1.4

Bound States in a Potential Well * . . . . . . . . . . . . . . . . . 192

9.1.5

The Potential Barrier . . . . . . . . . . . . . . . . . . . . . . . . 196

9.2

The 1D Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . 197

9.3

The Delta Function Potential * . . . . . . . . . . . . . . . . . . . . . . . 200

9.4

The Delta Function Model of a Molecule * . . . . . . . . . . . . . . . . . 202

9.5

The Delta Function Model of a Crystal * . . . . . . . . . . . . . . . . . 204

9.6

The Quantum Rotor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

9.7

Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 207

9.7.1

Probability Flux for the Potential Step * . . . . . . . . . . . . . 207

9.7.2

Scattering from a 1D Potential Well * . . . . . . . . . . . . . . . 208

9.7.3

Bound States of a 1D Potential Well * . . . . . . . . . . . . . . . 210

9.7.4

Solving the HO Differential Equation * . . . . . . . . . . . . . . 212

9.7.5

1D Model of a Molecule Derivation * . . . . . . . . . . . . . . . . 216

9.8

9.7.6 1D Model of a Crystal Derivation * . . . . . . . . . . . . . . . . 217

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

9.9

Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

9.10 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

10 Harmonic Oscillator Solution using Operators
10.1 Introducing A and A

230

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

10.2 Commutators of A, A and H . . . . . . . . . . . . . . . . . . . . . . . . 232

10.3 Use Commutators to Derive HO Energies . . . . . . . . . . . . . . . . . 233
10.3.1 Raising and Lowering Constants . . . . . . . . . . . . . . . . . . 235
10.4 Expectation Values of p and x . . . . . . . . . . . . . . . . . . . . . . . . 236
10.5 The Wavefunction for the HO Ground State . . . . . . . . . . . . . . . . 237
6

Contents

TOC

10.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

10.6.1 The Expectation Value of x in an Energy Eigenstate . . . . . . . 238
10.6.2 The Expectation Value of p in an Energy Eigenstate . . . . . . . 238
10.6.3 The Expectation Value of x in the State
10.6.4 The Expectation Value of
10.6.5 The Expectation Value of
10.6.6 Time Development of (t

1 (u0
2

+ u1 )

. . . . . . 238

1
2 2
2 m x in an Energy Eigenstate
p2
2m in an Energy Eigenstate . . .
= 0) = 12 (u1 + u2 ) . . . . . . . .

. . . 239
. . . 239
. . . 239

10.7 Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

10.8 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
11 More Fun with Operators
245
11.1 Operators in a Vector Space . . . . . . . . . . . . . . . . . . . . . . . . . 245
11.1.1 Review of Operators . . . . . . . . . . . . . . . . . . . . . . . . . 245
11.1.2 Projection Operators |jihj| and Completeness . . . . . . . . . . . 247
11.1.3 Unitary Operators . . . . . . . . . . . . . . . . . . . . . . . . . . 248
11.2 A Complete Set of Mutually Commuting Operators . . . . . . . . . . . . 248
11.3 Uncertainty Principle for Non-Commuting Operators . . . . . . . . . . . 248
11.4 Time Derivative of Expectation Values * . . . . . . . . . . . . . . . . . . 250
11.5 The Time Development Operator * . . . . . . . . . . . . . . . . . . . . . 251
11.6 The Heisenberg Picture * . . . . . . . . . . . . . . . . . . . . . . . . . . 252
11.7 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
11.7.1 Time Development Example

. . . . . . . . . . . . . . . . . . . . 253

11.8 Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

11.9 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
12 Extending QM to Two Particles and Three Dimensions

259

12.1 Quantum Mechanics for Two Particles . . . . . . . . . . . . . . . . . . . 259

12.2 Quantum Mechanics in Three Dimensions . . . . . . . . . . . . . . . . . 260
12.3 Two Particles in Three Dimensions . . . . . . . . . . . . . . . . . . . . . 260
12.4 Identical Particles

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

12.5 Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

12.6 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
13 3D Problems Separable in Cartesian Coordinates

266

13.1 Particle in a 3D Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

13.1.1 Filling the Box with Fermions . . . . . . . . . . . . . . . . . . . . 267
13.1.2 Degeneracy Pressure in Stars . . . . . . . . . . . . . . . . . . . . 268
7

1
2

. . . . . . . . . . . . 415

20.7.3 Adding ` = 4 to ` = 2 . . . . . . . . . . . . . . . . . . . . . . . . 415

20.7.4 Two electrons in an atomic P state . . . . . . . . . . . . . . . . . 416
20.7.5 The parity of the pion from d nn. . . . . . . . . . . . . . . . 417
20.8 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 418
20.8.1 Commutators of Total Spin Operators . . . . . . . . . . . . . . . 418
20.8.2 Using the Lowering Operator to Find Total Spin States . . . . . 418
20.8.3 Applying the S 2 Operator to 1m and 00 . . . . . . . . . . . . . 419
20.8.4 Adding any ` plus spin 12 . . . . . . . . . . . . . . . . . . . . . . . 421
20.8.5 Counting the States for |`1 `2 | j `1 + `2 . . . . . . . . . . . 423
20.9 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
20.10Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
21 Time Independent Perturbation Theory

427

21.1 The Perturbation Series . . . . . . . . . . . . . . . . . . . . . . . . . . . 427

21.2 Degenerate State Perturbation Theory . . . . . . . . . . . . . . . . . . . 428
21.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
21.3.1 H.O. with anharmonic perturbation (ax4 ). . . . . . . . . . . . . . 430
21.3.2 Hydrogen Atom Ground State in a E-field, the Stark Effect. . . . 430
21.3.3 The Stark Effect for n=2 Hydrogen. . . . . . . . . . . . . . . . . 432
21.4 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 434
21.4.1 Derivation of 1st and 2nd Order Perturbation Equations . . . . . 434
21.4.2 Derivation of 1st Order Degenerate Perturbation Equations . . . 436
21.5 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
21.6 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
22 Fine Structure in Hydrogen

439

22.1 Hydrogen Fine Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 439

22.2 Hydrogen Atom in a Weak Magnetic Field . . . . . . . . . . . . . . . . . 442
22.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444
22.4 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 444
22.4.1 The Relativistic Correction . . . . . . . . . . . . . . . . . . . . . 444
22.4.2 The Spin-Orbit Correction

. . . . . . . . . . . . . . . . . . . . . 445

22.4.3 Perturbation Calculation for Relativistic Energy Shift . . . . . . 445

Contents

TOC

22.4.4 Perturbation Calculation for H2 Energy Shift . . . . . . . . . . . 447

22.4.5 The Darwin Term . . . . . . . . . . . . . . . . . . . . . . . . . . 447
22.4.6 The Anomalous Zeeman Effect . . . . . . . . . . . . . . . . . . . 448
22.5 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
22.6 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
23 Hyperfine Structure

451

23.1 Hyperfine Splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451

23.2 Hyperfine Splitting in a B Field . . . . . . . . . . . . . . . . . . . . . . . 452
23.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
23.3.1 Splitting of the Hydrogen Ground State . . . . . . . . . . . . . . 454
23.3.2 Hyperfine Splitting in a Weak B Field . . . . . . . . . . . . . . . 455
23.3.3 Hydrogen in a Strong B Field . . . . . . . . . . . . . . . . . . . . 456
23.3.4 Intermediate Field . . . . . . . . . . . . . . . . . . . . . . . . . . 457
23.3.5 Positronium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
23.3.6 Hyperfine and Zeeman for H, muonium, positronium . . . . . . . 459
23.4 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 460
23.4.1 Hyperfine Correction in Hydrogen . . . . . . . . . . . . . . . . . 460
23.5 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
23.6 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
24 The Helium Atom

464

24.1 General Features of Helium States . . . . . . . . . . . . . . . . . . . . . 464

24.2 The Helium Ground State . . . . . . . . . . . . . . . . . . . . . . . . . . 466
24.3 The First Excited State(s) . . . . . . . . . . . . . . . . . . . . . . . . . . 466
24.4 The Variational Principle (Rayleigh-Ritz Approximation) . . . . . . . . 470
24.5 Variational Helium Ground State Energy . . . . . . . . . . . . . . . . . 472
24.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
24.6.1 1D Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . . . . 473
24.6.2 1-D H.O. with exponential wavefunction . . . . . . . . . . . . . . 474
24.7 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 475
24.7.1 Calculation of the ground state energy shift . . . . . . . . . . . . 475
24.8 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
24.9 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478

Contents

TOC

25 Atomic Physics
25.1 Atomic Shell Model

479
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479

25.2 The Hartree Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480

25.3 Hunds Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
25.4 The Periodic Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482
25.5 The Nuclear Shell Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 486
25.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
25.6.1 Boron Ground State . . . . . . . . . . . . . . . . . . . . . . . . . 488
25.6.2 Carbon Ground State . . . . . . . . . . . . . . . . . . . . . . . . 488
25.6.3 Nitrogen Ground State . . . . . . . . . . . . . . . . . . . . . . . . 489
25.6.4 Oxygen Ground State . . . . . . . . . . . . . . . . . . . . . . . . 491
25.7 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
25.8 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
26 Molecular Physics

493

26.1 The H+
2 Ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
26.2 The H2 Molecule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
26.3 Importance of Unpaired Valence Electrons . . . . . . . . . . . . . . . . . 496
26.4 Molecular Orbitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
26.5 Vibrational States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
26.6 Rotational States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
26.7 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
26.8 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 502
26.9 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
26.10Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
27 Time Dependent Perturbation Theory

504

27.1 General Time Dependent Perturbations . . . . . . . . . . . . . . . . . . 504

27.2 Sinusoidal Perturbations . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
27.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 509
27.3.1 Harmonic Oscillator in a Transient E Field . . . . . . . . . . . . 509
27.4 Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 510
27.4.1 The Delta Function of Energy Conservation . . . . . . . . . . . . 510
27.5 Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
27.6 Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 512

Contents

28 Radiation in Atoms

TOC

513

28.1 The Photon Field in the Quantum Hamiltonian . . . . . . . . . . . . . . 513

28.2 Decay Rates for the Emission of Photons

. . . . . . . . . . . . . . . . . 515

28.3 Phase Space: The Density of Final States . . . . . . . . . . . . . . . . . 516

28.4 Total Decay Rate Using Phase Space . . . . . . . . . . . . . . . . . . . . 517
28.5 Electric Dipole Approximation and Selection Rules . . . . . . . . . . . . 517
28.6 Explicit 2p to 1s Decay Rate . . . . . . . . . . . . . . . . . . . . . . . . 521
28.7 General Unpolarized Initial State . . . . . . . . . . . . . . . . . . . . . . 523
28.8 Angular Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
28.9 Vector Operators and the Wigner Eckart Theorem . . . . . . . . . . . . 528
28.10Exponential Decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
28.11Lifetime and Line Width . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
28.11.1 Other Phenomena Influencing Line Width . . . . . . . . . . . . . 531
28.12Phenomena of Radiation Theory . . . . . . . . . . . . . . . . . . . . . . 533
28.12.1 The M
ossbauer Effect . . . . . . . . . . . . . . . . . . . . . . . . 533
28.12.2 LASERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
28.13Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536
28.13.1 The 2P to 1S Decay Rate in Hydrogen . . . . . . . . . . . . . . . 536
28.14Derivations and Computations . . . . . . . . . . . . . . . . . . . . . . . 536
28.14.1 Energy in Field for a Given Vector Potential . . . . . . . . . . . 536
28.14.2 General Phase Space Formula . . . . . . . . . . . . . . . . . . . . 537
28.14.3 Estimate of Atomic Decay Rate . . . . . . . . . . . . . . . . . . . 537
28.15Homework Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 538
28.16Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
29 Scattering

540

29.1 The Born Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . 543

29.1.1 Scattering from a Screened Coulomb Potential . . . . . . . . . . 544
29.2 The Radial Equation and Constant Potentials * . . . . . . . . . . . . . . 546
29.3 Behavior at the Origin * . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
29.4 Spherical Bessel Functions * . . . . . . . . . . . . . . . . . . . . . . . . . 547
29.5 Particle in a Sphere * . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
29.6 Bound States in a Spherical Potential Well * . . . . . . . . . . . . . . . 550
29.7 Partial Wave Analysis of Scattering * . . . . . . . . . . . . . . . . . . . 552
29.8 Scattering from a Spherical Well * . . . . . . . . . . . . . . . . . . . . . 554
29.9 The Radial Equation for u(r) = rR(r) * . . . . . . . . . . . . . . . . . . 556
29.10Partial Wave Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
14

Contents

TOC

29.10.1 Scattering from a Hard Sphere . . . . . . . . . . . . . . . . . . . 559

29.11Homework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
29.12Sample Test Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
30 Classical Scalar Fields
563
30.1 Simple Mechanical Systems and Fields . . . . . . . . . . . . . . . . . . . 564
30.2 Classical Scalar Field in Four Dimensions . . . . . . . . . . . . . . . . . 565
31 Classical Maxwell Fields
31.1 Rationalized Heaviside-Lorentz Units . . . .
31.2 The Electromagnetic Field Tensor . . . . .
31.3 The Lagrangian for Electromagnetic Fields
31.4 Gauge Invariance can Simplify Equations .

.
.
.
.

575
575
576
578
581

32 Quantum Theory of Radiation

32.1 Transverse and Longitudinal Fields . . . . . . . . .
32.2 Fourier Decomposition of Radiation Oscillators . .
32.3 The Hamiltonian for the Radiation Field . . . . . .
32.4 Canonical Coordinates and Momenta . . . . . . . .
32.5 Quantization of the Oscillators . . . . . . . . . . .
32.6 Photon States . . . . . . . . . . . . . . . . . . . . .
32.7 Fermion Operators . . . . . . . . . . . . . . . . . .
32.8 Quantized Radiation Field . . . . . . . . . . . . . .
32.9 The Time Development of Field Operators . . . . .
32.10Uncertainty Relations and RMS Field Fluctuations
32.11Emission and Absorption of Photons by Atoms . .
32.12Review of Radiation of Photons . . . . . . . . . . .
32.12.1 Beyond the Electric Dipole Approximation
32.13Black Body Radiation Spectrum . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.

583
583
584
586
588
590
594
595
595
598
599
600
602
603
605

33 Scattering of Photons
33.1 Resonant Scattering
33.2 Elastic Scattering . .
33.3 Rayleigh Scattering .
33.4 Thomson Scattering
33.5 Raman Effect . . . .

.
.
.
.
.

607
613
614
616
616
618

.
.
.
.
.

34 Electron Self Energy Corrections

619
34.1 The Lamb Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 627
15

Contents

35 Dirac Equation

TOC

633

35.1 Diracs Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633

35.2 The Schr
odinger-Pauli Hamiltonian . . . . . . . . . . . . . . . . . . . . . 634
35.3 The Dirac Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
35.4 The Conserved Probability Current . . . . . . . . . . . . . . . . . . . . . 640
35.5 The Non-relativistic Limit of the Dirac Equation . . . . . . . . . . . . . 642
35.5.1 The Two Component Dirac Equation . . . . . . . . . . . . . . . 642
35.5.2 The Large and Small Components of the Dirac Wavefunction . . 643
35.5.3 The Non-Relativistic Equation . . . . . . . . . . . . . . . . . . . 643
35.6 Solution of Dirac Equation for a Free Particle . . . . . . . . . . . . . . . 647
35.6.1 Dirac Particle at Rest . . . . . . . . . . . . . . . . . . . . . . . . 649
35.6.2 Dirac Plane Wave Solution . . . . . . . . . . . . . . . . . . . . . 651
35.6.3 Alternate Labeling of the Plane Wave Solutions . . . . . . . . . . 659
35.7 Negative Energy Solutions: Hole Theory . . . . . . . . . . . . . . . . . 660
35.8 Equivalence of a Two Component Theory . . . . . . . . . . . . . . . . . 662
35.9 Relativistic Covariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
35.10Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
35.11Bilinear Covariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 671
35.12Constants of the Motion for a Free Particle . . . . . . . . . . . . . . . . 674
35.13The Relativistic Interaction Hamiltonian . . . . . . . . . . . . . . . . . . 677
35.14Phenomena of Dirac States . . . . . . . . . . . . . . . . . . . . . . . . . 677
35.14.1 Velocity Operator and Zitterbewegung . . . . . . . . . . . . . . . 677
35.14.2 Expansion of a State in Plane Waves . . . . . . . . . . . . . . . . 680
35.14.3 The Expected Velocity and Zitterbewegung . . . . . . . . . . . . 681
35.15Solution of the Dirac Equation for Hydrogen . . . . . . . . . . . . . . . 682
35.16Thomson Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693
35.17Hole Theory and Charge Conjugation . . . . . . . . . . . . . . . . . . . 697
35.18Charge Conjugate Waves . . . . . . . . . . . . . . . . . . . . . . . . . . 699
35.19Quantization of the Dirac Field . . . . . . . . . . . . . . . . . . . . . . . 702
35.20The Quantized Dirac Field with Positron Spinors . . . . . . . . . . . . . 707
35.21Vacuum Polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 708
35.22The QED LaGrangian and Gauge Invariance . . . . . . . . . . . . . . . 710
35.23Interaction with a Scalar Field . . . . . . . . . . . . . . . . . . . . . . . 711
36 Formulas with Hyperlinks to Course

712

Contents

TOC

Preface
These notes represent an experiment in the use of information technology in teaching an
advanced undergraduate physics course, Quantum Physics at UCSD. The experiment
has several goals.
To make all the class material including a complete set of lecture notes available
to students on the World-Wide Web.
To make use of some simple multimedia technology to enhance the class notes as
a learning tool compared to a conventional textbook.
To present a complex subject to students in several different ways so that each
student can use the learning techniques best suited to that individual.
To get some experience with the use of multimedia technologies in teaching advanced courses.
To produce course material that might be appropriate for distance learning or
self-paced courses in the future.
The current set of notes covers a 3 quarter course at UCSD, from the beginning of
Quantum Mechanics to the quantization of the electromagnetic field and the Dirac
equation. The notes for the last quarter should be considered to be a first draft.
At this time, the experiment is in progress. One quarter is not sufficient to optimize
the course material. While a complete set of html based notes has been produced, only
limited additional audio and visual material is now available.
It is my personal teaching experience that upper division physics students learn in
different ways. Many physics students get very little more than an introduction to the
material out of the lecture and prefer to learn from the textbook and homework. Some
students claim they cannot learn from the textbook and rely on lectures to get their
basic understanding. Some prefer a rather verbose exposition of the material in the
text, while others prefer a concise discussion largely based on equations. Modern media
have conditioned the students of today in a way that is often detrimental to learning
complex subjects from either a lecture or a textbook.
I chose to use html and the worldwide web as the primary delivery tool for enhanced
class notes. All of the standard software tools and information formats are usable from
html. Every computer can access this format using Internet browsers.
An important aspect of the design of the notes is to maintain a concise basic treatment
of the physics, with derivations and examples available behind hyperlinks. It is my goal,
17

Contents

TOC

not fully met at this time, to have very detailed derivations, with less steps skipped
than in standard textbooks. Eventually, this format will allow more examples than are
practical in a textbook.
Another important aspect is audio discussion of important equations and drawings.
The browser is able to concentrate on an equation while hearing about the details
instead of having to go back an forth between text and equation. The use of this needs
to be expanded and would benefit from better software tools.
Because of the heavy use of complex equations in this course, the html is generated
from LaTeX input. This has not proved to be a limitation so far since native html can
be included. LaTeX has the ability to produce high quality equations and input is fast
compared to other options. The LaTeX2html translator functions well enough for the
conversion.
Projecting the notes can be very useful in lecture for introductions, for review, and for
quick looks at derivations. The primary teaching though probably still works best at
the blackboard. One thing that our classrooms really dont facilitate is switching from
one mode to the other.
In a future class, with the notes fully prepared, I will plan to decrease the formal
lecture time and add lab or discussion session time, with students working moving at
their own pace using computers. Projects could be worked on in groups or individually.
Instructors would be available to answer questions and give suggestions.
Similar sessions would be possible at a distance. The formal lecture could be taped
and available in bite size pieces inside the lecture notes. Advanced classes with small
numbers of students could be taught based on notes, with less instructor support than
is usual. Classes could be offered more often than is currently feasible.
Jim Branson

1. Course Summary

1
1.1

TOC

Course Summary
Problems with Classical Physics

Around the beginning of the 20th century, classical physics, based on Newtonian
Mechanics and Maxwells equations of Electricity and Magnetism described nature
as we knew it. Statistical Mechanics was also a well developed discipline describing
systems with a large number of degrees of freedom. Around that time, Einstein introduced Special Relativity which was compatible with Maxwells equations but changed
our understanding of space-time and modified Mechanics.
Many things remained unexplained. While the electron as a constituent of atoms had
been found, atomic structure was rich and quite mysterious. There were problems
with classical physics including Black Body Radiation, the Photoelectric effect, basic
Atomic Theory, Compton Scattering, and eventually with the diffraction of all kinds
of particles. Planck hypothesized that EM energy was always emitted in quanta
E = h = ~
to solve the Black Body problem. Much later, de Broglie derived the wavelength for
particles.
h
=
p
Ultimately, the problems led to the development of Quantum Mechanics in which all
particles are understood to have both wave and a particle behavior.

1.2

Thought Experiments on Diffraction

Diffraction of photons, electrons, and neutrons has been observed (see the pictures)
and used to study crystal structure.
To understand the experimental input in a simplified way, we consider some thought
experiments on the diffraction of photons, electrons, and bullets through two slits. For
example, photons, which make up all electromagnetic waves, show a diffraction pattern
exactly as predicted by the theory of EM waves, but we always detect an integer number
of photons with the Plancks relation, E = h, between wave frequency and particle
energy satisfied.
Electrons, neutrons, and everything else behave in exactly the same way, exhibiting
wave-like diffraction yet detection of an integer number of particles and satisfying
= hp . This de Broglie wavelength formula relates the wave property to the particle
property p.
19

1. Course Summary

TOC

Figure 1: We can now do one of the thought experiments on diffraction. Time development of the accumulated hits in our single photon detector. Photons are detected one
at a time. Even though there is only one photon counted at a time, the interference
patter is eventually seen.

1.3

Probability Amplitudes

In Quantum Mechanics, we understand this wave-particle duality using (complex)

probability amplitudes which satisfy a wave equation.
~

(~x, t) = ei(k~xt) = ei(~p~xEt)/~

The probability to find a particle in a volume d3 x around position ~x and at some time
t is the absolute square of the probability amplitude (~x, t) times the volume.
2

P (~x, t)d3 x = |(~x, t)| d3 x

To compute the probability to find an electron at our thought experiment detector, we

add the probability amplitude to get to the detector through slit 1 to the amplitude to
get to the detector through slit 2 and take the absolute square.
2

Pdetector = |1 + 2 |

Quantum Mechanics completely changes our view of the world. Instead of a deterministic world, we now have only probabilities. We cannot even measure both
the position and momentum of a particle (accurately) at the same time. Quantum
Mechanics will require us to use the mathematics of operators, Fourier Transforms,
vector spaces, and much more.

1. Course Summary

1.4

TOC

Wave Packets and Uncertainty

The
p probability amplitude for a free particle with momentum p~ and energy E =
(pc)2 + (mc2 )2 is the complex wave function
free

x, t)
particle (~

= ei(~p~xEt)/~ .

Note that ||2 = 1 everywhere so this does not represent a localized particle. In fact
we recognize the wave property that, to have exactly one frequency, a wave must be
spread out over space.
We can build up localized wave packets that represent single particles by adding up
these free particle wave functions (with some coefficients).
1
(x, t) =
2~

(p)ei(pxEt)/~ dp

(We have moved to one dimension for simplicity.) Similarly we can compute the coefficient for each momentum
(p) =

1
2~

(x)eipx/~ dx.

These coefficients, (p), are actually the state function of the particle in momentum
space. We can describe the state of a particle either in position space with (x) or in
momentum space with (p). We can use (p) to compute the probability distribution
function for momentum.
2
P (p) = |(p)|
We will show that wave packets like these behave correctly in the classical limit, vindicating the choice we made for free particle (~x, t).
The Heisenberg Uncertainty Principle is a property of waves that we can deduce from
our study of localized wave packets.
px

~
2

It shows that due to the wave nature of particles, we cannot localize a particle into a
small volume without increasing its energy. For example, we can estimate the ground
state energy (and the size of) a Hydrogen atom very well from the uncertainty principle.
The next step in building up Quantum Mechanics is to determine how a wave function
develops with time particularly useful if a potential is applied. The differential
equation which wave functions must satisfy is called the Schrodinger Equation.
21

1. Course Summary

1.5

TOC

Operators

The Schr
odinger equation comes directly out of our understanding of wave packets. To
get from wave packets to a differential equation, we use the new concept of (linear)
operators. We determine the momentum and energy operators by requiring that, when
an operator for some variable v acts on our simple wavefunction, we get v times the
same wave function.
~
px =
i x
~

px ei(~p~xEt)/~ =
ei(~p~xEt)/~ = px ei(~p~xEt)/~
i x
= i~
E
t
i(~p~xEt)/~ = i~ ei(~p~xEt)/~ = Eei(~p~xEt)/~
Ee
t

1.6

Expectation Values

We can use operators to help us compute the expectation value of a physical variable.
If a particle is in the state (x), the normal way to compute the expectation value of
f (x) is

hf (x)i =
P (x)f (x)dx =
(x)(x)f (x)dx.

If the variable we wish to compute the expectation value of (like p) is not a simple
function of x, let its operator act on (x)

(x)
p(x)dx.

hpi =

We have a shorthand notation for the expectation value of a variable v in the state
which is quite useful.

h|v|i
(x)
v (x)dx.

We extend the notation from just expectation values to

(x)
v (x)dx

h|v|i

1. Course Summary

TOC

and

(x)(x)dx

h|i

We use this shorthand Dirac Bra-Ket notation a great deal.

1.7

Commutators

Operators (or variables in quantum mechanics) do not necessarily commute. We can

compute the commutator of two variables, for example
[p, x] px xp =

~
.
i

Later we will learn to derive the uncertainty relation for two variables from their commutator. We will also use commutators to solve several important problems.

1.8

The Schr
odinger Equation

Wave functions must satisfy the Schr

odinger Equation which is actually a wave equation.
~2 2
(~x, t)
(~x, t) + V (~x)(~x, t) = i~
2m
t
We will use it to solve many problems in this course. In terms of operators, this can
be written as
x, t) = E(~
x, t)
H(~
2

= p + V (~x) is the Hamiltonian operator. So the

where (dropping the (op) label) H
2m
Schr
odinger Equation is, in some sense, simply the statement (in operators) that the
kinetic energy plus the potential energy equals the total energy.

1.9

Eigenfunctions, Eigenvalues and Vector Spaces

For any given physical problem, the Schr

odinger equation solutions which separate
(between time and space), (x, t) = u(x)T (t), are an extremely important set. If
we assume the equation separates, we get the two equations (in one dimension for
simplicity)
(x, t) = u(x)T (t)
i~

T (t)
= E T (t)
t
23

1. Course Summary

TOC

Hu(x) = E u(x)
The second equation is called the time independent Schrodinger equation. For bound
states, there are only solutions to that equation for some quantized set of energies
Hui (x) = Ei ui (x).
For states which are not bound, a continuous range of energies is allowed.
The time independent Schr
odinger equation is an example of an eigenvalue equation.
Hi (~x) = Ei i (~x)
If we operate on i with H, we get back the same function i times some constant. In
this case i would be called and Eigenfunction, and Ei would be called an Eigenvalue.
There are usually an infinite number of solutions, indicated by the index i here.
Operators for physical variables must have real eigenvalues. They are called Hermitian
operators. We can show that the eigenfunctions of Hermitian operators are orthogonal
(and can be normalized).
hi |j i = ij
(In the case of eigenfunctions with the same eigenvalue, called degenerate eigenfunctions, we can must choose linear combinations which are orthogonal to each other.) We
will assume that the eigenfunctions also form a complete set so that any wavefunction
can be expanded in them,
X
(~x) =
i i (~x)
i

where the i are coefficients which can be easily computed (due to orthonormality) by
i = hi |i.
So now we have another way to represent a state (in addition to position space and
momentum space). We can represent a state by giving the coefficients in sum above.
(Note that p (x) = ei(pxEt)/~ is just an eigenfunction of the momentum operator and
x (p) = ei(pxEt)/~ is just an eigenfunction of the position operator (in p-space) so
they also represent and expansion of the state in terms of eigenfunctions.)
Since the i form an orthonormal, complete set, they can be thought of as the unit
vectors of a vector space. The arbitrary wavefunction would then be a vector in that
space and could be represented by its coefficients.

1
2

=
3
...
The bra-ket h|i i can be thought of as a dot product between the arbitrary vector
and one of the unit vectors. We can use the expansion in terms of energy eigenstates
24

1. Course Summary

TOC

to compute many things. In particular, since the time development of the energy
eigenstates is very simple,
i (~x, t) = i (~x)eiEi t/~
we can use these eigenstates to follow the time development of an arbitrary state

k
m.

The energy eigenvalues are

1
En = n +
~.
2
2

The energy eigenstates turn out to be a polynomial (in x) of degree n times emx
So the ground state, properly normalized, is just
m 14
2
u0 (x) =
emx /~ .
~

We developed a recursion relation from the differential equation to give all the HO
energy eigenfunctions.
We will later return the harmonic oscillator to solve the problem by operator methods.

1.13

Delta Function Potentials in One Dimension

The delta function potential is a very useful one to make simple models of molecules and
solids. First we solve the problem with one attractive delta function V (x) = aV0 (x).
Since the bound state has negative energy, the solutions that are normalizable are Cex
for x < 0 and Cex for x > 0. Making u(x) continuous and its first derivative have a
discontinuity computed from the Schr
odinger equation at x = 0, gives us exactly one
bound state with
ma2 V02
E=
.
2~2
Next we use two delta functions to model a molecule, V (x) = aV0 (x+d)aV0 (xd).
Solving this problem by matching wave functions at the boundaries at d, we find again
transcendental equations for two bound state energies. The ground state energy is more
negative than that for one delta function, indicating that the molecule would be bound.
A look at the wavefunction shows that the 2 delta function state can lower the kinetic
energy compared to the state for one delta function, by reducing the curvature of the
wavefunction. The excited state has more curvature than the atomic state so we would
not expect molecular binding in that state.
Our final 1D potential, is a model of a solid.
V (x) = aV0

(x na)

This has a infinite, periodic array of delta functions, so this might be applicable to a
crystal. The solution to this is a bit tricky but it comes down to
cos() = cos(ka) +

2maV0
sin(ka).
~2 k

1. Course Summary

TOC

Since the right hand side of the equation can be bigger than 1.0 (or less than -1), there
2 2
k
which do not have solutions. There are also bands of energies
are regions of E = ~2m
with solutions. These energy bands are seen in crystals (like Si).

1.14

Harmonic Oscillator Solution with Operators

We can solve the harmonic oscillator problem using operator methods. We write the
Hamiltonian in terms of the operator
r

p
m
A
x + i
2~
2m~
.
p2
1
1
H=
+ m 2 x2 = ~(A A + )
2m 2
2
We compute the commutators
[A, A ] =

i
([x, p] + [p, x]) = 1
2~

[H, A] = ~[A A, A] = ~[A , A]A = ~A

[H, A ] = ~[A A, A ] = ~A [A, A ] = ~A
If we apply the the commutator [H, A] to the eigenfunction un , we get [H, A]un =
~Aun which rearranges to the eigenvalue equation
H(Aun ) = (En ~)(Aun ).
This says that (Aun ) is an eigenfunction of H with eigenvalue (En ~) so it lowers
the energy by ~. Since the energy must be positive for this Hamiltonian, the lowering
must stop somewhere, at the ground state, where we will have
Au0 = 0.
This allows us to compute the ground state energy like this
1
1
Hu0 = ~(A A + )u0 = ~u0
2
2
showing that the ground state energy is 12 ~. Similarly, A raises the energy by
~. We can travel up and down the energy ladder using A and A, always in steps of
~. The energy eigenvalues are therefore

1
En = n +
~.
2
A little more computation shows that
Aun =

nun1
29

1. Course Summary

TOC

and that
A un =

n + 1un+1 .

These formulas are useful for all kinds of computations within the important harmonic oscillator system. Both p and x can be written in terms of A and A .
r
~
x=
(A + A )
2m
r
m~
p = i
(A A )
2

1.15

More Fun with Operators

It was shown that the uncertainty principle limit for two physical variables is proportional to the commutator of the two operators.
But the main results are on time developement. First, the time derivative of the
expectation value of an operator (that does not have an explicity time dependence),
can be computed from the commutator of that operator with the Hamiltonian.
d
i
h |A| i = h |[H, A]| i
dt
~
Thus an operator that commutes with the Hamiltonian will give a constant of the
motion.
We find the time development operator by solving the equation i~
t = H.
(t) = eiHt/~ (t = 0)
This implies that eiHt/~ is the time development operator. In some cases we can
calculate the actual operator from the power series for the exponential.
eiHt/~ =

X
(iHt/~)n
n!
n=0

We have been working in what is called the Schr

odinger picture in which the wavefunctions (or states) develop with time. There is the alternate Heisenberg picture in
which the operators develop with time while the states do not change. For example, if
we wish to compute the expectation value of the operator B as a function of time in
the usual Schr
odinger picture, we get
h(t)|B|(t)i = heiHt/~ (0)|B|eiHt/~ (0)i = h(0)|eiHt/~ BeiHt/~ |(0)i.
30

1. Course Summary

TOC

In the Heisenberg picture the operator B(t) = eiHt/~ BeiHt/~ .

We use operator methods to compute the uncertainty relationship between non-commuting
variables
i
(A)(B) h[A, B]i
2
which gives the result we deduced from wave packets for p and x.
Again we use operator methods to calculate the time derivative of an expectation value.

A
d
i

h|A|i = h|[H, A]|i +
dt
~
t

(Most operators we use dont have explicit time dependence so the second term is
usually zero.) This again shows the importance of the Hamiltonian operator for time
development. We can use this to show that in Quantum mechanics the expectation
values for p and x behave as we would expect from Newtonian mechanics (Ehrenfest
Theorem).
DpE
dhxi
i
i p2
= h[H, x]i = h[
, x]i =
dt
~
~ 2m
m

dhpi
i
i
~ d
dV (x)
= h[H, p]i =
[V (x),
] =
dt
~
~
i dx
dx
Any operator A that commutes with the Hamiltonian has a time independent
expectation value. The energy eigenfunctions can also be (simultaneous) eigenfunctions
of the commuting operator A. It is usually a symmetry of the H that leads to a
commuting operator and hence an additional constant of the motion.

1.16

Two Particles in 3 Dimensions

So far we have been working with states of just one particle in one dimension. The
extension to two different particles and to three dimensions is straightforward. The
coordinates and momenta of different particles and of the additional dimensions commute with each other as we might expect from classical physics. The only things
that dont commute are a coordinate with its momentum, for example,
[p(2)z , z(2) ] =

~
i

while
[p(1)x , x(2) ] = [p(2)z , y(2) ] = 0.
We may write states for two particles which are uncorrelated, like u0 (~x(1) )u3 (~x(2) ), or
we may write states in which the particles are correlated. The Hamiltonian for two
31

1. Course Summary

TOC

particles in 3 dimensions simply becomes

!
~2
2
2
2
~2
H=
+
+
+
2
2
2m(1) x2(1)
y(1)
z(1)
2m(2)
H=

2
2
2
+
+
2
2
x2(2)
y(2)
z(2)

!
+V (~x(1) , ~x(2) )

~2 2
~2 2
(1) +
+ V (~x(1) , ~x(2) )
2m(1)
2m(2) (1)

If two particles interact with each other, with no external potential,

~2 2
~2 2
(1) +
+ V (~x(1) ~x(2) )
2m(1)
2m(2) (1)

the Hamiltonian has a translational symmetry, and remains invariant under the
translation ~x ~x + ~a. We can show that this translational symmetry implies conservation of total momentum. Similarly, we will show that rotational symmetry
implies conservation of angular momentum, and that time symmetry implies conservation of energy.
For two particles interacting through a potential that depends only on difference on
the coordinates,
p~2
p~2
H = 1 + 2 + V (~r1 ~r2 )
2m 2m
we can make the usual transformation to the center of mass made in classical mechanics
~r ~r1 ~r2
~ m1~r1 + m2~r2
R
m1 + m2
and reduce the problem to the CM moving like a free particle
M = m1 + m2
~2 ~ 2

2M R
plus one potential problem in 3 dimensions with the usual reduced mass.
H=

1
1
1
=
+

m1
m2
H=

~2 ~ 2
+ V (~r)
2 r

So we are now left with a 3D problem to solve (3 variables instead of 6).

1. Course Summary

1.17

TOC

Identical Particles

Identical particles present us with another symmetry in nature. Electrons, for example,
are indistinguishable from each other so we must have a symmetry of the Hamiltonian
under interchange of any pair of electrons. Lets call the operator that interchanges
electron-1 and electron-2 P12 .
[H, P12 ] = 0
So we can make our energy eigenstates also eigenstates of P12 . Its easy to see (by
operating on an eigenstate twice with P12 ), that the possible eigenvalues are 1. It is a
law of physics that spin 12 particles called fermions (like electrons) always are antisymmetric under interchange, while particles with integer spin called bosons
(like photons) always are symmetric under interchange. Antisymmetry under
interchange leads to the Pauli exclusion principle that no two electrons (for example)
can be in the same state.

1.18

Some 3D Problems Separable in Cartesian Coordinates

We begin our study of Quantum Mechanics in 3 dimensions with a few simple cases
of problems that can be separated in Cartesian coordinates. This is possible when the
Hamiltonian can be written
H = Hx + Hy + Hz .
One nice example of separation of variable in Cartesian coordinates is the 3D harmonic oscillator
1
V (r) = m 2 r2
2
which has energies which depend on three quantum numbers.

3
E nx ny nz = n x + n y + n z +
~
2
It really behaves like 3 independent one dimensional harmonic oscillators.
Another problem that separates is the particle in a 3D box. Again, energies depend
on three quantum numbers

2 ~2
n2x + n2y + n2z
2
2mL
for a cubic box of side L. We investigate the effect of the Pauli exclusion principle by
filling our 3D box with identical fermions which must all be in different states. We can
use this to model White Dwarfs or Neutron Stars.
Enx ny nz =

In classical physics, it takes three coordinates to give the location of a particle in 3D.
In quantum mechanics, we are finding that it takes three quantum numbers to
label and energy eigenstate (not including spin).
33

1. Course Summary

1.19

TOC

Angular Momentum

For the common problem of central potentials V (r), we use the obvious rotational
~ = ~x p~, operators commute
symmetry to find that the angular momentum, L
with H,
[H, Lz ] = [H, Lx ] = [H, Ly ] = 0
but they do not commute with each other.
[Lx , Ly ] 6= 0
We want to find two mutually commuting operators which commute with H, so
we turn to L2 = L2x + L2y + L2z which does commute with each component of L.
[L2 , Lz ] = 0
We chose our two operators to be L2 and Lz .
Some computation reveals that we can write
p2 =

1
L2 + (~r p~)2 i~~r p~ .
2
r

With this the kinetic energy part of our equation will only have derivatives in r assuming
that we have eigenstates of L2 .
"
#
2

1
L2
~2 1
r
+

uE (~r) + V (r)uE (~r) = EuE (~r)

2 r2
r
r r ~2 r2
The Schr
odinger equation thus separates into an angular part (the L2 term)
and a radial part (the rest). With this separation we expect (anticipating the angular
solution a bit)
uE (~r) = RE` (r)Y`m (, )
will be a solution. The Y`m (, ) will be eigenfunctions of L2
L2 Y`m (, ) = `(` + 1)~2 Y`m (, )
so the radial equation becomes
"
#
2
~2 1

1
`(` + 1)
r
RE` (r) + V (r)RE` (r) = ERE` (r)
+

2 r2
r
r r
r2
We must come back to this equation for each V (r) which we want to solve.
We solve the angular part of the problem in general using angular momentum
operators. We find that angular momentum is quantized.
Lz Y`m (, ) = m~Y`m (, )
34

1. Course Summary

TOC

L2 Y`m (, ) = `(` + 1)~2 Y`m (, )

with ` and m integers satisfying the condition ` m `. The operators that raise
and lower the z component of angular momentum are

L Y`m

L = Lx iLy
p
= ~ `(` + 1) m(m 1)Y`(m1)

We derive the functional form of the Spherical Harmonics Y`m (, ) using the
differential form of the angular momentum operators.

1.20

Solutions to the Radial Equation for Constant Potentials

Solutions to the radial equation in a constant potential are important since they are the
solutions for large r in potentials of limitted range. They are therefore used in scattering
problems as the incoming and outgoing states. The solutions are the spherical Bessel
and spherical Neumann functions.
j` () = ()`

n` () = ()`

1 d
d

sin(
sin

1 d
d

cos(
cos

`
2 )

where = kr. The linear combination of these which falls off properly at large r is
called the Hankel function of the first type.
(1)
h` ()

= j` () + in` () = ()

1 d
d

`
sin i cos
i
ei( 2 )

We use these solutions to do a partial wave analysis of scattering, solve for bound
states of a spherical potential well, solve for bound states of an infinite spherical
well (a spherical box), and solve for scattering from a spherical potential well.

1.21

Hydrogen

The Hydrogen (Coulomb potential) radial equation is solved by finding the behavior
at large r, then finding the behavior at small r, then using a power series solution to
get

X
R() = `
ak k e/2
k=0

1. Course Summary

TOC

q
8E
with =
~2 r. To keep the wavefunction normalizable the power series must
terminate, giving us our energy eigenvalue condition.
En =

Z 2 2 mc2
2n2

Here n is called the principle quantum number and it is given by

n = nr + ` + 1
where nr is the number of nodes in the radial wavefunction. It is an odd feature of
Hydrogen that a radial excitation and an angular excitation have the same energy.
So a Hydrogen energy eigenstate n`m (~x) = Rn` (r)Y`m (, ) is described by three
integer quantum numbers with the requirements that n 1, ` < n and also an integer,
and l m `. The ground state of Hydrogen is 100 and has energy of -13.6 eV.
We compute several of the lowest energy eigenstates.

Figure 2: Diagram showing the lowest energy bound states of Hydrogen and their dominant (E1) decays.

1. Course Summary

1.22

TOC

Solution of the 3D HO Problem in Spherical Coordinates

As and example of another problem with spherical symmetry, we solve the 3D symmetric harmonic oscillator problem. We have already solved this problem in Cartesian
coordinates. Now we use spherical coordinates and angular momentum eigenfunctions.
The eigen-energies are

3
E = 2nr + ` +
~
2
where nr is the number of nodes in the radial wave function and ` is the total angular
momentum quantum number. This gives exactly the same set of eigen-energies as we
got in the Cartesian solution but the eigenstates are now states of definite total angular
momentum and z component of angular momentum.

1.23

Matrix Representation of Operators and States

We may define the components of a state vector as the projections of the state on
a complete, orthonormal set of states, like the eigenfunctions of a Hermitian operator.
i

|i =

hui |i
X
i |ui i
i

Similarly, we may define the matrix element of an operator in terms of a pair of

those orthonormal basis states
Oij hui |O|uj i.
With these definitions, Quantum Mechanics problems can be solved using the matrix
representation operators and states. An operator acting on a state is a matrix times a
vector.

(O)1
O11 O12 ... O1j ...
1
(O)2 O21 O22 ... O2j ... 2

... = ...

... ... ... ...

...
(O)i Oi1 Oi2 ... Oij ... j
...
...
... ... ... ...
...
The product of operators is the product of matrices. Operators which dont commute
are represented by matrices that dont commute.

1. Course Summary

1.24

TOC

A Study of ` = 1 Operators and Eigenfunctions

The set of states with the same total angular momentum and the angular momentum
operators which act on them are often represented by vectors and matrices. For example
the different m states for ` = 1 will be represented by a 3 component vector and the
angular momentum operators represented by 3X3 matrices. There are both practical
and theoretical reasons why this set of states is separated from the states with different
total angular momentum quantum numbers. The states are often (nearly) degenerate
and therefore should be treated as a group for practical reasons. Also, a rotation of
the coordinate axes will not change the total angular momentum quantum number so
the rotation operator works within this group of states.
We write our 3 component vectors as follows.

+
= 0

The matrices

0
~
Lx = 1
2 0

representing the angular momentum operators for ` = 1 are as follows.

1 0
0
1 0
1 0 0
~
0 1
Ly = 1 0 1
Lz = ~ 0 0 0
2i
1 0
0 1 0
0 0 1

The same matrices also represent spin 1, s = 1, but of course would act on a different
vector space.
The rotation operators (symmetry operators) are given by
Rz (z ) = eiz Lz /~

Rx (x ) = eix Lx /~

Ry (y ) = eiy Ly /~

for the differential form or the matrix form of the operators. For ` = 1 these are 3X3
(unitary) matrices. We use them when we need to redefine the direction of our coordinate axes. Rotations of the angular momentum states are not the same as rotations
of vectors in 3 space. The components of the vectors represent different quantities and
hence transform quite differently. The vectors we are using for angular momentum
actually should be called spinors when we refer to their properties under rotations and
Lorentz boosts.

1.25

Spin 1/2 and other 2 State Systems

The angular momentum algebra defined by the commutation relations between the
operators requires that the total angular momentum quantum number must either be
an integer or a half integer. The half integer possibility was not useful for orbital angular
38

1. Course Summary

TOC

momentum because there was no corresponding (single valued) spherical harmonic

function to represent the amplitude for a particle to be at some position.
The half integer possibility is used to represent the internal angular momentum of some
particles. The simplest and most important case is spin one-half.
There are just two
1
possible states with different z components of spin: spin up
, with z component
0

0
, with ~2 . The corresponding spin
of angular momentum + ~2 , and spin down
1
operators are

~ 0 1
~ 0 i
~ 1 0
Sx =
Sy =
Sz =
2 1 0
2 i 0
2 0 1
These satisfy the usual commutation relations from which we derived the properties of
angular momentum operators.
It is common to define the Pauli Matrices, i , which have the following properties.
Si
~
S

x =

0 1
1 0

~
i .
2
~
=
~
2

0 i
=
i 0

[i , j ]

2iijk k

x y + y x = x z + z x

z y + y z = 0

{i , j }

2ij

z =

1
0

0
1

The last two lines state that the Pauli matrices anti-commute. The matrices are the
Hermitian, Traceless matrices of dimension 2. Any 2 by 2 matrix can be written
as a linear combination of the matrices and the identity.

1.26

Quantum Mechanics in an Electromagnetic Field

The classical Hamiltonian for a particle in an Electromagnetic field is

1
e ~ 2
H=
e
p~ + A
2m
c
where e is defined to be a positive number. This Hamiltonian gives the correct Lorentz
force law. Note that the momentum operator will now include momentum in the field,
not just the particles momentum. As this Hamiltonian is written, p~ is the variable
~
conjugate to ~r and is related to the velocity by p~ = m~v ec A.
39

1. Course Summary

TOC

In Quantum Mechanics, the momentum operator is replaced in the same way to include
the effects of magnetic fields and eventually radiation.
e~
p~ p~ + A
c
Starting from the above Hamiltonian, we derive the Hamiltonian for a particle in
a constant magnetic field.

e ~ ~
e2 2 2
~2 2
~ 2 = (E + e)
+
B L +
r B (~r B)
2
2m
2mc
8mc
This has the familiar effect of a magnetic moment parallel to the angular momentum
vector, plus some additional terms which are very small for atoms in fields realizable
in the laboratory.
So, for atoms, the dominant additional term is
HB =

e ~ ~
~
B L = ~
B,
2mc

e ~
where
~ = 2mc
L. This is, effectively, the magnetic moment due to the electrons
orbital angular momentum.

The other terms can be important if a state is spread over a region much larger than
an atom. We work the example of a plasma in a constant magnetic field. A
charged particle in the plasma has the following energy spectrum

1
~2 k 2
eB~
n+
+
.
En =
me c
2
2me
which depends on 2 quantum numbers. ~k is the conserved momentum along the field
direction which can take on any value. n is an integer dealing with the state in x and
y. This problem can be simplified using a few different symmetry operators. We work
it two different ways: in one it reduces to the radial equation for the Hydrogen atom;
in the other it reduces to the Harmonic Oscillator equation, showing that these two
problems we can solve are somehow equivalent.

1.27

Local Phase Symmetry in Quantum Mechanics and the

Gauge Symmetry

There is a symmetry in physics which we might call the Local Phase Symmetry in
quantum mechanics. In this symmetry we change the phase of the (electron) wavefunction by a different amount everywhere in spacetime. To compensate for this change,
40

1. Course Summary

TOC

we need to also make a gauge transformation of the electromagnetic potentials. They

all must go together like this.
(~r, t)
~
A

ei ~c f (~r,t) (~r, t)
~ f
~ (~r, t)
A
+

1 f (~r, t)
c t

The local phase symmetry requires that Electromagnetism exist and have a gauge
symmetry so that we can keep the Schr
odinger Equation invariant under this phase
transformation.
We exploit the gauge symmetry in EM to show that, in field free regions, the
function f can be simply equal to a line integral of the vector potential (if we pick the
right gauge).
~r
~
f (~r) = d~r A.
~
r0

We use this to show that the magnetic flux enclosed by a superconductor is quantized.
We also show that magnetic fields can be used to change interference effects in quantum
mechanics. The Aharanov B
ohm Effect brings us back to the two slit diffraction
experiment but adds magnetic fields. The electron beams travel through two slits in
field free regions but we have the ability to vary a magnetic field enclosed by the path
of the electrons. At the screen, the amplitudes from the two slits interfere = 1 + 2 .
Lets start with B = 0 and A = 0 everywhere. When we change the B field, the
wavefunctions must change.
1

1 e

e
i ~c
e
i ~c

~
d~
r A

2 e

e
i ~c

~
d~
r A
i e
2
~c
+ 2 e
1 e
2

The relative phase from the two slits depends on the flux between the slits. By varying
the B field, we will shift the diffraction pattern even though B = 0 along the
whole path of the electrons.

1.28

Addition of Angular Momentum

It is often required to add angular momentum from two (or more) sources together to
get states of definite total angular momentum. For example, in the absence of external
fields, the energy eigenstates of Hydrogen (including all the fine structure effects) are
41

Assume we have already solved and an energy eigenvalue problem and now need to
include an additional term in the Hamiltonian. We can use time independent perturbation theory to calculate corrections to the energy eigenvalues and eigenstates. If the
Schr
odinger equation for the full problem is
(H0 + H1 )n = En n
and we have already solved the eigenvalue problem for H0 , we may use a perturbation
series, to expand both our energy eigenvalues and eigenstates in powers of the small
perturbation.
(2)

(1)

(0)

En = En + En + En + ...
!
P
n = N n +
cnk k
k6=n

cnk =

(1)
cnk

(2)

+ cnk + ...

where the superscript (0), (1), (2) are the zeroth, first, and second order terms in the
series. N is there to keep the wave function normalized but will not play an important
role in our results.
By solving the Schr
odinger equation at each order of the perturbation series, we
compute the corrections to the energies and eigenfunctions.
We just give
the first few terms above.
(1)

En = hn |H1 |n i
(1)

cnk =
(2)

En =

hk |H1 |n i
(0)
(0)
En Ek

P
k6=n

|hk |H1 |n i|2

(0)
(0)
En Ek

A problem arises in the case of degenerate states or nearly degenerate states. The
energy denominator in the last equation above is small and the series does not converge.
To handle this case, we need to rediagonalize the full Hamiltonian in the subspace of
nearly degenerate states.
X
(i)
h(j)
n |H|n ii = En j .
iN

This is just the standard eigenvalue problem for the full Hamiltonian in the subspace
of (nearly) degenerate states.
We will use time independent perturbation theory is used to compute fine structure
and hyperfine corrections to Hydrogen energies, as well as for many other calculations.
Degenerate state perturbation theory will be used for the Stark Effect and for hyperfine
splitting in Hydrogen.
44

1. Course Summary

1.30

TOC

The Fine Structure of Hydrogen

We have solved the problem of a non-relativistic, spinless electron in a coulomb potential exactly. Real Hydrogen atoms have several small corrections to this simple
solution. If we say that electron spin is a relativistic effect, they can all be called relativistic corrections which are off order 2 compared to the Hydrogen energies we have
calculated.
1. The relativistic correction to the electrons kinetic energy.
2. The Spin-Orbit correction.
3. The Darwin Term correction to s states from Dirac equation.
Calculating these fine structure effects separately and summing them we find that we
get a nice cancellation yielding a simple formula.
Enlm =

En(0)

(0) 2
En
4n
+
3

2mc2
j + 21

The correction depends only on the total angular quantum number and does not depend on ` so the states of different total angular momentum split in energy but there is
still a good deal of degeneracy. It makes sense, for a problem with spherical symmetry, that the states of definite total angular momentum are the energy
eigenstates and that the result depend on j.
We also compute the Zeeman effect in which an external magnetic field is applied to
Hydrogen. The external field is very important since it breaks the spherical symmetry
and splits degenerate states allowing us to understand Hydrogen through spectroscopy.
The correction due to a weak magnetic field is found to be

eB
e~B
1

E = n`jmj
(Lz + 2Sz ) n`jmj =
mj 1
2mc
2mc
2` + 1

1
The factor 1 2`+1
is known as the Lande g Factor because the state splits as if
it had this gyromagnetic ratio. We know that it is in fact a combination of the orbital
and spin g factors in a state of definite j. We have assumed that the effect of the field
is small compared to the fine structure corrections. We can write the full energy in a
weak magnetic field.

1
1
2
1
3
Enjmj `s = 2 mc2
+

+ gL B Bmj
2
n2
n3 j + 12
4n
Thus, in a weak field, the the degeneracy is completely broken for the states
njmj `s . All the states can be detected spectroscopically.
45

1. Course Summary

TOC

In the strong field limit we could use states of definite m` and ms and calculate the
effects of the fine structure, H1 + H2 , as a perturbation. In an intermediate strength
field, on the order of 500 Gauss, the combination of the Hydrogen fine structure Hamiltonian and the term to the B field must be diagonalized on the set of states with the
same principal quantum number n.

1.31

Hyperfine Structure

The interaction between the spin of the nucleus and the angular momentum
of the electron causes a further (hyperfine) splitting of atomic states. It is called
hyperfine because it is also order 2 like the fine structure corrections, but it is smaller
me
because of the mass dependence of the spin magnetic moment
by a factor of about m
p
for particles.
The magnetic moment of the nucleus is

~N =

ZegN ~
I
2MN c

where I~ is the nuclear spin vector. Because the nucleus, the proton, and the neutron
have internal structure, the nuclear gyromagnetic ratio is not just 2. For the proton,
it is gp 5.56.
We computed the hyperfine contribution to the Hamiltonian for ` = 0 states.
Hhf =

D e
E
~ ~
~ B
~ = 4 (Z)4 m (mc2 )gN 1 S I
S
3
mc
3
MN
n ~2

~ S,
~ spin-orbit interaction, we will define the total
Now, just as in the case of the L
angular momentum
~ + I.
~
F~ = S
It is in the states of definite f and mf that the hyperfine perturbation will be diagonal.
In essence, we are doing degenerate state perturbation theory. We could diagonalize
the 4 by 4 matrix for the perturbation to solve the problem or we can use what we
know to pick the right states to start with. Again like the spin orbit interaction, the
total angular momentum states will be the right states because we can write
the perturbation in terms of quantum numbers of those states.

~ I~ = 1 F 2 S 2 I 2 = 1 ~2 f (f + 1) 3 3
S
2
2
4 4

2
m
1
3
4
2
E = (Z)
(mc )gN 3 f (f + 1)
.
3
MN
n
2
46

1. Course Summary

TOC

For the hydrogen ground state we are just adding two spin 12 particles so the possible
values are f = 0, 1. The transition between the two states gives rise to EM waves with
= 21 cm.
We will work out the effect of an external B field on the Hydrogen hyperfine
states both in the strong field and in the weak field approximation. We also work the
problem without a field strength approximation. The always applicable intermediate field strength result is that the four states have energies which depend on the
strength of the B field. Two of the energy eigenstates mix in a way that also depends
on B. The four energies are

1.32

A~2
B B
4
s
2
A~2
A~2
2
En00

+ (B B) .
4
2

En00 +

The Helium Atom

The Hamiltonian for Helium has the same terms as Hydrogen but has a large perturbation due to the repulsion between the two electrons.
H=

p2
Ze2
p21
Ze2
e2
+ 2

+
2m 2m
r1
r2
|~r1 ~r2 |

Note that the perturbation due to the repulsion between the two electrons
is about the same size as the the rest of the Hamiltonian so first order perturbation
theory is unlikely to be accurate.
The Helium ground state has two electrons in the 1s level. Since the spatial
state is symmetric, the spin part of the state must be antisymmetric so s = 0 (as
it always is for closed shells). For our zeroth order energy eigenstates, we will use
product states of Hydrogen wavefunctions
u(~r1 , ~r2 ) = n1 `1 m1 (~r1 )n2 `2 m2 (~r2 )
and ignore the perturbation. The energy for two electrons in the (1s) state for Z = 2
is then 42 mc2 = 108.8 eV.
We can estimate the ground state energy in first order perturbation theory, using
the electron repulsion term as a (very large) perturbation. This is not very accurate.
We can improve the estimate of the ground state energy using the variational principle. The main problem with our estimate from perturbation theory is that we are not
accounting for changes in the wave function of the electrons due to screening.
47

1. Course Summary

TOC

We can do this in some reasonable approximation by reducing the charge of the nucleus
in the wavefunction (not in the Hamiltonian). With the parameter Z , we get a better
estimate of the energy.

Calculation
0th Order
1st Order perturbation theory
1st Order Variational
Actual

Energy
-108.8
-74.8
-77.38
-78.975

Zwf n
2
2
27
16

Note that the variational calculation still uses first order perturbation theory. It just
adds a variable parameter to the wavefunction which we use to minimize the energy.
This only works for the ground state and for other special states.
There is only one allowed (1s)2 state and it is the ground state. For excited states, the
spatial states are (usually) different so they can be either symmetric or antisymmetric
(under interchange of the two electrons). It turns out that the antisymmetric state
has the electrons further apart so the repulsion is smaller and the energy is lower.
If the spatial state is antisymmetric, then the spin state is symmetric, s=1. So the
triplet states are generally significantly lower in energy than the corresponding spin
singlet states. This appears to be a strong spin dependent interaction but is
actually just the effect of the repulsion between the electrons having a big
effect depending on the symmetry of the spatial state and hence on the symmetry of
the spin state.
The first exited state has the hydrogenic state content of (1s)(2s) and has s=1. We
calculated the energy of this state.
Well learn later that electromagnetic transitions which change spin are strongly
suppressed causing the spin triplet (orthohelium) and the spin singlet states (parahelium) to have nearly separate decay chains.

1.33

Atomic Physics

The Hamiltonian for an atom with Z electrons and protons has many terms representing
the repulsion between each pair of electrons.

X
Z 2
2
2
X
pi
Ze
e

= E.
+

2m
r
|~
r
r~j |
i
i
i=1
i>j
48

1. Course Summary

TOC

We have seen that the coulomb repulsion between electrons is a very large correction
in Helium and that the three body problem in quantum mechanics is only solved by
approximation.
The physics of closed shells and angular momentum enable us to make sense of even
the most complex atoms. When we have enough electrons to fill a shell, say the 1s or
2p, The resulting electron distribution is spherically symmetric because
`
X

|Y`m (, )| =

m=`

2` + 1
.
4

The closed shell screens the nuclear charge. Because of the screening, the potential
no longer has a pure 1r behavior. Electrons which are far away from the nucleus see less
of the nuclear charge and shift up in energy. We see that the atomic shells fill up in
the order 1s, 2s, 2p, 3s, 3p, 4s, 3d, 4p, 5s, 4d, 5p, 6s, 4f, 5d, 6p. The effect of screening
increasing the energy of higher ` states is clear. Its no wonder that the periodic table
is not completely periodic.
A set of guidelines, known as Hunds rules, help us determine the quantum numbers
for the ground states of atoms. The hydrogenic shells fill up giving well defined j = 0
states for the closed shells. As we add valence electrons we follow Hunds rules to
determine the ground state. We get a great simplification by treating nearly closed
shells as a closed shell plus positively charged, spin 21 holes. For example, if an atom
is two electrons short of a closed shell, we treat it as a closed shell plus two positive
holes.)

1. Couple the valence electrons (or holes) to give maximum total spin.
2. Now choose the state of maximum ` (subject to the Pauli principle. The Pauli
principle rather than the rule, often determines everything here.)
3. If the shell is more than half full, pick the highest total angular momentum state
j = ` + s otherwise pick the lowest j = |` s|.

1. Course Summary

1.34

TOC

Molecules

We can study simple molecules to understand the physical phenomena of molecules in

general. The simplest molecule we can work with is the H+
2 ion. It has two
nuclei (A and B) sharing one electron (1).
H0 =

p2e
e2
e2
e2

+
2m r1A
r1B
RAB

RAB is the distance between the two nuclei. We calculate the ground state energy
using the Hydrogen states as a basis.
The lowest energy wavefunction can be thought of as a (anti)symmetric linear combination of an electron in the ground state near nucleus A and the ground state near
nucleus B

~ = C (R) [A B ]
~r, R
q
where A = a1 3 er1A /a0 is g.s. around nucleus A. A and B are not orthogonal;
0

there is overlap. The symmetric (bonding) state has a large probability for the
electron to be found between nuclei. The antisymmetric (antibonding) state has
a small probability there, and hence, a much larger energy. Remember, this symmetry
is that of the wavefunction of one electron around the two nuclei.
The H2 molecule is also simple and its energy can be computed with the help of
the previous calculation. The space symmetric state will be the ground state.
2
e
e2

h|H|i = 2EH + (RAB )
+
2
RAB
r12
The molecule can vibrate in the potential created when the shared electron binds the
atoms together, giving rise to a harmonic oscillator energy spectrum.
Molecules can rotate like classical rigid bodies subject to the constraint that angular
momentum is quantized in units of ~.
Erot =

1.35

1 L2
`(` + 1)~2
~2
m 2 mc2
m
1
=

E
eV
2
2 I
2I
2M a0
M
2
M
1000

Time Dependent Perturbation Theory

We have used time independent perturbation theory to find the energy shifts of states
and to find the change in energy eigenstates in the presence of a small perturbation.
We now consider the case of a perturbation V that is time dependent. Such
50

1. Course Summary

TOC

a perturbation can cause transitions between energy eigenstates. We will

calculate the rate of those transitions.
We derive an equation for the rate of change of the amplitude to be in the nth
energy eigenstate.
i~

cn (t)
t

Vnk (t)ck (t)ei(En Ek )t/~

Assuming that at t = 0 the quantum system starts out in some initial state i , we
derive the amplitude to be in a final state n .
1
cn (t) =
i~

eini t Vni (t0 )dt0

An important case of a time dependent potential is a pure sinusoidal oscillating

(harmonic) perturbation. We can make up any time dependence from a
linear combination of sine and cosine waves. With some calculation, we derive the
transition rate in a harmonic potential of frequency .
in

2
dPn
2Vni
=
(En Ei + ~)
dt
~

This contains a delta function of energy conservation. The delta function may seem
strange. The transition rate would be zero if energy is not conserved and infinite if
energy is exactly conserved. We can make sense of this if there is a distribution function
of P () of the perturbing potential or if there is a continuum of final states that we
need to integrate over. In either case, the delta function helps us do the integral simply.

1.36

Radiation in Atoms

The interaction of atoms with electromagnetic waves can be computed using time dependent perturbation theory. The atomic problem is solved in the absence of EM waves,
then the vector potential terms in the Hamiltonian can be treated as a perturbation.
H=

e ~ 2
1
p~ + A
+ V (r).
2m
c

~ A
~ = 0, the perturbation is
In a gauge in which
V=

e ~
e2
A p~ +
A2 .
mc
2mc2

1. Course Summary

TOC

For most atomic decays, the A2 term can be neglected since it is much smaller than
~ p~ term. Both the decay of excited atomic states with the emission of radiation
the A
and the excitation of atoms with the absorption of radiation can be calculated.
An arbitrary EM field can be Fourier analyzed to give a sum of components of definite
~ r, t) 2A
~ 0 cos(~k
frequency. Consider the vector potential for one such component, A(~
2
2
~r t). The energy in the field is Energy = 2c2 V |A0 | . If the field is quantized (as
we will later show) with photons of energy E = ~, we may write field strength in
terms of the number of photons N .
~ r, t)
A(~
~ r, t)
A(~

2~c2 N
V

12

~
~
ei(k~rt) + ei(k~rt)

=
=

2 cos(~k ~r t)

The direction of the field is given by the unit polarization vector . The cosine term
has been split into positive and negative exponentials. In time dependent perturbation
theory, the positive exponential corresponds to the absorption of a photon and excitation of the atom and the negative exponential corresponds to the emission of a photon
and decay of the atom to a lower energy state.
Think of the EM field as a harmonic oscillator at each frequency, the negative exponential corresponds to a raising operator for the field and the positive exponential
to a
N
lowering
operator.
In
analogy
to
the
quantum
1D
harmonic
oscillator
we
replace

by N + 1 in the raising operator case.

~ r, t)
A(~

2~c2
V

~
~
N ei(k~rt) + N + 1ei(k~rt)

With this change, which will later be justified with the quantization of the field, there
is a perturbation even with no applied field (N = 0)
VN =0 = VN =0 eit =

1
e ~
e 2~c2 2 i(~k~rt)
e
p~
A p~ =
mc
mc V

which can cause decays of atomic states.

Plugging this N = 0 field into the first order time dependent perturbation equations,
the decay rate for an atomic state can be computed.
in

(2)2 e2
~
|hn |eik~r p~|i i|2 (En Ei + ~)
m2 V

The absolute square of the time integral from perturbation theory yields the delta
function of energy conservation.
52

1. Course Summary

TOC

To get the total decay rate, we must sum over the allowed final states. We can assume
that the atom remains at rest as a very good approximation, but, the final photon
states must be carefully considered. Applying periodic boundary conditions in a cubic
volume V , the integral over final states can be done as indicated below.
kx L = 2nx

dnx =

ky L = 2ny

dny =

kz L = 2nz

L
2 dkx
L
2 dky
L
2 dkz

dnz =
3

d n=

L3
3
(2)3 d k

tot =

V
3
(2)3 d k
3

in d n

With this phase space integral done aided by the delta function, the general formula
for the decay rate is

e2 (Ei En ) X
~
d |hn |eik~r () p~e |i i|2 .
tot =
2
2
3
2~ m c

This decay rate still contains the integral over photon directions and a sum over final
state polarization.
Computation of the atomic matrix element is usually done in the Electric Dipole approximation
~
eik~r 1
which is valid if the wavelength of the photon is much larger than the size of the atom.
With the help of some commutation relations, the decay rate formula becomes
3 X
in
tot =
d |
hn |~r|i i|2 .
2c2

The atomic matrix element of the vector operator ~r is zero unless certain constraints
on the angular momentum of initial and final states are satisfied. The selection rules
for electric dipole (E1) transitions are:
` = 1

m = 0, 1

s = 0.

This is the outcome of the Wigner-Eckart theorem which states that the matrix element
of a vector operator V q , where the integer q runs from -1 to +1, is given by
h0 j 0 m0 |V q |jmi = hj 0 m0 |j1mqih0 j 0 ||V ||ji
Here represents all the (other) quantum numbers of the state, not the angular momentum quantum numbers. In the case of a simple spatial operator like ~r, only the
orbital angular momentum is involved.
53

1. Course Summary

2p1s

TOC

2 5 2 1
3
3
4in
2in

4
(2)(4)
a
=
=
6

0

12
3c2
3
9c2

2 5 2

a0
4 6

3

We derive a simple result for the total decay rate of a state, summed over final photon
polarization and integrated over photon direction.
tot =

3
4in
|~rni |2
2
3c

This can be used to easily compute decay rates for Hydrogen, for example the 2p decay
rate.

5 2
3
2
4in

2p1s =
a0
4 6
2

9c
3
The total decay rate is related to the energy width of an excited state, as might be
expected from the uncertainty principle. The Full Width at Half Maximum (FWHM)
of the energy distribution of a state is ~tot . The distribution in frequency follows a
Breit-Wigner distribution.
Ii () = |i ()|2 =

1
( 0 )2 +

2
4

In addition to the inherent energy width of a state, other effects can influence measured
widths, including collision broadening, Doppler broadening, and atomic recoil.
The quantum theory of EM radiation can be used to understand many phenomena,
including photon angular distributions, photon polarization, LASERs, the Mossbauer
effect, the photoelectric effect, the scattering of light, and x-ray absorption.

1.37

Classical Field Theory

A review of classical field theory is useful to ground our development of relativistic

quantum field theories for photons and electrons. We will work with 4-vectors like the
coordinate vector below
(x1 , x2 , x3 , x4 ) = (x, y, z, ict)
using the i to get a in the time term in a dot product (instead of using a metric
tensor).
A Lorentz scalar Lagrangian density will be derived for each field theory we construct.
From the Lagrangian we can derive a field equation called the Euler-Lagrange equation.

L
L

=0
x (/x )

1. Course Summary

TOC

The Lagrangian for a massive scalar field can be deduced from the requirement that
it be a scalar

1
L=
+ 2 2 +
2 x x
where the last term is the interaction with a source. The Euler-Lagrange equation
gives

2 =
x x
which is the known as the Klein-Gordon equation with a source and is a reasonable
relativistic equation for a scalar field.
Using Fourier transforms, the field from a point source can be computed.
(~x)

Ger
4r

This is a field that falls off much faster than 1r . A massive scalar field falls off
exponentially and the larger the mass, the faster the fall off. This fits the form of
the force between nucleons fairly well although the actual nuclear force needs a much
more detailed study.

1.38

The Classical Electromagnetic Field

For the study of the Maxwell field, it is most convenient to make a small modification
to the system of units that are used. In
Rationalized Heaviside-Lorentz Units
the fields are all reduced by a factor of 4 and the charges are increased by the same
factor. With this change Maxwells equations, as well as the Lagrangians we use, are
simplified. It would have simplified many things if Maxwell had started off with this
set of units.
As is well known from classical electricity and magnetism, the electric and magnetic
field components are actually elements of a rank 2 Lorentz tensor.

0
Bz By iEx
Bz
0
Bx iEy

F =
By Bx
0
iEz
iEx iEy
iEz
0
This field tensor can simply be written in terms of the vector potential, (which is a
Lorentz vector).
A
F

~ i)
(A,
A
A
=

x
x

1. Course Summary

TOC

Note that F is automatically antisymmetric under the interchange of the indices.

With the fields so derived from the vector potential, two of Maxwells equations are
automatically satisfied. The remaining two equations can be written as one 4-vector
equation.
F
j
=
x
c
We now wish to pick a scalar Lagrangian. Since E&M is a well understood theory, the
Lagrangian that is known to give the right equations is also known.
1
1
L = F F + j A
4
c
Note that (apart from the speed of light not being set to 1) the Lagrangian does not
contain needless constants in this set of units. The last term is a source term which
provides the interaction between the EM field and charged particles. In working with
this Lagrangian, we will treat each component of A as an independent field. In this
case, the Euler-Lagrange equation is Maxwells equation as written above.
The free field Hamiltonian density can be computed according to the standard prescription yielding
L
(A /x4 )

(F4 )

1 2
(E + B 2 )
2

A
L
x4

if there are no source terms in the region.

Gauge symmetry may be used to put a condition on the vector potential.
A
= 0.
x
This is called the Lorentz condition. Even with this satisfied, there is still substantial gauge freedom possible. Gauge transformations can be made as shown
below.

x
2 = 0

A A +

1. Course Summary

1.39

TOC

Quantization of the EM Field

The Hamiltonian for the Maxwell field may be used to quantize the field in much the
same way that one dimensional wave mechanics was quantized. The radiation field can
~ while static charges give rise to Ak
be shown to be the transverse part of the field A
and A0 .
We decompose the radiation field into its Fourier components
2

XX
~
~
~ x, t) = 1
() ck, (t)eik~x + ck, (t)eik~x
A(~
V k =1

where () are real unit vectors, and ck, is the coefficient of the wave with wave vector
~k and polarization vector () . Once the wave vector is chosen, the two polarization
vectors must be picked so that (1) , (2) , and ~k form a right handed orthogonal
system.
Plugging the Fourier decomposition into the formula for the Hamiltonian density and
using the transverse nature of the radiation field, we can compute the Hamiltonian
(density integrated over volume).
H

X 2

ck, (t)ck, (t) + ck, (t)ck, (t)
c
k,

This Hamiltonian will be used to quantize the EM field. In calculating the Hamiltonian,
care has been taken not to commute the Fourier coefficients and their conjugates.
The canonical coordinate and momenta may be found
Qk, =

1
(ck, + ck, )
c

i
(ck, ck, )
c
for the harmonic oscillator at each frequency. We assume that a coordinate and its
conjugate momentum have the same commutator as in wave mechanics and that coordinates from different oscillators commute.
Pk, =

[Qk, , Pk0 ,0 ]

= i~kk0 0

[Qk, , Qk0 ,0 ]

[Pk, , Pk0 ,0 ]

1. Course Summary

TOC

As was done for the 1D harmonic oscillator, we write the Hamiltonian in terms of
raising and lowering operators that have the same commutation relations as in the 1D
harmonic oscillator.
1
(Qk, + iPk, )
ak, =
2~
1
ak, =
(Qk, iPk, )
2~

1
H =
ak, ak, +
~
2
h
i
ak, , ak0 ,0
= kk0 0
This means everything we know about the raising and lowering operators applies here.
Energies are in steps of ~ and there must be a ground state. The states can be labeled
by a quantum number nk, .

1
1
H =
ak, ak, +
~ = Nk, +
~
2
2
= ak, ak,

Nk,

The Fourier coefficients can now be written in terms of the raising and lowering operators for the field.
r
~c2
ak,
ck, =
2
r
~c2
a
ck, =
2 k,
r

1 X ~c2 ()
~
~
A =
ak, (t)eik~x + ak, (t)eik~x
2
V
k

=
=

h
i
1X
~ ak, ak, + ak, ak,
2
k,

X
1
~ Nk, +
2
k,

States of the field are given by the occupation number of each possible photon state.
|nk1 ,1 , nk2 ,2 , ..., nki ,i , ...i =

Y (ak , )nki ,i
i
i
p
|0i
n
ki ,i !
i

1. Course Summary

TOC

Any state can be constructed by operating with creation operators on the vacuum
state. Any state with multiple photons will automatically be symmetric under the
interchange of pairs of photons because the operators commute.
ak, ak0 ,0 |0i = ak0 ,0 ak, |0i
This is essentially the same result as our earlier guess to put an n + 1 in the emission
operator.
We can now write the quantized radiation field in terms of the operators at t = 0.
r

1 X ~c2 ()
ak, (0)eik x + ak, (0)eik x
A =
2
V k
~

Beyond the Electric Dipole approximation, the next term in the expansion of eik~x
is i~k ~x. This term gets split according to its rotation and Lorentz transformation
properties into the Electric Quadrupole term and the Magnetic Dipole term. The
interaction of the electron spin with the magnetic field is of the same order and
should be included together with the E2 and M1 terms.
e~ ~
(k () ) ~
2mc
The Electric Quadrupole (E2) term does not change parity and gives us the selection
rule.
|`n `i | 2 `n + `i
The Magnetic Dipole term (M1) does not change parity but may change the spin. Since
it is an (axial) vector operator, it changes angular momentum by 0, +1, or -1 unit.
The quantized field is very helpful in the derivation of Plancks black body radiation
formula that started the quantum revolution. By balancing the reaction rates proportional to N and N + 1 for absorption and emission in equilibrium the energy density
in the radiation field inside a cavity is easily derived.
U () = U ()

1.40

8
h 3
d
= 3 ~/kT
d
c e
1

Scattering of Photons

The quantized photon field can be used to compute the cross section for photon scattering. The electric dipole approximation is used to simplify the atomic matrix element
at low energy where the wavelength is long compared to atomic size.
59

1. Course Summary

TOC

To scatter a photon the field must act twice, once to annihilate the initial state photon
and once to create the final state photon. Since the quantized field contains both
creation and annihilation operators,
r

1 X ~c2 ()
ak, (0)eik x + ak, (0)eik x
A (x) =
2
V k
~ p~ term in second order can contribute to
either the A2 term in first order, or the A
scattering. Both of these amplitudes are of order e2 .
The matrix element of the A2 term to go from a photon of wave vector ~k and an atomic
state i to a scattered photon of wave vector k~0 and an atomic state n is particularly
simple since it contains no atomic coordinates or momenta.
0
e2
~ A|i;
~ ~k
hn; ~k 0 ( ) |A
() i =
2mc2

e2 1 ~c2 () (0 ) i(0 )t

e
ni
2mc2 V 0

The second order terms can change atomic states because of the p~ operator.
The cross section for photon scattering is then given by the

2 0
2
0
0
X hn|

d
e2

1

p
~
|jihj|

p
~
|ii
hn|

p
~
|jihj|

p
~
|ii

=
ni 0
+

2
0
d
4mc

m~ j
ji
ji +

Kramers-Heisenberg Formula. The three terms come from the three Feynman
diagrams that contribute to the scattering to order e2 .
This result can be specialized for the case of elastic scattering, with the help of some
commutators.

2
2

delas
e2
m 2 X
hi|
0 ~x|ji hj|
~x|ii hi|
~x|ji hj|
0 ~x|ii
=
ji

d
4mc2
~
ji
ji +
j

Lord Rayleigh calculated low energy elastic scattering of light from atoms using
classical electromagnetism. If the energy of the scattered photon is less than the energy
needed to excite the atom, then the cross section is proportional to 4 , so that blue
light scatters more than red light does in the colorless gasses in our atmosphere.
If the energy of the scattered photon is much bigger than the binding energy of the
atom, >> 1 eV. then the cross section approaches that for scattering from a free
electron, Thomson Scattering.

2
e2
d
2
=
|
0 |
d
4mc2
The scattering is roughly energy independent and the only angular dependence is on
polarization. Scattered light can be polarized even if incident light is not.
60

1. Course Summary

1.41

TOC

Electron Self Energy

Even in classical electromagnetism, if one can calculates the energy needed to assemble
an electron, the result is infinite, yet electrons exist. The quantum self energy correction
is also infinite although it can be rendered finite if we accept the fact that out theories
are not valid up to infinite energies.
The quantum self energy correction has important, measurable effects. It causes observable energy shifts in Hydrogen and it helps us solve the problem of infinities due
to energy denominators from intermediate states.
The coupled differential equations from first order perturbation theory for the state
under study n and intermediate states j may be solved for the self energy correction.
En

~
k,

|Hnj |2

1 ei(nj )t
~(nj )

The result is, in general, complex. The imaginary part of the self energy correction is
directly related to the width of the state.
2
=(En ) = n
~
The time dependence of the wavefunction for the state n is modified by
the self energy correction.
n (~x, t) = n (~x)ei(En +<(En ))t/~ e

n t
2

This gives us the exponential decay behavior that we expect, keeping resonant
scattering cross sections from going to infinity.
The real part of the correction should be studied to understand relative energy shifts
of states. It is the difference between the bound electrons self energy and
that for a free electron in which we are interested. The self energy correction for a
free particle can be computed.
Ef ree

2Ecutof f 2
p
3m2 c2

We automatically account for this correction by a change in the observed mass of the
electron. For the non-relativistic definition of the energy of a free electron, an increase
in the mass decreases the energy.
mobs

(1 +

4Ecutof f
)mbare
3mc2
61

1. Course Summary

TOC

If we cut off the integral at me c2 , the correction to the mass is only about
0.3%,
Since the observed mass of the electron already accounts for most of the self energy
correction for a bound state, we must correct for this effect to avoid double counting
of the correction. The self energy correction for a bound state then is.
2Ecutof f
En(obs) = En +
hn|p2 |ni
3m2 c2

In 1947, Willis E. Lamb and R. C. Retherford used microwave techniques to determine

the splitting between the 2S 21 and 2P 12 states in Hydrogen. The result can
be well accounted for by the self energy correction, at least when relativistic quantum
mechanics is used. Our non-relativistic calculation gives a qualitative explanation of
the effect.

45
mc2
11 1

En(obs) =
log
+
mc2
3n3
2~
nj
24 5

1.42

The Dirac Equation

Our goal is to find the analog of the Schr

odinger equation for relativistic spin onehalf particles, however, we should note that even in the Schrodinger equation, the
interaction of the field with spin was rather ad hoc. There was no explanation of the
gyromagnetic ratio of 2. One can incorporate spin into the non-relativistic equation by
using the Schr
odinger-Pauli Hamiltonian which contains the dot product of the
Pauli matrices with the momentum operator.
2
e~
1
~ [~
p + A(~
r, t)] e(~r, t)
H=
2m
c
A little computation shows that this gives the correct interaction with spin.
1
e~
e~
~ r, t)
H=
[~
p + A(~
r, t)]2 e(~r, t) +
~ B(~
2m
c
2mc
This Hamiltonian acts on a two component spinor.
We can extend this concept to use the relativistic energy equation. The idea
is to replace p~ with ~ p~ in the relativistic energy equation.
2
E
p2 = (mc)2
c

E
E
~ p~
+ ~ p~ = (mc)2
c
c

~
~
i~
+ i~~
i~
i~~ = (mc)2
x0
x0
62

1. Course Summary

TOC

Instead of an equation which is second order in the time derivative, we can make a
first order equation, like the Schr
odinger equation, by extending this equation to four
components.
(L)
(R)

1
~ (L)
i~
i~~
=
mc
x0

Now rewriting in terms of A = (R) + (L) and B = (R) (L) and ordering it as
a matrix equation, we get an equation that can be written as a dot product between
4-vectors.
!

~
~
i~ x
i~~

0
0
i~
x4
0
= ~
+

~
~
0
x
i~~
i~ x
i~
0
4
0

0 ii
1 0
= ~
+
= ~
ii
0
0 1 x4
xi
x
Define the 4 by 4 matrices are by.
i

0
ii

1
0

ii
0

0
1

With this definition, the relativistic equation can be simplified a great deal

mc

=0
+

x
~
where the

0
0
1 =
0
i
0
0
3 =
i
0
relations.

gamma matrices are given by

0 0 i
0 0 0 1

0 i 0
2 = 0 0 1 0

i 0
0
0 1 0 0
0 0
0
1 0 0 0 and they satisfy anti-commutation
0 i 0
1 0 0
0
0 1 0
0
0 i
0
4 =

0
0 0
0 0 1 0
i 0 0
0 0 0 1
{ , } = 2

In fact any set of matrices that satisfy the anti-commutation relations would yield
equivalent physics results, however, we will work in the above explicit representation
of the gamma matrices.
63

1. Course Summary

Defining = 4 ,

TOC

j = ic

satisfies the equation of a conserved 4-vector current

j = 0
x
and also transforms like a 4-vector. The fourth component of the vector shows that the
probability density is . This indicates that the normalization of the state includes
all four components of the Dirac spinors.
For non-relativistic electrons, the first two components of the Dirac spinor are large
while the last two are small.

A
=
B

e ~
pc
c
~ p~ + A
A
A
B
2
2mc
c
2mc2
We use this fact to write an approximate two-component equation derived from the
Dirac equation in the non-relativistic limit.
!
~ S
~
Ze2
p4
Ze2 L
Ze2 ~2 3
p2

+
+
(~r) = E (N R)
2m
4r
8m3 c2
8m2 c2 r3
8m2 c2
This Schr
odinger equation, derived from the Dirac equation, agrees well
with the one we used to understand the fine structure of Hydrogen. The first two terms
are the kinetic and potential energy terms for the unperturbed Hydrogen Hamiltonian.
The third term is the relativistic correction to the kinetic energy. The fourth
term is the correct spin-orbit interaction, including the Thomas Precession
effect that we did not take the time to understand when we did the NR fine structure.
The fifth term is the so called Darwin term which we said would come from the
Dirac equation; and now it has.
For a free particle, each component of the Dirac spinor satisfies the Klein-Gordon
equation.
p~ = up~ ei(~p~xEt)/~
This is consistent with the relativistic energy relation.

1. Course Summary

TOC

The four normalized solutions for a Dirac particle at rest are.

1
0
eimc2 t/~
(1) = E=+mc2 ,+~/2 =

0
V
0

0

1
1
eimc2 t/~
(2) = E=+mc2 ,~/2 =

0
V
0

0

1
0
e+imc2 t/~
(3) = E=mc2 ,+~/2 =

1
V
0

0

1
0
e+imc2 t/~
(4) = E=mc2 ,~/2 =

0
V
1
The first and third have spin up while the second and fourth have spin down.
The first and second are positive energy solutions while the third and fourth are
negative energy solutions, which we still need to understand.
The next step is to find the solutions with definite momentum. The four plane wave
solutions to the Dirac equation are
s
mc2 (r) i(~p~xEt)/~
(r)
p~
u e
|E|V p~
where the four spinors are given by.

1
r
2
0
E
+
mc
(1)
pz c
up~ =

2
2mc
E+mc2

(px +ipy )c
E+mc2

pz c
r
E+mc2
(px +ipy )c
E + mc2
E+mc2

2mc2

(2)

up~ =

1
0

E + mc2
(px ipy )c

2
2mc
E+mc2
pz c
E+mc2

(px ipy )c

(3)
up~

0
1

r
(4)
up~

E + mc2

2mc2

E+mc2
pz c
E+mc2

0
1

E is positive for solutions 1 and 2 and negative for solutions 3 and 4. The spinors are
orthogonal
|E|
(r) (r 0 )
up~ up~ =
rr0
mc2
65

1. Course Summary

TOC

and the normalization constants have been set so that the states are properly normalized and the spinors follow the convention given above, with the normalization
proportional to energy.
The solutions are not in general eigenstates of any component of spin but are eigenstates
of helicity, the component of spin along the direction of the momentum.
Note that with E negative, the exponential ei(~p~xEt)/~ has the phase velocity, the
group velocity and the probability flux all in the opposite direction of the momentum
as we have defined it. This clearly doesnt make sense. Solutions 3 and 4 need to be
understood in a way for which the non-relativistic operators have not prepared us. Let
us simply relabel solutions 3 and 4 such that
p~ ~
p
E E

so that all the energies are positive and the momenta point in the direction of the
velocities. This means we change the signs in solutions 3 and 4 as follows.

1
r

0
E + mc2
(1)
pz c ei(~p~xEt)/~
p~
=

2
2EV
E+mc
(px +ipy )c
E+mc2

r
(2)

r
(3)
p~

0
1

E + mc2
(px ipy )c ei(~p~xEt)/~

2EV
E+mc2

pz c
E+mc2
pz c
E+mc2
(px +ipy )c
mc2
E+mc2 ei(~p~xEt)/~

E+
2EV

1
0

(px ipy )c
s
(4)
p~

|E| + mc2

2|E|V

E+mc2
pz c
E+mc2

0
1

i(~p~xEt)/~
e

We have plane waves of the form

eip x /~
with the plus sign for solutions 1 and 2 and the minus sign for solutions 3 and 4. These
sign in the exponential is not very surprising from the point of view of possible
solutions to a differential equation. The problem now is that for solutions 3 and 4 the
momentum and energy operators must have a minus sign added to them and the phase
66

1. Course Summary

TOC

of the wave function at a fixed position behaves in the opposite way as a function of
time than what we expect and from solutions 1 and 2. It is as if solutions 3 and 4 are
moving backward in time.
If we change the charge on the electron from e to +e and change the sign of the
exponent, the Dirac equation remains the invariant. Thus, we can turn the negative
exponent solution (going backward in time) into the conventional positive exponent
solution if we change the charge to +e. We can interpret solutions 3 and 4 as positrons.
We will make this switch more carefully when we study the charge conjugation operator.
The Dirac equation should be invariant under Lorentz boosts and under rotations, both
of which are just changes in the definition of an inertial coordinate system. Under
Lorentz boosts, x transforms like a 4-vector but the matrices are constant. The
Dirac equation is shown to be invariant under boosts along the xi direction if we
transform the Dirac spinor according to
0

Sboost

cosh + ii 4 sinh
2
2

with tanh = .
The Dirac equation is invariant under rotations about the k axis if we transform
the Dirac spinor according to
0
Srot

= Srot

= cos + i j sin
2
2

with ijk is a cyclic permutation.

Another symmetry related to the choice of coordinate system is parity. Under a parity
inversion operation the Dirac equation remains invariant if
0 = SP = 4

1 0 0
0
0 1 0
0

Since 4 =
0 0 1 0 , the third and fourth components of the spinor change
0 0 0 1
sign while the first two dont. Since we could have chosen 4 , all we know is that
components 3 and 4 have the opposite parity of components 1 and 2.
From 4 by 4 matrices, we may derive 16 independent components of covariant objects.
We define the product of all gamma matrices.
5 = 1 2 3 4
67

1. Course Summary

TOC

which obviously anticommutes with all the gamma matrices.

{ , 5 } = 0
For rotations and boosts, 5 commutes with S since it commutes with the pair of
gamma matrices. For a parity inversion, it anticommutes with SP = 4 .
The simplest set of covariants we can make from Dirac spinors and matrices are
tabulated below.
Classification
Scalar
Pseudoscalar
Vector
Axial Vector
Rank 2 antisymmetric tensor
Total

Covariant Form

no. of Components

1
1
4
4
6
16

Products of more matrices turn out to repeat the same quantities because the square
of any matrix is 1.
For many purposes, it is useful to write the Dirac equation in the traditional form
H = E. To do this, we must separate the space and time derivatives, making the
equation less covariant looking.

+
=0
x
~

ic4 j pj + mc2 4 = ~
t
Thus we can identify the operator below as the Hamiltonian.
H = ic4 j pj + mc2 4
The Hamiltonian helps us identify constants of the motion. If an operator commutes
with H, it represents a conserved quantity.
Its easy to see the pk commutes with the Hamiltonian for a free particle so that momentum will be conserved. The components of orbital angular momentum do not
commute with H.
[H, Lz ] = ic4 [j pj , xpy ypx ] = ~c4 (1 py 2 px )
The components of spin also do not commute with H.
[H, Sz ] = ~c4 [2 px 1 py ]

1. Course Summary

TOC

But, from the above, the components of total angular momentum do commute
with H.
[H, Jz ] = [H, Lz ] + [H, Sz ] = ~c4 (1 py 2 px ) + ~c4 [2 px 1 py ] = 0
The Dirac equation naturally conserves total angular momentum but not the
orbital or spin parts of it.
We can also see that the helicity, or spin along the direction of motion does commute.
~ p~] = [H, S]
~ p~ = 0
[H, S
For any calculation, we need to know the interaction term with the Electromagnetic
field. Based on the interaction of field with a current
1
Hint = j A
c
and the current we have found for the Dirac equation, the interaction Hamiltonian is.
Hint = ie4 k Ak
This is simpler than the non-relativistic case, with no A2 term and only one power of
e.
The Dirac equation has some unexpected phenomena which we can derive. Velocity
eigenvalues for electrons are always c along any direction. Thus the only values of
velocity that we could measure are c.
Localized states, expanded in plane waves, contain all four components of the plane
wave solutions. Mixing components 1 and 2 with components 3 and 4 gives rise to
Zitterbewegung, the very rapid oscillation of an electrons velocity and position.
hvk i =

4
XX

|cp~,r |2

p
~ r=1

pk c2
E

2 X
4
XX
mc3 h
(r)
(r 0 )
(r)
(r 0 )
cp~,r0 cp~,r up~
i4 k up~ e2i|E|t/~ cp~,r0 cp~,r up~ i4 k up~ e2i|E|t
|E|
r=1 0
p
~

r =3

The last sum which contains the cross terms between negative and positive energy
represents extremely high frequency oscillations in the expected value of the
velocity, known as Zitterbewegung. The expected value of the position has similar
rapid oscillations.
It is possible to solve the Dirac equation exactly for Hydrogen in a way very similar to
the non-relativistic solution. One difference is that it is clear from the beginning that
69

1. Course Summary

TOC

the total angular momentum is a constant of the motion and is used as a basic quantum
number. There is another conserved quantum number related to the component of spin
~ With these quantum numbers, the radial equation can be
along the direction of J.
solved in a similar way as for the non-relativistic case yielding the energy relation.
E=s
1+

mc2

Z 2 2
2
q
2
nr + (j+ 12 ) Z 2 2

We can identify the standard principle quantum number in this case as n = nr + j + 21 .

This result gives the same answer as our non-relativistic calculation to order 4 but is
also correct to higher order. It is an exact solution to the quantum mechanics
problem posed but does not include the effects of field theory, such as the Lamb
shift and the anomalous magnetic moment of the electron.
A calculation of Thomson scattering shows that even simple low energy photon scattering relies on the negative energy or positron states to get a non-zero answer. If the
calculation is done with the two diagrams in which a photon is absorbed then emitted
by an electron (and vice-versa) the result is zero at low energy because the interaction Hamiltonian connects the first and second plane wave states with the third and
fourth at zero momentum. This is in contradiction to the classical and non-relativistic
calculations as well as measurement. There are additional diagrams if we consider the
possibility that the photon can create and electron positron pair which annihilates with
the initial electron emitting a photon (or with the initial and final photons swapped).
These two terms give the right answer. The calculation of Thomson scattering makes
it clear that we cannot ignore the new negative energy or positron states.
The Dirac equation is invariant under charge conjugation, defined as changing electron
states into the opposite charged positron states with the same momentum and spin
(and changing the sign of external fields). To do this the Dirac spinor is transformed
according to.
0 = 2
Of course a second charge conjugation operation takes the state back to the original

1. Course Summary

TOC

. Applying this to the plane wave solutions gives

s
s
s
mc2 (1) i(~p~xEt)/~
mc2 (4) i(~p~x+Et)/~
mc2 (1) i(~p~x+Et
(1)
p~ =
up~ e
u~p e
v e

|E|V
|E|V
|E|V p~
s
s
s
2
2
mc
mc
mc2 (2) i(~p~x+Et)/
(2)
(3)
(2)
up~ ei(~p~xEt)/~
u~p ei(~p~x+Et)/~
v e
p~ =
|E|V
|E|V
|E|V p~
s
s
mc2 (3) i(~p~x+|E|t)/~
mc2 (2) i(~p~x|E|t)/~
(3)
p~ =
up~ e

u e
|E|V
|E|V ~p
s
s
mc2 (4) i(~p~x+|E|t)/~
mc2 (1) i(~p~x|E|t)/~
(4)
p~ =
up~ e

u e
|E|V
|E|V ~p
(1)

(2)

(1)

which defines new positron spinors vp~ and vp~ that are charge conjugates of up~ and
(2)

up~ .

1.43

The Dirac Equation

To proceed toward a field theory for electrons and quantization of the Dirac field we
wish to find a scalar Lagrangian that yields the Dirac equation. From the study of
is a scalar and that we can form a scalar from
Lorentz covariants we know that
the dot product of two 4-vectors as in the Lagrangian below. The Lagrangian cannot
depend explicitly on the coordinates.

L = c~

mc2
x

(We could also add a tensor term but it is not needed to get the Dirac equation.)
The independent fields are considered to be the 4 components of and the four
The Euler-Lagrange equation using the independent fields is
components of .
simple since there is no derivative of in the Lagrangian.

L
L
=0

x ( /x

)
L
=0

c~
mc2 = 0
x

+
=0
x
~
71

1. Course Summary

TOC

This gives us the Dirac equation indicating that this Lagrangian is the right
one. The Euler-Lagrange equation derived using the fields is the Dirac adjoint
equation,
The Hamiltonian density may be derived from the Lagrangian in the standard way and
the total Hamiltonian computed by integrating over space. Note that the Hamiltonian
density is the same as the Hamiltonian derived from the Dirac equation directly.

H = ~c4 k
+ mc2 4 d3 x
xk
We may expand in plane waves to understand the Hamiltonian as a sum of oscillators.
s
4
XX
mc2
(r)
cp~,r up~ ei(~p~xEt)/~
(~x, t) =
|E|V
r=1
p
~

(~x, t) =

4
XX

p
~ r=1

mc2
(r) i(~
c u
e p~xEt)/~
|E|V p~,r p~

Writing the Hamiltonian in terms of these fields, the formula can be simplified
yielding
H

4
XX

E cp~,r cp~,r .

p
~ r=1

By analogy with electromagnetism, we can replace the Fourier coefficients for the Dirac
plane waves by operators.
H

4
XX

(r) (r)
bp~

E bp~

p
~ r=1

(~x, t)

4
XX

mc2 (r) (r) i(~p~xEt)/~

b up~ e
|E|V p~

mc2 (r) (r) i(~p~xEt)/~

b
up~ e
|E|V p~

p
~ r=1

(~x, t)

4
XX
p
~ r=1

(r)

The creation an annihilation operators bp~

relations.

(r)

and bp~ satisfy anticommutation

1. Course Summary

TOC

(r 0 )

(r)

{bp~ , bp~0

rr0 p~p~0

{bp~ , bp~ }

(r)

bp~

(r)

{bp~

(r)

, bp~

Np~

(r) (r)
bp~

(r)

Np~ is the occupation number operator. The anti-commutation relations constrain the
occupation number to be 1 or 0.
The Dirac field and Hamiltonian can now be rewritten in terms of electron and
positron fields for which the energy is always positive by replacing the operator to
annihilate a negative energy state with an operator to create a positron state with
the right momentum and spin.
(4)

(1)

bp~

(2)

bp~

dp~

(3)

These anti-commute with everything else with the exception that

(s)

(s0 )

{dp~ , dp~0

} = ss0 p~p~0

Now rewrite the fields and Hamiltonian.

r
2

XX
mc2 (s) (s) i(~p~xEt)/~
(s) (s)
bp~ up~ e
+ dp~ vp~ ei(~p~xEt)/~
(~x, t) =
EV
p
~ s=1
r
2

XX
mc2 (s) (s) i(~p~xEt)/~
(s) (s)

(~x, t) =
+ dp~ vp~ ei(~p~xEt)/~
bp~ up~ e
EV
s=1
p
~

2
XX
p
~

s=1

2
XX
p
~

(s) (s)
(s) (s)
E bp~ bp~ dp~ dp~

(s) (s)
(s) (s)
E bp~ bp~ + dp~ dp~ 1

s=1

All the energies of these states are positive.

1. Course Summary

TOC

commuting, there would have been no lowest energy ground state so this Energy subtraction would not have been possible. Fermi-Dirac statistics are required for
particles satisfying the Dirac equation.
Since the operators creating fermion states anti-commute, fermion states must
be antisymmetric under interchange. Assume br and br are the creation and annihilation operators for fermions and that they anti-commute.
{br , br0 } = 0
The states are then antisymmetric under interchange of pairs of fermions.
br br0 |0i = br0 br |0i
Its not hard to show that the occupation number for fermion states is either
zero or one.
Note that the spinors satisfy the following slightly different equations.
(s)

(i p + mc)up~ = 0
(s)

(i p + mc)vp~ = 0

2. The Problems with Classical Physics

TOC

The Problems with Classical Physics

By the late nineteenth century the laws of physics were based on Mechanics and the
law of Gravitation from Newton, Maxwells equations describing Electricity and Magnetism, and on Statistical Mechanics describing the state of large collection of matter.
These laws of physics described nature very well under most conditions, however, some
measurements of the late 19th and early 20th century could not be understood. The
problems with classical physics led to the development of Quantum Mechanics and
Special Relativity.
Some of the problems leading to the development of Quantum Mechanics are listed
here.
Black Body Radiation: Classical physics predicted that hot objects would instantly radiate away all their heat into electromagnetic waves. The calculation,
which was based on Maxwells equations and Statistical Mechanics, showed that
the radiation rate went to infinity as the EM wavelength went to zero, The
Ultraviolet Catastrophe. Planck solved the problem by postulating that EM
energy was emitted in quanta with:
E = h
The Photoelectric Effect: When light was used to knock electrons out of solids,
the results were completely different than expected from Maxwells equations.
The measurements were easy to explain (for Einstein) if light is made up of
particles with the energies Planck postulated.
Te = h W
Atoms: After Rutherford found that the positive charge in atoms was concentrated in a very tiny nucleus, classical physics predicted that the atomic electrons
orbiting the nucleus would radiate their energy away and spiral into the nucleus.
This clearly did not happen. The energy radiated by atoms also came out in
quantized amounts in contradiction to the predictions of classical physics. The
Bohr Atom postulated an angular momentum quantization rule, L = n~ for
n = 1, 2, 3..., that gave the right result for hydrogen, but turned out to be wrong
since the ground state of hydrogen has zero angular momentum.
En =

13.6
Z 2 2 c2
= 2 eV
2n2
n

It took a full understanding of Quantum Mechanics to explain the detailed Hydrogen energy spectrum. The full atomic energy spectrum was computable based
on the understanding of Hydrogen.
75

2. The Problems with Classical Physics

TOC

Compton Scattering: When light was scattered off electrons, it behaved just like a
particle but changes wave length in the scattering; more evidence for the particle
nature of light and Plancks postulate.
Waves and Particles: In diffraction experiments, light was shown to behave like
a wave while in experiments like the Photoelectric effect, light behaved like a
particle. More difficult diffraction experiments showed that electrons (as well as
the other particles) also behaved like a wave, yet we can only detect an integer
number of electrons (or photons). In some (thought) experiments on diffraction,
we can conclude what the role of the wavefunction is.
Quantum Mechanics incorporates a wave-particle duality and explains all of the
above phenomena. In doing so, Quantum Mechanics changes our understanding of
nature in fundamental ways. While the classical laws of physics are deterministic, QM
is probabilistic. We can only predict the probability that a particle will be found in
some region of space.
There was no theoritical motivation to move in the direction of Quantum Physics. At
every step, we were driven by experiment, often with the great theorists protesting. The
formulation of the Schr
odinger equation, in particular, was driven by experiment.
The Schr
odinger equation is consistent with measurement and with the deductions of
Planck and de Broglie. Physicists were also driven to the theory of the collapse of
the wave function when measurements are made. Finally with the step to the Dirac
equation for electrons, theory found a way to incorporate electron spin naturally
and predicted the existence of antiparticles.
Electromagnetic waves like light are made up of particles we call photons. Einstein,
based on Plancks formula, hypothesized that the particles of light had energy proportional to their frequency.
E = h
The new idea of Quantum Mechanics is that every particles probability (as a function
of position and time) is equal to the square of a probability amplitude function and
that these probability amplitudes obey a wave equation. This is much like the case in
electromagnetism where the energy density goes like the square of the field and hence
the photon probability density goes like the square of the field, yet the field is made up
of waves. So probability amplitudes are like the fields we know from electromagnetism
in many ways.
De Broglie assumed E = h for photons and other particles and used Lorentz
invariance (from special relativity) to derive the wavelength for particles like electrons.
=

h
p
76

2. The Problems with Classical Physics

TOC

The rest of wave mechanics was built around these ideas, giving a complete picture
that could explain the above measurements and could be tested to very high accuracy,
particularly in the hydrogen atom. We will spend several chapters exploring these
ideas.
Example: Assume the photon is a particle with the standard de Broglie
wavelength. Use kinematics to derive the wavelength of the scattered photon as a function of angle for Compton Scattering.
Gasiorowicz Chapter 1
Griffiths does not really cover this.

2.1

Black Body Radiation *

A black body is one that absorbs all the EM radiation (light...) that strikes it. To
stay in thermal equilibrium, it must emit radiation at the same rate as it absorbs it so
a black body also radiates well. (Stoves are black.)
Radiation from a hot object is familiar to us. Objects around room temperature radiate
mainly in the infrared as seen the the graph below.

Figure 5: Intensity of black body radiation as a function of wavelength for three different
temperatures, ranging from room T, to solar surface T.
77

2. The Problems with Classical Physics

TOC

If we heat an object up to about 1500 degrees we will begin to see a dull red glow and
we say the object is red hot. If we heat something up to about 5000 degrees, near the
temperature of the suns surface, it radiates well throughout the visible spectrum and
we say it is white hot.
By considering plates in thermal equilibrium it can be shown that the emissive power
over the absorption coefficient must be the same as a function of wavelength, even for
plates of different materials.
E2 (, T )
E1 (, T )
=
A1 ()
A2 ()
It there were differences, there could be a net energy flow from one plate to the other,
violating the equilibrium condition.

Figure 6: Two bodies at the same temperature in equilibrium exchange energy in EM

radiation. To be in equilibrium, the rate of energy flow in must be the same as the rate
of flow out. A Black Body absorbs all incident radiation at any wave lenth (A() 1),
so it must also have the highest possible rate of emission of radiation.
A black body is one that absorbs all radiation incident upon it.
ABB = 1
Thus, the black body Emissive power, E(, T ), is universal and can be derived from
first principles.
A good example of a black body is a cavity with a small hole in it. Any light incident
upon the hole goes into the cavity and is essentially never reflected out since it would
have to undergo a very large number of reflections off walls of the cavity. If we make
the walls absorptive (perhaps by painting them black), the cavity makes a perfect black
body.
78

2. The Problems with Classical Physics

TOC

Figure 7: A small hole in a cavity approaches a perfect Black Body since any energy
entering the hole will not be reflected back out. We can compute the EM energy density
inside the cavity as a function of wavelength using Statistical Mechanics. This is very
simply related to the energy flowing out of the small hole and hence to the emission
spectrum of a Black Body.
There is a simple relation between the energy density in a cavity, u(, T ), and the black
body emissive power of a black body which simply comes from an analysis of how much
radiation, traveling at the speed of light, will flow out of a hole in the cavity in one
second.
c
E(, T ) = u(, T )
4
The only part that takes a little thinking is the 4 in the equation above.
Rayleigh and Jeans calculated the energy density (in EM waves) inside a cavity
and hence the emission spectrum of a black body. Their calculation was based on
simple EM theory and equipartition. It not only did not agree with data; it said that
all energy would be instantly radiated away in high frequency EM radiation. This was
called the ultraviolet catastrophe.
u(, T ) =

2.2

The Photoelectric Effect

The Photoelectric Effect shows that Plancks hypothesis, used to fit the Black Body
data, is actually correct for EM radiation. Einstein went further and proposed, in 1905,
that light was made up of particles with energy related to the frequency of the light,
E = h. (He got his Nobel prize for the Photoelectric effect, not for Special or General
Relativity.) When light strikes a polished (metal) surface electrons are ejected.

Figure 11: Basic Photoelectric proces: photon is absored by electron in a polished metal
plate, giving it enough energy to leave the solid.
82

2. The Problems with Classical Physics

TOC

Solve for velocity.

n~
mr
Plug into force equation to get formula for r.
v=

mn2 ~2 r2
m2 r 2 r
n 2 ~2
mr
1
r

=
=
=

e2
40
e2
40
me2 1
40 ~2 n2

Now we just want to plug v and r into the energy formula. We write the Hydrogen
1 e2
1
potential in terms of the fine structure constant SI = 4
137
.
0 ~c
~c
r
1
mc 1
=
r
~ n2

V (r) =

We now compute the energy levels.

The constant mc
=
2
energy spectrum.

511000
2(137)2

1
mv 2 + V (r)
2

2
1
n~
~c
m

2
mr
r

2
1
n~c
2 ~mc2
m

2
n2 ~
n2 ~

2
1
c 2 mc2
m

2 n
n2
2
2
1 mc
2 mc2
2 mc2

=
2
2
2
n
n
2n2
= 13.6 eV. Bohrs formula gives the right Hydrogen

We can also compute the ground state radius of the Bohr orbit.
1
r

mc2
~c
(137)(197.3)
= 0.053 nm
511000
91

2. The Problems with Classical Physics

TOC

This is also about the right radius.

Bohr Radius
a0 =

mc2
= 0.053 nm
~c

Although angular momentum is quantized in units of ~, the ground state of Hydrogen has zero angular momentum. This would put Bohrs electron in the nucleus.
Bohr fit the data, with some element of truth, but his model is WRONG.

2.5
2.5.1

Derivations and Computations

Black Body Radiation Formulas *

(Not yet available.)

2.5.2

The Fine Structure Constant and the Coulomb Potential

We will now grapple for the first time with the problem of which set of units to use.
Advanced texts typically use CGS units in which the potential energy is
V (r) =

e2
r

while the Standard International units

V (r) =

1 e2
40 r

We can circumvent the problem by defining the dimensionless fine structure constant .
Fine Structure Constant
SI =

1 e2
1

40 ~c
137

CGS =

e2
1

~c
137
92

2. The Problems with Classical Physics

TOC

So in either set of units the Hydrogen potential is

Coulomb Potential in Terms of
V (r) =

2.6

~c
r

Examples

2.6.1

The Solar Temperature *

Estimate the solar temperature using

the solar radiation intensity on the earth of 1.4 kilowatts per square meter.
(rsun = 7 108 m, dsun = 1.5 1011 m)
and the solar spectrum which peaks at about 500 nm.
First we compute the power radiated per unit area on the solar surface.
2
R = (1400 W/m2 )(4d2sun )/(4rsun
) = 6.4 107 W/m2

We compare this to the expectation as a function of temperature.

R(T ) = (5.67 108 W/m2 / K4 ) T 4
and get
T4

6.4 107
5.67 108
5800 K

We can work from the energy density computed by Planck u(, T ), to compute the
Emission as a function of . First we have stated that
c
E(, T ) = u(, T )
4
Well leave proof of this to the homework. Then we need to transform between E(, T )
and E(, T ). This type of transformation of probability functions is useful to understand.
93

2. The Problems with Classical Physics

TOC

Lets assume that E(, T ) peaks at 500 nm as one of the graphs shows. We need to
transform E(, T ). Remember P ()d = P ()d for distribution functions.
h
2 2
c2 eh/kT 1

d 2 2
h
E(, T ) = 2 h/kT
d c e
1
2
2
2
h
h
2 4
=
= 3 h/kT
2
h/kT
c c e
c e
1
1
E(, T )

This peaks when

5
eh/kT 1
is maximum.
5 4
eh/kT 1

5 (h/kT )eh/kT
=0
(eh/kT 1)2

5
eh/kT 1

(h/kT )eh/kT
(eh/kT 1)2

5(eh/kT 1)
= h/kT
eh/kT
5(1 eh/kT ) = h/kT

Lets set y = h/kT and solve the equation

y = 5(1 ey )
I solved this iteratively starting at y=5 and got y = 4.97 implying
h = 4.97kT
If we take = 500 nm, them = 6 1014 .
T =

h
(6.6 1034 )(6 1014 )
=
= 6 103 = 5700
5k
(5)(1.4 1023 )

Thats agrees well.

2.6.2

Black Body Radiation from the Early Universe *

Find the frequency at which the the Emissive E(, T ) is a maximum for the 2.7 degree
cosmic background radiation. Find the wavelength for which E(, T ) is a maximum.
94

2. The Problems with Classical Physics

TOC

The cosmic background radiation was produced when the universe was much hotter
than it is now. Most of the atoms in the universe were ionized and photons interacted
often with the ions or free electrons. As the universe cooled so that neutral atoms
formed, the photons decoupled from matter and just propagated through space. We see
these photons today as the background radiation. Because the universe is expanding,
the radiation has been red shifted down to a much lower temperature. We observe
about 2.7 degrees. The background radiation is very uniform but we are beginning to
observe non-uniformities at the 105 level.
From the previous problem, we can say that the peak occurs when
h

5kT

5kT /h

ch/(5kT ) =

(3 108 )(6.6 1034 )

= 1mm
(5)(1.4 1023 )(2.7)

Similarly the peak in occurs when

= 2.8kT /h =

2.6.3

(1.4 1023 )(2.7)

= 6 1010 Hz
(6.6 1034 )

Compton Scattering *

Compton scattered high energy photons from (essentially) free electrons

in 1923. He measured the wavelength of the scattered photons as a function of the
scattering angle. The figure below shows both the initial state (a) and the final state,
with the photon scattered by an angle and the electron recoiling at an angle . The
photons were from nuclear decay and so they were of high enough energy that it didnt
matter that the electrons were actually bound in atoms. We wish to derive the formula
for the wavelength of the scattered photon as a function of angle. We solve
the problem using only conservation of energy and momentum. Lets work in units
in which c = 1 for now. Well put the c back in at the end. Assume the photon is
initially moving in the z direction with energy E and that it scatters in the yz plane
so that px = 0.

2. The Problems with Classical Physics

TOC

Figure 20: Diagram of Compton scattering, the scattering of photons by (essentially)

free electrons.
Conservation of momentum gives
E = E 0 cos + pe cos
and
E 0 sin = pe sin .
Conservation of energy gives
E + m = E0 +

p2e + m2

Our goal is to solve for E 0 in terms of cos so lets make sure we eliminate the .
Continuing from the energy equation
p
E E 0 + m = p2e + m2
squaring and calculating p2e from the components
E 2 + E 02 + m2 2EE 0 + 2mE 2mE 0 = (E E 0 cos )2 + (E 0 sin )2 + m2
and writing out the squares on the right side
E 2 + E 02 + m2 2EE 0 + 2mE 2mE 0 = E 2 + E 02 2EE 0 cos + m2
and removing things that appear on both sides
2EE 0 + 2mE 2mE 0 = 2EE 0 cos
96

2. The Problems with Classical Physics

TOC

and grouping
m(E E 0 ) = EE 0 (1 cos )
(E E 0 )
(1 cos )
=
EE 0
m
1
(1 cos )
1

=
E0
E
m
Since = h/p = h/E in our fine units,
h
(1 cos ).
m
We now apply the speed of light to make the units come out to be a length.
0 =

Compton Scattering
0 =
C =

hc
(1 cos )
mc2

hc
= 2.43 1012 m
mc2

These calculations can be fairly frustrating if you dont decide which variables you
want to keep and which you need to eliminate from your equations. In this case we
eliminated by using the energy equation and computing p2e .

2.6.4

Rutherfords Nuclear Size *

If the positive charge in gold atoms were uniformly distributed over a sphere or radius
5 Angstroms, what is the maximum particle kinetic energy for which the can be
scattered right back in the direction from which it came?
To solve this, we need to compute the potential at the center of the charge distribution
relative to the potential at infinity (which we will say is zero). This tells us directly
the kinetic energy in eV needed to plow right through the charge distribution.
1 Ze
where Z is the number of protons
The potential at the surface of the nucleus is 4
0 R
in the atom and R is the nuclear radius. Thats the easy part. Now we need to integrate
our way into the center.

1 Ze
V =

40 R

0
R

1 r3 Ze
dr
40 R3 r2
97

2. The Problems with Classical Physics

The

r3
R3

TOC

gives the fraction of the nuclear charge inside a radius r.

0
1
1 Ze
V =

Zerdr
40 R
40 R3
R

3Ze
1
Ze Ze
V =
+
=
40 R
2R
80 R

V =

1.7 107
(3)(79)(1.6 1019 C)
=
Nm/C
8(8.85 1012 C2 /Nm2 )R
R

The is then the kinetic energy in eV needed for a particle of charge +e to plow right
through the center of a spherical charge distribution. The particle actually has charge
+2e so we need to multiply by 2. For a nuclear radius of 5
Aor 5 1010 meters, we
need about 680 eV to plow through the nucleus. For the actual nuclear radius of about
5 Fermis or 5 1015 meters, we need 68 MeV to plow through.
Lets compare the above SI units numbers to my suggested method of using
the fine structure constant... Putting in the alpha charge of 2e.
U=

3Z~c
(3)(79)(1973)
6Ze2
=
=
eV = 683 eV
80 R
R
(137)(5)

This is easier.

2.7

Homework Problems

1. A polished Aluminum plate is hit by beams of photons of known energy. It is

measured that the maximum electron energy is 2.3 0.1 eV for 200 nm light and
0.90 0.04 eV for 258 nm light. Determine Plancks constant and its error based
on these measurements.
h W = Emax
initial energy = final energy
h(2 1 ) = E2 E1 subtract equations for 2 measurements
E1
solve for h
h = E22
1
= c
h = E12 E11
c

2. A 200 keV photon collides with an electron initially at rest. The photon is
observed to scatter at 90 degrees in the electron rest frame. What are the kinetic
98

2. The Problems with Classical Physics

TOC

energies of the electron and photon after the scattering?

0 0 = mhe c (1 cos )
Compron Scattering
h
12
Compton wavelength (fyi)
C = me c = 2.43 10
E = h
for the photon
= c
= hc
compute initial lambda
E
hc
0 = hc
+
(1

cos
)
plug
2
E
me c
hc
0 = hc
+
(1

0)
90 degrees
E me c 2

0 = hc
0

E =

hc
0

1
E

1
me c 2
E
E
1+ mc
2
0

wavelength of scattered photon

Energy of scattered photon

Te = E E = E
Te =

200
(200) 200+511

E
E
1+ mc
2

E
E E+mc
2

keV

get Te by conservation of Energy

put in numbers

3. Use the energy density in a cavity as a function of frequency and T

u(, T ) =

8h
3
3
h/kT
c e
1

to calculate the emissive power of a black body E(, T ) as a function of wavelength and temperature.
(a) First show that E(, T ) = 4c u(, T ) based on the flow of energy out of a small
hole in a cavity.
(b) Second transform from E(, T ) to E(, T ) using the fact (from transformations of integrals for example) that P ()d = P ()d.
u(, T )d = u(, T )d
transform from energy density in to
d
sign of derivative doesnt matter
u(, T ) = u(, T ) d
= c
relationship between freq. and wavelength
d
c
=

take
derivative
2
d

3
c
u(, T ) = 8h
plug
c3 eh/kT 1 2
u(, T ) =
u(, T ) =

c /
8h
c
c3 ehc/kT 1 2
8hc
5 ehc/kT 1

write in terms of
simplify

4. Rutherford derived the differential cross section for Coulomb scattering from
a point charge
d
=
d

zZe2
160

1
2
E sin4

zZ~c
4E sin2 2

(review calculations in this chapter). To go from the differential cross section (for
scattering from nuclei) to a rate into a detector for a plane wave beam incident
upon a target with N nuclei one computes.
d
Rate= d
(N) (incident flux)()
where the incident flux is in alpha particles per square meter per second and
is the solid angle covered by the detector.
99

2. The Problems with Classical Physics

TOC

For a narrow beam incident upon a (wider) thin foil one computes.
d
(nuclei per m2 ) (incident particles per second)().
Rate= d
Calculate the rate of scatters into a 0.1 steradian detector at = 160 degrees for
104 , 5.5 MeV alpha particles per second in a narrow beam incident upon a gold
foil with 0.2 grams per cm2 . You will need to use the density of gold.
5. Review the derivation (based on relativistic kinematics) of the shift in wavelength
in Compton scattering. If 100 keV x-rays are scattered from free electrons, what
will be the energy of x-rays scattered through an angle of 60 degrees? What will
be the direction of the recoil electron?

2.8

Sample Test Problems

1. What is the maximum wavelength of electromagnetic radiation which can eject

electrons from a metal having a work function of 3 eV? (3 points)
Answer
The minimum photon energy needed to knock out an electron is 3 eV. We just
need to convert that to wavelength.
E

h = 3eV
c

3eV

1240 ev nm
413 nm
3eV

hc
=
3eV

2. * Based on classical electromagnetism and statistical mechanics, Rayleigh computed the energy density inside a cavity. He found that, at a temperature T, the
energy density as a function of frequency was
8 2
kB T.
c3
Why is this related to black body radiation? Why was this in obvious disagreement with observation?
u(, T ) =

3. What is the maximum wavelength of electromagnetic radiation which can eject

electrons from a metal having a work function of 2 eV?
4. * State simply what problem with black-body radiation caused Planck to propose
the relation E = h for light.
5. The work function of a metal is 2 eV. Assume a beam of light of wavelength is
incident upon a polished surface of the metal. Plot the maximum electron energy
(in eV) of electrons ejected from the metal versus in Angstroms. Be sure to
label both axes with some numerical values.
100

3. Diffraction

TOC

Diffraction

Feynman Lectures, Volume III, Chapter I.

Gasciorawicz does not really this.
Rohlf Chapters 5
Griffiths doesnt cover this.

3.1

Diffraction from Two Slits

Water waves will exhibit a diffractive interference pattern in a 2 slit experiment as

diagrammed below. The diagram shows the crests of the water waves at some time.
Downstream from the slits, we will see constructive interference where the waves from
the slits are in phase and destructive interference where they are 180 degrees out of
phase, for example where a crest from one slit meets a trough from the other. The plot
labeled I12 shows the interference pattern at the location of the absorber. Diffraction
is a simple wave phenomenon.

Figure 21: Diffraction of water waves in a ripple tank.

The diffraction of light was a well known phenomenon by the end of the 19th
century and was well explained in classical ElectroMagnetic theory since light was held
to be a EM wave. The diffraction pattern from two narrow slits is particularly easy to
understand in terms of waves. The setup is shown in the diagram below.

101

3. Diffraction

TOC

Figure 22: At the angle shown, the path from the upper slit is longer by d sin for the
screen very far away compared to the distance between the slits.
EM waves of wavelength are emitted from a single light-source, like a laser. They
travel to two narrow slits, (for simplicity) equidistant from the source, and a distance
d apart. Light travels from the slits to a detection screen. A diffraction pattern can be
seen on the detection screen, like the one shown in the picture below.

Figure 23: Diffraction pattern due to interference between the path through slit-1 and
the path through slit-2.
102

3. Diffraction

TOC

The center of the diffraction pattern occurs at the location on the screen equidistant
from each slit where the waves from the two slits are in phase (because they
have traveled exactly the same distance) and the fields add, so the waves interfere
constructively and there is an intensity maximum. Some distance off this center of
the diffraction pattern, there will be destructive interference between waves from the
two slits and the intensity will be zero. This will occur when the distance traveled by
two waves differs by /2, so the waves are 180 degrees out of phase and the
fields from the two slits cancel.
We can compute this location by looking at the above diagram. We assume that the
distance to the screen is much greater than d. For light detected at an angle , the
extra distance traveled from slit 1 is just d sin . So the angle of the first minimum
(or null) can be found from the equation d sin = 2 .
More generally we will get a maximum if the paths from the slits differ by an
integer number of wavelengths d sin = n and we will get a null when the
paths differ by a half integer number wavelengths. d sin null = (n+1)
.
2
Although it is very difficult because electrons are charged, 2 slit electron diffraction
has also been observed.

Figure 24: Distribution from 2-slit electron diffraction.

So, all kinds of particles seem to diffract indicating there is some kind of wave involved.
We will now continue with some thought experiments on diffraction to illustrate the
physics that Quantum Mechanics needed to match.
103

3. Diffraction

TOC

Example: Derive the location of the nodes in the diffraction pattern from
two narrow slits a distance d apart. Now try to compute the intensity distribution.

3.2

Diffraction from Crystals

Electron waves were first demonstrated by measuring diffraction from crystals. Davison and Germer observed diffraction of electrons from a Nickel crystal in
1928. They varied the electron energy to measure the electron wavelength, agreeing
well with the de Broglie expectation.
First we see electron diffraction from a single crystal in the Davison-Germer experiment. Electron diffraction off polycrystalline material gives concentric rings
instead of spots.

Figure 25: Electron diffraction from a single perfect crystal by Davison and Germer(left) and electron diffraction from a polycrystaline sample (right).
X-ray diffraction from crystals is a powerful tool and x-rays easily pentrate a small sample. Neutrons also penetrate samples well and can be used to stucy crystals. Charged
particles easily loose energy and coherence when traversing a sample so results are not
as clear.

104

3. Diffraction

TOC

Figure 26: X-ray diffraction from a single Sodium Crystal which has periodic
locations of the atoms (left) and neutron diffraction from a single Sodium crystal
(right).

3.3

The de Broglie Wavelength

The Lorentz transformation had been postulated for ElectroMagnetic waves before
Einstein developed Special Relativity. The EM waves were entirely consistent with
Relativity. For example, the phase of an EM wave at some point is the same as at the
Lorentz transformed point.
De Broglie applied this Lorentz invariance requirement on the phase of
matter waves to determine what the wavelength must be. Its easy for us to derive
the wavelength using 4-vectors. Position and time form one 4-vector.
x = (ct, ~x)
Energy and momentum form another.

p =

E
, p~
c

Recall that Lorentz vectors must be transformed but Lorentz scalars are automatically invariant under transformations. For example the scalar formed by dotting the
4-momentum into itself is
p p c2 = E 2 + p2 c2 = m2 c4 .
105

3. Diffraction

TOC

The mass of a particle is clearly Lorentz invariant.

To compute the wavelength for our matter waves, lets use the scalar
p x = Et + p~ ~x
Its already easy to see that the phase in a wave like
ei(~p~xEt)/~
is Lorentz invariant.
To read off the wavelength, lets pick off the part of the expression that corresponds to
p
e2i(x/t) . We see that 2
= ~ and therefore the de Broglie wavelength is.
de Broglie Wavelength
=

2~
h
=
p
p

De Broglie derived this result in 1923.

We can also read off the frequency .
E
= E/h
2~
This was in some sense the input to our calculation.
=

The de Broglie wavelength will be our primary physics input for the development
of Quantum Mechanics. Its not that this work was the most significant, but, this
wavelength summarizes most of what happened before 1923.

3.3.1

Computing de Broglie Wavelengths

We usually quote the energy of a particle in terms of its kinetic energy in electron
Volts, eV (or Million electron Volts, MeV). The reason for this is that particles are
usually accelerated to some energy by an electric field. If I let an electron (or proton...)
be accelerated through a 100 Volt potential difference, it will have a kinetic energy of
100eV.
The whole problem of computing a de Broglie wavelength is to convert from kinetic
energy to momentum. If you always want to be correct without any need for thinking,
use the relativistically correct formula for the kinetic energy
p
T = (mc2 )2 + p2 c2 mc2
106

3. Diffraction

TOC

and solve it for pc,

pc =

p
(T + mc2 )2 (mc2 )2

then use this handy formula to get the answer.

h
2~c
=
p
pc

I remember that ~c = 197.3 eV nm allowing me to keep the whole calculation in eV. I

also know the masses of the particles.
me c2 = 0.51 MeV
mp c2 = 938.3 MeV
(If T <<< mc2 , make sure the precision of your calculator sufficient or use the
non-relativistic method below.)
If you know that the particle is super-relativistic, so that T >> mc2 , then just use
pc = T and life is easy.
If you know that the particle is highly non-relativistic, T << mc2 , then you can use

(pc)2
p2
= 2mc
2mc2 T .
T = 2m
2 giving pc =
So, for example, compute the wavelength
of a 100 eV electron. This is non-relativistic
since 100 eV << 510000 eV. So pc = 106 100 eV or 10000 eV.
=

3.4

2197.3
12000
=
= 0.12 nm
pc
10000

Single Slit Diffraction

There are many other examples of diffraction. The picture below shows diffraction
from a single slit where waves from different parts of the slit interfere with each other.
To get the field at some point on the detection screen, one should integrate over the
slit.

107

3. Diffraction

TOC

Figure 27: Distribution for single-slit diffraction of light. Note that the central maximun
is twice as wide as the other peaks and that the envelope for the distribution falls of
more quickly than in 2-slit diffraction.
Example: Derive the location of the nodes in the diffraction pattern from
one slit of width a. Now try to compute the intensity distribution for single
slit diffraction. (Advanced Link)

3.5

Wave Particle Duality (Thought Experiments)

Richard Feynman (Nobel Prize for Quantum ElectroDynamics...) presents several

thought experiments in his Lectures on Physics, third volume.

108

No matter what distance between the slits we choose, we never observe diffraction
for the bullets. They always give the P1 +P2 pattern (probably different for the bullets).
111

3. Diffraction

TOC

Can you explain this?

The bullets wavelength is much much smaller than the actual size of the
bullet.
Why didnt we see diffraction for the bullets. Bullets are macroscopic objects with
huge momenta compared to individual particles. High momentum implies a small de
Broglie wavelength. So to see diffraction, we must make the distance between the slits
much smaller than we did for the photons (perhaps 1020 times smaller). But bullets
are also big. They only fit through big slits which must them be further apart than
the slits used for photons. For bullets, the wavelength is tiny compared to any
slit a bullet would fit through. Therefore no bullet diffraction is possible.
Feynman Lectures on Physics, Vol. III Chapter 1

3.6

Doing the Critical (Diffraction) Thought Experiment

With modern equipment, we can actually perform the critical thought experiment described above. It is now possible to detect single photons with fairly low noise
using an image intensifier. The equipment is interesting but were trying to understand
quantum mechanics.
So if we make a singel photon detector where the position of the photon on the detector
can be read out, we can do the 2 slit diffraction experiment with very low intensity so
there is only one photon in the apparatus at a time. We can detect single photons over
some long time period and build up a measured diffraction distribution. We will see
that the individual photons do interfere with themselves. That is the different terms
in the wavefunction for a single photon interfere giving the standard 2 slit diffraction
distribution.

Figure 33: Time development of the accumulated hits in our single photon detector.
Even though there is only one photon counted at a time, the interference patter is
eventually seen.
112

3. Diffraction

TOC

The video can be downloaded here.

3.7
3.7.1

Examples
Intensity Distribution for Two Slit Diffraction *

Derive the location of the nodes in the diffraction pattern from two narrow slits a
distance d apart. Now try to compute the intensity distribution.

E0 sin(t)

2d sin
E2 = E0 sin t +

ab
a+b
sin(a) + sin(b) = 2 cos
sin
2
2

d sin
d sin
sin t +
E = 2E0 cos

d
sin

I = 2E02 cos2

3.8

Homework

1. What is the de Broglie wavelength for each of the following particles? The energies
given are the kinetic energies.
a 1 eV electron
a 104 MeV proton
a 1 gram lead ball moving with a velocity of 100 cm/sec.

113

3. Diffraction

=
2

TOC

h
p

deBroglie wavelength is easy

p
2m

=
T
p = 2mT
pc = 2mc2 T
= hc
pc
hc
= 2mc
2T
= 1240

2(511000)(1)

13 eV mc2 = 0.5 MeV is very NR

solve for pc
use convenient units
use hc =1240 eV nm
the NR answer
put in numbers in eV and nm

p
T = p(mc2 )2 + (pc)2 mc2
pc = (T + mc2 )2 (mc2 )2
hc
=
2 2
2 2

T = 104 MeV is 10 times larger than mp c2 = 938 MeV

solve for pc
solve for wavelength

10% in quadrature makes a visible correction

(T +mc ) (mc )
1240 M eV f m
10898 M eV

p = mv
h
= mv
27
erg sec
= 1.0510
(1 g)(50000 cm/s)
= 2.1 1032 cm

lead ball is very nonrelativistic

no eV input so just use Plancks constant in SI
put in cgs numbers
wavelength is much much smaller than size

2. Use the Rydberg formula to calculate the three longest wavelenghts in the Lyman
series, the three longest wavelenghts in the Balmer series, and the three longest
in the Paschen-Bach series. For each wavelength, state what part of the EM
spectrum it belongs too.
2
mc2
En = 2n
= 13.6
Hydrogen Energies
2
n2 eV

Ephoton = E1 E2 = 13.6
+
n2
=

2 2
hc
2~c n1 n2
E
13.6 n21 n22
2 2
2197.3 n1 n2
13.6 n21 n22

13.6
n22

= 13.6

1
n22

1
n21

nm
=
Lyman n2 = 1, Balmer n2 = 2, Paschen-Bach n2 = 3.

Conservation of E

wavelength of photon
wavelength of photon

3. Light of wavelength 300 nm is incident upon two narrow slits separated by 200
microns. At what angle with the first null in the intensity distribution appear?
The first intensity maximum is at zero angle. At what angle will the second
maximum appear?
d sin = 2 first minimum
d sin = second maximum (first past 0)

3.9

Sample Test Problems

1. What is the de Broglie wavelength of an electron with 13.6 eV of kinetic energy?

What is the de Broglie wavelength of an electron with 10 MeV of kinetic energy?
114

3. Diffraction

TOC

Answer
13.6 eV is much less than mc2 = 0.511M eV so this is non-relativistic.
p2
=
2m
p2 c2
=
2mc2
pc =

13.6

13.6
p
2mc2 (13.6)
h
2~
2~c
2~c
=
=
=
=p
p
p
pc
2mc2 (13.6)
2(197.3 eV nm)
= p
0.333 nm
2(0.511 106 eV )(13.6 eV )

10 MeV is much bigger than mc2 for an electron so it is super-relativistic and we

can use E = pc.
=

2~c
2(197.3) M eV F )
=
120 F
pc
10 M eV

2. What is the de Broglie wavelength for each of the following particles? The energies
given are the kinetic energies.
a) a 10 eV electron
b) a 1 MeV electron
c) a 10 MeV proton
3. A 2 slit electron diffraction experiment is set up as (not) shown below. The observed electron intensity distribution is plotted in the figure. Now an intense light
source is introduced near the two slits. With this light, it can be seen which
slit each electron goes through. Qualitatively plot the new electron intensity distribution from each slit and from the 2 slits combined. What is the condition on
the wavelength of the light for this effect to occur?
4. What is the de Broglie wavelength for each of the following particles? The energies
given are the kinetic energies.
a) a 1 eV electron
b) a 104 MeV proton
5. What K.E. must a Hydrogen atom have so that its de Broglie wavelength is
smaller than the size of the atom? (Factors of 2 are not important.)
6. Calculate the de Broglie wavelength for (a) a proton with 10 MeV kinetic energy,
(b) An electron with 10 MeV kinetic energy, and (c) a 1 gram lead ball moving
with a velocity of 10 cm/sec (one erg is one gram cm2 /sec2 ). Be sure to take
account of relativity where needed.
115

4. The Solution: Probability Amplitudes

TOC

The Solution: Probability Amplitudes

For EM waves, the intensity, and hence the probability to find a photon, is proportional to the square of the fields. The fields obey the wave equation. The fields from
two slits can add constructively or destructively giving interference patterns.
We will use the same ideas for electrons, although the details of the field will
vary a bit because electrons and photons are somewhat different kinds of particles. For
both particles the wavelength is given by
=

h
p

and the frequency by

E = h = ~.
We will use a complex probability amplitude (x, t) for the electron. (See the
reiview of complex numbers .) The real and imaginary parts are out of phase like
the EM fields. The traveling wave of electrons with momentum p in the x direction
and energy E then is
(x, t) ei(kxt) = ei(pxEt)/~
(See the review of traveling waves ). The probability to find an electron
in an interval dx is equal to the absolute square of the complex probability
amplitude.
P (x, t)dx = |(x, t)|2 dx
(We will overcome the problem that this probability is 1 everywhere for our simple
wavefunction.) So |(x, t)|2 dx is the probability density function at time t.
We have just put in most of the physics of Quantum Mechanics. Much of
what we do for the rest of the course will be deduced from the paragraph
above. Our input came from de Broglie and Planck, with support from
experiments.
Lets summarize the physics input again.
Free particles are represented by complex wave functions with a relationship between their particle properties energy and momentum, and their wave properties
frequency and wavelength given by Planck and de Broglie.
The absolute square of the wavefunction gives the probability density function.
Quantum Mechanics only tells us the probability.
We can make superpositions of our free particle wave functions to make states
that do not have definite momentum. We will find that any state can be made
from the superposition of free particle states with different momentum.
116

4. The Solution: Probability Amplitudes

TOC

We now have a wave-particle duality for all the particles, however, physics now
only tells us the probability for some quantum events to occur. We have lost the
complete predictive power of classical physics.
Gasiorowicz Chapter 1
Rohlf Chapter 5
Griffiths 1.2, 1.3
Cohen-Tannoudji et al. Chapter

4.1
4.1.1

Derivations and Computations

Review of Complex Numbers

This is a simple review, but, you must make sure you use complex numbers correctly.
One of the most common mistakes in test problems is to forget to take the complex
conjugate when computing a probability.
A complex number c = a + ib consists of a real part a and an imaginary part ib. (We
choose a and b to be real numbers.) i is the square root of -1.
The complex conjugate of c is c = a ib. (Just change the sign of all the i.)
The absolute square of a complex number is calculated by multiplying it by its complex
conjugate.
|c|2 = c c = (a ib)(a + ib) = a2 + iab iab + b2 = a2 + b2
This give the magnitude squared of the complex number. The absolute square is always
real.
We will use complex exponentials all the time.

cos + i sin

cos i sin

You can verify that the absolute square of these exponentials is always 1. They are
often called a phase factor.
We can write sin =

ei ei
2i

and cos =

ei +ei
.
2

117

4. The Solution: Probability Amplitudes

TOC

As with other exponentials, we can multiply them by adding the exponents.

eikx eit = ei(kxt)

4.1.2

Review of Traveling Waves

A normal traveling wave may be given by

cos(kx t).
The phase of the wave goes through 2 in one wavelength in x. So the wavelength
satisfies
k = 2.
Similarly the phase goes through 2 in one period in time.
= 2
is the angular frequency. It changes by 2 every cycle. The frequency increases by
1 every cycle so
= 2.
There is no reason to memorize these equations. They should be obvious.
Lets see how fast one of the peaks of the wave moves. This is called the phase velocity.
At time t = 0, there is a peak at x = 0. This is the peak for which the argument of
cosine is 0. At time t = 1, the argument is zero when kx = t or at x = k . If we
compute the phase velocity by taking x
t , we get

vphase = .
k
That is, one of the peaks of this wave travels with a velocity of k .
v=

2
= 2 =
k

In non-relativistic QM, we have ~k = p, E = ~, and E =

2 2

p2
2m ,

E
~ k
~k
=
=
~
2m~
2m
You may remember that a pulse will move at the group velocity which is given by

d
2~k
~k
p
vg =
=
=
= .
dk
2m
m
m
(k) =

(The phase velocity for the non-relativistic case is vp =

p
2m .)

118

4. The Solution: Probability Amplitudes

4.2

TOC

Sample Test Problems

1. Write down the two (unnormalized) free particle wave functions for a particle of
kinetic energy E. Include the proper time dependence and expressions for other
constants in terms of E.

119

5. Wave Packets

TOC

Wave Packets

Gasiorowicz Chapter 2
Rohlf Chapters 5
Griffiths Chapter 2

5.1

Building a Localized Single-Particle Wave Packet

We now have a wave function for a free particle with a definite momentum p
(x, t) = ei(pxEt)/~ = ei(kxt)
where the wave number k is defined by p = ~k and the angular frequency satisfies
E = ~. It is not localized since P (x, t) = |(x, t)|2 = 1 everywhere.
We would like a state which is localized and normalized to one particle.

Normalization Condition

(x, t)(x, t)dx = 1

To make a wave packet which is localized in space, we must add components of

different wave number. Recall that we can use a Fourier Series to compose any
function f (x) when we limit the range to L < x < L. We do not want to limit our
states in x, so we will take the limit that L . In that limit, every wave number
is allowed so the sum turns into an integral. The result is the very closely related
Fourier Transform

1
A(k)eikx dk
f (x) =
2

with coefficients which are computable,

1
A(k) =
2

f (x)eikx dx.

120

5. Wave Packets

TOC

The normalizations of f (x) and A(k) are the same (with this symmetric form) and
both can represent probability amplitudes.

A (k)A(k)dk

f (x)f (x)dx =

We understand f (x) as a wave packet made up of definite momentum terms eikx .

The coefficient of each term is A(k). The probability for a particle to be found in a
region dx around some value of x is |f (x)|2 dx. The probability for a particle to have
wave number in region dk around some value of k is |A(k)|2 dk. (Remember that p = ~k
so the momentum distribution is very closely related. We work with k for a while for
economy of notation.)

5.2

Two Examples of Localized Wave Packets

Lets now try two examples of a wave packet localized in k and properly normalized at t = 0.
1
a

1. A square packet: A(k) =

2. A Gaussian packet: A(k) =

for k0

a
2

< k < k0 +

a
2

and 0 elsewhere.

2 1/4 (kk0 )2
e
.

These are both localized in momentum about p = ~k0 .

Check the normalization of (1).

k0 + a
2

1
|A(k)| dk =
a

dk =

1
a=1
a

k0 a
2

Check the normalization of (2) using the result for a definite integral of a Gaussian

p
2
dx eax = a .

r
|A(k)|2 dk =

e2(kk0 ) dk =

=1
2

121

5. Wave Packets

TOC

So now we take the Fourier Transform of (1) right here.

f (x)

k0 + a
2

1 1
A(k)eikx dk =
2 a

eikx dk
k0 a
2

i
1 1 ikx
1 1 ik0 x h iax/2
e
=
e
e
eiax/2
2a ix
2a ix

r
h

i
1 1 ik0 x
a ik0 x 2 sin ax
ax
2

e
2i sin
=
e
2
2
ax
2a ix
k0 + a
2
k0 a
2

Fourier Transform of Square Packet

A(k) =

1
a

a2 < k <
|x| > a2

a
2

a ik0 x 2 sin ax
2
e
2
ax

r
f (x) =

k =

x =

a
2
2
a

2 sin( ax
2 )
is equal to 1 at x = 0 and that it decreases from there. If you square
Note that
ax
this, it should remind you of a single slit diffraction pattern! In fact, the single slit
gives us a square localization in position space and the F.T. is this sin(x)
function.
x

The

Fourier Transform of a Gaussian wave packet A(k) =

f (x) =

1
2

1/4

2 1/4 (kk0 )2
e

eik0 x e 4

also a Gaussian. We will show later that a Gaussian is the best one can do to localize
a particle in position and momentum at the same time.

Fourier Transform of a Gaussian Packet

A(k) =

f (x) =

1/4

1
2

e(kk0 )

1/4

eik0 x e 4

1
k =
2
x =

122

5. Wave Packets

TOC

In both of these cases of f (x) (transformed from a normalized A(k) localized in momentum space) we see
A coefficient which correctly normalizes the state to 1,
eik0 x a wave corresponding to momentum ~k0 ,
and a packet function which is localized in x.
We have achieved our goal of finding states that represent one free particle. We
see that we can have states which are localized both in position space and momentum
space. We achieved this by making wave packets which are superpositions of states
with definite momentum. The wave packets, while localized, have some width in x and
in p.

5.3

The Heisenberg Uncertainty Principle

The wave packets we tried above satisfy an uncertainty principle which is a

property of waves. That is kx 12 .
For the square packet the full width in k is k = a. The width in x is a little hard
2
to define, but, lets use the first node in the probability found at ax
2 = or x = a . So
4
the width is twice this or x = a . This gives us
kx = 4
which certainly satisfies the limit above. Note that if we change the width of A(k), the
width of f (x) changes to keep the uncertainty product constant.
For the Gaussian wave packet, we can rigorously read the RMS width of the probability distribution as was done at the end of the section on the Fourier Transform
of a Gaussian.

x =

1
k =
4
We can again see that as we vary the width in k-space, the width in x-space varies to
keep the product constant.
1
x k =
2
The Gaussian wave packet gives the minimum uncertainty. We will prove this
later.
123

5. Wave Packets

TOC

If we translate into momentum p = ~k, then

p = ~k.
So the Heisenberg Uncertainty Principle states.

Heisenbrg Uncertainty Principle

~
2

It says we cannot know the position of a particle and its momentum at the same time
and tells us the limit of how well we can know them.
If we try to localize a particle to a very small region of space, its momentum becomes
uncertain. If we try to make a particle with a definite momentum, its probability
distribution spreads out over space.
Example: Determine how we could see inside atoms using EM waves.
Example: Determine how we could see inside nuclei using EM waves.

5.4

Position Space and Momentum Space

We can represent a state with either (x) or with (p). We can (Fourier) transform
from one to the other.
We have the symmetric Fourier Transform.
f (x)

A(k)

2
1

A(k)eikx dk

f (x)eikx dx

When we change variable from k to p, we get the Fourier Transforms in terms of

x and p.

124

5. Wave Packets

TOC

Fourier Transform in 1D

(x) =

(p) =

1
2~

(p)eipx/~ dp

(x)eipx/~ dx

With this Fourier Transform we can try p0 (p) = (p p0 ), the transform of which
1
would be p0 (x) = 2~
eip0 x/~ which is the same as above. Similarly we could use
1
eipx0 /~ .
x0 (x) = (x x0 ) to get x0 (p) = 2~

(p p0 )(p p1 )dp = (p1 p0 )

This is the correct normalization for the momentum eigenfunction.

We can add this as a third wave packet for which we took the Fourier
Transform, finding that the width of the delta function is zero and its transform has
infinite witdth.
These formulas are worth a little study. If we define up (x) to be the state with
definite momentum p, (in position space) our formula for it is
up (x) =

1
eipx/~ .
2~

Similarly, the state (in momentum space) with definite position x is

vx (p) =

1
eipx/~
2~

These states cannot be normalized to 1 but they do have a normalization convention

which is satisfied due to the constant shown.
Our Fourier Transform can now be read to say that we add up states of definite
momentum to get (x)

(x) =
(p)up (x)dp

and we add up states of definite position to get (p).

(p) =

(x)vx (p)dx

125

5. Wave Packets

TOC

There is a more abstract way to write these states. Using the notation of Dirac, the
1
state with definite momentum p0 , up0 (x) = 2~
eip0 x/~ might be written as
|p0 i
and the state with definite position x1 , vx1 (p) =

1
2~

eipx1 /~ might be written

|x1 i.
The arbitrary state represented by either (x) or (p), might be written simple as
|i.

The actual wave function (x) would be written as

(x) = hx|i.
This gives us the amplitude to be at x for any value of x.
We will find that there are other ways to represent Quantum states. This was a preview.
We will spend more time on Dirac Bra-ket notation later.

5.5

Time Development of a Gaussian Wave Packet *

So far, we have performed our Fourier Transforms at t = 0 and looked at the result
only at t = 0. We will now put time back into the wave function and look at the wave
packet at later times. We will see that the behavior of photons and non-relativistic
electrons is quite different.
Assume we start with our Gaussian (minimum uncertainty) wavepacket A(k) =
2
e(kk0 ) at t = 0. We can do the Fourier Transform to position space, including the
time dependence.

(x, t) =
A(k)ei(kx(k)t) dk

We write explicitly that depends on k.

For our free particle, this just means that the energy depends on the momentum. For
a photon, E = pc = h = ~, and p = ~k so ~ = ~kc, and hence = kc. For the
photon we have assumed a relationship between the energy and momentum given by
the kinematics of massless particles.
126

5. Wave Packets

TOC

For an free electron, we will also assume the kinematic relationship between
2 2
p2
k
energy and momentum. For an non-relativistic electron, E = 2m
, so ~ = ~2m
,
2
.
Later
we
will
make
a
similar
input
for
an
electron
in
potential
and hence = ~k
2m
(that is, in some force field) using the Schr
odinger equation.
To cover the general case, lets expand (k) around the center of the wave packet in
k-space.

d
1 d2
(k) = (k0 ) +
(k k0 ) +
(k k0 )2
dk
2 dk 2
k0

We anticipate the outcome a bit and name the coefficients.

(k) = 0 + vg (k k0 ) + (k k0 )2
For the photon, vg = c and = 0. For the NR electron, vg =

~k0
m

and =

~
2m .

Performing the Fourier Transform, we get

1/4
(xvg t)2

i(k0 x0 t)
4(+it)
(x, t) =
e
e
2( + it)2
r
(xvg t)2

2
2(2 + 2 t2 )
e
|(x, t)| =
2(2 + 2 t2 )

We see that the photon will move with the velocity of light and that the wave packet
will not disperse, because = 0.
p
,
For the NR electron, the wave packet moves with the correct group velocity, vg = m
q

~t 2
but the wave packet spreads with time. The RMS width is = + 2m /.

Time Development of Gaussian WP

(x, t) =

vg =

2( + it)2

~k0
d
=

dk k0
m

1/4

ei(k0 x0 t) e

(xvg t)2
4(+it)

~
1 d2
=

2
2 dk k0
2m

A wave packet naturally spreads because it contains waves of different momenta and
hence different velocities. Wave packets that are very localized in space spread rapidly.
127

5. Wave Packets

5.6
5.6.1

TOC

Derivations and Computations

Fourier Series *

Fourier series allow us to expand any periodic function on the range (L, L), in
terms of sines and cosines also periodic on that interval.
f (x) =

An cos

nx
L

n=0

Bn sin

n=1

nx
L

Since the sines and cosines can be made from the complex exponentials, we can equally
well use them for our basis for expansion. This has the nice simplification of having
only one term in the sum, using negative n to get the other term.

f (x) =

an e

inx
L

The exponentials are orthogonal and normalized over the interval (as were the sines
and cosines)
L
imx
inx
1
e L e L dx = nm
2L
L

so that we can easily compute the coefficients.

1
an =
2L

f (x)e

inx
L

In summary, the Fourier series equations we will use are:

Fourier Series
f (x) =

an e

inx
L

1
an =
2L

L
f (x)e

inx
L

We will expand the interval to infinity in the following section.

128

5. Wave Packets

5.6.2

TOC

Fourier Transform *

To allow wave functions to extend to infinity, we will expand the interval used
L .
As we do this we will use the wave number
k=

n
.
L

As L ., k can take on any value, implying we will have a continuous distribution

of k. Our sum over n becomes an integral over k.
dk =
If we define A(k) =
we want.

f (x) =
an =
An =
An =

2
Lan ,

an e

n=
L

1
2L

f (x)e

2
1
L 2L

1
2

f (x) =
f (x) =
f (x) =
A(k) =

1
2 L

1
2

f (x)e

1
2

Standard Fourier Series

inx
L

Standard Fourier Series

redefine coefficient

we can make the transform come out with the constants

inx
L

dn.
L

inx
L

An e

dx
inx
L

f stays the same

A(k)eikx L
dk

but is rewritten in new A and dk

A(k)eikx dk

result

f (x)eikx dx

result

This is just the extension of the Fourier series to all x.

If f (x) is normalized, then A(k) will also be normalized with this (symmetric) form of
the Fourier Transform. Thus, if f (x) is a probability amplitude in position-space, A(k)
can be a probability amplitude (in k-space).

129

5. Wave Packets

5.6.3

TOC

Integral of Gaussian

This is just a slick derivation of the definite integral of a Gaussian from minus infinity
to infinity. With other limits, the integral cannot be done analytically but is tabulated.
Functions are available in computer libraries to return this important integral.
The answer is

ax2

dx e

r
=

.
a

Define

dx eax .

Integrate over both x and y so that

dx eax

I2 =

dy eay =

dxdy ea(x

+y 2 )

Transform to polar coordinates.

I = 2

rdr e

ar 2

2
2

1
=
d(r2 ) ear = ear
a
a
0

Now just take the square root to get the answer above.

Definite Integral of a Gaussian

ax2

dx e

r
=

Other forms can be obtained by differentiating with respect to a.

130

5. Wave Packets

TOC

More Gaussian Definite Integrals

ax2

dx e

=
a

2 ax2

dx x e

1
=
2a

5.6.4

Fourier Transform of Gaussian *

We wish to Fourier transform the Gaussian wave packet in (momentum) k1/4 (kk )2
0
space A(k) = 2
e
to get f (x) in position space. The Fourier Transform

formula is
1/4
2
1
2
f (x) =
e(kk0 ) eikx dk.
2

Now we will transform the integral a few times to get to the standard definite integral
of a Gaussian for which we know the answer. First,
k 0 = k k0
which does nothing really since dk 0 = dk.
f (x)

1/4
2
ik0 x
e
e(kk0 ) ei(kk0 )x dk
3
2

1/4
02
0
ik0 x
e
ek eik x dk 0
3
2

f (x)

Now we want to complete the square in the exponent inside the integral. We plan
002
a term like ek so we define
ix
k 00 = k 0
.
2
Again dk 00 = dk 0 = dk. Lets write out the planned exponent to see what we are missing.

ix
k
2
0

= k 02 + ik 0 x +

x2
4
131

5. Wave Packets

TOC

We need to multiply by e 4 to cancel the extra term in the completed square.

1/4
ix 2
x2
ik0 x
(k0 2
) 4
dk 0
f (x) =
e
e
e
2 3

That term can be pulled outside the integral since it doesnt depend on k.

1/4
002
x2
ik0 x 4
e
e
f (x) =
ek dk 00
3
2

So now we have the standard Gaussian integral which just gives us

1/4 r
x2
eik0 x e 4
f (x) =
3
2

1/4
x2
1
f (x) =
eik0 x e 4
2

Fourier Transform of a Gaussian

A(k) =

f (x) =

1/4

1
2

e(kk0 )

1/4

eik0 x e 4

Lets check the normalization.

r
r

x2
1
1
2
2
|f (x)| dx =
e
dx =
2 = 1
2
2

Given a normalized A(k), we get a normalized f (x).

The RMS deviation, or standard deviation of a Gaussian can be read from the distribution. The Standard Normal Distribution (Gaussian) has two parameters,
the mean X and the standard deviation , the RMS deviation from the mean. It is a
normalized probability distribution function.
Standard Normal Distribution
PX, (x) =

1
2 2

(xX)2
2 2

132

5. Wave Packets

TOC

Squaring f (x), we get

1 x2
e 2 .
2
Reading from either the coefficient or the exponential we see that

x =
P (x) =

For the width in k-space,

2 2(kk0 )2
e
.

Reading from the coefficient of the exponential, we get

P (k) =

1
k = .
4
We can see that as we vary the width in k-space, the width in x-space varies to keep
the product constant.
1
x k =
2
Translating this into momentum, we get the limit of the Heisenberg Uncertainty
Principle.

Uncertainty of Gaussian Wavepackets

x p =

~
2

In fact the Uncertainty Principle states that

x p

~
2

so the Gaussian wave packets seem to saturate the bound!

5.6.5

Time Dependence of a Gaussian Wave Packet *

Assume we start with our Gaussian (minimum uncertainty) wavepacket A(k) = e(kk0 )
at t = 0.
r

1
(x, t) =
A(k)ei(kx(k)t) dk
2

133

5. Wave Packets

TOC

We write explicitly that w depends on k. For our free particle, this just means that
the energy depends on the momentum. To cover the general case, lets expand (k)
around the center of the wave packet in k-space.
(k) = (k0 ) +

1 d2
d
|k (k k0 )2
|k0 (k k0 ) +
dk
2 dk 2 0

We anticipate the outcome a bit and name the coefficients.

(k) = 0 + vg (k k0 ) + (k k0 )2
We still need to do the integral as before. Make the substitution k 0 = k k0 giving
02
A(k 0 ) = ek . Factor out the constant exponential that has no k 0 dependence.
r
(x, t)

1 i(k0 xw0 t)
e
2

A(k 0 )ei(k xvg t) eik

dk 0

(x, t)

1/4
02
0
02
i(k0 xw0 t)
e
ek ei(k xvg t) eik t dk 0
2 3
1/4
ei(k0 xw0 t)
2 3

e[it]k ei(k xvg t) dk 0

We now compare this integral to the one we did earlier (so we can avoid the
work of completing the square again). Keeping the constants, we had
f (x) =

1/4 r
1/4
x2
ik0 x
k02 ik0 x
0
e
e
e
dk
=
eik0 x e 4
3
3
2
2

Our new integral is the same with the substitutions k0 x k0 x 0 t, k 0 x

k 0 (x vg t), and ( + it) (except in the orignal constant in A(k)). We can then
write down the answer

1/4
(xvg t)2

i(k0 x0 t)
4(+it)
(x, t) =
e
e
2( + it)2

1/4
1/4
(xvg t)2
(xvg t)2

2
4(+it) e 4(it)
|(x, t)| =
e
2( + it)2
2( it)2
r
(xvg t)2

2(2 + 2 t2 )
|(x, t)|2 =
e
2(2 + 2 t2 )
134

5. Wave Packets

TOC

We can again compare this to the Standard Normal Distribution: P (x) =

and read off that:
2 + 2 t2
2 =

5.6.6

1
e
2 2

(xX)2
2 2

Numbers

The convenient unit of energy (mass and momentum too) is the electron Volt. It
is the energy an electron (or other unit charged particle) gets when it goes through a
potential difference of 1 Volt.

The Electron Volt is Our Preferred Energy Unit

1 eV = 1.602 1012 erg = 1.602 1019 Joule

Since the Volt is an SI unit, we might think the eV is also SI, but it is not really. The
adjustements made between cgs units, SI units, and rationalized Heaviside
Lorentz units are made mainly to make the constants appearing in the EM field
equations go away. To do this they adjust the units of the fields and the units of charge,
keeping the force the same. Using a unit which is charge times potential largely avoids
the problem, although cgs does use cm. Anyway the use of eV as an energy unit:
is convenient for use for particles with the fundamental charge e accelerated by
fields,
gives us a unit where we can effectively set c = 1 for energy and momentum
calculations,
can be agnostic on which set of units we want to use,
and pick units that are sized right for atomic physics or nuclear physics problems,
rather than for kilogram masses.
We will work mainly in these units. The primary exception is when we are showing
that macroscopic objects do not show quantum effects.
Use the fine structure constant to avoid CGS units which are used in many advanced
textbooks.
e2
=
= 1/137
~c
135

5. Wave Packets

TOC

The combination ~c saves a lot of work.

~c = 197.3 eV nm = 197.3 MeV fm
1 Fermi = 1 fm = 1.0 1015 m
We can write the Coulomb potential between an electron and a nucleus using these
constants to make it independent of units.
VCoulomb =

Z~c
r

The Bohr radius gives the size of the Hydrogen atom.

a0 =

~
= 0.529 1010 m
me c
mp = 938.3 MeV/c

mn = 939.6 MeV/c

me = 0.511 MeV/c

5.6.7

The Dirac Delta Function

The Dirac delta function is zero everywhere except at the point where its
argument is zero. At that point, it is just the right kind of infinity so that

Basic Integral of a Delta Function

dx f (x) (x) = f (0)

This is the definition of the delta function. It picks of the value of the function f (x)
at the point where the argument of the delta function vanishes.
The transformation of an integral allows us to compute other integrals containing delta
functions.

136

5. Wave Packets

TOC

Other Integrals of Delta Functions

dx f (x) (x a) = f (a)

"
dx f (x) (g(x)) =

1
dg
|
| dx

#
f (x)
g(x)=0

If we make a wave packet in p-space using the delta function, and we transform to
position space,
(x) =

1
2~

(p p0 )eipx/~ dp =

1
eip0 x/~
2~

we just get the state of definite p.

This is a state of definite momentum written in momentum space. (p p0 )
1
Its Fourier transform is p (x, t) = 2~
ei(pxEt)/~
This is a state of definite position written in position space. (x x0 )

5.7
5.7.1

Examples
Can I See inside an Atom

To see inside an atom, we must use light with a wavelength smaller than the size of
the atom. With normal light, once a surface is polished down to the .25 micron level,
it looks shiny. You can no longer see defects. So to see inside the atom, we would need
light with = hp = 0.1
A.
p

pc =

2~
0.01
2~c
2197.3
=
= 120000 eV
0.1
0.01

This is more than enough kinetic energy to blow the atom apart. You cant see
inside.
137

5. Wave Packets

TOC

A similar calculation can be made with the uncertainty principle.

~
2
~c
(pc)x
2
~c
E
2x
~c
= 10000eV
E
2(0.1
A)
px

So we need 10 keV photons while the binding energy is 13 eV, so this will
still blow the atom apart.
So we cant watch the inside of an atom.
We can probe atoms with high energy photons (for example). These will blow the
atoms apart, but we can use many atoms of the same kind. We learn about the
internal structure of the atoms by scattering particles off them, blowing them apart.

5.7.2

Can I See inside a Nucleus

~c
In a similar fashion to the previous section, E 2(0.1F
) = 1000 MeV.
The binding energy per nucleon is a few MeV, so, we will also blow nuclei apart to look
carefully inside them. We again can just use lots of nuclei to allow us to learn about
internal nuclear structure.

5.7.3

Estimate the Hydrogen Ground State Energy

The reason the Hydrogen atom (and other atoms) is so large is the essentially uncertainty principle. If the electron were confined to a smaller volume, p would increase,
causing p to increase on average. The energy would increase not decrease.
Lets assume there is no angular momentum in the ground state adn that we want
to localize the electron as well as we can around the nucleus. We can use the
uncertainty principle to estimate the minimum energy for Hydrogen. This
is not a perfect calculation but it is more correct than the Bohr model. The idea is
that the radius must be larger than the spread in position, and the momentum must

138

5. Wave Packets

TOC

be larger than the spread in momentum.

px ~

estimate in 1D

r x

typical r bigger than spread

p p

typical p bigger than spread

This is our formula for the potential energy in terms of the dimensionless fine
structure constant .
e2
~c
V (r) = =
r
r

The total energy is:

p2
~c

2m
r

Use the uncertainty principle.

pr
E
E

~
~cp
p2

2m
~
p2

cp
2m

Differentiate with respect to p and set equal to zero to get the minimum.
dE
dp
p
E

=
=

p
c = 0
m
mc
2 mc2
2 mc2
2 mc2 =
= 13.6 eV
2
2

Note that the potential energy is just (-2) times the kinetic energy (as we expect from
the Virial Theorem). The ground state energy formula is exactly correct, even
though this calculation should only be good to within a factor of 5 or so.
We can also estimate the radius.
r=

~
~
~c
197.3eV nm(137)
=
=
=
= 0.053nm
p
mc
mc2
511000eV

The ground state of Hydrogen has zero (orbital) angular momentum. It is not moving
in a circular orbit as Bohr hypothesized. The electron just has a probability distribution
that is spread out over about 1
A. If it were not spread out, the energy would go up.
139

5. Wave Packets

5.8

TOC

Homework

1. The Dirac delta function has the property that

f (x)(x x0 ) dx = f (x0 )

Find the momentum space wave function (p) if (x) = (x x0 ).

1
(p) = 2~
(x)eipx/~
Fourier Transform
1
2~

(p) =

(x x0 )eipx/~

of delta function

ipx0 /~

1 e
2~

(p) =

delta function does integral

2. Use the calculation of a spreading Gaussian wave packet to find the fractional
change in size of a wave packet between t = 0 and t = 1 second for an electron
localized to 1 nanometer. Now find the fraction change for a 1 gram weight localized to 1 nanometer.
2
|(x, t)|2 =

2 + 2 t2

(xvg t)

e 2(2 +2 t2 )

spreading gaussian

~
= 12 ddk2 |k0 = 2m
2
2
= = 2 nm
2 + 2 t2
2
q =

f
f
f
f

for NR particle
sigma width at t=0
at time t

2 + 2 t2

q 2
2 2
= 1 + 2t
q
~2 t2
= 1 + 4m
2 4
q
~2 c2 (c2 t2 )
= 1 + 4m2 c4 4
q
nm)2 (31018
= 1 + (197.3 eV
4m2 c4 (1 nm)4

width increase factor

simplify
plug in
units OK
nm)2

all lengths in nm, use eV

(197.3 eV nm)(31018 nm)

106 eV (1 nm)2

f 1+

1
2

(197.3)(310

1.81014 J
1.61019 J/eV

big for electron; neglect 1

1 + 3 1013

very small for 1g mass, aprox.

3. Use the uncertainty principle to estimate the energy of the ground state of a
p2
+ 12 kx2 .
harmonic oscillator with the Hamiltonian H = 2m
2

p
H = 2m
+ 21 m 2 x2
px px ~2
~2
p2 4x
2
~2
1
2 2
E 8mx
2 + 2 m x
2
dE
~
2
= 4mx
3 + m x = 0
dx
~2
2 4
4m 2= m x
~
4
4m2 2 = x
~
2
x = 2m
~2
~2 2m
p2 = 4x
= ~m
2 =
4~
2
~m
1
2 ~
E = 4m + 2 m 2m = ~

1
4

= 12 ~

H.O.
uncertainty: HO is gaussian
square above, solve for p2
plug in p2 in terms of x
find minimum by setting deriv. to 0
algebra
algebra
solve for x2
get p2
plug in to find E
140

5. Wave Packets

TOC

4. Estimate the kinetic energy of an electron confined to be inside a nucleus of radius 5 Fermis. Estimate the kinetic energy of a neutron confined inside the same
nucleus.
5. Show that

(x)x(x)dx =

(p) i~
p

(p)dp.

Remember that the wave functions go to zero at infinity.

I
call the integral I
(p) i~ p (p)dp

ipx/~
1

I = 2~
(p) i~ p
e
(x)dx dp
expand phi(p) using FT
I=

1
2~

(p) =
I=
I

1
2~

(p)

ipx/~

(x)e

(x)dx dp

ipx0 /~ 0
1
e
(x )dx0
2~

ipx0 /~ 0

0
e

(x )dx

(FT of phi)*
xeipx/~ (x)dx dp

expand phi*(p)

!
0

eip(x x)/~ dp (x0 )x(x)dx0 dx

(x x) (x0 )x(x)dx0 dx

take derivative

1
=
2~

(x)x(x)dx

switch order
ID delta func

QED

do integral with delta func

6. Directly calculate the the RMS uncertainty in x for the state (x) =
by computing
p
x = h|(x hxi)2 |i.
(x)2 = h|(x hxi)2 |i =h|x2 |i
(x)

x2 e2x / dx
= e2x2 / dx

= 2

1
2a

=
a

1
2a

ax2
2

ave x=0 by symmetry

(x)
note that /4 has the place of 2 in the gaussian.

use definite integral formulas

take sqrt
as a check

7. Calculate hpn i for the state in the previous problem. Use this to calculate p in
a similar way to the x calculation.

141

5. Wave Packets

(x) = Ce

TOC

original wfn

p2
4~2

F.T.
by symmetry
constant in gauss integral

(p) = C e
hpn i = 0 for odd n
a = /2~2

p
2

dp eap = a
a
a

dp p2 eap =

dp p4 eap =

formula to get p2 integral

3
12
a
a
= 12 a 2

1 23
a 2 a

formula to apply again

32 12 a 2

apply formula second time

p
hp i = 135...(n1)
a(n+1)/2 a
2n/2
hpn i = 135...(n1)
an/2
2n/2 2
1
~
2
(p) = 2a q
=
1
p = 2a
=
px = ~2

apply formula once

compute integal and divide by norm.

n even
take the p2 case

compute p
uncertainty principle saturated

8. The momentum space wave function is given by (p) = (5p ~k0 ). What is the
wavefunction in position space?
i
h

1
f
(x)
delta function integral
dx f (x) (g(x)) = | dg
|

(x) =
(x) =

5.9

1
2~

dp (p) e

ipx/~

1
eik0 x/5
5 2~

g(x)=0

FT formula
compute

Sample Test Problems

1. A nucleus has a radius of 4 Fermis. Use the uncertainty principle to estimate

the kinetic energy for a neutron localized inside the nucleus. Do the same for an
electron.
Answer
px ~
pr ~
Try non-relativistic formula first and verify approximation when we have the
energy.
E=

~2
(~c)2
(197.3M eV F )2
p2
=
=
=
1.3M eV
2m
2mr2
2mc2 r2
2(940M eV )(4F )2
142

5. Wave Packets

TOC

This is much less than 940 MeV so the non-relativistic approximation is very
good.
The electron energy will be higher and its rest mass is only 0.51 MeV so it WILL
be relativistic. This makes it easier.
pr ~
E = pc =

~c
197.3M eV F
=
50M eV
r
4F

2. * Assume that the potential for a neutron near a heavy nucleus is given by
V (r) = 0 for r > 5 Fermis and V (r) = V0 for r < 5 Fermis. Use the uncertainty
principle to estimate the minimum value of V0 needed for the neutron to be bound
to the nucleus.
3. Use the uncertainty principle to estimate the ground state energy of Hydrogen.
Answer
px ~
pr ~
p2
e2
p2
e2
E=

=
p
2m
r
2m
~
(We could have replaced p equally well.) Minimize.
dE
p
e2
=

=0
dp
m
~
me2
p=
~
e2
m 2 e4
p2
me4
me4
me4
me4
p=
E=
2 =
2 = 2
2m
~
2m~2
~
2~2
~
2~
e2
=
~c
e2 = ~c
1
E = 2 mc2
2
4. * Given the following one dimensional probability amplitudes in the position
variable x, compute the probability distribution in momentum space. Show that
the uncertainty principle is roughly satisfied.
a (x) =

1
2a

for a < x < a, otherwise (x) = 0.

1
4

x
b (x) = (
) e

143

5. Wave Packets

TOC

c (x) = (x x0 )
5. Use the Heisenberg uncertainty principle to estimate the ground state energy for
a particle of mass m in the potential V (x) = 12 kx2 .
6. * Find the one dimensional wave function in position space (x) that corresponds
to (p) = (p p0 ).
7. * Find the one dimensional wave function in position space (x) that corresponds
to (p) = 12b for b < p < b, and (p) = 0 otherwise.
8. * Assume that a particle is localized such that (x) = 1a for 0 < x < a and that
(x) = 0 elsewhere. What is the probability for the particle to have a momentum
between p and p + dp?
9. A beam of photons of momentum p is incident upon a slit of width a. The
resulting diffraction pattern is viewed on screen which is a distance d from the slit.
Use the uncertainty principle to estimate the width of the central maximum
of the diffraction pattern in terms of the variables given.
10. * The wave-function of a particle in position space is given by (x) = (x a).
Find the wave-function in momentum space. Is the state correctly normalized?
Explain why.
2

11. * A particle is in the state (x) = Aex /2 . What is the probability for the
particle to have a momentum between p and p + dp?
12. A hydrogen atom has the potential V (r) =
estimate the ground state energy.

e2
r .

Use the uncertainty principle to

13. * Assume that (p) = 12a for |p| < a and (p) = 0 elsewhere. What is (x)?
What is the probability to find the particle between x and x + dx?
14. The hydrogen atom is made up of a proton and an electron bound together by
the Coulomb potential, V (r) = ~c
r . It is also possible to make a hydrogen-like
atom from a proton and a muon. The force binding the muon to the proton is
identical to that for the electron but the muons mass is 106 MeV/c2 . Use the
uncertainty principle to estimate the energy and the radius of the ground state
of muonic hydrogen.
15. * Given the following one dimensional probability amplitudes in the momentum
representation, compute the probability amplitude in the position representation,
(x). Show that the uncertainty principle is satisfied.

(a) (p)
= 12a for a < p < a, (p)
= 0 elsewhere.

(b) (p) = (p p0 )
2
1

4 p /2
(c) (p)
= (
) e

16. * Assume that (p) = (p p0 ). What is (x)? What is < p2 >? What is
< x2 >?
144

6. Operators

TOC

Operators

Operators will be used to help us derive a differential equation that our wave-functions
must satisfy. They will also be used in almost any Quantum Physics calculation.

An example of a linear operator is a simple differential operator like x

, which we
understand to differentiate everything to the right of it with respect to x.

6.1

Operators in Position Space

To find operators for physical variables in position space, we will look at wave functions
with definite momentum. Our state of definite momentum p0 (and definite energy E0 )
is
1
up0 (x, t) =
ei(p0 xE0 t)/~ .
2~
We can build any other state from superpositions of these states using the Fourier
Transform.

(x, t) =
(p) up (x, t)dp

6.1.1

The Momentum Operator

(op)

We determine the momentum operator by requiring that, when we operate with px

on up0 (x, t), we get p0 times the same wave function.
p(op) up0 (x, t) pup0 (x, t) = p0 up0 (x, t)
(op)

This means that for these definite momentum states, multiplying by px is the same as
multiplying by the variable p. We find that this is true for the following momentum
operator.

Momentum Operator in x-Space

p =

~
i x

We can verify that this works by explicit calculation.

145

6. Operators

TOC

If we take our momentum operator and act on a arbitrary state,

p(x, t) = p

(p) up (x, t)dp =

(p) p up (x, t)dp

it gives us the right p for each term in the integral. This will allow us to compute
expectation values for any variable we can represent by an operator.

6.1.2

The Energy Operator

We can deduce and verify the energy operator in the same way.
Total Energy Operator
= i~
E (op) E
t

6.1.3

The Position Operator

What about the position operator, x(op) x

? The answer is simply
Position Operator in x-Space
x
=x

when we are working in position space with up0 (x, t) =

have been above).

6.1.4

1 ei(p0 xE0 t)/~

(as we

The Hamiltonian Operator

We can develop other operators using the basic ones. We will use the Hamiltonian
operator which, for our purposes, is the sum of the kinetic and potential energies.
This is the non-relativistic case.
H=

p2
+ V (x)
2m
146

6. Operators

TOC

Hamiltonian Operator in x-Space

2
2
= ~ + V (x)
H
2m x2

Since the potential energy just depends on x, its easy to use. Angular momentum
operators will later be simply computed from position and momentum operators.

6.2

Operators in Momentum Space

If we want to work in momentum space, we need to look at the states of definite

position to find our operators. The state (in momentum space) with definite position
x0 is
1
vx0 (p) =
eipx0 /~
2~
The operators are

Position Operator in Momentum Space

x
= i~

and
p = p.
The (op) notation used above is usually dropped. If we see the variable p, use of the
operator is implied (except in state written in terms of p like (p)). We have introduced
to economically clarify when use of the operator is required.
the (LaTex) O
Gasiorowicz Chapter 3
Griffiths doesnt cover this.
Cohen-Tannoudji et al. Chapter

147

6. Operators

6.3

TOC

Expectation Values

Operators allow us to compute the expectation value of some physics quantity given
the wavefunction. If a particle is in the state (x, t), the normal way to compute
the expectation value of f (x) is

hf (x)i =

(x)(x)f (x)dx.

P (x)f (x)dx =

We can move the f (x) between just before anticipating the use of linear operators.

(x) f (x) (x)dx

hf (x)i =

If the variable we wish to compute the expectation value of (like p) is not a simple
function of x, let its operator act on (x). The expectation value of p in the state
is

Expectation Value using and Operator

(x)
p(x)dx

hpi = h|p|i =

The Dirac Bra-ket notation shown above is a convenient way to represent the expectation value of a variable given some state.
Example: A particle is in the state (x) =
expectation value of p?

1/4 ik x x2
1
e 0 e 4 .
2

What is the

For any physical quantity v, the expectation value of v in an arbitrary state is

(x)
v (x)dx

h|v|i =

The expectation values of physical quantities should be real.

Gasiorowicz Chapter 3
148

6. Operators

TOC

Griffiths Chapter 1
Cohen-Tannoudji et al. Chapter

6.4

Dirac Bra-ket Notation

A state with definite momentum p. |pi

A state with definite position x. |xi
The dot product between two abstract states 1 and 2 is defined to be:
Dot Product Between States

1 2 dx

h1 |2 i

This dot product projects the state 2 onto 1 and represents the amplitude to go
from 2 to 1 .
To find the probability amplitude for our particle to by at any position x, we dot the
state of definite x into our state . (x) = hx|i

To find the probability amplitude for our particle to have a momentum p, we dot the
state of definite x into our state . (p) = hp|i

6.5

Commutators

Operators (or variables in quantum mechanics) do not necessarily commute. We can

see our first example of that now that we have a few operators. We define the
commutator to be
Commutator of p and x
[p, x] px xp

149

6. Operators

TOC

(using p and x as examples.)

We will now compute the commutator between p and x. Because p is represented
by a differential operator, we must do this carefully. Lets think of the commutator as
a (differential) operator too, as generally it will be. To make sure that we keep all the

x that we need, we will compute [p, x](x) then remove the (x) at the end to see
only the commutator.
~
~
= px(x) xp(x) =
x(x) x
(x)
i
x
i x

~
(x)
(x)
~
[p, x](x) =
(x) + x
x
= (x)
i
x
x
i

[p, x](x)

So, removing the (x) we used for computational purposes, we get the commutator.
Commutator of p and x
[p, x] =

~
i

Later we will learn to derive the uncertainty relation for two variables from their commutator. Physical variable with zero commutator have no uncertainty principle and
we can know both of them at the same time.
We will also use commutators to solve several important problems.
We can compute the same commutator in momentum space.

d
d
~
d
[p, x] = [p, i~ ] = i~ p p = i~() =
dp
dp
dp
i
~
[p, x] =
i
The commutator is the same in any representation.
Example:
Example:
Example:
Example:
[Lx , Ly ].

Compute
Compute
Compute
Compute

the commutator [E, t].

the commutator [E, x].
the commutator [p, xn ].
the commutator of the angular momentum operators

Gasiorowicz Chapter 3
150

6. Operators

TOC

Griffiths Chapter 3
Cohen-Tannoudji et al. Chapter

6.6
6.6.1

Derivations and Computations

Verify Momentum Operator
1
~
1

ei(p0 xE0 t)/~ =

ei(p0 xE0 t)/~
i x 2~
2~
1 ~ ip0 i(p0 xE0 t)/~
1

e
= p0
ei(p0 xE0 t)/~
2~ i ~
2~
p

6.6.2

Verify Energy Operator

1 ei(p0 xE0 t)/~ = 1 i~ iE0 ei(p0 xE0 t)/~
E
~
2~
2~
1
= E0
ei(p0 xE0 t)/~
2~

6.7
6.7.1

Examples
Expectation Value of Momentum in a Given State
1/4 ik x x2
1
e 0 e 4 .
2

A particle is in the state (x) =

What is the expectation value of

We will use the momentum operator to get this result.

hpi

(x)
p(x)dx

h|p|i =

1
2

1/4

x
ik0 x 4

~
i x

1
2

1/4

eik0 x e 4 dx

1
2

1/2

~
i

eik0 x e 4

ik0 x x2
e
e 4 dx
x

151

6. Operators

TOC

hpi

=

=

1
2
1
2

1/2

~
i
~
i

x
ik0 x 4

ik0 e

x
2

x2
2x ik0 x x2
ik0 x 4
4

ik0 e
e
e
e
dx
4

2x x2

e 2
4

The second term gives zero because the integral is odd about x = 0.

1/2

x2
1
~
h|p|i =
ik0 e 2 dx
2
i

h|p|i =

1
2

1/2

~k0 2 = ~k0

Excellent. Our wavepacket centered at k0 has an expected momentum of ~k0 computed

using the momentum operator.

6.7.2

Commutator of E and t

Again use the crutch of keeping a wave function on the right to avoid mistakes.

[E, t](x, t) =
i~ t ti~
(x, t)
t
t

(x, t)
= i~(x, t) + i~t i~t
t
t
= i~(x, t)
Removing the wave function, we have the commutator.
[E, t] = i~

6.7.3

Commutator of E and x

Again use the crutch of keeping a wave function on the right to avoid mistakes.

[E, x](x, t) =
i~ x xi~
(x, t)
t
t

=
i~x i~x
(x, t) = 0
t
t
Since

x
t

= 0.
152

6. Operators

6.7.4

TOC

Commutator of p and xn

We can use the commutator [p, x] to help us. Remember that px = xp + [p, x].
[p, xn ]

= pxn xn p
=

(px)xn1 xn p

= xpxn1 + [p, x]xn1 xn p

= x(px)xn2 + [p, x]xn1 xn p
= x2 pxn2 + x[p, x]xn2 + [p, x]xn1 xn p
= x2 pxn2 + 2[p, x]xn1 xn p
= x3 pxn3 + 3[p, x]xn1 xn p
= xn p + n[p, x]xn1 xn p
~
= n[p, x]xn1 = n xn1
i

It is usually not wise to use the differential operators and a wave function crutch to
compute commutators like this one. Use the known basic commutators when
you can. Nevertheless, we can compute it that way.
[p, xn ] =

~ n
~
~
x xn
= nxn1
i x
i x
i

~ n1
nx
i
It works pretty well for this particular case, but not if I have p to some power...
[p, xn ] =

6.7.5

Commutator of Lx and Ly

Angular momentum is defined by

~ = ~r p~.
L
So the components of angular momentum are
Lz = xpy ypx
Lx = ypz zpy
Ly = zpx xpz .
We wish to compute [Lx , Ly ] which has all the coordinates and momenta in it.
153

6. Operators

TOC

The only operators that do not commute are the coordinates and their conjugate momenta.
[x, y] = 0
[px , py ] = 0
[pi , rj ] =

~
ij
i

So now we just need to compute. First write out the commutator then use the distributive property of commutation over addition.
[Lx , Ly ]

[ypz zpy , zpx xpz ]

[ypz , zpx ] [ypz , xpz ] [zpy , zpx ] + [zpy , xpz ]

= y[pz , z]px 0 0 + x[z, pz ]py

~
(ypx xpy ) = i~Lz
=
i
More generally, since we can orient our right handed coordinate system any way.
[Li , Lj ] = i~ijk Lk
It is not necessary (or wise) to use the differential operators and a wave function crutch
to compute commutators like this one. Use the known basic commutators when
you can.

154

6. Operators

6.8

TOC

Homework

1. Calculate the commutator [p2 , x2 ].

d2
2
2 d2
2
[p2 , x2 ] = [~2 dx
2 , x ] = ~ [ dx2 , x ]

2
2
d
d
2
2
2 d
2d
~2 [ dx
2x
+
x

x
2 , x ] = ~
2
dx
dx
dx

2
2
d
2d
+
2x
+
x

x2 ddx2
[p2 , x2 ] = ~2 2 + 2x d
dx
dx
dx2

d
[p2 , x2 ] = ~2 2 + 4x dx

2
d
2
2
2
[p , x ] = ~ 2 + 4x dx = 2 ~i + 4 ~i xp
[p2 , x2 ] = p2 x2 x2 p2
[p2 , x2 ] = p(xp + ~i )x x2 p2 = pxpx + ~i px x2 p2
[p2 , x2 ] = (xp + ~i )px + ~i px x2 p2
[p2 , x2 ] = xppx + 2 ~i px x2 p2
[p2 , x2 ] = xp(xp + ~i ) + 2 ~i px x2 p2
[p2 , x2 ] = xpxp + 2 ~i px x2 p2
[p2 , x2 ] = x(xp + ~i )p + ~i xp + 2 ~i px x2 p2
[p2 , x2 ] = x2 p2 + 2 ~i xp + 2 ~i px x2 p2
[p2 , x2 ] = 2 ~i (xp + px)
2
[p2 , x2 ] = 4 ~i xp + 2 ~i

use p =

~ d
i dx

differentiate once
second time
simplify
answer by method 1
definition
commute middle px
now commute first px
simplify
now commute last px
simplify
now commute middle px
simplify
reasonable form of answer
compare answer-2

2. Find the commutator [p, eik0 x ] where k0 is a constant and the second operator

P
(ik0 x)n
can be expanded as eik0 x =
.
n!
n=0

3. Calcutlate the commutators [p, xn ] and [pm , xn ].

We have already computed [p, xn ] (two ways) in an example.
~
[p, xn ] = n xn1
i
Now use that to efficiently compute the nasty commutator [pm , xn ]. Try to understand this step since its the way to compute efficently.
[pm , xn ] = pm1 [p, xn ] + pm2 [p, xn ]p + pm3 [p, xn ]p2 + ... + [p, xn ]pm1

~ m1 n1
p
x
+ pm2 xn1 p + pm3 xn1 p2 + ... + xn1 pm1
i
The answer is not simple but can be expressed as a sum.
[pm , xn ] = n

[pm , xn ] = n

~ X mk n1 k1
p
x
p
i
k=1

155

6. Operators

6.9

TOC

Sample Test Problems

1. The absolute square of a wave function for a free particle is given as:
r
2
2
2 2
a
2
|(x, t)| =
ea(xvg t) /2(a +b t )
2(a2 + b2 t2 )
Find the expected value of x as a function of time. Find the expected value of
x2 as a function of time. Compute the RMS x-width of this wave packet as a
function of time.
2. Find the commutator [p, eik0 x ] where k0 is a constant and the second operator

P
(ik0 x)n
can be expanded as eik0 x =
.
n!
n=0

3. Which of the following are linear operators?

O1 (x) = 1/(x)
O2 (x) =

(x)
x
2

O3 (x) = x (x)
O4 (x) = (x + a)
4. For a free particle, the total energy operator H is given by H = p2 /2m. Compute
the commutators [H,x] and [H,p]. If a particle is in a state of definite energy, what
do these commutators tell you about how well we know the particles position
and momentum?
5. Find the commutator [x, p3 ].
6. Compute the commutator [H, x2 ] where H is the Hamiltonian for a free particle.

156

7. The Schr
odinger Equation

TOC

The Schr
odinger Equation

Schr
odinger developed a differential equation for the time development of a
wave function. Since the Energy operator has a time derivative, the kinetic energy
operator has space derivatives, and we expect the solutions to be traveling waves, it
is natural to try an energy equation. The Schr
odinger equation is the operator
statement that the kinetic energy plus the potential energy is equal to the
total energy.

7.1

Deriving the Equation from Operators

For a free particle, we have

p2
=E
2m
~2 2

= i~
2
2m x
t
Lets try this equation on our states of definite momentum.

~2 2
1
1

ei(p0 xE0 t)/~ = i~

ei(p0 xE0 t)/~
2
2m x
t 2~
2~
157

7. The Schr
odinger Equation

TOC

The constant in front of the wave function can be removed from both sides. Its there
for normalization, not part of the solution. We will go ahead and do the differentiation.

~2 p20 i(p0 xE0 t)/~

iE0 i(p0 xE0 t)/~
e
e
= i~
2m ~2
~
p20 i(p0 xE0 t)/~
e
= E0 ei(p0 xE0 t)/~
2m

Our wave function will be a solution of the free particle Schrodinger equation provided
p20
E0 = 2m
. This is exactly what we wanted. So we have constructed an equation that
has the expected wave-functions as solutions. It is a wave equation based on the total
energy.
Adding in potential energy, we have the Schr
odinger Equation

Schr
odinger Equation in 1D
~2 2 (x, t)
(x, t)
+ V (x)(x, t) = i~
2
2m
x
t

or
H(x, t) = E(x, t)
where
H=

p2
+ V (x)
2m

is the Hamiltonian operator.

In three dimensions, this becomes.

Schr
odinger Equation in 3D
H(~x, t) =

~2 2
(~x, t)
(~x, t) + V (~x)(~x, t) = i~
2m
t

We will use it to solve many problems in this course.

So the Schr
odinger Equation is, in some sense, simply the statement (in operators)
that the kinetic energy plus the potential energy equals the total energy.
158

7. The Schr
odinger Equation

7.2

TOC

Schr
odinger Gives Time Development of Wavefunction

The Schr
odinger equation lets us compute the time derivative of the wave function
from the current wavefunction, so in principal, we can integrate to get at a later
(or even earlier) time. There are also other methods that can be used, based on the
Schr
odinger equation.
Since all that we can know about a system is its wavefunction, this is all that
we can do to predict the future based on current knowledge. Note that a normal
wave equation which is second order in time is not as predictive.
We can also learn about the energy eigenstates using the Schrodinger equation.
Because of the conjugate relationship between energy and time, time development
is very closely related to the energy.

7.3

The Flux of Probability *

In analogy to the Poynting vector for EM radiation, we may want to know the probability current in some physical situation. For example, in our free particle solution,
the probability density is uniform over all space, but there is a net flow along the
direction of the momentum.
We can derive an equation showing conservation of probability by differentiating P (x, t) =
and using the Schr
odinger Equation. We can derive the probability conservation
equation identifying j(x, t) as the probability current.

Probability Flux in 1D
P (x, t) j(x, t)
+
=0
t
x

~

j(x, t) =

2mi
x
x

This current can be computed from the wave function.

If we integrate if over some interval in x
b
a

P (x, t)
dx =
t

b
a

j(x, t)
dx
x
159

7. The Schr
odinger Equation

d
dt

TOC

b
P (x, t)dx = j(x = a, t) j(x = b, t)
a

the equation says that the rate of change of probability in an interval is equal to the
probability flux into the integral minus the flux out.
Extending this analysis to 3 dimensions, we get.
Probability Flux in 3D
P (x, t) ~ ~
+ j(~r, t) = 0
t
h
i
~ r, t) (~r, t)
~ (~r, t)
~j(~r, t) = ~ (~r, t)(~
2mi

7.4

The Schr
odinger Wave Equation

The normal equation we get, for waves on a string or on water, relates the second
space derivative to the second time derivative. The Schrodinger equation uses only
the first time derivative, however, the addition of the i relates the real part of the
wave function to the imaginary part, in effect shifting the phase by 90 degrees as the
2nd derivative would do.
~2 2 (x, t)
(x, t)
+ V (x)(x, t) = i~
2
2m
x
t
The Schr
odinger equation is built for complex wave functions.
When Dirac tried to make a relativistic version of the equation, where the energy
relation is a bit more complicated, he discovered new physics.
Gasiorowicz Chapter 3
Griffiths Chapter 1
Cohen-Tannoudji et al. Chapter

7.5

The Time Independent Schr

odinger Equation

Second order differential equations, like the Schr

odinger Equation, can be solved by
separation of variables. These separated solutions can then be used to solve the
160

7. The Schr
odinger Equation

TOC

problem in general.
Assume that we can factorize the solution between time and space.
(x, t) = u(x)T (t)
Plug this into the Schr
odinger Equation.
2 2

~ u(x)
T (t)
+ V (x)u(x) T (t) = i~u(x)
2m x2
t
Put everything that depends on x on the left and everything that depends on t on the
right.
2 2

~ u(x)
+
V
(x)u(x)
2
i~ Tt(t)
2m x
=
= const. = E
u(x)
T (t)
Since we have a function of only x set equal to a function of only t, they both
must equal a constant. In the equation above, we call the constant E, (with some
knowledge of the outcome). We now have an equation in t set equal to a constant
i~

T (t)
= E T (t)
t

which has a simple general solution,

T (t) = CeiEt/~
and an equation in x set equal to a constant
~2 2 u(x)
+ V (x)u(x) = E u(x)
2m x2
which depends on the problem to be solved (through V (x)).
The x equation is often called the Time Independent Schr
odinger Equation.
Here, E is a constant. The full time dependent solution is.

Time Independent Schr

odinger Equation
~2 2 u(x)
+ V (x)u(x) = E u(x)
2m x2
(x, t) = u(x)eiEt/~

The solutions of this equation are called Energy Eigenstates. We see light from
transitions between these states. They are also referred to as the Stationary States
161

7. The Schr
odinger Equation

TOC

since their only time dependence is the complex phase so that the probability distribution is time independent or stationary. They are also referred to at the Definite
Energy States since they only have one energy.
One can write any state as a linear combination of these eigenstates (indluding both
bound and unbound states).
Example: Solve the Schr
odinger equation for a constant potential V0 .

7.6
7.6.1

Derivations and Computations

Linear Operators

Linear operators L satisfy the equation

L(a + b) = aL + bL

where a and b are arbitrary constants and and are arbitrary wave-functions. A
multiplicative constant is a simple linear operator. Differential operators clearly are
linear also.
An example of a non-linear operator (which we will not use) is N which has the property
N = 2 .

7.6.2

Probability Conservation Equation *

Start from the probability and differentiate with respect to time.

P (x, t)

=
( (x, t)(x, t)) =
+
t
t
t
t
Use the Schr
odinger Equation
~2 2

+ V (x) = i~
2m x2
t
and its complex conjugate
~2 2

+ V (x) = i~
.
2
2m x
t

162

7. The Schr
odinger Equation

TOC

(We assume V (x) is real. Imaginary potentials do cause probability not to be conserved.)
Now we need to plug those equations in.

1 ~2 2
~2 2
P (x, t)

V
(x)

+
V
(x)

t
i~ 2m x2
2m
x2

2
1 ~2 2
~

=

=
i~ 2m x2
x2
2mi x x
x
This is the probability conservation equation if j(x, t) is identified as the
probability current.
P (x, t) j(x, t)
+
=0
t
x

~

j(x, t) =
2mi
x
x

7.7
7.7.1

Examples
Solution to the Schr
odinger Equation in a Constant Potential

Assume we want to solve the Schr

odinger Equation in a region in which the potential
is constant and equal to V0 . We will find two solutions for each energy E.
We have the equation.
~2 2 (x)
+ V0 (x) = E(x)
2m x2
2 (x)
2m
= 2 (E V0 )(x)
x2
~
Remember that x is an independent variable in the above equation while E is a constant
to be determined in the solution.
For E > V0 , there are solutions
eikx
and
eikx
p
if we define k by the equation ~k = + 2m(E V0 ). Any energy is allowed as for
free particles. These are waves traveling in opposite directions with the same energy
163

7. The Schr
odinger Equation

TOC

(and magnitude of momentum). A general solution with energy E will be a linear

combination of the allowed solutions:
E (x) = Aeikx + Beikx
We could also use the linear combinations of the two solutions
sin(kx)
and
cos(kx).
There are only two linearly independent solutions with a given E. We need to
choose either the exponentials or the trig functions, not both. The sin and cos solutions
represent states of definite energy but contain particles moving to the left and to the
right. They are not definite momentum states. They will be useful to us for some
solutions.
The solutions are also technically correct for
p E < V0 but k becomes imaginary. Lets
write the solutions in terms of ~ = i~k = 2m(V0 E) The solutions are
ex
and
ex .
These are not waves at all, but real exponentials. Note that these are solutions for
regions where the particle is not allowed classically, due to energy conservation; the
total energy is less than the potential energy. We will use these solutions in Quantum
Mechanics.

7.8

Homework

1. The wave function for a particle is initially (x) = Ceikx + Deikx . What is the
probability flux j(x)?
ikx
(x) = Ceikx
given
+ De

d
~
j(x) = 2mi
d
dx dx
= C eikx + D eikx

d
ikx
+ D eikx
dx = ik C e
d
ikx
+ Deikx
dx = ik Ce

flux formula
compute
compute
compute

Now we plug all these into the flux formula and simplify.

i~k
j = 2mi
|C|2 C De2ikx + CD e2ikx |D|2 + |C|2 + C De2ikx CD e2ikx |D|2
j=

~k
|C|2 |D|2
m
164

8. Eigenfunctions, Eigenvalues and Vector Spaces

8
8.1

TOC

Eigenfunctions, Eigenvalues and Vector Spaces

Eigenvalue Equations

The time independent Schr

odinger Equation is an example of an Eigenvalue equation.
H u(x) = E u(x)
The Hamiltonian operates on u(x) the eigenfunction, giving a constant E the
eigenvalue, times the same function. (Eigen just means the same in German.)
Usually, for bound states, there are many eigenfunction solutions (denoted here
by the index i).

Energy Eigenstates
Hi = Ei i

For states representing one particle (particularly bound states) we must require that
the solutions be normalizable. Solutions that are not normalizable must be discarded. A normalizable wave function must go to zero at infinity.

(x)(x)dx = 1

() 0

In fact, all the derivatives of must go to zero at infinity in order for the wave
function to stay at zero.
It is generally the normalizability condition that
requires discrete Energy eigenvalues in solving a problem. In particular, for
many important problems, it is the requirement that the wave function goes to zero at
infinity.
There can be an eigenvalue equation for any operator, particularly those representing
physical variables. We have already discussed the momentum and position eigenstates.
Here are some eigenvalue equations for various physical operators, other than energy.
165

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

Eigenvalue Equations for Some Physical Operators

pup1 (x) = p1 up1 (x)

continuous

~
Sz =
2

discrete

z n`m = m~n`m
L

discrete

If a physical variable is measured, the only possible outcomes of the measurement

are the eigenvalues of the operator. Thus for operators with discrete eigenvalues, like
Energy or Spin, physical variables are quantized.
We will prove later that the eigenfunctions of Hermitian operators are orthogonal to
each other.
Eigenstates Orthonormal
hi |j i = ij

We will assume that the eigenfunctions form a complete set so that any function
can be written as a linear combination of them.
= 1 1 + 2 2 + 3 3 + ...
Expansion of in Eigenstates
=

i i

i=1

i = hi |i

(This can be proven for many of the eigenfunctions we will use.)

Since the eigenfunctions are orthogonal, we can easily compute the coefficients in
the expansion of an arbitrary wave function .
*
+
X
X
X

i = hi |i = i
j j =
hi |j ij =
ij j = i
j
j
j
166

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

We will later think of the eigenfunctions as unit vectors in a vector space. The arbitrary
wave function is then a vector in that space.

1
2

i j Ej hi |j i

i j Ej ij

i i Ei

|i | Ei

We can see that the coefficients of the eigenstates represent probability amplitudes to be in those states, since the absolute squares of the coefficients i i
obviously give the probability.
We now have three equivalent ways to represent a abstract state |i:

Three Ways to Specify an Abstract State

(x)
(p)
i

All of these are the probability amplitudes to have some eigenvalue of a physical
variable.

8.2

Measurement in Quantum Mechanics

The lines seen in atomic spectra indicated some kind of energy quantization
before atoms were understood. With the Bohr model, it was postulated by Bohr that
167

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

transitions in Hydrogen were between a set of quantized states and that only energies
that were the difference between state energies are allowed for photons.

Figure 34: Energy Spectra of photons emitted by atoms due to transitions between
Energy Eigenstates. Photons of other energies are not observed to be emitted by atoms.

We now know that when a measurement of any physical variable is made,

the only possible results are the eigenvalues of the operator for that physical
variable. So for momentum, any value is possible. This is also true for position. We
will see that for bound states, only a certain set of discrete energies are possible. Born
hypothesized that even though a system can be in a mixed state, only the eignevalues
can be measured and that the probability is proportional amplitude squared.
This view was bolstered in 1923 when Stern and Gerlach measured the z component of the magnetic moment of neutral Silver atoms using a large gradient
in a magnetic field. Silver atoms are understood to have zero total orbital angular
momentum, so that effectively, they measured the magnetic moment (component) of a
single electron. They found that the value of the magnetic moment along the
168

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

direction of the field gradient was quantized, with only two possible values.
The picture of the beam is shown below, with the field off on the left and with the
field on on the right. In the center of the magnet, the beam is clearly split into two
possibilities.

Figure 35: To do this early measurement, Stern and Gerlach columnated the beam with
a horizontal slit and applied a B field varying in the veritcal direction. This exerts
a force on the electrons magnetic moment. The field gradient isnt uniform in the
horixontal direction. It is maximum in the middle. This experiment would be harder
with electrons which are charged and would both bend a lot and also diffract more in
the horizontal slit.

The electrons magnetic moment is proportional to its spin and if we measure the
electron spin along any direction the possible results are ~2 .
Once the spin along the z-axis has been measured, the state of the atom collapses
to be in the measured state, so that if the upper beam is passed through another
apparatus, all of the atoms will go into the upper beam.

169

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

Figure 36: The figure shows diagrams of three possible Stern-Gerlach experiments. In
all three the beam is first separated according to the spin along the z direction then the
bottom beam is blocked. Lets assume that this means that the positive spin component
eigenvalue has been measured and selected. The beam after the first block is in the state
with the positive spin component along z, so that if the z component is measured again,
all the atomes will be found to be in this positive state. On the other hand if the beam
is analyzed along the x direction, both spin posibilities will be seen. In the third figure
the positive spin along z is first selected, then the positive x state. When analyzed in
the z direction both spin states will be seen. But if the second block is removed, only
the positive z spin will be seen.

This collapse of the wavefunction upon measurement is not entirely new. For
example, an unpolarized light beam going through a linear polarizer has the electric
field projected onto the direction of the polarizer. Thus the wavefunction
which is the field, collapses to the measured direction. The Stern-Gerlach is a cleaner
quantum system because we keep the other polarization in the lower beam while the
polarizer simply absorbs the other linear polarization.

8.3

Hermitian Conjugate of an Operator

First let us define the Hermitian Conjugate of an operator A to be A . The meaning

of this conjugate is given in the following equation.

(x)A(x)dx = h|Ai hA |i

h|A|i =

That is, A must operate on the conjugate of and give the same result for the integral
as when H operates on .
170

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

The definition of the Hermitian Conjugate of an operator can be simply written

in Bra-Ket notation.
Hermitian Conjugate of Operator
hA |i = h|Ai

Starting from this definition, we can prove some simple things. Taking the complex
conjugate
h|A i = hA|i
Now taking the Hermitian conjugate of A .
h A

|i = hA|i
A

If we take the Hermitian conjugate twice, we get back to the same operator.
Its easy to show that

(A) = A
and

(A + B) = A + B
just from the properties of the dot product.
We can also show that

(AB) = B A .
h|ABi = hA |Bi = hB A |i
Example: Find the Hermitian conjugate of the operator a + ib.

Example: Find the Hermitian conjugate of the operator

8.4

x .

Hermitian Operators

A physical variable must have real expectation values (and eigenvalues). This
implies that the operators representing physical variables have some special properties.
171

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

By computing the complex conjugate of the expectation value of a physical variable, we

can easily show that physical operators are their own Hermitian conjugate.

(x)(H(x)) dx = hH|i
h|H|i =
(x)H(x)dx =

hH|i = h|Hi = hH |i
H = H
Operators that are their own Hermitian Conjugate are called Hermitian Operators.

Hermitian Operator
H = H

8.5

Eigenfunctions and Vector Space

Wavefunctions are analogous to vectors in 3D space. The unit vectors of our vector
space are eigenstates.
In normal 3D space, we represent a vector by its components.
~r = x
x + y y + z z =

3
X

ri u
i

i=1

The unit vectors u

i are orthonormal,
u
i u
j = ij
where ij is the usual Kroneker delta, equal to 1 if i = j and otherwise equal to zero.
Eigenfunctions the unit vectors of our space are orthonormal.
hi |j i = ij
We represent our wavefunctions the vectors in our space as linear combinations of the eigenstates (unit vectors).
=

X
i=1

i i
172

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

j j

j=1

The eigenstates generally form a complete set so that any state can be written as
a linear combination. This was originally shown by Dirichlet for the Fourier series
with very minimal requirements for the function. The main requirement is that its
integrable, which should also be true for physical functions. Other solutions can also
be shown to be complete but we will not do that here.
In normal 3D space, we can compute the dot product between two vectors using
the components.
~r1 ~r2 = x1 x2 + y1 y2 + z1 z2
In our vector space, we define the dot product to be

+
*
X
X

X
X

j j =
i j hi |j i
i i
h|i =

j=1
i=1 j=1
i=1
=

i j ij =

i=1 j=1

i i

i=1

We also can compute the dot product from the components of the vectors. Our vector
space is a little bit different because of the complex conjugate involved in the definition
of our dot product.
From a more mathematical point of view, the square integrable functions form a
(vector) Hilbert Space. The scalar product is defined as above.

d3 r

h|i

The properties of the scalar product are easy to derive from the integral.
h|i = h|i
h|1 1 + 2 2 i = 1 h|1 i + 2 h|2 i
h1 1 + 2 2 |i = 1 h1 |i + 2 h2 |i
h|i is real and greater than 0. It equals zero iff = 0. We may also derive the
Schwartz inequality.
p
h1 |2 i h1 |1 ih2 |2 i
Linear operators take vectors in the space into other vectors.
0 = A
173

8. Eigenfunctions, Eigenvalues and Vector Spaces

8.6

TOC

The Particle in a 1D Box

As a simple example, we will solve the 1D Particle in a Box problem. That is a

particle confined to a region 0 < x < a. We can do this with the (unphysical) potential
which is zero with in those limits and + outside the limits.

0 0<x<a
V (x) =
elsewhere
Because of the infinite potential, this problem has very unusual boundary conditions. (Normally we will require continuity of the wave function and its first derivative.)
The wave function must be zero at x = 0 and x = a since it must be continuous and it
is zero in the region of infinite potential. The first derivative does not need to be continuous at the boundary (unlike other problems), because of the infinite discontinuity
in the potential.
The time independent Schr
odinger equation (also called the energy eigenvalue
equation) is
Huj = Ej uj
with the Hamiltonian (inside the box)
H=

~2 d2
2m dx2

Our solutions will have

uj = 0
outside the box.
The solution inside the box could be written as
uj = eikx
where k can be positive or negative. We do need to choose linear combinations
that satisfy the boundary condition that uj (x = 0) = uj (x = a) = 0.
We can do this easily by choosing
uj = C sin(kx)
which automatically satisfies the BC at 0. To satisfy the BC at x = a we need the
argument of sine to be n there.
nx
un = C sin
a
174

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

Plugging this back into the Schr

odinger equation, we get
2 2
~2
n
C sin(kx) = EC sin(kx).
2
2m
a
There will only be a solution which satisfies the BC for a quantized set of energies.
En =

n2 2 ~2
2ma2

We have solutions to the Schr

odinger equation that satisfy the boundary conditions.
Now we need to set the constant C to normalize them to 1.
a
sin2

hun |un i = |C|

nx
a

dx = |C|2

a
2

Remember that the average value of sin2 is one half (over half periods). So we set C
giving us the eigenfunctions
r
nx
2
un =
sin
a
a
The first four eigenfunctions are graphed below. The ground state has the least curvature and the fewest zeros of the wavefunction.

175

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

Note that these states would have a definite parity if x = 0 were at the center of the
box.
The expansion of an arbitrary wave function in these eigenfunctions is essentially our
original Fourier Series. This is a good example of the energy eigenfunctions being
orthogonal and covering the space.

Particle in a 1D Box
r
un =

nx
2
sin
a
a

En =

8.6.1

n 2 2 ~2
2ma2

The Same Problem with Parity Symmetry

If we simply redefine the position of the box so that a2 < x <

problem has symmetry under the Parity operation.

a
2,

then our

x x
The Hamiltonian remains unchanged if we make the above transformation. The Hamiltonian commutes with the Parity operator.
[H, P ] = 0
This means that (P ui ) is an eigenfunction of H with the same energy eigenvalue.
H(P ui ) = P (Hui ) = P Ei ui = Ei (P ui )
Thus, it must be a constant times the same energy eigenfunction.
P ui = cui
The equations says the energy eigenfunctions are also eigenfunctions of the
parity operator.
If we operate twice with parity, we get back to the original function,
P 2 ui = ui
so the parity eigenvalues must be 1.
176

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

If V (x) has Parity Symmetry

[H, P ] = 0
P ui = 1ui

P ui = 1ui
The boundary conditions are

a
( ) = 0.
2

This gives two types of solutions.

Particle in a Parity Symmetric Box
r
u+
n (x)

2
cos
a
r

u
n (x)

(2n 1)x
a

2
sin
a

En+ = (2n 1)2

En = (2n)2

2nx
a

2 ~2
2ma2

Together, these are exactly equivalent to the set of solutions we had with the
box defined to be from 0 to a. The u+
n (x) have eigenvalue +1 under the parity
operator. The u
(x)
have
eigenvalue
-1
under
the parity operator.
n
This is an example of a symmetry of the problem, causing an operator to commute
with the Hamiltonian. We can then have simultaneous eigenfunctions of that operator
and H. In this case all the energy eigenfunctions are also eigenstates of parity. Parity
is conserved.
An arbitrary wave function can be written as a sum of the energy eigenfunctions recovering the Fourier series in its standard form.
(x) =

+

[A+
n un (x) + An un (x)]

n=1

177

8. Eigenfunctions, Eigenvalues and Vector Spaces

8.7

TOC

Momentum Eigenfunctions

We can also look at the eigenfunctions of the momentum operator.

pop up (x) = pup (x)
~ d
up (x) = pup (x)
i dx
The eigenstates are
up (x) = Ceipx/~
with p allowed to be positive or negative.
These solutions do not go to zero at infinity so they are not normalizable to one particle.
hp|pi = hup |up i =
This is a common problem for this type of state. In general, for states which are not
bound, the allowed energies or momenta are continuous. The eigenstates are orthonormal but we cannot use a Kronecker delta for a continuous variable like we can for
bound state wich are countable.
We will use a different type of normalization for the momentum eigenstates
(and the position eigenstates).

ei(pp )x/~ dx = 2~|C|2 (p p0 )

hp |pi = |C|

Instead of the Kronecker delta, we use the Dirac delta function. The momentum
eigenstates have a continuous range of eigenvalues so that they cannot be indexed like
the energy eigenstates of a bound system. This means the Kronecker delta could not
work anyway.
If we set |C|2 =

1 ,
2~

then we have the normalization:

1
hp |pi =
2~

ei(pp )x/~ dx = (p p0 )

So the momentum eigenstates are

up (x) =

1
eipx/~
2~
178

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

satisfying the normalization condition

hp0 |pi = (p p0 )

We have used the result derived from the Fourier Transform t hat:
An Important Integral Yielding a Function
1
2~

ei(pp )x/~ dx (p p0 )

to determine the constant in front of the momentum eigenfunction, which we could

also read off from the Fourier Transform.
We can then also write the momentum eigenfunction in momentum space, with eigenvalue p0 for example as:
p0 (p) = (p p0 )
If we take the Fourier transform of that, we get the momentum eigenfunction in position
space
1
up0 (x) =
eip0 x/~
2~
which agrees with the momentum eigenfunction derived above.
For a free particle Hamiltonian, both momentum and parity commute with H. So
we can make simultaneous eigenfunctions.
[H, p] = 0
[H, P ] = 0

We cannot make eigenfunctions of all three operators since

[P, p] 6= 0.
So we have the choice of the eikx states which are eigenfunctions of H and of p, but
contain positive and negative parity components. or we have the sin(kx) and cos(kx)
states which contain two momenta but are eigenstates of H and Parity. These are just
different linear combinations of the same solutions.
179

8. Eigenfunctions, Eigenvalues and Vector Spaces

8.8
8.8.1

TOC

Derivations and Computations

Eigenfunctions of Hermitian Operators are Orthogonal

We wish to prove that eigenfunctions of Hermitian operators are orthogonal. In fact

we will first do this except in the case of equal eigenvalues.
Assume we have a Hermitian operator A and two of its eigenfunctions such
that
A1 = a1 1
A2 = a2 2 .

Now we compute h2 |A|1 i two ways.

h2 |A1 i = a1 h2 |1 i
h2 |A1 i = hA2 |1 i = a2 h2 |1 i

Remember the eigenvalues are real so theres no conjugation needed.

Now we subtract the two equations. The left hand sides are the same so they give
zero.
0 = (a2 a1 )h2 |1 i
If a1 6= a2 then
h2 |1 i = 0.
The eigenfunctions are orthogonal.
What if two of the eigenfunctions have the same eigenvalue? Then, our proof doesnt
work. Assume h2 |1 i is real, since we can always adjust a phase to make it so. Since
any linear combination of 1 and 2 has the same eigenvalue, we can use any linear
combination. Our aim will be to choose two linear combinations which are
orthogonal. Lets try
+

1
(1 + 2 )
2
1
(1 2 )
2

180

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

so
h+ | i =
=

1
(1 1 + (h1 |2 i h2 |1 i))
2
1
(h1 |2 i h2 |1 i) = 0.
2

This is zero under the assumption that the dot product is real.
We have thus found an orthogonal set of eigenfunctions even in the case that
some of the eigenvalues are equal (degenerate). From now on we will just assume
that we are working with an orthogonal set of eigenfunctions.

8.8.2

Continuity of Wavefunctions and Derivatives

We can use the Schr

odinger Equation to show that the first derivative of the wave
function should be continuous, unless the potential is infinite at the boundary.
~2 d2
= (E V (x))
2m dx2
d2
2m
= 2 (V (x) E)
2
dx
~
Integrate both sides from just below a boundary (assumed to be at x = 0) to just
above.
+ 2
+
d
2m
dx = 2
(V (x) E)dx 0
dx2
~

Let go to zero and the right hand side must go to zero for finite potentials.

d
d

0
dx +
dx
Infinite potentials are unphysical but often handy. The delta function potential is very
handy, so we will derive a special continuity equation for it. Assume V (x) = V0 (x).
Integrating the Schr
odinger Equation, we get
+

2m
d2
dx = 2
dx2
~

+
(V0 (x) E)dx

As before, finite terms in the right hand integral go to zero as 0, but now the delta
function gives a fixed contribution to the integral.

d
d
2m

= 2 V0 (0)
dx +
dx
~
181

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

There is a discontinuity in the derivative of the wave function proportional to

the wave function at that point (and to the strength of the delta function potential).

8.8.3

Delta Function Implicit in Fourier Transforms

If we take the Fourier Transform of some (x) then transform back again we can
identify a Dirac Delta Function that is implicit in the Fourier Transform formulas.

0
1
1
1

(x)eipx/~ dx =
(p0 )eip x/~ dp0 eipx/~ dx
(p) =
2~
2~
2~

Interchanging the order of integration, we can identify the Dirac Delta function based
on its defining property.

0
1
0
i(p p)x/~
0

(p) =
(p )
e
dx dp =
(p0 )(p0 p)dp0
2~

The quantity in square brackets must be the Dirac Delta function according to this
result.

0
1
0
ei(p p)x/~ dx
(p p) =
2~

This clearly goes to infinity for p0 = p and we assume just the right infinity to behave
like the delta function, based on this Fourier Transform calculation.
For p 6= p0 , we can do the integral to see what we are getting.
1
(p p) =
2~

ei(p p)x/~ dx =

0
1
[ei(p p)x/~ ]

0
2i(p p)

While this may not obviously go to zero, if its applied to a square integrable
function, the oscillation at infinity is not important.

8.9
8.9.1

Examples
Hermitian Conjugate of a Constant Operator

If we have the operator O = a + ib where a and b are real, what is its Hermitian
conjugate? By the definition of the Hermitian conjugate
h|Oi = hO |i.
182

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

It is easy to see from the integral that

h(a ib)|i = h|(a + ib)i = (a + ib)h|i
So the Hermitian conjugate of a constant operator is its complex conjugate.

8.9.2

Hermitian Conjugate of

We wish to compute the Hermitian conjugate of the operator

integral to derive the result.

x .

We will use the

(x)
=
(x)
dx

x
x

We can integrate this by parts, differentiating the and integrating to get .

(x)

= [ (x)(x)]
(x)dx =

x
x
x

So the Hermitian conjugate of

is x
.

~
Note that the Hermitian conjugate of the momentum operator is i
x which is the
same as the original operator. So the momentum operator is Hermitian.

8.10

Homework

1. Prove that the parity operator defined by P (x) = (x) is a hermitian operator
and find its possible eigenvalues.
h|P i = hP |i
definition of H.C.

h|P i =
(x)(x)dx
x0 = x
h|P i =

h|P i =
h|P i =

dx0 = dx
(x0 )(x0 )(dx0 )

(x)(x)(dx)

(x)(x)dx = hP |i

transform variable of integration

interchange limits; change dummy name
Q.E.D.

P = a
P P = a2 =
a = 1

eigenvalue
Parity twice gives same coordinates
two possible eigenvalues
183

8. Eigenfunctions, Eigenvalues and Vector Spaces

TOC

2. Prove the Schwartz inequality |h|i| h|ih|i. (Start from the fact that
h + C| + Ci 0 for any C.
h + C| + Ci 0
h|i + Ch|i + C h|i + |C|2 h|i 0
choose
C = h|i
h|i
h|i

|h|i|2
h|i
|h|i|2
h|i

|h|i|2
h|i

h|i
h|ih|i |h|i|2

|h|i|2
h|i

0
Q.E.D.

3. A particle is in the first excited state of a box of length L. What is that state?
Now one wall of the box is suddenly moved outward so that the new box has
length 3L. What is the probability for the particle to be in the ground state of
the new box? What is the probability for the particle to be in the first excited
state of the new box? You may find it useful to know that

sin ((A B)x) sin ((A + B)x)

sin(Ax) sin(Bx)dx =

.
2(A B)
2(A + B)
(L)

2
initial state in box of length L
(3L)
1
ground state of new box length 3L
(3L)
2
first excited state of new box length 3
(3L) (L)
h1 |2 i
amplitude to be in ground state of new box
(3L) (L) 2
|h1 |2 i|
probability to be in ground state of new box
2

q

L q

2
2
x
sin 3L
sin 2x
dx
initial wfn is zero for x > L
Pg.s. =
3L
L
L

0

L 2
q q
7
5
sin( 3L
x)
2
2 sin( 3L x)
Pg.s. = 3L L 2 5 2 7
use integral given
3L
3L

0

2

sin( 5
sin( 7 )
3 )

2 73
Pg.s. = 23L
5

2 3L
3L
q

2
2

1
Pg.s. = 3L 34 21

5
7
2 3L
1 3L 3L

3L 2

Pg.s. = 2L 5 + 7
3 1 1 2
Pg.s. = 2
+
36 25 7
Pg.s. = 70 = 0.0268
Pf irst = 0.032
algebra is similar
4. Model a nucleus with a diameter of 2 fm, as a 1D infinite square well (Particle in
a Box). What would be the energy difference between the ground state and the
first excited state for a neutron bound in the box?
E = (22 12 )

2 ~2
3h2 c2
3(1240 M eV f m)2
=
=
= 153 M eV
2mn L2
8mn c2 L2
8(939.7 M eV )(2 f m)2
184

8. Eigenfunctions, Eigenvalues and Vector Spaces

8.11

TOC

Sample Test Problems

1. A particle is confined to a box of length L in one dimension. It is initially in

the ground state. Suddenly, one wall of the box is moved outward making a new
box of length 3L. What is the probability that the particle is in the ground state
of the new box? You may find it useful to know that sin(Ax) sin(Bx)dx =
sin((AB)x)
sin((A+B)x)
2(AB)
2(A+B)
Answer
(L)

(3L)

= |hu0 |u0 i|2

r
L
L r
2
2
x
x
2
x
x
(L) (3L)
sin
sin
=
sin
sin
dx
hu0 |u0 i =
L
L 3L
3L
L
3L
3L
P

=
=
P

sin 2x
3L
2
2 3L

sin 4x
3L
4
2 3L

2
2 3L

=
3L
3L 4
0

!
3
3 1 3
9

=
2
2
2 2
8

2 1
4
sin
sin
3
2
3

81
64 2

2. A particle of mass m is in a 1 dimensional box of length L. The particle is in

the ground state. The size of the box is suddenly (symmetrically) expanded to
length 3L. Find the probability for the particle to be in the ground state of the new
potential. (Your answer may include an integral which you need not evaluate.)
Find the probability to be in the first excited state of the new potential.
3. Two degenerate eigenfunctions of the Hamiltonian are properly normalized and
have the following properties.
H1

E0 1

E0 2

P 1

P 2

What are the properly normalized states that are eigenfunctions of H and P?
What are their energies?
4. Find the first (lowest) three Energy eigenstates for a particle localized in a box
such that 0 < x < a. That is, the potential is zero inside the box and infinite
outside. State the boundary conditions and show that your solutions satisfy
them. Normalize the solutions to represent one particle in the box.
185

8. Eigenfunctions, Eigenvalues and Vector Spaces

(Note that a beam coming from the right, would have given a eik x term for x > 0.)

Figure 37: Energy diagram for a potential step with the general solution form for a beam
of particles incident from the left. This problem displays the quantum phenomona of
reflection where classical physics would predict 100% transmission.
There are two unknown coefficients R and T which will be determined by matching
boundary conditions. We will not require normalization to one particle, since we have
a beam with definite momentum, which cannot be so normalized. (A more physical
problem to solve would use an incoming wave packet with a spread in momentum.)
Continuity of the wave function at x = 0 implies
1 + R = T.
The exponentials are all equal to 1 there so the equation is simple.
Continuity of the derivative of the wavefunction at x = 0 gives
0

[ikeikx ikReikx ]x=0 = [ik 0 T eik x ]x=0

Evaluate and plug in T from the equation above. We can solve the problem.
k(1 R) = k 0 (1 + R)
(k + k 0 )R = (k k 0 )

The coefficients are

k k0
k + k0
2k
T =1+R=
.
k + k0
R=

189

9. One Dimensional Potentials

TOC

We now have the full solution, given our assumption of particles incident from the
left.
(
0
ikx
x<0
eikx + kk
k+k0 e
u(x) =
0
2k
ik x
x>0
k+k0 e
Classically, all of the particles would be transmitted, continuing on to infinity.
In Quantum Mechanics, some probability is reflected.

Reflection Probability for Potential Step with E > V0

Preflection = |R|2 =
r
k=

2mE
~2

k k0
k + k0

r
0

k =

2m(E V0 )
~2

(Note that we can simply use the coefficient of eikx because the incoming term has
a coefficient of 1 and because the reflected particles are moving with the same
velocity as the incoming beam.)
If we wish to compute the transmission probability, the easy way to do it is to say
that its
4kk 0
Ptransmission = 1 Preflection =
.
(k + k 0 )2
Well get the same answers for the reflection and transmission coefficients using the
probability flux to solve the problem.
The transmission probability goes to 1 one k = k 0 (since there is no step). The
transmission probability goes to 0 for k 0 = 0 (since the kinetic energy is zero).

9.1.3

The Potential Well with E > 0 *

With positive energy, this is again a scattering type problem, now with three regions
of the potential, all with E > V .

x < a
0
V (x) = V0 a < x < a

0
x>a
190

9. One Dimensional Potentials

TOC

Numbering the three regions from left to right,

u1 (x) = eikx + Reikx
0

u2 (x) = Aeik x + Beik x

u3 (x) = T eikx

Again we have assumed a beam of definite momentum incident from the left
and no wave incident from the right.

Figure 38: Energy diagram for scattering from a 1D potential well with the general
solution form for a beam incident from the left. This problem displays the quantum
phenomena common to scattering, like specific energies where the scattering amplitude
goes to zero.
There are four unknown coefficients. We now match the wave function and its
first derivative at the two boundaries yielding 4 equations.
Some hard work yields the reflection and transmission amplitudes.

191

9. One Dimensional Potentials

TOC

Reflection and Transmission Amplitudes for 1D Potential Well

2kk 0

(k 02 k 2 ) sin(2k 0 a)
cos(2k 0 a) i(k 02 + k 2 ) sin(2k 0 a)

2kk 0

cos(2k 0 a)

R = ie2ika
T = e2ika

r
k=
r
0

k =

2kk 0
i(k 02 + k 2 ) sin(2k 0 a)

2mE
~2

2m(E V0 )
~2

The squares of these give the reflection and transmission probability, since the
potential is the same in the two regions.
Again, classically, everything would be transmitted because the energy is larger
than the potential. Quantum mechanically, there is a probability to be transmitted
and a probability to be reflected. The reflection probability will go to zero for
certain energies: R 0 if
2k 0 a = n
2m(E V0 )
k0 =
~2
n2 2 ~2
E = V0 +
8ma2
r

There are analogs of this in 3D. The scattering cross section often goes to zero for certain
particular energies. For example, electrons scattering off atoms may have nearly zero
cross section at some particular energy. Again this is a wave property.

9.1.4

Bound States in a Potential Well *

We will work with the same potential well as in the previous section but assume that
V0 < E < 0, making this a bound state problem. Note that this potential has a
Parity symmetry. In the left and right regions the general solution is
u(x) = Aex + Bex
192

9. One Dimensional Potentials

TOC

with

r
=

2mE
.
~2

The ex term will not be acceptable at and the ex term will not be acceptable
at + since they diverge and we could never normalize to one bound particle.
u1 (x)

C1 ex

u3 (x)

C3 ex

In the center well use the sine and cosine solutions anticipating parity eigenstates.
u2 (x) = A cos(kx) + B sin(kx)
r
2m(E + V0 )
k=
~2
Again we will have 4 equations in 4 unknown coefficients.

Figure 39: Energy diagram for bound states in a 1D potential well with the general
solution form. This problem shows how the number of bound states behaves for finite
binding potentials of limited extent.
The calculation shows that either A or B must be zero for a solution. This means that
the solutions separate into even parity and odd parity states. We could have
guessed this from the potential.

193

9. One Dimensional Potentials

TOC

The even states have the (quantization) constraint on the energy that

= tan(ka)k
!r
2mE
2m(E + V0 )
2m(E + V0 )
= tan
a
2
2
~
~
~2
!
r
r
E
2m(E + V0 )
a
= tan
E + V0
~2
r

and the odd states have the constraint

E
= cot
E + V0

= cot(ka)k
!
2m(E + V0 )
a
~2

These are transcendental equations, so we will solve them graphically. The plot below
compares the square root on the left hand side of the transcendental equations to the
tangent on the right for the even states and to -cotangent on the right for odd states.
Where the curves intersect (not including the asymptote), is an allowed energy. There
is always one even solution for the 1D potential well. In the graph shown, there are 2
even and one odd solution. The wider and deeper the well, the more solutions.

194

9. One Dimensional Potentials

TOC

Figure 40: Graphic solution to the equations for Energy eigenvalues for the 1D potential
well. The blue line represents the sqrt on the left side of both equations. It starts at
+ for E = V0 and goes to zero for E = 0. The violet line represents the right side
of the equation for parity even states. The yellow line represents the right side of the
equation for odd parity states. Allowed energies are at the intesections of one of the
RHS lines with the blue line. Since the tangent will start from zero for E = V0 and
go to some finite value for E = 0, this first tangent curve must cross the blue curve at
least once. There will always be at least one intersection for the even parity states. The
yellow line does not have to cross the blue line so there can be zero parity odd states.
The bigger the size of the well, the more solutions there will be.

Transcendental Equations for Eigen-Energies (1D Potential Well)

= tan(ka)k
= cot(ka)k

Even Parity
Odd Parity

2m(E + V0 )
~2
r
2mE
=
~2

Try this 1D Potential Applet. It allows you to vary the potential and see the eigenstates.
195

9. One Dimensional Potentials

9.1.5

TOC

The Potential Barrier

With an analysis of the Potential Barrier problem, we can understand the phenomenon
of quantum tunneling.

0
V (x) = +V0

x < a
a < x < a
x>a

Numbering the three regions from left to right,

u1 (x) = eikx + Reikx
u2 (x) = Aex + Bex
u3 (x) = T eikx
Again we assume a beam of definite momentum incident from the left and no wave
incident from the right. For the solutions outside the barrier,
r
2mE
k=
.
~2
Inside the barrier,
r
=

2m(V0 E)
.
~2

Figure 41: Energy diagram for scattering from a 1D potential barrier with the general
solution form for a beam incident from the left. This problem displays the quantum
phenomonon of tunneling.
This is actually the same as the (unbound) potential well problem with the substitution
k 0 i
196

9. One Dimensional Potentials

TOC

in the center region.

The amplitude to be transmitted is
T = e2ika

2k
.
2k cosh(2a) i(k 2 2 ) sinh(2a)

We can compute the probability to be transmitted.

Potential Barrier Transmission Probability
|T |2 =

(2k)2

(k 2 + 2 )2 sinh2 (2a) + (2k)2

r
k=

2mE
~2

4k
k 2 + 2

e4a

2m(V0 E)
~2

An approximate probability is sometimes useful.

|T | e

2(2a)

a q 2m

[V (x)E]

Classically the transmission probability would be zero. In Quantum Mechanics,

the particle is allowed to violate energy conservation for a short time and so
has a chance to tunnel through the barrier.
Tunneling can be applied to cold emission of electrons from a metal, alpha decay of
nuclei, semiconductors, and many other problems.

9.2

The 1D Harmonic Oscillator

The harmonic oscillator is an extremely important physics problem. Many

potentials look like a harmonic oscillator near their minimum. This is the first nonconstant potential for which we will solve the Schr
odinger Equation.
The harmonic oscillator Hamiltonian is given by
H=

1
p2
+ kx2
2m 2

which makes the Schr

odinger Equation for energy eigenstates.
~2 d2 u 1 2
+ kx u = Eu
2m dx2
2

197

9. One Dimensional Potentials

TOC

Note that this potential also has a Parity symmetry. The potential is unphysical
because it does not go to zero at infinity, however, it is often a very good approximation,
and this potential can be solved exactly.
It is standard to remove the spring constant k from the Hamiltonian, replacing it with
the classical oscillator frequency.
r
k
=
m
The Harmonic Oscillator Hamiltonian becomes.

p2
1
+ m 2 x2
2m 2

The differential equation to be solved is

~2 d2 u 1
+ m 2 x2 u = Eu.
2m dx2
2
To solve the Harmonic Oscillator equation, we will:
first change to dimensionless variables e.g. y =

p m
~

second find the form of the solution for y ,

third multiply that large y solution by a polynomial and plug it into the differential equation obtaining a new differential equation in h(y),
then plug in h(y) =
of the polynomial

am y m and derive a recursion relation in the coefficients

m=0
am ,

and finally show that the polynomial series must terminate if the solutions are
to be normalizable.
The energy eigenvalues are derived from the termination condition. The energy eigenfunctions are derived from the recurssion relation with the knowlege of where they
terminate.
The energy eigenvalues are
198

9. One Dimensional Potentials

TOC

En =

1
n+
~
2

for n = 0, 1, 2, .... There are a countably infinite number of solutions with equal
energy spacing. We have been forced to have quantized energies by the requirement
that the wave functions be normalizable.
The ground state wave function is.

u0 (x) =

m 14
~

emx

/2~

This is a Gaussian (minimum uncertainty) distribution. Since the HO potential has

a parity symmetry, the solutions either have even or odd parity. The ground
state is even parity.
The first excited state is an odd parity state, with a first order polynomial multiplying the same Gaussian.

u1 (x) =

m 14 r 2m
2
xemx /2~
~
~

The second excited state is even parity, with a second order polynomial multiplying
the same Gaussian.

mx2
u2 (x) = C 1 2
~

emx

/2~

Note that n is equal to the number of zeros of the wavefunction. This is a common
trend. With more zeros, a wavefunction has more curvature and hence more kinetic
energy.
The general solution can be written as
199

9. One Dimensional Potentials

TOC

un (x) =

ak y k ey

k=0

with the coefficients determined by the recursion relation

ak+2 =

2(k n)
ak
(k + 1)(k + 2)

and the dimensionless variable y given by.

r
y=

m
x
~

The series terminates with the last nonzero term having k = n.

two states is that the odd parity state has a substantially smaller probability
to be between the atoms.

9.5

The Delta Function Model of a Crystal *

The Kronig-Penny model of a solid crystal contains an infinite array of repulsive

delta functions.

X
V (x) = aV0
(x na)
n=

Our states will have positive energy.

This potential has a new symmetry, that a translation by the lattice spacing a
leaves the problem unchanged. The probability distributions must therefore have this
symmetry
|(x + a)|2 = |(x)|2 ,
which means that the wave function differs by a phase at most.
(x + a) = ei (x)
The general solution in the region (n 1)a < x < na is
n (x) = An sin(k[x na]) + Bn cos(k[x na])
r
2mE
k=
~2
By matching the boundary conditions and requiring that the probability be periodic,
we derive a constraint on k similar to the quantized energies for bound states.
Allowed Electron Energies in 1D Solid
cos() = cos(ka) +
r
k=

maV0
sin(ka)
~2 k

2mE
~2

Recall that the phase shift can be anything, however, since cos() can only take on
values between -1 and 1, there are allowed bands of k and gaps between those
bands.
204

9. One Dimensional Potentials

TOC

Figure 44: The graph shows the RHS of the above energy equation: cos(ka) +
maV0
~2 k sin(ka) as a function of k. Also shown is the energy which depends simply on
k. Since the LHS of the above equation cos must clearly be between -1 and +1, there
is no solution in some regions of k (and E). This gives rise to the quantum solid
phenomonon of energy bands. In 3D there are more symmetries but energy bands are
still the dominant phenomonon of crystaline solids.

205

9. One Dimensional Potentials

TOC

Figure 45: Energy bands as a function of lattice spacing in solids with the structure of
silicon or diamond.
This energy band phenomenon is found in solids with a periodic lattice. In 3D there
are 3 (smallest) vectors that give translations to a symmetric place in the crystal. This
larger translation symmetry gives rise to Bloch Waves as the 3D energy eigenstates.
The Bloch Wave solution is:
~
(~r) = eik~r u(~r)
~

where u(~r) is a function periodic in the lattice and the plane wave factor eik~r has real
wave vector ~k. All distinct Bloch waves occur for k-values within the first Brillouin
Zone of the reciprocal lattice. The first Brillouin zone is that nearest zero in k space.
Larger wave vectors simply reproduce the same states. The Bloch states are spread
out over the crystal and electrons exhibit very high mobility because the states already
account for the periodic array of atoms and electrons thus only scatter from impuities
of vibratonal exitations of the lattice, rather than from the atoms themselves.
Solids with partially filled bands or ovelapping bands are conductors. Solids with filled
bands and a large band gaps are insulators. Semiconductors also have a filled band
but have a small Energy gap between the last filled band and the higher energy unfilled
band. With the introduction of impurities, a small number of charge carriers (of either
charge) can be introduced into the otherwise filled band.

206

9. One Dimensional Potentials

9.6

TOC

The Quantum Rotor

It is useful to simply investigate angular momentum with just one free rotation
angle. This might be the quantum plane propeller. We will do a good job of this in 3
dimensions later.
Lets assume we have a mass m constrained to move in a circle of radius r.
Assume the motion in the circle is free, so there is no potential. The kinetic energy is
p2
d
1
2
2 mv = 2m for p = mr dt .
If we measure distance around the circle, then x = r and the one problem we have is
that once I go completely around the circle, I am back to x = 0. Lets just go ahead
and write our wavefunction.
ei(pxEt)/~ = ei(prEt)/~
Remembering angular momentum, lets call the combination pr = L. Our wave is
ei(LEt)/~ .
This must be single valued so we need to require that
ei(2LEt)/~ = ei(0Et)/~
ei(2L)/~ = 1
L = n~
n = 0, 1, 2, 3...

The angular momentum must be quantized in units of ~.

This will prove to be true for 3 dimensions too, however, the 3 components of angular
momentum do not commute with each other, leading to all kinds of fun.

9.7
9.7.1

Derivations and Computations

Probability Flux for the Potential Step *

The probability flux is given by

j(x, t) =

.
2mi
x
x

207

9. One Dimensional Potentials

TOC

We can save some effort by noticing that this contains an expression minus its complex
conjugate. (This assures that term in brackets is imaginary and the flux is then real.)

~
du du
~
du
j=
u

u =
u
CC
2im
dx
dx
2im
dx
For x < 0
j
j
j

~
[(eikx + R eikx )(ikeikx ikReikx ) CC]
2im
i~k
=
[1 Re2ikx + R e2ikx R R] + CC
2im
~k
= [1 |R|2 ] .
m

The probability to be reflected is the reflected flux divided by the incident flux. In this
case its easy to see that its |R|2 as we said. For x > 0
j = |T |2

~k 0
.
m

The probability to be transmitted is the transmitted flux divided by the incident flux.

Transmission Probability Computed from Flux

Ptrans = |T |2

~k 0 m
4k 2 k 0
4kk 0
=
=
m ~k
(k + k 0 )2 k
(k + k 0 )2

again as we had calculated earlier.

9.7.2

Scattering from a 1D Potential Well *

0
V (x) = V0

x < a
a < x < a
x>a

Numbering the three regions from left to right,

u1 (x) = eikx + Reikx
0

u2 (x) = Aeik x + Beik x

u3 (x) = T eikx
208

9. One Dimensional Potentials

TOC

Again we have assumed no wave incident from the right (but we could add that solution
if we wanted).
We now match the wave function and its first derivative at the two boundaries yielding
4 equations. Thats good since we have 4 constants to determine. At x = a we have 2
equations which we can use to eliminate A and B.
0

T eika = Aeik a + Beik a

ikT eika = ik 0 Aeik a ik 0 Beik a

0
0
k ika
T e = Aeik a Beik a
k0

k
1 ika
ik0 a
1+ 0
Ae
= Te
2
k

1 ika
k
ik0 a
Be
= Te
1 0
2
k
At x = a we have 2 equations which can now be written in terms of R and T by
using the above.
0

eika + Reika = Aeik a + Beik a

ikeika ikReika = ik 0 Aeik a ik 0 Beik a

0
0
k
1
k
eika + Reika = T eika 1 + 0 e2ik a + 1 0 e2ik a
2
k
k

0
0
1
k
k
keika kReika = T eika k 0 1 + 0 e2ik a 1 0 e2ik a
2
k
k
0

0

0
0
1
k
k
eika Reika = T eika
+ 1 e2ik a
1 e2ik a
2
k
k
We can add equations to eliminate R.

0
0
k
1
k
eika + Reika = T eika 1 + 0 e2ik a + 1 0 e2ik a
2
k
k
0

0

0
0
1
k
k
eika Reika = T eika
+ 1 e2ik a
1 e2ik a
2
k
k

0
0
0
0
1
k
k
k
k
2eika = T eika 2 +
+ 0 e2ik a + 2
0 e2ik a
2
k
k
k
k

0
k
k
0
4e2ika = T 4 cos(2k 0 a) 2i
+
sin(2k
a)
k0
k
T =

2e2ika
2 cos(2k 0 a) i kk0 +

k0
k

sin(2k 0 a)

0 2ika

T =

2kk 0

2kk e
i (k 2 + k 02 ) sin(2k 0 a)

cos(2k 0 a)

209

9. One Dimensional Potentials

TOC

We can subtract the same equations to most easily solve for R.

0
1
k0
k
k
k
2ik0 a
2ik0 a
+
2Reika = T eika

e
e

2
k0
k
k
k0

1
k0
k
R = T 2i 0 sin(2k 0 a) + 2i sin(2k 0 a)
4
k
k
0

i
k
k
R = T sin(2k 0 a)
0
2
k
k
h 0
i
ikk 0 e2ika sin(2k 0 a) kk kk0
R=
2kk 0 cos(2k 0 a) i (k 2 + k 02 ) sin(2k 0 a)

i k 02 k 2 sin(2k 0 a)e2ika
R=
2kk 0 cos(2k 0 a) i (k 2 + k 02 ) sin(2k 0 a)

We have solved the boundary condition equations to find the reflection and transmission
amplitudes.
Reflection and Transmission Amplitudes for a 1D Potential Well
R = ie2ika
T = e2ika
r
k=

(k 02 k 2 ) sin(2k 0 a)
2kk 0 cos(2k 0 a) i(k 02 + k 2 ) sin(2k 0 a)

2kk 0
2kk 0 cos(2k 0 a) i(k 02 + k 2 ) sin(2k 0 a)

2mE
~2

r
0

k =

2m(E + V0 )
~2

The squares of these give the reflection and transmission probability, since the potential
is the same in the two regions.

9.7.3

Bound States of a 1D Potential Well *

In the two outer regions we have solutions

u1 (x) = C1 ex
u3 (x) = C3 ex
r
2mE
=
.
~2
210

9. One Dimensional Potentials

TOC

In the center we have the same solution as before.

u2 (x) = A cos(kx) + B sin(kx)
r
2m(E + V0 )
k=
~2
(Note that we have switched from k 0 to k for economy.) We will have 4 equations in 4
unknown coefficients.
At a we get
C1 ea = A cos(ka) B sin(ka)
C1 ea = kA sin(ka) + kB cos(ka).

At a we get
C3 ea = A cos(ka) + B sin(ka)
C3 ea = kA sin(ka) + kB cos(ka).

Divide these two pairs of equations to get two expressions for .

=
=

kA sin(ka) + kB cos(ka)
A cos(ka) B sin(ka)
kA sin(ka) + kB cos(ka)
A cos(ka) + B sin(ka)

Factoring out the k, we have two expressions for the same quantity.
A sin(ka) + B cos(ka)

=
k
A cos(ka) B sin(ka)

A sin(ka) B cos(ka)
=
k
A cos(ka) + B sin(ka)

If we equate the two expressions,

A sin(ka) + B cos(ka)
A sin(ka) B cos(ka)
=
A cos(ka) B sin(ka)
A cos(ka) + B sin(ka)
211

9. One Dimensional Potentials

TOC

and cross multiply, we have

(A sin(ka) + B cos(ka))(A cos(ka) + B sin(ka))
= (A cos(ka) B sin(ka))(A sin(ka) B cos(ka)).

The A2 and B 2 terms show up on both sides of the equation and cancel. Whats left is
AB(sin2 (ka) + cos2 (ka))

= AB( cos2 (ka) sin2 (ka))

= AB

Either A or B, but not both, must be zero. We have parity eigenstates, again, derived
from the solutions and boundary conditions.
This means that the states separate into even parity and odd parity states. We could
have guessed this from the potential.
Now lets use one equation.
=

A sin(ka) + B cos(ka)
k
A cos(ka) B sin(ka)

k If we set B = 0, the even states have the constraint on the energy that
= tan(ka)k
and, if we set A = 0, the odd states have the constraint
= cot(ka)k.

9.7.4

Solving the HO Differential Equation *

The differential equation for the 1D Harmonic Oscillator is.

~2 d2 u 1
+ m 2 x2 u = Eu.
2m dx2
2
By working with dimensionless variables and constants, we can see the basic
equation and minimize the clutter. We use the energy in terms of ~.
=

2E
~

We define a dimensionless coordinate.

r
y=

m
x
~

212

9. One Dimensional Potentials

TOC

The equation becomes.

d2 u 2m
1
+ 2 (E m 2 x2 )u = 0
dx2
~
2
d2 u
+ ( y 2 )u = 0
dy 2

(Its probably easiest to just check the above equation by substituting as below.

2E
~ d2 u
m 2
+

x
u=0
m dx2
~
~
d2 u 2m
1
+ 2 (E m 2 x2 )u = 0
2
dx
~
2
It works.)
Now we want to find the solution for y .
d2 u
+ ( y 2 )u = 0
dy 2
becomes

d2 u
y2 u = 0
dy 2

which has the solution (in the large y limit)

u = ey

This exponential will dominate a polynomial as y so we can write our general

solution as
2
u(y) = h(y)ey /2
where h(y) is a polynomial.
Take the differential equation
d2 u
+ ( y 2 )u = 0
dy 2
and plug
u(y) = h(y)ey

213

9. One Dimensional Potentials

TOC

into it to get
2
2
2
d2
h(y)ey /2 + h(y)ey /2 y 2 h(y)ey /2 = 0
dy 2
2
2
d2 h(y) y2 /2
dh(y) y2 /2
e
2
ye
h(y)ey /2 + h(y)y 2 ey /2
2
dy
dy

+h(y)ey

y 2 h(y)ey

dh(y)
d h(y)
2y
h(y) + y 2 h(y) + h(y) y 2 h(y) = 0
dy 2
dy
d2 h(y)
dh(y)
2y
+ ( 1)h(y) = 0
dy 2
dy
This is our differential equation for the polynomial h(y).
Write h(y) as a sum of terms.
h(y) =

9. One Dimensional Potentials

TOC

But, lets see what we have. For large m,

am+2 =

2m + 1
2
am am
(m + 1)(m + 2)
m

The series for

y 2 ey
has the coefficient of y
If m = 2n,

2n+2

equal to

1
2n n!

am+2 =

X y 2n+2
2n n!

and the coefficient of y 2n equal to

1
2n1 (n1)! .

1
1
am = am .
2n
m
2

So our polynomial solution will approach y 2 ey /2 and our overall solution

2
will not be normalizable. (Remember u(y) = h(y)ey /2 .) We must avoid this.
We can avoid the problem if the series terminates and does not go on to infinite
m.
2m + 1
am
am+2 =
(m + 1)(m + 2)
The series will terminate if
= 2n + 1
for some value of n. Then the last term in the series will be of order n.
an+2 =

0
an = 0
(n + 1)(n + 2)

The acceptable solutions then satisfy the requirement

2E
= 2n + 1
~

(2n + 1)
1
E=
~ = n +
~
2
2
=

Again, we get quantized energies when we satisfy the boundary conditions at

infinity.
We can summarize the results for the solutions that were obtained here.

215

9. One Dimensional Potentials

TOC

Harmonic Oscillator Solutions from Diff. Eq.

un (x) =

ak y k ey

k=0

r
y=

ak+2 =

m
x
~

2(k n)
ak
(k + 1)(k + 2)

The ground state wavefunction is particularly simple, having only one term.
u0 (x) = a0 e

y 2
2

= a0 emx

/2~

Lets find a0 by normalizing the wavefunction.

|a0 |2 emx

dy = 1

r
2

|a0 |

~
=1
m

Harmonic Oscillator Ground State

u0 (x) =

9.7.5

m 14
~

emx

/2~

1D Model of a Molecule Derivation *

ex
x < d
A (ex + ex ) d < x < d

ex
x>d
r
2mE
.
=
~2
Since the solution is designed to be symmetric about x = 0, the boundary conditions
at d are the same as at d. The boundary conditions determine the constant A and
constrain .
216
(x) =

9. One Dimensional Potentials

TOC

Continuity of gives.
ed = A ed + ed
A=

ed
ed + ed

The discontinuity in the first derivative of at x = d is

2maV0 d
ed A ed ed =
e
~2
ed ed
2maV0
1 A
=
ed
~2
d
d
e e
2maV0
1 d
=
e + ed
~2
d
2maV0
e ed
=
1
+
~2
ed + ed
2maV0
= 1 + tanh(d)
~2
Well need to study this transcendental equation to see what the allowed energies are.

9.7.6

1D Model of a Crystal Derivation *

We are working with the periodic potential

V (x) = aV0

(x na).

Our states have positive energy. This potential has the symmetry that a translation
by the lattice spacing a leaves the problem unchanged. The probability distributions
must therefore have this symmetry
|(x + a)|2 = |(x)|2 ,
which means that the wave function differs by a phase at most.
(x + a) = ei (x)
The general solution in the region (n 1)a < x < na is
n (x) = An sin(k[x na]) + Bn cos(k[x na])
217

9. One Dimensional Potentials

TOC

r
k=

2mE
~2

Now lets look at the boundary conditions at x = na. Continuity of the wave function
gives
n (na)
An sin(0) + Bn cos(0)
Bn
Bn+1

= n+1 (na)
=

An+1 sin(ka) + Bn+1 cos(ka)

= An+1 sin(ka) + Bn+1 cos(ka)

Bn + An+1 sin(ka)
=
.
cos(ka)

The discontinuity in the first derivative is

dn+1
dn
2maV0

=
n (na)
dx na
dx na
~2
2maV0
k[An+1 cos(ka) Bn+1 sin(ka)] k[An cos(0) Bn sin(0)] =
Bn
~2
2maV0
Bn
k[An+1 cos(ka) + Bn+1 sin(ka) An ] =
~2

Substituting Bn+1 from the first equation

2maV0
Bn
~2
2maV0
An+1 (cos(ka) + sin(ka) tan(ka)) + Bn tan(ka) An =
Bn
~2 k
cos2 (ka) + sin2 (ka)
2maV0
An+1 =
Bn Bn tan(ka) + An
cos(ka)
~2 k
2maV0
An+1 =
Bn cos(ka) Bn sin(ka) + An cos(ka)
~2 k
k[An+1 cos(ka) + [Bn + An+1 sin(ka)] tan(ka) An ] =

218

9. One Dimensional Potentials

TOC

Plugging this equation for An+1 back into the equation above for Bn+1 we get
Bn+1

Bn+1

Bn + An+1 sin(ka)
cos(ka)
2maV0
~2 k B n

cos(ka) Bn sin(ka) + An cos(ka) sin(ka)
cos(ka)

Bn
2maV0
sin2 (ka)
+
Bn sin(ka) Bn
+ An sin(ka)
cos(ka)
~2 k
cos(ka)

Bn
2maV0
1
+
Bn sin(ka) Bn
cos(ka) + An sin(ka)
cos(ka)
~2 k
cos(ka)
2maV0
Bn sin(ka) + Bn cos(ka) + An sin(ka).
~2 k
Bn +

We now have two pairs of equations for the n+1 coefficients in terms of the n coefficients.

An+1

2maV0
Bn cos(ka) Bn sin(ka) + An cos(ka)
~2 k
2maV0
=
Bn sin(ka) + Bn cos(ka) + An sin(ka)
~2 k
i
= e An

Bn+1

= ei Bn

An+1
Bn+1

Using the second pair of equations to eliminate the n + 1 coefficients, we have

2maV0
cos(ka)

sin(ka)
Bn
(ei cos(ka))An =
~2 k

2maV0
ei cos(ka)
sin(ka)
Bn = sin(ka)An .
~2 k

219

9. One Dimensional Potentials

TOC

Now we can eliminate all the coefficients.

2maV0
(ei cos(ka))(ei cos(ka)
sin(ka))
~2 k

2maV0
cos(ka) sin(ka) sin(ka)
=
~2 k

2maV0
2i
i
sin(ka) + cos(ka) + cos(ka)
e e
~2 k
2maV0
+ 2
sin(ka) cos(ka) + cos2 (ka)
~ k
2maV0
=
sin(ka) cos(ka) sin2 (ka)
2k
~

2maV0
e2i ei
sin(ka)
+
2
cos(ka)
+1=0
~2 k
Multiply by ei .
ei + ei

2maV0
sin(ka)
+
2
cos(ka)
=0
~2 k
maV0
cos() = cos(ka) + 2 sin(ka)
~ k

This relation puts constraints on k, like the constraints that give us quantized
energies for bound states. Recall that phi is the phase shift in the wave function when
translated to the nearest symmetric point, and can take on any value. Since cos() can
only take on values between -1 and 1, there are allowed bands of k and gaps between
those bands.

9.8

Examples

This whole section is examples.

9.9

Homework

1. The odd bound state solution to the potential well problem bears many similarities to the zero angular momentum solution to the 3D spherical potential well.
Assume the range of the potential is 2.3 1013 cm, the binding energy is -2.9
MeV, and the mass of the particle is 940 MeV. Find the depth of the potential
in MeV. (The equation to solve is transcendental.)
220

9. One Dimensional Potentials

TOC

= cot(ka)k

q
q
2mc2 (E+V0 )
E
=

cot
a
E+V0
~2 c2
E = 1.1
mc2= 940
a = 2.7 fm

q
q
2(940)(V0 1.1)(2.7)
1.1
=

cot
V0 1.1
197.32

2(940)(V0 1.1)(2.7)
1.1 = 1.049 = V0 1.1 cot
197.3

Vary V0 to get right energy

V0 = 25.41

gives solution

2. Find the three lowest energy wave-functions for the harmonic oscillator.
p
y 2
u0 = a0 e 2
y = m
for u0 , only a0 , normalize
~ x
1

m 4 mx2 /2~
u0 = ~
e
y 2

u1 = a1 ye 2
2 ax2
p
1
x e
dx = 2a
4 ax2
pa
x e
dx = 4a32 a
a = m
~
41 q 2m mx2 /2~
u1 = m
~
~ xe
a2 = 2a0
2 mx2 /2~
u2 =a0 (1 2 m
~ x )e
mx2 /~

m 4 2
2
a20
1 4 m
x
+
4
x
e
dx
~
q
~

2
2
~
~
m
3~
1 4 m
= a20 m
~ 2m + 4
~
4m2 2
q
q
~
~
= a20 m
(1 2 + 3) = a20 2 m
1
2
m 4
u2 = 4~
(1 2 m
x2 )emx /2~
~
1
m 4
2 mx2 /2~
u2 = 4~
(1 + 2 m
~ x )e

for u1 , only a1 , normalize

compute from formula sheet
compute from formula sheet
read of a
for u2 , a0 and a2 , normalize

Hermite polynomial convention

3. Assume the potential for particle bound inside a nucleus is given by

(
V0
x<R
V (x) = ~2 `(`+1)
x>R
2mx2
and that the particle has mass m and energy E > 0. Estimate the lifetime of the
particle inside this potential.
|T |2 e2(2a) = e
2

|T |2 e
2

a q 2m

~2 `(`+1)
E]
2mx2

[V (x)E] dx

from pot. barrier

~ `(`+1)
2ma2 2 = E
`(`+1)
2
a = ~ 2mE
q
a q ~2 `(`+1)
2 2mE
[ 2mEx2 1] dx
~2
2
R

|T | e

limits from problem

E = V at x = a
solve for a2
factor out const.
221

9. One Dimensional Potentials

|T |2 e

TOC

q
a q a2
2 2mE
1 dx
~2
x2

use a to simplify

q
a a2 x2
2 2mE
dx
x
~2

|T |2 e
a2 x2 dx

to integral found in table

a+ a2 x2
=

a
ln

x
h

iax

2 x2
a+
a
=
a2 x2 a ln
x
R

2
2
2
2
= a R a ln (1) + a ln a+ aR R

2
2
= a2 R2 + a ln a+ aR R
|T |2 e

int. table; for |x| |a|

R < x < a this problem
evaluate
algebra

2
2
2 R2 +a ln a+ a R
2 2mE

a
2
R
~
q

8mE(a2 R2 )
~2

|T |2 e
|T |2 e

8mEa2
~2

plug integral

a+ a2 R2
ln
R

a+ a2 R2
R

split sum in exponent

8mEa2
~2

algebra

=
q

R 2m
V0
|T |2

1
2R
2Rm 2 =
v|T |2 /2R = v|T |2 =
2mV0 |T |
q
q
2

8mEa
8mE(a2 R2 )
2 R2
~2
a+
a
2m
2
~
e
V0
R

decay rate=prob. times attemps p

=R
plug in T 2
This is the answer that was requested but lets do an example to see if it makes
sense. Consider a 1 MeV ` = 2 neutron in a nucleus with R = 1 fm.
s
r
~2 c2 `(` + 1)
6
a=
= 197.3
= 11 f m
2mc2 E
2(940)(1)
Then we can compute |T |2 to get an idea is this seems correct.
|T |2 e4.81 (131)4.83 = 7.3 109
Lets assume the potential inside the nucleus, where the
neutron is bouncing
back
q
q
R

2mc2

(1)

2(939)

50
and forth, is -50 MeV. Then the lifetime is = c|T |V20 = 31023 (7.310
9 ) =
15
23
2.8 10
seconds. (Note that we have used c = 3 10 fm/s.)

4. The 1D model of a crystal puts the following constraint on the wave number k.
cos() = cos(ka) +
2

ma2 V0 sin(ka)
~2
ka

Assume that ma~2V0 = 3

2 and plot the constraint as a function of ka. Plot the
allowed energy bands on an energy axis assuming V0 = 2 eV and the spacing
between atoms is 5 Angstroms.
(~c)2 2
E = 2mc
the Energy
2k
2
| cos(ka) + ma~2V0 sin(ka)
|
<
1
the
constraint
ka
3 sin(ka)
| cos(ka) + 2 ka | < 1
plot this vs ka
| cos(ka) + 6.56 sin(ka)
|
<
1
plot
this vs ka too and find energy bands
ka
222

9. One Dimensional Potentials

TOC

5. In a 1D square well, there is always at least one bound state. Assume the width
of the square well is a. By the uncertainty principle, the kinetic energy of an
~2
electron localized to that width is 2ma
2 . How can there be a bound state even
for small values of V0 ?
It is not true that the wavefunction will be localized to the width of the square
well. It will extend beyond the size of the well, decaying exponentially. The
square well will contribute negatively to the potential energy and by making x
large enough, the kinetic energy does not have to grow much so it is possible to
always have a bound state with slightly negative energy.
6. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =
0) = 12 (u0 + u1 ). Is correctly normalized? Compute the expected value of x
as a function of time by doing the integrals in the x representation.
u(x) are real
(t) = 12 ei5t/2 (u1 (x)eit + u2 (x))

h(t)|x|(t)i = 12
u1 xu2 eit + u2 xu1 eit dx odd functions give 0

h(t)|x|(t)i = cos(t)
u1 xu2 dx
simplify to one integral

14 q 2m mx2 /2~ m 14

2 mx2 /2~
dx
(1 + 2 m
= cos(t)
x m
~
~ xe
4~
~ x )e

4 mx2 /~
= cos(t) 1 m
(x2 + 2 m
dx
~
~ x )e

q
q
~
3~2
= cos(t) 1 m
( ~ + 2 m
~ 4m2 2 )
q ~ q m 2m
~
~
3~
= cos(t) 1 m
m ( 2m + 2m )
q ~
~
= cos(t) m

plug

7. A particle is in the first excited state of a box of length L. What is that state?
Now one wall of the box is suddenly moved outward so that the new box has
length D > L. What is the probability for the particle to be in the ground state
of the new box? What is the probability for the particle to be in the first excited
state of the new box? You may find it useful to know that

sin ((A B)x) sin ((A + B)x)

.
sin(Ax) sin(Bx)dx =
2(A B)
2(A + B)

This answer is an example of the calculation for D = 3L, which shows the method
(L)
2
initial state in box of le
(3L)
1
ground state of new bo
and allows us to get a numerical result as well. 2(3L)
first excited state of ne
(3L) (L)
h1 |2 i
amplitude to be in grou
(3L) (L)
|h1 |2 i|2 probability to be in gro

223

9. One Dimensional Potentials

TOC

2
q
L q

2
x
2
Pg.s. =
sin 3L
sin 2x
dx
3L
L
L
0

2

L
q q
7
5
x)
sin( 3L
2
2 sin( 3L x)

Pg.s. = 3L

5
7
L

2 3L
2 3L

0

2

sin( 5
sin( 7 )
3 )

Pg.s. = 23L
2 73
5

2 3L
3L
q

2
2

1
Pg.s. = 3L 34 21

5
7
2 3L
1 3L 3L

2
3L

Pg.s. = 2L 5 + 7
3 1 1 2
Pg.s. = 2
+
36 25 7

Pg.s. = 70 = 0.0268
Pf irst = 0.032

initial wfn is zero for x > L

use integral given

algebra is similar

8. A particle is initially in the nth eigenstate of a box of length L. Suddenly the

walls of the box are completely removed. Calculate the probability to find that
the particle has momentum between p and p + dp. Is energy conserved?
L ipx/~
1
e
n dx
F.T.
(p) = 2~
0

L = 2a
q

p = ~kq
k0 =

2
nx
2
n = L sin L = L sin(k0 x)
L ikx
1
(p) = ~L
e
sin(k0 x)dx

n
L

define some variables

the initial state

(p) =
(p) =

1
~L

L
0

1
2i ~L

eikx e

ik0 x

eik0 x
dx
2i

eikx eik0 x eik0 x dx

limits form 0 to L implies...

0
L

ei(kk0 )x ei(k+k0 )x dx not function
0 i(kk )x

i(k+k0 )x L

0
1
(p) = 2~L e kk0 e k+k0
do the integral
0

i(kk0 )L
i(k+k0 )L
1
1e
1e k+k0
(p) = 2~L
kk0

n ikL
1(1) e
1+(1)n eikL
1
eik0 L = ein = sin(n) = (1)n
(p) = 2~L

k+k0
0
kkika

n ika
e
)
(eika +(1)n eika )
1
(p) = 2~L
eika (e (1)

let a = L2
kk0
k+k0
ika

n ika
e
)
1
(p) = 2~L
2k0 eika (e (1)
combine fractions
k2 k02

2
kL
k0 cos( 2 )
4
P (p)dp = ~L
dp
for n odd
k2 k02

2
kL
k0 sin( 2 )
4
P (p)dp = ~L
dp
for n even
k2 k02
Before the walls were removed, the particle had a definite energy. Afterward,
there is a spread, so the energy distribution has changed.
(p) =

1
2i ~L

9. A particle is in a box with solid walls at x = a2 . The state at t = 0 is constant

224

9. One Dimensional Potentials

TOC

q
(x, 0) = a2 for a4 < x < a4 and the (x, 0) = 0 everywhere else. Write this
state as a sum of energy eigenstates of the particle in a box. Write (x, t) in
terms of the energy eigenstates. Write the state at t = 0 as (p). Would it be
correct (and why) to use (p) to compute (x, t)?
So that we can use one formula, work the same problem with walls at 0 and L.
3L/4

3L/4

n = hn |i = 2
sin( nx )dx = 2 L cos( nx )
n from dot prod.
L

L/4
cos( 3n
4 )

L n

L/4

cos( n
4 )

2
n = n
n = 0
n = (1)(n1)/2 n42
q

P
n(x+ L
2)
)eiEn t/~
(x, t) = L2
(1)(n1)/2 n42 sin(
L

evaluate integral
n even (odd parity)
n odd, (even parity)

n(odd)

The expected energy of this state goes to infinity since the energy of the nth eigenstate goes like n2 and the probability goes like n12 . I expect the infinity because
of the discontinuity in the derivative of the initial state.
The F.T. of a square wavepacket has been done in the notes
No, you cant do the time time dependence with the F.T. since the momentum
eigenstates are not energy eigenstates of the box.
10. Find the correctly normalized energy eigenfunction u5 (x) for the 1D harmonic
oscillator.
We can do the similar problem of u4 (y) which also has 3 terms.
u4 has 3 terms, a0 , a2 , a4
2(kn)
ak+2 = (k+1)(k+2)
ak
recursion relation
a2 =
a4 =

2(04)
(0+1)(0+2) a0
2(24)
(2+1)(2+2) a2

un = hn (y)e

= 4a0

k=0

13 a2

k=2

y 2
2

= 34 a0
p m
~ x

definitions from notes

y
u4q
= a0 (1 4y + 43 y 4 )e 2

2
~
a20 m
(1 4y 2 + 43 y 4 )2 ey
q

dy =

a20

a1/2

q m

1 3/2
a
2

(1 8 12 +

83
34

~ 24
1
m ( 9 ) =q

m 1/4
3
a0 = ~
8
q

m 1/4
3
u4 = ~
8 (1

a20

+ 16 43

plug in an

dy = 1

~
a20 m
(1 8y 2 + 83 y 4 + 16y 4
The integrals are:

p m

32 6
3 y

tranform x to y, norma
+

16 8 y
9 y )e

3 5/2
a
4
32 15
3 8

16 105
9 16 )

dy = 1

15 7/2
a
8
=1

square term
105 9/2

a
16

plug integrals
add terms
solve for norm. const.

4y 2 + 43 y 4 )e

y 2
2

write normalized u4
225

9. One Dimensional Potentials

TOC

11. A beam of particles of energy E > 0 coming from is incident upon a double
delta function potential in one dimension. That is V (x) = (x + a) (x a).
a) Find the solution to the Schr
odinger equation for this problem.
b) Determine the coefficients needed to satisfy the boundary conditions.
c) Calculate the probability for a particle in the beam to be reflected by the
potential and the probability to be transmitted.
ikx
e + Reikx x < b/2
= Aeikx + Beikx |x| < b/2

T eikx
x > b/2
ikb/2
for economy of notation
Let = e
= 2m
~2
d
2m
dx = ~2 (a)
formula
1/ + R = A/ + B
continity at -b/2
1 + R2 = A + B2
(1) continity at -b/2
T 2 = A2 + B
(2) continuity at b/2
ik((A B2 ) (1 R2 )) = 2m
(A
+ B2 ) d/dx at -b/2
2
~
2m
2
2
2
ik(T (A B)) = ~2 T
d/dx at b/2

T 2 A2 + B = ik
T 2
(4)
1 + R2 = A + B2
(1) of 4 eq
A2 + B =T 2
(2) of 4 eq

A 1+
1 R2 = 1 ik

2
A B = 1 + ik T 2

A2 = 1 + 2ik
T 2

2
B = 2ik
T

1 = 1 2ik A 2ik
B2

R2 = 2ik
A + 1 + 2ik
B2

1=
T =

4
1 2ik
1 + 2ik
T + 2ik
2ik T
1
1
=
2
2
2 2ikb
1+( 2k
1+( 2k
) (14 )
) ( 2k
)e

(3) of 4 eq
(4) of 4 eq
(2)+(4)
(2)-(4)
(1)+(3)
(1)-(3)

plug (2+4) and (2-4) into (1)+(3)

solve for T

226

9. One Dimensional Potentials

TOC

2
1 + 2ik
T /2 1 + 2ik
2ik T

2
2
2( )(1+
( )(1+ 2ik )(1/ )
) sin(kb)
R = 2ik
= 2k 2 2ik 2 2ikb
2
1+( 2k
1+( 2k ) ( 2k ) e
) (14 )
1

|T |2 =
2 2
2
2
4
1+( 2k
) +( 2k
) ( 2k
) 1+( 2k
) (e2ikb +e2ikb )
1

|T |2 =
2 2
2
2
4
1+( 2k
) +( 2k
) 2( 2k
) 1+( 2k
) cos(2kb)
1

|T |2 =
2 2
2
4
2
1+( 2k
) +( 2k
) 2( 2k
) 1+( 2k
) (12 sin2 (kb))
1

|T |2 =
2
2
2
2 2
1+( 2k
) ( 2k
) +4( 2k
) 1+( 2k
) sin2 (kb)
1

|T |2 =
2
2
1+4( 2k
) 1+( 2k
) sin2 (kb)
2
2
4( 2k
) 1+( 2k
) sin2 (kb)

|R|2 =
2
2
1+4( 2k ) 1+( 2k ) sin2 (kb)
|R|2 + |T |2 = 1
R=

9.10

2ik

solve for R

simplified

it checkes!

Sample Test Problems

m
p
x + i
2~
2m~

and its Hermitian conjugate A .

H = ~ A A +
2

We will use the commutators between A, A and H to solve the HO problem.

[A, A ] = 1

The commutators with the Hamiltonian are easily computed.

[H, A]

[H, A ]

= ~A
= ~A

From these commutators we can show that A is a raising operator for Harmonic
Oscillator states
230

10. Harmonic Oscillator Solution using Operators

A u n =

TOC

n + 1un+1

and that A is a lowering operator.

Aun =

nun1

Because the lowering must stop at a ground state with positive energy, we can show
that the allowed energies are

1
En = n +
~.
2
The actual wavefunctions can be deduced by using the differential operators for A and
A , but often it is more useful to define the nth eigenstate in terms of the ground state
and raising operators.

1
un = (A )n u0
n!

Almost any calculation of interest can be done without actual functions since we
can express the operators for position and momentum.
r
~
x =
(A + A )
2m
r
m~
p = i
(A A )
2

10.1

Introducing A and A

The Hamiltonian for the 1D Harmonic Oscillator

p2
1
+ m 2 x2
2m 2

can be rewritten in terms of the operator A

231

10. Harmonic Oscillator Solution using Operators

TOC

Definition of HO Lowering Operator

m
p
x + i
2~
2m~

m
p
x i
2~
2m~

and its Hermitian conjugate

A =

Both terms in the Harmonic Oscillator Hamiltonian are squares of operators. Note that
A is chosen so that A A is close to the Hamiltonian. First just compute the quantity
A A =
A A =
~(A A)

m 2
p2
i
x +
+
(xp px)
2~
2m~ 2~
m 2
p2
i
x +

[p, x]
2~
2m~ 2~
p2
1
1
+ m 2 x2 ~.
2m 2
2

From this we can see that the Hamiltonian can be written in terms of A A and
some constants.
HO Hamiltonian in Terms of Operators
1
p2
+ m 2 x2
2m 2

1

H = ~ A A +
2
H=

10.2

Commutators of A, A and H

We will use the commutator between A and A to solve the HO problem. The operators
are defined to be
r

m
p
A =
x + i
2~
2m~
r

m
p

A =
x i
.
2~
2m~
232

10. Harmonic Oscillator Solution using Operators

TOC

The commutator is
[A, A ]

=
=

m
1
i
i
[x, x] +
[p, p]
[x, p] +
[p, x]
2~
2m~
2~
2~
i
i
([x, p] + [p, x]) = [p, x] = 1.
2~
~

Lets use this simple commutator

The Commutator of A and A
[A, A ] = 1

to compute commutators with the Hamiltonian. This is easy if H is written in

terms of A and A .
[H, A]

~[A A, A] = ~[A , A]A = ~A

[H, A ]

~[A A, A ] = ~A [A, A ] = ~A

The Commutators of H with A and A

[H, A] = ~A
[H, A ] = ~A

10.3

Use Commutators to Derive HO Energies

We have computed the commutators

[H, A]

[H, A ]

= ~A
= ~A

233

10. Harmonic Oscillator Solution using Operators

TOC

Apply [H, A] to the energy eigenfunction un .

[H, A]un = ~Aun
HAun AHun = ~Aun
H(Aun ) En (Aun ) = ~Aun
H(Aun ) = (En ~)(Aun )

This equation shows that Aun is an eigenfunction of H with eigenvalue En ~.

Therefore,A lowers the energy by ~.
Now, apply [H, A ] to the energy eigenfunction un .
[H, A ]un = ~A un
HA un A Hun = ~A un
H(A un ) En (A un ) = ~(A un )
H(A un ) = (En + ~)(A un )
A un is an eigenfunction of H with eigenvalue En + ~. A raises the energy by
~.
Raising and Lowering Gives Energy Eigenstates
H(A un ) = (En + ~)(A un )
H(Aun ) = (En ~)(Aun )

We cannot keep lowering the energy because the HO energy cannot go below
zero.
1
1
h|H|i =
hp |p i + m 2 hx |x i 0
2m
2
The only way to stop the lowering operator from taking the energy negative, is for the
lowering to give zero for the wave function. Because this will be at the lowest energy,
this must happen for the ground state. When we lower the ground state, we
must get zero.

Lowering the Ground State Gives Zero

Au0 = 0
234

10. Harmonic Oscillator Solution using Operators

TOC

Since the Hamiltonian contains A in a convenient place, we can deduce the ground
state energy.
1
1
Hu0 = ~(A A + )u0 = ~u0
2
2
1
The ground state energy is E0 = 2 ~ and the states in general have energies

1
E = n+
~
2
since we have shown raising and lowering in steps of ~. Only a state with energy
E0 = 12 ~ can stop the lowering so the only energies allowed are:
HO Energies Derived Using Operators

E=

1
2

It is interesting to note that we have a number operator for n

1
H =
A A +
~
2

10.3.1

Nop

A A

1
(Nop + )~
2

Raising and Lowering Constants

We know that A raises the energy of an eigenstate but we do not know what
coefficient it produces in front of the new state.
A un = Cun+1
We can compute the coefficient using our operators.
|C|2

= hA un |A un i = hAA un |un i
= h(A A + [A, A ])un |un i = (n + 1)hun |un i = n + 1

The effect of the raising operator is

A un =

n + 1un+1 .
235

10. Harmonic Oscillator Solution using Operators

TOC

Similarly, the effect of the lowering operator is

Aun = nun1 .
These are extremely important equations for any computation in the HO problem.

HO Raising and Lowering Operators

A un =

Aun =

n + 1 un+1

n un1

We can also write any energy eigenstate in terms of the ground state and
the raising operator.
1
un = (A )n u0
n!

10.4

Expectation Values of p and x

It is important to realize that we can just use the definition of A to write x and p in
terms of the raising and lowering operators.
Write x and p Operators in Terms of A and A
r

~
(A + A )
2m
r
m~
p = i
(A A )
2
x=

This will allow for any computation.

Example: The expectation value of x for any energy eigenstate is zero.
Example: The expectation value of p for any energy eigenstate is zero.
Example: The expectation value of x in the state 12 (u0 + u1 ).
Example:
The expectation value of 12 m 2 x2 for any energy eigenstate is

1
1
2 n + 2 ~.

p2
Example: The expectation value of 2m
for any energy eigenstate is 21 n + 12 ~.
236

10. Harmonic Oscillator Solution using Operators

Example:
(t = 0) =

10.5

TOC

The expectation
value of p as a function of time for the state
+ u2 ) is m~ sin(t).

1 (u1
2

The Wavefunction for the HO Ground State

The equation
Au0 = 0
can be used to find the ground state wavefunction. Write A in terms of x and p
and try it.
r

m
p

u0 = 0
x+i
2~
2m~

d
mx + ~
u0 = 0
dx
du0
mx
=
u0
dx
~
This first order differential equation can be solved to get the ground state wavefunction.

u0 = Cemx

/2~

We could continue with the raising operator to get excited states.

!
r
r

m
~ d
1u1 =
x
u0
2~
2m dx

Usually we will not need the actual wave functions for our calculations.

237

10. Harmonic Oscillator Solution using Operators

10.6
10.6.1

TOC

Examples
The Expectation Value of x in an Energy Eigenstate

We can compute the expectation value of x simply.

r
r
~
~

hun |x|un i =
hun |A + A |un i =
(hun |Aun i + hun |A un i)
2m
2m
r

~
( nhun |un1 i + n + 1hun |un+1 i) = 0
=
2m
We should have seen that coming. Since each term in the x operator changes the
eigenstate, the dot product with the original (orthogonal) state must give zero.

10.6.2

The Expectation Value of p in an Energy Eigenstate

(See the previous example is you want to see all the steps.) The expectation value of
p also gives zero.
r
m~
hun |p|un i = i
hun |A A |un i = 0
2

10.6.3

The Expectation Value of x in the State

1
hu0 + u1 |x| u0 + u1 i =
2
=
=
=
+
+
=
=

1 (u0
2

+ u1 )

r
1
~
hu0 + u1 |A + A |u0 + u1 i
2 2m
r
~
hu0 + u1 |Au0 + Au1 + A u0 + A u1 i
8m
r

~
hu0 + u1 |0 + 1u0 + 1u1 + 2u2 i
8m
r

~
( 1hu0 |u0 i + 1hu0 |u1 i
8m

2hu0 |u2 i + 1hu1 |u0 i

1hu1 |u1 i + 2hu1 |u2 i)

r
~
(1 + 1)
8m
r
~
2m
238

10. Harmonic Oscillator Solution using Operators

10.6.4

The Expectation Value of

1
2 2
2 m x

TOC

in an Energy Eigenstate

The expectation of x2 will have some nonzero terms.

~
hun |AA + AA + A A + A A |un i
2m
~
hun |AA + A A|un i
2m

hun |x2 |un i =

We could drop the AA term and the A A term since they will produce 0 when the
dot product is taken.

~
hun |x2 |un i =
(hun | n + 1Aun+1 i + hun | nA un1 i)
2m

~
=
(hun | n + 1 n + 1un i + hun | n nun i)
2m

~
1
~
=
((n + 1) + n) = n +
2m
2 m
With this we can compute the expected value of the potential energy.

1
1
~
1
1
1
1
2 2
2
=
n+
~ = En
hun | m x |un i = m n +
2
2
2 m
2
2
2

10.6.5

The Expectation Value of

The expectation value of

p2
2m

in an Energy Eigenstate

hun |

p2
2m

p
|un i =
2m
=
=

1 m~
hun | AA A A|un i
2m 2
~
hun |AA + A A|un i
4
~
1
((n + 1) + n) = En
4
2

(See the previous section for a more detailed computation of the same kind.)

10.6.6

Time Development of (t = 0) =

1 (u1
2

+ u2 )

Start off in the state at t = 0.

1
(t = 0) = (u1 + u2 )
2

239

10. Harmonic Oscillator Solution using Operators

TOC

Now put in the simple time dependence of the energy eigenstates, eiEt/~ .

3
5
3
1
1
(t) = (u1 ei 2 t + u2 ei 2 t ) = ei 2 t (u1 + eit u2 )
2
2

Factoring out one complex exponential will simplify the subsequent algebra.
We can compute the expectation value of p.
r
m~ 1
hu1 + eit u2 |A A |u1 + eit u2 i
h|p|i = i
2 2
r

m~ 1
=
hu1 |A|u2 ieit hu2 |A |u1 ieit
2 2i
r
m~ 1 it it
=
2e
2e
2 2i

= m~ sin(t)

h(t)|p|(t)i = m~ sin(t)

10.7

Homework

1. At t = 0, a 1D harmonic oscillator is in a linear combination of the energy

eigenstates
r
r
2
3
=
u3 + i
u4
5
5
Find the
of p as a function of time using operator methods.
pexpected value
p
A = ( m
x
+
i
)
start from A
2m~
q 2~

p = i m~
get p operator
2 (A A)
q
q

4
3
i9t/2
it
=e
write (t)
7 u4 + i
7 u3 e
q
q
q
q
q
4
3
4
3
it

it
h(t)|p|(t)i = i m~
h
u
+
i
u
e
|A

A|
u
+
i
i
4
3
4
2
7
7
7
7 u3 e
q
q

12
h(t)|p|(t)i = i m~
i 4eit + i 4eit
compute
2
49
q
q
q
m~
12
96m~
h(t)|p|(t)i =
cos(t) algebra
2
49 4 cos(t) =
49
240

10. Harmonic Oscillator Solution using Operators

TOC

p
2. Evaluate the uncertainty in x for the 1D HO ground state hu0 |(x x
)2 |u0 i
where x
= hu0 |x|u0 i. Similarly, evaluate the uncertainty in p for the ground state.
What is the product px? Now do the same for the first excited state. What
isp
the product px for
p this state?
2 |u i =
hu0 |(x

)
hu0 |x2 |u0 i
0
p m
p
A = ( 2~ x + i 2m~ )
q
~
x = 2m
(A + A)
q
q
p
~
~
hu0 |AA |u0 i = 2m
x = 2m
q
p = i m~
(A A)
q2 p
q
m~
|u i =
p = m~
hu
|AA
2 minus signs
0
0
2
2
q
q
~
m~
~
ground state
px = 2m
2 = 2
q
q
p

~
~
x = 2m
hu1 |AA + A A|u1 i = 2m
2+1
q
q
p

2+1
p = m~
1 |AA + A A|u1 i =
2q huq
2
~
m~
~
px = 3 2m
first excited state
2 = 32
3. Calculate hui |x|uj i and hui |p|uj i.
hui |x|u
qji =

~
(A + A)
q2m

p = i m~
A)
2 (A
q
q

~
~
hui | j + 1uj+1 + juj1 i
hui |x|uj i = 2m hui |A + A|uj i = 2m
q

~
hui |x|uj i = 2m
j + 1i(j+1) + ji(j1)
q

hui |p|uj i = i m~
j + 1i(j+1) ji(j1)
2

4. P
Calculate hui |xp|uj i by direct calculation. Now calculate the same thing using
hui |x|uk ihuk |p|uj i.
k
q
q
~
m~

1st way
hui |xp|uj i = i 2m
2 hui |(A + A)(A A)|uj i
~

P ~

=
i2
k + 1i(k+1) + ki(k1)
j + 1k(j+1) jk(j1)
k

p
p
= i ~2
(j + 1)(j + 2)i(j+2) jij + (j + 1)ij j(j 1)i(j2)
careful with s

)
5. If h(A ) is a polynomial in the operator A , show that Ah(A )u0 = dh(A
u0 . As
dA
a result of this, note that since any energy eigenstate can be written as a series
d
of raising operators times the ground state, we can represent A by dA
.
241

10. Harmonic Oscillator Solution using Operators

TOC

We can simply commute the operator A through the polynomial to u0 which then
gives 0.
Ah(A )u0 =

an A(A )n u0 =

an (n)(A )n1 u0 =

dh(A )
u0
dA

6. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =

0) = 12 (u0 + u1 ).
Compute the expected value of x as a function of time using A and A in

the Schrodinger
picture.
Now do the same in the Heisenberg picture.
q
~
x = 2m
(A + A)
(t) = 12 e3it/2 (u0 eit + u1 )
q
~ 1
it
+ u1 |A + A|u0 eit + u1 i
h|x|i = 2m
2 hu0 e
q
q ~
~ 1
it
h|x|i = 2m
+ eit = 2m
cos(t)
2 e
1 (u0
2
it

(t) =
A(t) = q
e
x(t) =

+ u1 )

A
A (t) = eit Aq

~ 1
~
it
+ eit = 2m
cos(t)
2m 2 e

Schrodinger Picture
operator for x
algebra

Heisenberg picture
time dependent operators
algebra

7. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =

0) = 12 (u1 + u3 ). Compute the expected value of x2 as a function of time.

~
x2 = 2m
(A A + AA + AA + A A )
1
(t) = 2 e7it/2 (u1 e2it + u3 )
~

hx2 i = 21 2m
hu1 e2it + u3 |A A + AA + AA
|u1 e2it + u3 i
+ A A2it

~
2
2it
hx i = 4m hu1 e + u3 | 1u
6u3 e
+
3u3 + 4u3 + 6u1 + 20u5 i
1 + 2u1 + 0 +

~
~
hx2 i = 4m
3 + 6e2it + 6e2it + 4 = 4m
7 + 2 6 cos(2t)

10.8

Sample Test Problems

1. A 1D harmonic oscillator is in a linear combination of the energy eigenstates

r
r
2
1
u0 i
u1
=
3
3
Find the expected value of p.
Answer
242

10. Harmonic Oscillator Solution using Operators

=
=
=

+
r
r
r
2
1
2
1

i
u0 i
u1 |A A |
u0 i
u1
3
3
3
3
*r
+
r
r
r
r
m~
2
1
2
1
u0 i
u1 |
u1 i
u0
i
2
3
3
3
3
r
r
r
m~
2
2
(i
i
)
i
2
9
9
r r
8 m~
2

=
m~
9
2
3
r

h|p|i =

TOC

m~
2

2. Assuming un represents the nth 1D harmonic oscillator energy eigenstate, calculate hun |p|um i.
Answer
r

m~ A A
2
i
r
m~
hun |p|um i = i
hun |A A |um i
2
r

m~
= i
( mn(m1) m + 1n(m+1) )
2
p

p
3. Evaluate the uncertainty in x for the 1D HO ground state hu0 |(x x
)2 |u0 i
where x
= hu0 |x|u0 i. Similarly, evaluate the uncertainty in p for the ground state.
What is the product px?
Answer
Its easy to see that x
= 0 either from the integral or using operators. Ill use

243

10. Harmonic Oscillator Solution using Operators

TOC

operators to compute the rest.

r
x =
hu0 |x2 |u0 i =
=
p

hu0 |p2 |u0 i =

x =
p

px =

~
(A + A )
2m
~
hu0 |AA + AA + A A + A A )|u0 i
2m
~
~
~
hu0 |AA |u0 i =
1=
2m
2m
2m
r
1 m~
(A A )
i
2
m~
m~

hu0 | AA |u0 i =
2
2
r
~
2m
r
m~
2
~
2

4. Use the commutator relation between A and A to derive [H, A]. Now show that
A is the lowering operator for the harmonic oscillator energy.
5. q
At t = 0, q
a one dimensional harmonic oscillator is in the state (t = 0) =
3
1
4 u0 + i
4 u1 . Calculate the expected value of p as a function of time.
6. At t = 0, a harmonic oscillator is in a linear combination of the n = 1 and n = 2
states.
r
r
2
1
=
u1
u2
3
3
Find hxi and hx2 i as a function of time.
7. A 1D harmonic oscillator is in a linear combination of the energy eigenstates
r
r
2
1
=
u0 +
u2 .
3
3
Find hx2 i.

244

11. More Fun with Operators

11
11.1
11.1.1

TOC

More Fun with Operators

Operators in a Vector Space
Review of Operators

First, a little review. Recall that the square integrable functions form a vector
space, much like the familiar 3D vector space.
~r = av~1 + bv~2
in 3D space becomes
|i = 1 |1 i + 2 |2 i.
The scalar product is defined as

dx (x)(x)

h|i =

and many of its properties can be easily deduced from the integral.
h|i = h|i
As in 3D space,
~a ~b |a| |b|
the magnitude of the dot product is limited by the magnitude of the vectors.
p
h|i h|ih|i
This is called the Schwartz inequality.
Operators are associative but not commutative.
AB|i = A(B|i) = (AB)|i
An operator transforms one vector into another vector.
|0 i = O|i
Eigenfunctions of Hermitian operators
H|ii = Ei |ii
245

11. More Fun with Operators

TOC

form an orthonormal
hi|ji = ij
complete set
|i =

hi|i|ii =

|iihi|i.

Note that we can simply describe the j th eigenstate at |ji.

Expanding the vectors |i and |i,
|i =

bi |ii

|i =

ci |ii

we can take the dot product by multiplying the components, as in 3D space.

X
h|i =
bi ci
i

The expansion in energy eigenfunctions is a very nice way to do the time development of a wave function.
X
|(t)i =
|iihi|(0)ieiEi t/~
i

The basis of definite momentum states is not in the vector space, yet we can use
this basis to form any state in the vector space.
1
|i =
2~

dp(p)|pi

Any of these amplitudes can be used to define the state.

Three ways to Write a State
ci = hi|i
(x) = hx|i
(p) = hp|i
246

11. More Fun with Operators

11.1.2

TOC

Projection Operators |jihj| and Completeness

Now we move on a little with our understanding of operators. A ket vector followed by
a bra vector is an example of an operator. For example the operator which projects
a vector onto the j th eigenstate is
|jihj|
First the bra vector dots into the state, giving the coefficient of |ji in the state, then
its multiplied by the unit vector |ji, turning it back into a vector, with the right length
to be a projection. An operator maps one vector into another vector, so this is an
operator.
The sum of the projection operators is 1, if we sum over a complete set of states,
like the eigenstates of a Hermitian operator.
Completeness Identity
X

|iihi| = 1

This is an extremely useful identity for solving problems. We could already see this in
the decomposition of |i above.
X
|i =
|iihi|i.
i

The same is true for definite momentum states.

Completeness of Momentum Eigenstates

|pihp| dp = 1

We can form a projection operator into a subspace.

X
P =
|iihi|
subspace

We could use this to project out the odd parity states, for example.
247

11. More Fun with Operators

11.1.3

TOC

Unitary Operators

Unitary operators preserve the scalar products of state vectors.

h|i = hU |U i = h|U U i
This means that
U U = 1.
Unitary operators will be important for the matrix representation of operators. The
will allow us to change from one orthonormal basis to another.

11.2

A Complete Set of Mutually Commuting Operators

If an operator commutes with H, we can make simultaneous eigenfunctions of

energy and that operator. This is an important tool both for solving the problem and
for labeling the eigenfunctions.
A complete set of mutually commuting operators will allow us to define a
state in terms of the quantum numbers of those operators. Usually, we will need one
quantum number for each degree of freedom in the problem.
For example, the Hydrogen atom in three dimensions has 3 coordinates for the internal
problem, (the vector displacement between the proton and the electron). We will need
three quantum numbers to describe the state. We will use an energy index, and two
angular momentum quantum numbers to describe Hydrogen states. The operators
will all commute with each other. The Hydrogen atom also has 3 coordinates for the
position of the atom. We will might use px , py and pz to describe that state. The
operators commute with each other.
If we also consider the spin of the electron in the Hydrogen atom, we find that we need
to add one more commuting operator to label the states and to compute the energies
accurately. If we also add the spin of the proton to the problem, the we need still one
more quantum number to describe the state.
If it is possible, identifying the commuting operators to be used before solving the
problem will usually save time.

11.3

Uncertainty Principle for Non-Commuting Operators

Let us now derive the uncertainty relation for non-commuting operators

A and B. First, given a state , the Mean Square uncertainty in the physical
248

11. More Fun with Operators

TOC

quantity represented is defined as

(A)2

h|(A hAi)2 i = h|U 2 i

(B)2

h|(B hBi)2 i = h|V 2 i

where we define (just to keep our expressions small)

= A h|Ai

= B h|Bi.

Since hAi and hBi are just constants, notice that

[U, V ] = [A, B]
OK, so much for the definitions.
Now we will dot U + iV into itself to get some information about the uncertainties.
The dot product must be greater than or equal to zero.
hU + iV |U + iV i 0
2

h|U i + h|V i + ihU |V i ihV |U i 0

This expression contains the uncertainties, so lets identify them.

(A)2 + 2 (B)2 + ih|[U, V ]|i 0
Choose a to minimize the expression, to get the strongest inequality.

2(B)2 + ih|[U, V ]|i = 0

ih|[U, V ]|i
=
2(B)2

Plug in that .
1 h|[U, V ]|i2
h|[U, V ]|i2
+
0
2
4
(B)
2(B)2
1
i
(A)2 (B)2 h|[U, V ]|i2 = h| [U, V ]|i2
4
2
(A)2

This result is the uncertainty for non-commuting operators.

249

11. More Fun with Operators

TOC

Uncertainty for Non-Commuting Operators

i
(A)2 (B)2 h| [A, B]|i2
2

Note that for commutators like [Lx , Ly ] = i~Lz , the expectation value and square are
needed. If the commutator is a constant, as in the case of [p, x], the expectation values
can be removed and the square root taken.
(A)(B)

i
[A, B]
2

For momentum and position, this agrees with the uncertainty principle we know.

(p)(x)

i
~
h[p, x]i =
2
2

(Note that we could have simplified the proof by just stating that we choose to dot
]i
(U + ih[U,V
2(B)2 V ) into itself and require that its positive. It would not have been clear
that this was the strongest condition we could get.)

11.4

Time Derivative of Expectation Values *

We wish to compute the time derivative of the expectation value of an operator A in

the state . Thinking about the integral, this has three terms.

A
d
d

+ |A| d
h |A| i =
|A| +
dt
dt
t
dt

A
1
1

=
hH |A| i +
h |A| Hi +
i~
i~
t

A
i

=
h |[H, A]| i +
~
t

This is an important general result for the time derivative of expectation values.
250

11. More Fun with Operators

TOC

Time Derivative of Expectation Values

A
d
i

h |A| i = h |[H, A]| i +
dt
~
t

which becomes simple if the operator itself does not explicitly depend on time.
d
i
h |A| i = h |[H, A]| i
dt
~
Expectation values of operators that commute with the Hamiltonian are constants of
the motion.
We can apply this to verify that the expectation value of x behaves as we would expect
for a classical particle.

Ehrenfest Theorem
d hxi
i
i
= h[H, x]i =
dt
~
~

D E
p2
p
,x
=
2m
m

This is a good result. This is called the Ehrenfest Theorem.

For momentum,

Time Derivative of Expected Momentum

dhpi
i
i
= h[H, p]i =
dt
~
~

~ d
dV (x)
V (x),
=
i dx
dx

which Mr. Newton told us a long time ago.

11.5

The Time Development Operator *

We can actually make an operator that does the time development of a wave
function. We just make the simple exponential solution to the Schrodinger equation
251

11. More Fun with Operators

TOC

using operators.
i~

= H
t

Time Development Operator

(t) = eiHt/~ (0)

where H is the operator. We can expand this exponential to understand its meaning
a bit.

X
(iHt/~)n
eiHt/~ =
n!
n=0
This is an infinite series containing all powers of the Hamiltonian. In some cases, it
can be easily computed.
eiHt/~ is the time development operator. It takes a state from time 0 to time t.

11.6

The Heisenberg Picture *

To begin, lets compute the expectation value of an operator B.

h(t)|B|(t)i = heiHt/~ (0)|B|eiHt/~ (0)i
= h(0)|eiHt/~ BeiHt/~ |(0)i

According to our rules, we can multiply operators together before using them. We can
then define the operator that depends on time.

Operator in Heisenberg Picture

B(t) = eiHt/~ BeiHt/~

If we use this operator, we dont need to do the time development of the wavefunctions!
This is called the Heisenberg Picture. In it, the operators evolve with time
and the wavefunctions remain constant.
252

11. More Fun with Operators

TOC

The usual Schr

odinger picture has the states evolving and the operators constant.
We can now compute the time derivative of an operator.
d
B(t)
dt

=
=

iH iHt/~ iHt/~
iH iHt/~
e
Be
eiHt/~ B
e
~
~
i iHt/~
i
e
[H, B]eiHt/~ = [H, B(t)]
~
~

It is governed by the commutator with the Hamiltonian.

As an example, we may look at the HO operators A and A . We have already computed
the commutator.
[H, A] = ~A
i
dA
= ~A = iA
dt
~
We can integrate this.
A(t) = eit A(0)
Take the Hermitian conjugate.
A (t) = eit A (0)
We can combine these to get the momentum and position operators in the Heisenberg
picture.

11.7
11.7.1

p(t)

x(t)

p(0) cos(t) mx(0) sin(t)

p(0)
x(0) cos(t) +
sin(t)
m

Examples
Time Development Example

Start off in the state.

1
(t = 0) = (u1 + u2 )
2

In the Schr
odinger picture,
3
5
3
1
1
(t) = (u1 ei 2 t + u2 ei 2 t ) = ei 2 t (u1 + eit u2 )
2
2

253

11. More Fun with Operators

TOC

We can compute the expectation value of x.

r
1
~
h|x|i =
hu1 + eit u2 |A + A |u1 + eit u2 i
2 2m
r

~
1
hu1 |A|u2 ieit + hu2 |A |u1 ieit
=
2 2m
r
1
~ it it
=
2e
+ 2e
2 2m
r
~
=
cos(t)
m

In the Heisenberg picture

1
h|x(t)|i =
2

~
h|eit A + eit A |i
2m

This gives the same answer with about the same amount of work.

11.8

Homework

1. Consider the functions of one angle () with and () = ().

d
Show that the angular momentum operator L = ~i d
values.
!has real expectation !

~ d

d
d
= ~i
h|Li =
i d d = ~i | |
d d
d d

!
~ d
h|Li =
d = hL|i
i d

2. Prove that the parity operator defined by P (x) = (x) is a hermitian operator
and find its possible eigenvalues.
h|P i = hP |i
definition of H.C.

h|P i =
(x)(x)dx
x0 = x
h|P i =

h|P i =
h|P i =

dx0 = dx
(x0 )(x0 )(dx0 )

(x)(x)(dx)

(x)(x)dx = hP |i

transform variable of integration

interchange limits and rename dummy var
Q.E.D.

P = a
P P = a2 =
a = 1

eigenvale
Parity twice gives same coordinates
two possible eigenvalues
254

11. More Fun with Operators

TOC

3. The hyper-parity operator H has the property that H 4 = for any state .
Find the eigenvalues of H for the case that it is not Hermitian and the case that
it is Hermitian.
H = b
eigenvalue
H 4 = = b4
b4 = 1
b = 1, 1, i, i
4 possible values in general
h|H 2 i = hH|Hi > 0
b = 1 2 vals for Hermitian operator
4. An operator is Unitary if U U = U U = 1. Prove that a unitary operator preserves inner products, that is hU |U i = h|i. Show that if the states |ui i are
orthonormal, that the states U |ui i are also orthonormal. Show that if the states
|uP
i i form a complete set, then the states U |ui i also form a complete set.
|ui ihui | = 1
closure identity
i P
U |ui ihui |U = U 1U
left and right multiply
Pi

U |ui ihui |U = 1
use unitary premise
i
P
|U ui ihU ui | = 1
closure identity for transformed basis
i

dot product preserved

orthonormal

5. A general one dimensional scattering problem could be characterized by an (arbitrary) potential V (x) which is localized by the requirement that V (x) = 0 for
|x| > a. Assume that the wave-function is
ikx
Ae + Beikx
x < a
(x) =
Ceikx + Deikx
x>a
Relating the outgoing waves to the incoming waves by the matrix equation

C
S11 S12
A
=
B
S21 S22
D
show that
|S11 |2 + |S21 |2 = 1
|S12 |2 + |S22 |2 = 1

S11 S12
+ S21 S22
=0

Use this to show that the S matrix is unitary.

255

11. More Fun with Operators

TOC

There should be no source or sink of probability

and probabilty should not build up in any region.
So outside
region of the potential, the flux should be constant.
the
Aeikx + (S21 A + S22 D)eikx
x < a
(x) =
(S11 A + S12 D)eikx + Deikx
x>a
|A|2 |S21 A + S22 D|2 = |S11 A + S12 D|2 |D|2

|A|2 |S21 |2 |A|2 |S22 |2 |D|2 S21

S22 A D S21 S22
AD
2
2
2
2

= |S11 | |A| + |S12 | |D| + S11 S12 A D + S11 S12 AD |D|2

|S11 |2 + |S21 |2 = 1
|S12 |2 + |S22 |2 = 1

S11 S12
+ S21 S22
=0

S11
S21
S11 S12

S S=

S21 S22
S12 2S22

|S11 | + |S21 |2
S11
S12 + S21
S22
1 0
=
=

S11 S12
+ S21 S22
|S12 |2 + |S22 |2
0 1
6. Calculate the S matrix for the potential

V0
V (x) =
0

v is same
for
any A and D
A*A term
D*D term
A*D term
AD* term

Unitary

|x| < a
|x| > a

and
that the above
show

conditions are satisfied.
C
T R
A
=
B
R
T
D

T
R
T R
T T + R R T R + R T
SS =
=
R T
R T
R T + T R T T + R R
2k
2ika
T =e
2k cosh(2a)i(k2 2 ) sinh(2a)
(k02 k2 ) sin(2k0 a)
R = ie
2kk0 cos(2k0 a)i(k02 +k2 ) sin(2k0 a)
(2 +k2 ) sinh(2a)
R = ie2ika 2k cosh(2a)i(k
2 2 ) sinh(2a)
4k2 2 +(k4 +2k2 2 +4 ) sinh2 (2a)

T T + R R = 4k2 2 cosh2 (2a)+(k4 2k2 2 +4 ) sinh2 (2a)

2 2
(2+sinh2 (2a))+(k4 +4 ) sinh2 (2a)
T T + R R = 2k2 22k
(2 cosh2 (2a)sinh2 (2a))+(k4 +4 ) sinh2 (2a)
2 2
(2+sinh2 (2a))+(k4 +4 ) sinh2 (2a)
T T + R R = 2k
=1
2k2 2 (2+sinh2 (2a))+(k4 +4 ) sinh2 (2a)
2
2
2k(
+k
)
sinh(2a)
T R + R T = i |2k cosh(2a)i(k2 2 ) sinh(2a)|2 + C.C. = 0
2ika

from potential barrier

from potential well
transform k 0 i

cosh2 sinh2 = 1
Q.E.D.

7. Show that if an operator H is hermitian, then the operator eiH is unitary.

P
(iH)j
eiH =
j!
j=0

j=0

eiH eiH = 1

(iH)j
j!

= eiH

assume H = H
unitary

256

11. More Fun with Operators

11.9

TOC

Sample Test Problems

1. Calculate the commutator [Lx , Lz ] where Lx = ypz zpy and Lz = xpy ypx .
State the uncertainty principle for Lx and Lz .
Answer

[Lx , Lz ]

=
=

Lx Lz

[ypz zpy , xpy ypx ] = x[y, py ]pz + z[py , y]px

~
(xpz + zpx ) = i~(xpz zpx ) = i~Ly
i
i
i
~
h[Lx , Lz ]i = (i~)hLy i = hLy i
2
2
2

2. * A particle of mass m is in a 1 dimensional potential V (x). Calculate the

The kinetic energy can simply be added and the potential now depends on 3 coordinates. The Hamiltonian in 3D is
H=

p2y
p2x
p2
p2
~2 2
+
+ z + V (~r) =
+ V (~r) =
+ V (~r).
2m 2m 2m
2m
2m

This extension is really very simple.

12.3

Two Particles in Three Dimensions

The generalization of the Hamiltonian to three dimensions is simple.

p~21
p~2
+ 2 + V (~r1 ~r2 )
2m 2m

We define the vector difference between the coordinates of the particles.

~r ~r1 ~r2

We also define the vector position of the center of mass.

~ m1~r1 + m2~r2
R
m1 + m2
260

12. Extending QM to Two Particles and Three Dimensions

TOC

We will use the chain rule to transform our Hamiltonian. As a simple example,
if we were working in one dimension we might use the chain rule like this.
d
r
R
=
+
dr1
r1 r r1 R
In three dimensions we would have.
~1 =
~r+

m1
~R

m1 + m2

~ 2 =
~r+

m2
~R

m1 + m2

Putting this into the Hamiltonian we get

~2 ~ 2
~2 ~ 2
1
+ V (~r1 ~r2 )
2m1
2m2 2

"
#

2
m1
2m
~2 ~ 2
1
2
~R+
~r
~R
r +

H=
2m1
m1 + m2
m1 + m2
"
#

2
~2 ~ 2
m2
2m
2
~ 2R
~r
~ R + V (~r)
r +
+

2m2
m1 + m2
m1 + m2

1
1
~ 2 + m1 + m2
~ 2 + V (~r).
H = ~2
+

r
2m1
2m2
2(m1 + m2 )2 R
Defining the reduced mass
1
1
1
=
+

m1
m2
and the total mass
M = m1 + m2
we get.

261

12. Extending QM to Two Particles and Three Dimensions

TOC

Center of Mass Transformation

~r ~r1 ~r2
~ m1~r1 + m2~r2
R
m1 + m2
1
1
1
=
+

m1
m2
M = m1 + m2
H=

~2 ~ 2
~2 ~ 2
R
+ V (~r) = HCM + Hinternal
2M
2 r

The Hamiltonian actually separates into two problems: the motion of the center
of mass as a free particle
~2 ~ 2
HCM =

2M R
and the internal interaction between two particles

Hinternal =

~2 ~ 2
+ V (~r)
2 r

with a reduced mass.

This is exactly the same separation that we would make in classical physics.

12.4

Identical Particles

It is not possible to tell the difference between two electrons. They are identical in
every way. Hence, there is a clear symmetry in nature under the interchange
of any two electrons.
We define the interchange operator P12 . By our symmetry, this operator commutes
with H so we can have simultaneous eigenfunctions of energy and interchange.
262

12. Extending QM to Two Particles and Three Dimensions

TOC

If we interchange twice, we get back to the original state,

P12 (x1 , x2 ) = (x2 , x1 )
P12 P12 (x1 , x2 ) = (x1 , x2 )
so the possible eigenvalues of the interchange operator are just +1 and -1.
Identical Particle Interchange Eigenvalue Equation

P12 =

bosons
fermions

It turns out that both possibilities exist in nature. Some particles like the electron,
always have the -1 quantum number. They are spin one-half particles and are called
fermions. The overall wavefunction changes sign whenever we interchange any pair of
fermions. Some particles, like the photon, always have the +1 quantum number. They
are integer spin particles, called bosons.
There is an important distinction between fermions and bosons which we can derive
from the interchange symmetry. If any two fermions are in the same state, the wave
function must be zero in order to be odd under interchange.
Antisymmetrize 2 Particle Wavefunction
= ui (x1 )uj (x2 ) ui (x1 )uj (x2 ) uj (x1 )ui (x2 )

(Usually we write a state like ui (x1 )uj (x2 ) when what we mean is the antisymmetrized
version of that state ui (x1 )uj (x2 ) uj (x1 )ui (x2 ).) Thus, no two fermions can be in
the same state. This is often called the Pauli exclusion principle.
In fact, the interchange symmetry difference makes fermions behave like matter and
bosons behave like energy. The fact that no two fermions can be in the same
state means they take up space, unlike bosons. It is also related to the fact that
fermions can only be created in conjunction with anti-fermions. They must be made
in pairs. Bosons can be made singly and are their own anti-particle as can be
seen from any light.

12.5

Homework

1. The energy spectrum of hydrogen can be written in terms of the principal quan2
2
tum number n to be E = 2nc2 . What are the energies (in eV) of the photons
263

12. Extending QM to Two Particles and Three Dimensions

TOC

from the n = 2 n = 1 transition in hydrogen and deuterium? What is the

difference in photon energy between the two isotopes of hydrogen?

2
2
2
2
2
2
N me c
E = E2 E1 = 2c 212 112 = 3 8c = 38 m
mN +me
1
1
1
hydrogen
= mp + me
E =
E =
Ed

32
8
32
8

(938.3)(0.511)
938.3+0.511 = 10.204 eV
(938.3+939.6)(0.511)
=
938.3+939.6+0.511

2
938.3+939.62.2
EH = 3 8me 938.3+939.62.2+0.511

hydrogen
deuterium

938.3
938.3+0.511

= 0.00278 eV

2. Prove that the operator that exchanges two identical particles is Hermitian.
We have shown that the possible eigenvalues are 1, so it has real eigenvalues.
h|P12 i = h|i = hP12 |i
3. Two identical, non-interacting spin 12 particles are in a box. Write down the full
lowest energy wave function for both particles with spin up and for one with spin
up and the other spin down. Be sure your answer has the correct symmetry under
the interchange of identical particles.
= 12 (u1 (x1 )u2 (x2 ) u1 (x2 )u2 (x1 ))
= 12 u1 (x1 )u1 (x2 ) ( )

4. The momentum operator conjugate to any cooridinate xi is ~i x

. Calculate the
i
commutators of the center of mass coordinates and momenta [Pi , Rj ] and of the
internal coordinates and momenta [pi , rj ]. Calculate the commutators [Pi , rj ] and
[pi , Rj ].
~r ~r1 ~r2
~ m1 ~r1 +m2 ~r2
R
m1 +m2

[Pi , Rj ] = ~i [ R
, Rj ] = ~i ij
trivially
i
~
[pi , rj ] = i ij
also trivially
m1
~ m2 ~r2
~
r
=
R
1
m1 +m2
m1 +m2

2
~ m2 ~r2
R
~r1 = m1m+m
m1 +m2
1

m1 +m2
2
~
~r1 = m1
R m1m+m
(
r
~

~
r
)
1
2

m1 +m2 ~
m1 +m2
2
~
r
=
r
R
+m
1
m1
m1
m1 ~
m1 m2
m2
~
~
~r1 = R + (m1 +m2 )2 ~r = R + M ~r
~ m1 ~r
~r2 = R
M
~
~r +
~r
R =
1
2
m
~ r m1
~r
~ r = 2

2
[pi , Rj ] = ~i [ m
M r1i

12.6

m1 r1j +m2 r2j

m1
M r2i ,
M

]=0

independent as expected

Sample Test Problems

1. * Calculate the Fermi energy for N particles of mass m in a 3D cubic box of

264

12. Extending QM to Two Particles and Three Dimensions

TOC

side L. Ignore spin for this problem.

Answer
The energy levels are given in terms of three quantum numbers.
E=

2 ~2 2
(n + n2y + n2z )
2mL2 x

The number of states with inside some (n2x + n2y + n2z )max ( 18 of a a sphere in n
space) is
3
14
2
N=
(n2x + n2y + n2z )max
83
So for N particles filling the levels,
3

6N
.

2

6N 3
=

2
(n2x + n2y + n2z )max
=

(n2x + n2y + n2z )max

The energy corresponding to this is the Fermi energy.

EF =

2 ~2
2mL2

2. * We put N fermions of mass m into a (cold) one dimensional box of length

L. The particles go into the lowest energy states possible. If the Fermi energy is
defined as the energy of the highest energy particle, what is the Fermi energy of
this system? You may assume that there are 2 spin states for these fermions.

265

13. 3D Problems Separable in Cartesian Coordinates

TOC

3D Problems Separable in Cartesian Coordinates

We will now look at the case of potentials that separate in Cartesian coordinates. These will be of the form.
V (~r) = V1 (x) + V2 (y) + V3 (z)
In this case, we can solve the problem by separation of variables.
H = Hx + Hy + Hz
(Hx + Hy + Hz )u(x)v(y)w(z) = Eu(x)v(y)w(z)
[Hx u(x)] v(y)w(z) + u(x) (Hy + Hz ) v(y)w(z) = Eu(x)v(y)w(z)
(Hy + Hz )v(y)w(z)
Hu(x)
=E
= x
u(x)
v(y)w(z)
The left hand side of this equation depends only on x, while the right side depends
on y and z. In order for the two sides to be equal everywhere, they must both be
equal to a constant which we call x .
The x part of the solution satisfies the equation
Hx u(x) = x u(x).
Treating the other components similarly we get
Hy v(y) = y v(y)
Hz w(z) = z w(z)
and the total energy is
E = x + y + z .
There are only a few problems which can be worked this way but they are important.

13.1

Particle in a 3D Box

An example of a problem which has a Hamiltonian of the separable form is the particle
in a 3D box. The potential is zero inside the cube of side L and infinite outside.
It can be written as a sum of terms.
H = Hx + Hy + Hz
The energies and eigenfunctions depend on three quantum numbers, (since there
are 3 degrees of freedom).
266

13. 3D Problems Separable in Cartesian Coordinates

TOC

Particle in a 3D Cubic Box

Enx,ny,nz =

2 ~2 2
(n + n2y + n2z )
2mL2 x

32
n y
n z
n x
2
y
z
x
sin
sin
unx,ny,nz (~r) =
sin
L
L
L
L

For a cubic box like this one, there will often be degenerate states.

13.1.1

Filling the Box with Fermions

If we fill a cold box with N fermions, they will all go into different low-energy states. In
fact, if the temperature is low enough, they will go into the lowest energy N states. If
we fill up all the states up to some energy, that energy is called the Fermi energy. All
the states with energies lower than EF are filled, and all the states with energies larger
than EF are empty. (Non zero temperature will put some particles in excited states,
but, the idea of the Fermi energy is still valid.) Since the energy goes like n2x + n2y + n2z ,
it makes sense to define a radius rn in n-space out to which the states are filled.
2 ~2 2
2 ~2 2
2
2
(n
+
n
+
n
)
=
r < EF
x
y
z
2mL2
2mL2 n

Figure 46: A picture of the Fermi surface in n-space. The surface is spherical and the
number of states can be estimated from the volume of one eighth of the sphere.
267

13. 3D Problems Separable in Cartesian Coordinates

TOC

This is an exapmple of a Fermi Surface which in this case is the surface of one eighth
of a sphere. The Fermi Surface separates the filled states from the unfilled states
(here) in n-space. In other problems the surface may be in k-space.
The number of states within the radius is
N = (2)spin

14 3
r
83 n

where we have added a factor of 2 because fermions have two spin states. This is an
approximate counting of the number of states based on the volume of a sphere
in n-space. The factor of 81 indicates that we are just using one eighth of the sphere in
n-space because all the quantum numbers must be positive.
We can now relate the Fermi energy to the number of particles in the box.
2 ~2 2
2 ~2
EF =
r
=
n
2mL2
2mL2

2 ~2
=
2m

3N
L3

2 ~2
=
2m

3 ~2
=
10m

We can also integrate to get the total energy of all the fermions.
Etot

1
=2
8

~
3 ~2 rn5
3 ~2
4r
dr
=
=
2mL2
2mL2 5
10mL2
2r

2 2 2

where the last step shows how the total energy depends on the number of particles per
unit volume n. It makes sense that this energy is proportional to the volume.
The step in which EF and Etot is related to N is often useful.
Fermi Energy and Total Energy of N Particles in a Cubic Box
2 ~2
EF =
2m
Etot

13.1.2

3 ~2
=
10mL2

3N
L3

Degeneracy Pressure in Stars

The 3D Particle in a Box problem, filled with identical fermions can be used
can be used to understand how the Degeneracy Pressure of the fermions depends
268

13. 3D Problems Separable in Cartesian Coordinates

TOC

on volume. Volume is clearly defined in the Box problem, unlike other problems we
will solve.
The pressure exerted by fermions squeezed into a small box is what keeps cold stars
from collapsing. White Dwarfs are held up by electrons and Neutron Stars are
held up by neutrons in a much smaller box.
We can compute the pressure from the dependence of the energy on the volume,
for a fixed number of fermions.
dE
P
Etot
P
P

= F~ d~s = P Ads = P dV
Etot
=
V

5
3 2
~
3N 3 2
=
V 3
10m

5
3 ~2 3N 3 5
=
V 3
15m

5
5
3 ~2 3n 3
3 ~2 3N 3 5
L =
=
15m

15m

The last step verifies that the pressure only depends on the density, not the
volume and the N independently, as it should. We will use:
3 ~2
P =
15m

3 ~2
=
15m

35
V

5
3

Fermi Pressure Computed for N Particles in Cubic Box

Pf ermi

3 ~2
=
15m

V 3

To understand the collapse of stars, we must compare this to the pressure of gravity.
We compute this approximately, ignoring general relativity and, more significantly, the
variation of gravitational pressure with radius.

269

13. 3D Problems Separable in Cartesian Coordinates

R
=

TOC

GMinside 4r2
dr
r

G( 34 r3 )4r2
dr
r

(4)2 2 5
3GM 2

G R =
15
5R

The mass of the star is dominated by nucleons.

M = N MN
Putting this into our energy formula, we get.
3
Eg = G(N MN )2
5

4
3

V 3

We can now compute the pressure due to gravity near the center of the dwarf.
13
4
1
4
E
2
= G(N MN )
Pg =
V 3
V
5
3
The pressures must balance: Pe = Pg . For a white dwarf, the pressure ballance
equation is:

5
13
4
3 ~2 3Ne 3 5
1
4
V 3 = G(N MN )2
V 3
15me

5
3
We can solve for the radius.

R=

3
4

3 ~2
2
3Gme MN

53 53
3
Ne

There are about two nucleons per electron

N 2Ne
so the radius becomes.
Radius of a White Dwarf Based on Fermi Pressure

R=

81 2
512

~2
31
2 N
Gme MN
270

13. 3D Problems Separable in Cartesian Coordinates

TOC

The radius decreases as we add mass. For one solar mass, N = 1057 , we get a radius
of 7200 km, the size of the earth. The Fermi energy is about 0.2 MeV.
A white dwarf is the remnant of a normal star. It has used up its nuclear fuel, fusing
light elements into heavier ones, until most of what is left is Fe56 which is the most
tightly bound nucleus. Now the star begins to cool and to shrink. It is stopped by the
pressure of electrons. Since the pressure from the electrons grows faster than
the pressure of gravity, the star will stay at about earth size even when it cools.
If the star is more massive, the Fermi energy goes up and it becomes possible to
absorb the electrons into the nucleons, converting protons into neutrons. The Fermi
energy needs to be above 1 MeV. If the electrons disappear this way, the star collapses
suddenly down to a size for which the Fermi pressure of the neutrons stops the collapse
(with quite a shock). Actually some white dwarfs stay at earth size for a long time as
they suck in mass from their surroundings. When they have just enough, they collapse
forming a neutron star and making a supernova. The supernovae are all nearly identical
since the dwarfs are gaining mass very slowly. The brightness of this type of supernova
has been used to measure the accelerating expansion of the universe.
We can estimate the neutron star radius.
RR

MN 1 5
N 3 2 3 = 10
me

Its about 10 kilometers. If the pressure at the center of a neutron star becomes too
great, it collapses to become a black hole. This collapse is probably brought about by
general relativistic effects, aided by strange quarks.

13.2

The 3D Harmonic Oscillator

The 3D harmonic oscillator can also be separated in Cartesian coordinates. For the
case of a central potential, V = 12 m 2 r2 , this problem can also be solved nicely in
spherical coordinates using rotational symmetry. The cartesian solution is easier and
better for counting states though.
Lets assume the central potential so we can compare to our later solution. We
could have three different spring constants and the solution would be as simple. The
Hamiltonian is
H
H
H

p2
1
+ m 2 r2
2m 2
p2y
p2x
1
1
p2
1
=
+ m 2 x2 +
+ m 2 y 2 + z + m 2 z 2
2m 2
2m 2
2m 2
= Hx + Hy + Hz

271

13. 3D Problems Separable in Cartesian Coordinates

TOC

The problem separates nicely, giving us three independent harmonic oscillators.

3D Harmonic Oscillator Energy Eigenstates

3
nx + ny + nz +
2

nx,ny,nz (x, y, z) = unx (x)uny (y)unz (z)

This was really easy. The un are just the 1D HO eigenstates.

This problem has a different Fermi surface in n-space than did the particle in a
box. The boundary between filled and unfilled energy levels is a plane defined by
nx + ny + nz =

3
EF

~
2

We will leave it to the homework to find the Fermi Energy. In this problem, there is
no defined volume so the Fermi Pressure will not be computed.

13.3

Homework

1. Calculate the Fermi energy of a gas of massless fermions with n particles per unit
volume.
q
H = pc = p2x + p2y + p2z c
in the box, infinity out
H 2 = (p2x + p2y + p2z )c2 E 2
can separate this equation
kx L = nx
wfn goes to zero at walls
2 h2 c2 n2x
2
2 2 2
x part of E 2
(E )x = ~ kx c =
L2
2 2 2
2 2 2
E 2 = L~2 c (n2x + n2y + n2z ) = L~2 c rn2 total E 2
E = ~c
total E in rn
L rn

14
3
3N 1/3
3
N = 2 8 3 rn = 3 rn rn =
radius in n-space

3N 1/3
EF = ~c
EF in relativistic limit
L

r
4
n
2 ~c rn
Etot = 2 18 4 r2 r~c
Total energy: integrate
L dr = L 4
0

2
2
4/3
4/3 1/3
Etot = 4L~c 3N
= 4~c 3N
V
plug

2
1/3
Etot
~c 3N 4/3
3
4/3 ~c 4/3
P = V = 12 V
=
n
4/3 power of density (not 5/3)
4
2. The number density of conduction electrons in copper is 8.5 1022 per cubic
centimeter. What is the Fermi energy in electron volts?
272

13. 3D Problems Separable in Cartesian Coordinates

EF =

2 ~2 3n 3
2m

2
2 ~2 c2 3n 3
2mc2

22
3

EF =
8.5 10
EF =

(197.3)
2(511000)

TOC

Fermi energy from notes

= 8.5 101 nm3

2
3(85) 3
= 7.05 eV

3. The radius of a nucleus is approximately 1.1A 3 Fermis, where A = N + Z, N is

the number of neutrons, and Z is the number of protons. A Lead nucleus consists
of 82 protons and 126 neutrons. Estimate the Fermi energy of the protons and
neutrons separately.
2
2 2
3n 3
~
EF = 2m
Fermi energy from notes

V = 34 r3 = 43 1.13 A = 43 1.13 (208) = 1160 fm3

nneutron = 126
fm3
V = 0.109
2

2 2 2
~ c
3n 3
EF = 2mc
= 45.2 MeV
for neutrons
2

82
nproton = V = 0.0707 fm3
2
2 2 2
~ c
3n 3
EF = 2mc
= 33.9 MeV
for protons
2

2
81~c
=
=
17.9M
eV
Max Coulomb energy for p
VCoulomb 81e
r
6.52
4. A particle of mass m in 3 dimensions is in a potential V (x, y, z) = 12 k(x2 + 2y 2 +
3z 2 ). Find the energy eigenstates in terms of 3 quantum numbers. What is
the energy of the ground state and first excited state?

q k
2 + nz + 12
3 ~ m separates in xyz
E = nx + 12 + ny + 12
~q k

ground state
E = 1+ 2+ 3 2 m

~q k

E = 3+ 2+ 3 2 m
first excited
5. N identical fermions are bound (at low temperature) in a potential V (r) =
1
2 2
2 m r . Use separation in Cartesian coordinates to find the energy eigenvalues in terms of a set of three quantum numbers (which correspond to 3 mutually
commuting operators). Find
the Fermi energy of the system.

E = nx + ny + nz + 32 ~ < EF
E = sn + 23 ~ < EF
s3
s3
s3
N = 2 6n = 3n
2 for spin, do integral for 6n
1
3
sn = (3N
boundary is a plane with nx + ny + nz < sn

)
1

EF = (3N ) 3 +

13.4

3
2

H.O. has no relation to Volume

Sample Test Problems

1. A particle of mass m in 3 dimensions is in a potential V (x, y, z) = 21 k(x2 + 2y 2 +

3z 2 ). Find the energy eigenstates in terms of 3 quantum numbers. What is
the energy of the ground state and first excited state?
273

13. 3D Problems Separable in Cartesian Coordinates

TOC

2. * N identical fermions are bound (at low temperature) in a potential V (r) =

1
2 2
2 m r . Use separation in Cartesian coordinates to find the energy eigenvalues
in terms of a set of three quantum numbers (which correspond to 3 mutually
commuting operators). Find the Fermi energy of the system. If you are having
trouble finding the number of states with energy less than EF , you may assume
that it is (EF /~)3 .
3. A particle of mass m is in the potential V (r) = 21 m 2 (x2 + y 2 ). Find operators
that commute with the Hamiltonian and use them to simplify the Schrodinger
equation. Solve this problem in the simplest way possible to find the eigenenergies in terms of a set of quantum numbers that describe the system.
4. A particle is in a cubic box. That is, the potential is zero inside a cube of side
L and infinite outside the cube. Find the 3 lowest allowed energies. Find the
number of states (level of degeneracy) at each of these 3 energies.
5. A particle of mass m is bound in the 3 dimensional potential V (r) = kr2 .
a) Find the energy levels for this particle.
b) Determine the number of degenerate states for the first three energy levels.
6. A particle of mass m is in a cubic box. That is, the potential is zero inside a
cube of side L and infinite outside.
a) Find the three lowest allowed energies.
b) Find the number of states (level of degeneracy) at each of these three energies.
c) Find the Fermi Energy EF for N particles in the box. (N is large.)
7. A particle is confined in a rectangular box of length L, width W , and tallness T .
Find the energy eigenvalues in terms of a set of three quantum numbers (which
correspond to 3 mutually commuting operators). What are the energies of the
three lowest energy states if L = 2a, W = 1a, and T = 0.5a.
8. A particle of mass m is bound in the 3 dimensional potential V (r) = kr2 .
9. a) Find the energy levels for this particle.
10. b) Determine the number of degenerate states for the first three energy levels.
11. In 3 dimensions, a particle of mass m is bound in a potential V (r) =

e
.
x2 +z 2

a) The definite energy states will, of course, be eigenfunctions of H. What

other operators can they be eigenfunctions of?
b) Simplify the three dimensional Schr
odinger equation by using these operators.
274

14. Angular Momentum

14
14.1

TOC

Angular Momentum
Rotational Symmetry

If the potential only depends on the distance between two particles,

V (~r) = V (r)
the Hamiltonian has Rotational Symmetry. This is true for the coulomb (and
gravitational) potential as well as many others. We know from classical mechanics
that these are important problems.
If a the Hamiltonian has rotational symmetry, we can show that the Angular Momentum operators commute with the Hamiltonian.
[H, Li ] = 0
~ to be conserved,
We therefore expect the expectation value of each component of L
however, since the angular momentum operators dont commute with each other, we
will not be able to label our states with the quantum numbers for the three components
of angular momentum. Recall that we are looking for a set of mutually commuting
operators to label our energy eigenstates. We actually want two operators plus H to
give us three quantum numbers for states in three dimensions.
The components of angular momentum do not commute with each other
[Lx , Ly ] = i~Lz
Commutators of Angular Momentum Components
[Li , Lj ] = i~ijk Lk

but the square of the angular momentum commutes with any of the components
Commutators of L2
[L2 , Li ] = 0

These commutators lead us to choose the mutually commuting set of operators

to be H, L2 , and Lz . We could have chosen any component, however, it is most
convenient to choose Lz given the standard definition of spherical coordinates.
275

14. Angular Momentum

TOC

The Schr
odinger equation now can be rewritten with only radial derivatives and L2 .
~2 2
uE (~r) + V (r)uE (~r)
2
#
"
2

1
L2
~2 1
uE (~r) + V (r)uE (~r)
r
+

2 r2
r
r r ~2 r2

= EuE (~r)
= EuE (~r)

This leads to a great simplification of the 3D problem.

It is possible to separate the Schr
odinger equation since r and L2 appear separately. Write the solution as a product
Separation of Variables with Spherical Symmetry
uE (~r) = RE` (r)Y`m (, )
L2 Y`m (, ) = `(` + 1)~2 Y`m (, )
Lz Y`m (, ) = m~Y`m (, )

2
`(` + 1)~2
~2 2
+
R
(r)
+
V
(r)
+
RE` (r) = ERE` (r)
E`
2 r2
r r
2r2

where ` labels the eigenvalue of the L2 operator and m labels the eigenvalue of the
Lz operator. Since Lz does not appear in the Schr
odinger equation, we only label the
radial solutions with the energy and the eigenvalues of `. We have grouped the term
due to angular momentum with the potential. It is often called a pseudo-potential.
For ` 6= 0, it is like a repulsive potential.
By assuming the eigenvalues of L2 have the form `(` + 1)~2 , we have anticipated the
solution but not constrained it, since the units of angular momentum are those of
~ and since we expect L2 to have positive eigenvalues.
hY`m |L2 |Y`m i = hLx Y`m |Lx Y`m i + hLy Y`m |Ly Y`m i + hLz Y`m |Lz Y`m i 0
The assumption that the eigenvalues of Lz are some (dimensionless) number times ~
does not constrain our solutions at all.
We will use the algebra of the angular momentum operators to help us solve the
angular part of the problem in general.
For any given problem with rotational symmetry, we will need to solve a particular
differential equation in one variable r.
276

14. Angular Momentum

14.2

TOC

Angular Momentum Algebra: Raising and Lowering Operators

We have already derived the commutators of the angular momentum operators

[Lx , Ly ]

= i~Lz

[Li , Lj ]

= i~ijk Lk

[L2 , Li ]

We have shown that angular momentum is quantized for a rotor with a single angular
variable. To progress toward the possible quantization of angular momentum variables
in 3D, we define the operator L+ and its Hermitian conjugate L .
L Lx iLy .
Since L2 commutes with Lx and Ly , it commutes with these operators.
[L2 , L ] = 0
The commutator with Lz is.
[L , Lz ] = [Lx , Lz ] i[Ly , Lz ] = i~(Ly iLx ) = ~L .
From the commutators [L2 , L ] = 0 and [L , Lz ] = ~L , we can derive the effect of
the operators L on the eigenstates Y`m , and in so doing, show that ` is an integer
greater than or equal to 0, and that m is also an integer
Raising and Lowering Lz
` = 0, 1, 2, ...
` m `
m = `, ` + 1, ..., `
L Y`m = ~

`(` + 1) m(m 1)Y`(m1)

Therefore, L+ raises the z component of angular momentum by one unit of ~ and L

lowers it by one unit. The raising stops when m = ` and the operation gives zero,
L+ Y`` = 0. Similarly, the lowering stops because L Y`` = 0.
277

14. Angular Momentum

TOC

Angular momentum is quantized. Any measurement of a component of angular

momentum will give some integer times ~. Any measurement of the total angular
momentum gives the somewhat curious result
p
|L| = `(` + 1)~
where ` is an integer.
Note that we can easily write the components of angular momentum in terms of the
raising and lowering operators.
Lx

1
(L+ + L )
2
1
(L+ L )
2i

We will also find the following equations useful (and easy to compute).
[L+ , L ]
2

= i[Ly , Lx ] i[Lx , Ly ] = ~(Lz + Lz ) = 2~Lz

= L+ L + L2z ~Lz .

Example: What
is the expectation
value of Lz in the state
q
q
2
1
(~r) = R(r)( 3 Y11 (, ) i 3 Y11 (, ))?
Example: What
is the expectation
value of Lx in the state
q
q
2
1
(~r) = R(r)( 3 Y11 (, ) 3 Y10 (, ))?
278

14. Angular Momentum

TOC

Example: What are the eigenfunctions and eigenvalues of the Ly operator

for the ` 1 subspace?

14.3

The Angular Momentum Eigenfunctions

The angular momentum eigenstates are eigenstates of two operators.

Spherical Harmonics are Eigenfunctions of L2 and Lz
Lz Y`m (, ) = m~Y`m (, )
L2 Y`m (, ) = `(` + 1)~2 Y`m (, )

All we know about the states are the two quantum numbers ` and m. We have no
additional knowledge about Lx and Ly since these operators dont commute with Lz .
The raising and lowering operators L = Lx iLy raise or lower m, leaving `
unchanged.

L Y`m = ~

`(` + 1) m(m 1)Y`(m1)

The differential operators take some work to derive.

Angular Momentum Differential Operators

Lz =

L = ~e

~
i

+ i cot

279

14. Angular Momentum

TOC

Its easy to find functions that give the eigenvalue of Lz .

Y`m (, ) = ()() = ()eim
~
~
Lz Y`m (, ) =
()eim = im()eim = m~Y`m (, )
i
i
To find the dependence, we will use the fact that there are limits on m. The state
with maximum m must give zero when raised.

i
L+ Y`` = ~e
+ i cot
` ()ei` = 0

This gives us a differential equation for that state.

d()
+ i() cot (i`) = 0
d
d()
= ` cot ()
d
The solution is
() = C sin` .
Check the solution.

d
= C` cos sin`1 = ` cot
d

Its correct.
Here we should note that only the integer value of ` work for these solutions. If we
were to use half-integers, the wave functions would not be single valued, for example at
= 0 and = 2. Even though the probability may be single valued, discontinuities
in the amplitude would lead to infinities in the Schr
odinger equation. We will find later
that the half-integer angular momentum states are used for internal angular
momentum (spin), for which no or coordinates exist.
Therefore, the eigenstate Y`` is.
Y`m Derived from Operators
Y`` = C sin` ()ei`
Y`(`1) = CL Y``

We can compute the next state down by operating with L . We can continue to lower
m to get all of the eigenfunctions.
280

14. Angular Momentum

TOC

We call these eigenstates the Spherical Harmonics. The spherical harmonics are
normalized.
1

d Y`m
Y`m = 1

d(cos )
1

d Y`m
Y`m = 1

Since they are eigenfunctions of Hermitian operators, they are orthogonal.

Spherical Harmonics are Orthonormal

d Y`m
Y`0 m0 = ``0 mm0

We will use the actual function in some problems.

Some of the Spherical Harmonics
1
Y00 =
4
r
3 i
Y11 =
e sin
8
r
3
Y10 =
cos
4
r
15 2i
e
sin2
Y22 =
32
r
15 i
Y21 =
e sin cos
8
r
5
Y20 =
(3 cos2 1)
16

The spherical harmonics with negative m can be easily compute from those with positive m.
281

14. Angular Momentum

TOC

Y`m Formula

Y`(m) = (1)m Y`m

Any function of and can be expanded in the spherical harmonics.

f (, ) =

X
`
X

C`m Y`m (, )

`=0 m=`

The spherical harmonics form a complete set.

Spherical Harmonics form a Complete Set

X
`
X

|Y`m i hY`m | =

`=0 m=`

X
`
X

|`mi h`m| = 1

`=0 m=`

When using bra-ket notation, |`mi is sufficient to identify the state.

The spherical harmonics are related to the Legendre polynomials which are
functions of .
Spherical Harmonics are Related to Legendre Polynomials

Y`0 (, ) =

Y`m = (1)

2` + 1
4

2` + 1 (` m)!
4 (` + m)!

P` (cos )
12

P`m (cos )eim

282

14. Angular Momentum

14.3.1

TOC

Parity of the Spherical Harmonics

In spherical coordinates, the parity operation is

+ .

The radial part of the wavefunction, therefore, is unchanged and the

R(r) R(r)
parity of the state is determined from the angular part. We know the state
Y`` in general. A parity transformation gives.
Y`` (, ) Y`` ( , + ) = ei` ei` sin` () = ei` Y`` = (1)` Y``
The states are either even or odd parity depending on the quantum number `.
parity = (1)`
The angular momentum operators are axial vectors and do not change sign under a
parity transformation. Therefore, L does not change under parity and all the Y`m
with have the same parity as Y``
L Y`` (1)` L Y``
Y`m ( , + ) = (1)` Y`m (, )

Parity of Spherical Harmonics

P Y`m = (1)` Y`m

14.4
14.4.1

Derivations and Computations

Rotational Symmetry Implies Angular Momentum Conservation

In three dimensions, this means that we can change our coordinates by rotating about
any one of the axes and the equations should not change. Lets try and infinitesimal
283

14. Angular Momentum

TOC

rotation about the z axis. The x and y coordinates will change.

x0 = x + dy
y 0 = y dx

The original Schr

odinger Equation is
H(x, y, z) = E(x, y, z)
and the transformed equation is
H(x + dy, y dx, z) = E(x + dy, y dx, z).

Now we Taylor expand this equation.

H(x, y, z) + Hd
y
x = E(x, y, z) + Ed
y
x
x
y
x
y
Subtract off the original equation.

H
y
x =
y
x H
x
y
x
y
We find an operator that commutes with the Hamiltonian.

~
x
y
=0
H,
i
y
x
Note that we have inserted the constant ~i in anticipation of identifying this operator
as the z component of angular momentum.
~ = ~r p~
L

~

Lz =
x
y
= xpy ypx
i
y
x
We could have done infinitesimal rotations about the x or y axes and shown that all
the components of the angular momentum operator commute with the Hamiltonian.
[H, Lz ] = [H, Lx ] = [H, Ly ] = 0
Remember that operators that commute with the Hamiltonian imply physical quantities that are conserved.
284

14. Angular Momentum

14.4.2

TOC

The Commutators of the Angular Momentum Operators

[Lx , Ly ] 6= 0,

however, the square of the angular momentum vector commutes with all the components.
[L2 , Lz ] = 0
This will give us the operators we need to label states in 3D central potentials.
Lets just compute the commutator.
[Lx , Ly ]

[ypz zpy , zpx xpz ] = y[pz , z]px + [z, pz ]py x

~
[ypx xpy ] = i~Lz
i

=
=

Since there is no difference between x, y and z, we can generalize this to

[Li , Lj ] = i~ijk Lk
where ijk is the completely antisymmetric tensor and we assume a sum over repeated
indices.
ijk = jik = ikj = kji
The tensor is equal to 1 for cyclic permutations of 123, equal to -1 for anti-cyclic
permutations, and equal to zero if any index is repeated. It is commonly used for a
cross product. For example, if
~ = ~r p~
L
then
Li = rj pk ijk
where we again assume a sum over repeated indices.
Now lets compute commutators of the L2 operator.
L2
[Lz , L2 ]

= L2x + L2y + L2z

[Lz , L2x ] + [Lz , L2y ]

[Lz , Lx ]Lx + Lx [Lz , Lx ] + [Lz , Ly ]Ly + Ly [Lz , Ly ]

= i~(Ly Lx + Lx Ly Lx Ly Ly Lx ) = 0

We can generalize this to

[Li , L2 ] = 0.
~
L2 commutes with every component of L.
285

14. Angular Momentum

14.4.3

Rewriting

p2
2

TOC

Using L2

We wish to use the L2 and Lz operators to help us solve the central potential problem.
If we can rewrite H in terms of these operators, and remove all the angular derivatives,
problems will be greatly simplified. We will work in Cartesian coordinates for a while,
where we know the commutators.
First, write out L2 .
L2

=
=

(~r p~)2
"
2
2
2 #

2
~
y
z
x
y
+ z
+ x
z
y
x
z
y
x

Group the terms.

L2 =

2

2

2

2
2
2

2
2
+
+
+
~2 x 2
+
y
+
z
z 2
y 2
x2
z 2
x2
y 2

2
2

2
+ 2yz
+ 2xz
+ 2x
+ 2y
+ 2z
2xy
xy
yz
xz
x
y
z

We expect to need to keep the radial derivatives so lets identify those by dotting ~r into
p~. This will also make the units match L2 .

2

2
2
(~r p~)
=
~ x
+y
+z
x
y
z

2
2

2
2

2
2
=
~2 x2 2 + y 2 2 + z 2 2 + 2xy
+ 2yz
+ 2xz
x
y
z
xy
yz
xz

+x
+y
+z
x
y
z

By adding these two expressions, things simplify a lot.

L2 + (~r p~)2 = r2 p2 + i~~r p~
We can now solve for p2 and we have something we can use in the Schrodinger equation.
p2

=
=

1
L2 + (~r p~)2 i~~r p~
2
r

2
1 2 ~2

1
L 2 r
~2
2
r
r
r
r r
286

14. Angular Momentum

TOC

The Schr
odinger equation now can be written with only radial derivatives and L2 .
"
#
2

~2 1
1
L2
r
+

uE (~r) + V (r)uE (~r) = EuE (~r)

2 r2
r
r r ~2 r2

14.4.4

Spherical Coordinates and the Angular Momentum Operators

The transformation from spherical coordinates to Cartesian coordinate is.

= r sin cos

= r sin sin

= r cos

The transformation from Cartesian coordinates to spherical coordinates is.

p
x2 + y 2 + z 2
r =
z
cos = p
2
x + y2 + z2
y
tan =
x

We now proceed to calculate the angular momentum operators in spherical coordinates.

The first step is to write the x

in spherical coordinates. We use the chain rule and the
i
above transformation from Cartesian to spherical. We have used d cos = sin d and
d tan = cos12 d. Ultimately all of these should be written in the sperical cooridnates
but its convenient to use x for example in intermediate steps of the calculation.
287

14. Angular Momentum

=
=
=
=

TOC

r
cos
tan
+
+
x r
x cos
x tan
x
xz 1
y

+ 3

cos2
r r
r sin x2

1
1
1 sin sin

sin cos
+ sin cos cos

cos2
2
2
r r
sin
r sin cos

1 sin
sin cos
+ cos cos

r r

r sin
r
cos
tan
+
+
y r
y cos
y tan
y
yz 1
1

+ 3
+ cos2
r r
r sin x

1
1
1
1

sin sin
+ sin sin cos
+
cos2
r r
sin
r sin cos

1 cos

+ sin cos
+
sin sin
r r

r sin
r
cos
tan
+
+
z r
z cos
z tan

1 z 2 1
z
+
3
r r
r
r
sin
1
1

+
1 cos2
cos
r r
sin

cos
sin
r r

Bringing together the above results, we have.

1 sin
+ cos cos

r r

r sin

1 cos
= sin sin
+ sin cos
+
r r

r sin

= cos
sin
r
r

sin cos

288

14. Angular Momentum

TOC

Now simply plug these into the angular momentum formulae.

~
Lz
=
x
y
=
i
y
x
i

~
y
z
i z
x
L
=
Lx iLy =
i
z
y
x
z

~

=
(y ix)
z
i
i
z
y
x

z
i
=
~ (x iy)
z
x
y

i
=
~r sin e
cos
i
z
x
y

i
=
~ sin e
r cos
sin
r

iei
i
i
cos r sin e
+ cos e

sin

i cos

2
2
i
cos

=
~e
sin

sin

+ i cot
=
~ei

We will use these results to find the actual eigenfunctions of angular momentum.

Lz
L

14.4.5

~
i

= ~e

+ i cot

The Operators L

The next step is to figure out how the L operators change the eigenstate Y`m . What
eigenstates of L2 are generated when we operate with L+ or L ?
L2 (L Y`m ) = L L2 Y`m = `(` + 1)~2 (L Y`m )
Because L2 commutes with L , we see that we have the same `(` + 1) after operation.
This is also true for operations with Lz .
L2 (Lz Y`m ) = Lz L2 Y`m = `(` + 1)~2 (Lz Y`m )
289

14. Angular Momentum

TOC

The operators L+ , L and LZ do not change `. That is, after we operate, the new
state is still an eigenstate of L2 with the same eigenvalue, `(` + 1).
The eigenvalue of Lz is changed when we operate with L+ or L .
Lz (L Y`m )

= L Lz Y`m + [Lz , L ]Y`m

= m~(L Y`m ) ~L Y`m = (m 1)~(L Y`m )

(This should remind you of the raising and lowering operators in the HO solution.)
From the above equation we can see that (L Y`m ) is an eigenstate of Lz .
L Y`m = C (`, m)Y`(m1)
These operators raise or lower the z component of angular momentum by one unit of
~.
Since L = L , its easy to show that the following is greater than zero.
hL Y`m |L Y`m i 0
hY`m |L L Y`m i 0

Writing L+ L in terms of our chosen operators,

L L

(Lx iLy )(Lx iLy ) = L2x + L2y iLx Ly iLy Lx

= L2x + L2y i[Lx , Ly ] = L2 L2z ~Lz

we can derive limits on the quantum numbers.
hY`m |(L2 L2z ~Lz )Y`m i 0
(`(` + 1) m2 m)~2 0
`(` + 1) m(m 1)

We know that the eigenvalue `(` + 1)~2 is greater than zero. We can assume that
`0
because negative values just repeat the same eigenvalues of `(` + 1)~2 .
The condition that `(` + 1) m(m 1) then becomes a limit on m.
` m `
290

14. Angular Momentum

TOC

Now, L+ raises m by one and L lowers m by one, and does not change `. Since m is
limited to be in the range ` m `, the raising and lowering must stop for m = `,
L Y`(`) = 0
L+ Y`` = 0
The raising and lowering operators change m in integer steps, so, starting from m = `,
there will be states in integer steps up to `.
m = `, ` + 1, ..., ` 1, `
Having the minimum at ` and the maximum at +` with integer steps only works if `
is an integer or a half-integer. There are 2` + 1 states with the same ` and different
values of m. We now know what eigenstates are allowed.
The eigenstates of L2 and Lz should be normalized
hY`m |Y`m i = 1.
The raising and lowering operators acting on Y`m give
L Y`m = C (`, m)Y`(m1)
The coefficient C (`, m) can be computed.
hL Y`m |L Y`m i = |C(`, m)|2 hY`(m1) |Y`(m1) i
= |C(`, m)|2
hL Y`m |L Y`m i = hY`m |L L Y`m i
= hY`m |(L2 L2z ~Lz )Y`m i
(`(` + 1) m2 m)~2

|C(`, m)|2 = `(` + 1) m2 m ~2
p
C (l, m) = ~ `(` + 1) m(m 1)
=

We now have the effect of the raising and lowering operators in terms of the normalized
eigenstates.
p
L Y`m = ~ `(` + 1) m(m 1)Y`(m1)

14.5
14.5.1

Examples
The Expectation Value of Lz

What is the expectation value of Lz in the state (~r) = R(r)(

2
3 Y11 (, )i

1
3 Y11 (, ))?

291

14. Angular Momentum

TOC

The radial part plays no role. The angular part is properly normalized.
*r
+
r
r
r
2
1
2
1
h|Lz |i =
Y11 i
Y11 |Lz |
Y11 i
Y11
3
3
3
3
+
*r
r
r
r
2
1
2
1
Y11 i
Y11 |
~Y11 + i
~Y11
=
3
3
3
3

2 1
1
=

~= ~
3 3
3

14.5.2

The Expectation Value of Lx

What is the expectation value of Lx in the state (~r) = R(r)(

2
3 Y11 (, )

1
3 Y10 (, ))?

We need Lx = (L+ + L )/2.

*r
+
r
r
r
1
2
1
2
1
Y11
Y10 |L+ + L |
Y11
Y10
h|Lx |i =
2
3
3
3
3
*r
+
r
r
r
1
2
1
2
1
=
Y11
Y10 |
L Y11
L+ Y10
2
3
3
3
3
+
*r
r
r
r
~
2
1
2
1
2Y10
2Y11
=
Y11
Y10 |
2
3
3
3
3
=

14.5.3

~2
2
(1 1) = ~
23
3

The Eigenstates of Ly

Find the eigenstates of the Ly operator in the subspace with ` = 1.

There are three spherical harmonics with ` = 1 and the eigenstates must be linear
combinations of these. The possible eigenvalues must be the same as the eigenvalues
of Lz , since there is no physical difference between y an z. We can write Ly = (L+
L )/2i. For example, the state with an eigenvalue of +~ must satisfy the equation:
(y)

(y)

Ly X+ = ~X+
The possible linear combinations are:
(y)

X+ = aY11 + bY10 + cY11

292

14. Angular Momentum

TOC

We then write the equation we can use to find the coefficients a, b and c.
(L+ L )(aY11 + bY10 + cY11 )

L Y`m
2~(a0 + bY11 + cY10 aY10 bY11 c0)
bY11 + (c a)Y10 bY11

+2i~(aY11 + bY10 + cY11 )

p
= ~ `(` + 1) m(m 1)Y`(m1)
=

+2i~(aY11 + bY10 + cY11 )

(aY11 + bY10 + cY11 )

2i
1
i
1
Y11 + Y10 Y11
2
2
2

a = c =
(y)

X+

bY11 + (c a)Y10 bY11

2i
(y)

bY11 + (c a)Y10 bY11

=
=

1
1
Y11 + 0Y10 + Y11
2
2

= (aY11 + bY10 + cY11 )

b
a = c =
2i
i
1
1
(y)
Y11 Y10 Y11
X
=
2
2
2
We will find this type of problem more convenient in matrix and vector notation.

14.6

Homework

1. A particle is in the state = R(r)

1
3 Y21

1
3 Y20

1
3 Y22

. Find the

expected values of L , Lz , Lx , and Ly .

important fundamental easy
hL2 i = 2(2 + 1)~2 = 6~2
since ` = 2 for all terms
hLz i = (1 +p
0 + 2) 13 ~ = ~
each term has prob=1/3
L Y`m = ~ `(` + 1) m(m 1) Y`,m1
hLx i = 12 hL+ + L i
write Lx in terms of L

11
hLx i = 2 3 ~hY
Y22 |2Y
6Y20 + i 6Y21 2Y21 i
21 + iY20
22 + i 6Y21 0 +

hLx i = 16 ~ i 6 2 i 6 2 = 23 ~
1
hLy i = 2i
hL+ L i

1 1
= 2i 3 ~hY21 + iY20 Y22 |2Y22 + i 6Y

i
6Y21 + 2Y21 i
21
20

1
hLy i = 6i
~ i 6 2 + i 6 + 2 = 36 ~
q
q

1
2
2. A particle is in the state = R(r)
Y
+
i
3 11
3 Y10 . If a measurement of the
293

14. Angular Momentum

TOC

x component of angular momentum is made, what are the possilbe outcomes and
what are the probabilities of each?
important fundamental medium
Of course since these are ` = 1 states, the possible outcomes will be ~, 0, and +~.
Find eigenstates of Lx for ` = 1
1
L )(aY11
+ bY10 + cY11 ) = m~(aY
11 + bY10 + cY11 )
2 (L+ +

1
~(0
+
2bY
+
2cY
+
a
2Y
+
2bY
11
10
10
11 + 0) = m~(aY11 + bY10 + cY11 )
2
1 (bY11 + (a + c)Y10 + bY11 + 0) = m(aY11 + bY10 + cY11 )
2
q

(x)
m=1 = R(r) 12 Y11 + 12 Y10 + 12 Y11
solve and normalize
q
q

(x)
1
m=0 = R(r)
Y11 + 0Y10 12 Y11
q
2

(x)
m=1 = R(r) 21 Y11 12 Y10 + 12 Y11
q q 2

2 q

(x)

5
dot product= amplitude
measure +~: P = hm=1 |i = 13 21 + i 23 12 = 12
q 2
q q

measure 0~: P = 13 12 + i 23 0 = 16
q q 2
q

5
measure ~: P = 13 12 i 23 12 = 12
3. Calculate the matrix elements hY`m1 |Lx |Y`m2 i and hY`m1 |L2x |Y`m2 i
computation medium
Lx = 12 (L+ p
+ L )
L Y`m = ~ `(` + 1) m(m 1) Y`,m1
hY`m1 |Lx |Y`m
i = 12 hY`m1 |L+ + L |Y`m2 i
2
p
p
`(` + 1) m2 (m2 1) Y`m2 1 i
= 12 ~hY
`m1 | `(` + 1) m2 (m2 + 1) Y`m2 +1 +
p
= 21 ~ `(` + 1) m1 m2 (m1 ,m2 +1 + m1 ,m2 1 )
= 14 hY`m1 |L+ L+ + L+ L + L L+ + L L |Y`m2 i
1
2
2
= 41 hY
2 `(` + 1) m2 ~ m1 m2
p`m1 |L+ L+ + L L |Y`m2 i + p
1
= 4p `(` + 1) (m2 + 1)(m2 + 2)p `(` + 1) m2 (m2 + 1)~2 m1 ,m2 +2
+ 41 `(` + 1) (m2 1)(m2 2) `(` + 1) m2 (m2 1)~2 m1 ,m2 2
+ 21 `(` + 1) m22 ~2 m1 m2
L2 +L2

`(`+1)m2
2I1

m2
2I2

E`m = ~2 m2 2I12 2I11 + `(`+1)

2I
1
2

`(`+1)
I2
m
E`m = ~2 2I
1

+
I1
2I1
2

spectrum dominated by m
with small changes due to `

5. Prove that hL2x i = hL2y i = 0 is only possible for ` = 0.

294

14. Angular Momentum

TOC

proof easy

hL2x + L2y i = hL2 L2z i = `(` + 1) m2 ~2
= `(` + 1) `2 ~2 = `~2

min value for m = `

only zero for ` = 0

6. Write the spherical harmonics for ` 2 in terms of the Cartesian coordinates x,

y, and z.
computation
medium q
q
q
2
15
15
15 (x+iy)
2i
sin2 = 32
(cos + i sin )2 sin2 = 32
Y22 = 32 e
2
r
q
2
2
15 x y +2ixy
Y22 = 32
r2
q
q
15 i
15 (x+iy)z
Y21 = 8
e sin cos = 8
r2
q
q
2
5
5
Y20 = 16
(3 cos2 1) = 16
(3 zr2 1)
7. A particle in a spherically symmetric potential has the wave-function (x, y, z) =
2
C(xy + yz + zx)er . A measurement of L2 is made. What are the possible
results and the probabilities of each? If the measurement of L2 yields 6~2 , what
are the possible measured values of Lz and what are the corresponding probabilities?
computaton hard
All of these terms come from a Y2m since they are all second order in a coordinate.
So we will always measure L2 = 6~2 .

Y`(m)q= (1)m Y`m

xy =

32 1
4i (Y22

q15

Y22 )

8 1
(Y21 + Y21 )
q 15 2i
8 1
zx = 15 2 (Y21 Y21 )
(x, y, z) iY22 + iY22 + iY21 + iY21 Y21 + Y21
2~:
P=1/6
2~:
P=1/6
~:
P=1/3
~:
P=1/3

yz =

14.7

Sample Test Problems

1. Derive the commutators [L2 , L+ ] and [Lz , L+ ]. Now show that L+ Y`m = CY`(m+1) ,
that is, L+ raises the Lz eigenvalue but does not change the L2 eigenvalue.
Answer
L+ = Lx + iLy
Since L2 commutes with both Lx and Ly ,
[L2 , L+ ] = 0.
295

14. Angular Momentum

TOC

[Lz , L+ ] = [Lz , Lx +iLy ] = [Lz , Lx ]+i[Lz , Ly ] = i~(Ly iLx ) = ~(Lx +iLy ) = ~L+
We have the commutators. Now we apply them to a Y`m .
[L2 , L+ ]Y`m = L2 L+ Y`m L+ L2 Y`m = 0
L2 (L+ Y`m ) = `(` + 1)~2 (L+ Y`m )
So, L+ Y`m is also an eigenfunction of L2 with the same eigenvalue. L+ does not
change `.
[Lz , L+ ]Y`m = Lz L+ Y`m L+ Lz Y`m = ~L+ Y`m
Lz (L+ Y`m ) m~(L+ Y`m ) = ~(L+ Y`m )
Lz (L+ Y`m ) = (m + 1)~(L+ Y`m )
So, L+ raises the eigenvalue of Lz .
2. Write the (normalized) state which is an eigenstate of L2 with eigenvalue `(` +
1)~2 = 2~2 and also an eigenstate of Lx with eigenvalue 0~ in terms of the
usual Y`m .
Answer
An eigenvalue of `(`+1)~2 = 2~2 implies ` = 1. We will need a linear combination

of the Y1m to get the eigenstate of Lx = L+ +L

.
2
L+ + L
(AY11 + BY10 + CY11 ) = 0
2
~
(AY10 + BY11 + BY11 + CY10 ) = 0
2
(BY11 + (A + C)Y10 + BY11 ) = 0
Since this is true for all and , each term must be zero.
B=0
A = C
The state is

1
(Y11 Y11 )
2

The trivial solution that A = B = C = 0 is just a zero state, not normalizable to

1.
3. Write the (normalized) state which is an eigenstate of L2 with eigenvalue `(` +
1)~2 = 2~2 and also an eigenstate of Ly with eigenvalue 1~ in terms of the
usual Y`m .
4. Calculate the commutators [pz , Lx ] and [L2x , Lz ].
p
5. Derive the relation L+ Y`m = ~ `(` + 1) m(m + 1)Y`(m+1) .

296

14. Angular Momentum

TOC

6. A particle is in a l = 1 state and is known to have angular momentum in the x

direction equal to +~. That is Lx = ~. Since we know l = 1, must have
the form = R(r)(aY1,1 + bY1,0 + cY1,1 ). Find the coefficients a, b, and c for
normalized.
7. Calculate the following commutators: [x, Lz ], [L+ , L2 ], [ 21 m 2 r2 , px ].
8. Prove that, if the Hamiltonian is symmetric under rotations, then [H, Lz ] = 0.
9. In 3 dimensions, a particle is in the state:
(r) = C(iY33 2Y30 + 3Y31 ) R(r)
where R(r) is some arbitrary radial wave function normalized such that

R (r)R(r)r2 dr = 1.
0

a) Find the value of C that will normalize this wave function.

b) If a measurement of Lz is made, what are the possible measured values and
what are probabilities for each.
c) Find the expected value of < Lx > in the above state.
10. Two (different) atoms of masses M1 and M2 are bound together into the ground
state of a diatomic molecule. The binding is such that radial excitations can be
neglected at low energy and that the atoms can be assumed to be a constant
distance a = 3
A apart. (We will ignore the small spread around r = a.)
a) What is the energy spectrum due to rotations of the molecule?
b) Assuming that R(r) is given, write down the energy eigenfunctions for the
ground state and the first excited state.
c) Assuming that both masses are about 1000 MeV, how does the excitation
energy of the first excited state compare to thermal energies at 300 K.

297

15. Hydrogen

TOC

Hydrogen

The Hydrogen atom consists of an electron bound to a proton by the Coulomb

potential. We would like to use the dimensionless fine structure constant and
the standard atomic (nuclear) units of eV and nm (or MeV and fm) to describe the
potential.

Vcgs (r) =
VSI (r) =

e2
r

e2
40 r

VQM (r) =

~c
r

We can generalize the potential to a nucleus of charge Ze without complication of the

problem, so the Hamiltonian for the Hydrogenic Problem is:
H=

~2 2 Z~c

2
r

which has spherical symmetry.

Since the potential is spherically symmetric, the problem separates and the solutions will be a product of a radial wavefunction and one of the spherical harmonics.
n`m (~r) = Rn` (r)Y`m (, )
We have already studied the spherical harmonics.
The radial wavefunction satisfies the differential equation that depends on the
angular momentum quantum number `.

298

15. Hydrogen

TOC

The Hydrogenic Problem

~2 2 Z~c

2
r

n`m (~r) = Rn` (r)Y`m (, )

d2
2 d
+
2
dr
r dr

RE` (r) +

2
~2

me mN
me + mN

E+

Z~c `(` + 1)~2

r
2r2

RE` (r) = 0

The differential equation can be solved using techniques similar to those used to
solve the 1D harmonic oscillator equation. We find the eigen-energies and the
radial wavefunctions written in the dimensionless variable where the coefficients
of the polynomials can be found from the recursion relation.

The Solution to the Hydrogenic Problem

1 2 2 2
Z c
2n2

Rn` () =

ak k e/2

k=0

ak+1 =

k+`+1n
k nr
ak =
ak
(k + 1)(k + 2` + 2)
(k + 1)(k + 2` + 2)
r
=

8E
2Z
r=
r
~2
na0

n = nr + ` + 1
n = 1, 2, 3, ...

The principle quantum number n is an integer from 1 to infinity. This principle

quantum number is actually the sum of the radial quantum number plus ` plus 1, and
therefore, the total angular momentum quantum number ` must be less than n.
` = 0, 1, 2, ..., n 1
299

15. Hydrogen

TOC

This choice of n predated the Hydrogen solution and obviously gives a simple formula
for the Energies. This unusual way of labeling the states comes about because a radial
excitation has the same energy as an angular excitation for Hydrogen. This is often
referred to as an accidental degeneracy.

15.1

The Radial Wavefunction Solutions

Defining the Bohr radius,

Define Bohr Radius

~
~

mc
c

we can compute the radial wave functions. Here is a list of the first several radial wave
functions Rn` (r).

Some Radial Wavefunction Solutions

R10 = 2

R21

1
=
3

R20 = 2

R32

R31

Z
2a0

Z
a0

eZr/a0
Zr
a0

Zr
1
2a0

eZr/2a0

3 2
2 2
Z 2 Zr
=
eZr/3a0
a0
27 5 3a0

3

Z 2 Zr
Zr
4 2
=
1
eZr/3a0
3
3a0
a0
6a0

R30 = 2

Z
3a0

Zr
2 (Zr)
1
+
3a0
27a20

!
eZr/3a0

300

15. Hydrogen

TOC

Note that the radial wavefunctions can be real functions for hydrogen bound states.
The ` = 0 wavefunctions are finite at the origin while for ` > 0 the wavefunctions
go to zero at the origin. For a given principle quantum number n,the largest ` radial
wavefunction is given by
Rn,n1 rn1 eZr/na0
The radial wavefunctions should be normalized as below.

r2 Rn`
Rn` dr = 1
0

Example: Compute the expectedvalues of E, L2 , Lz , and Ly in the Hydrogen state 16 (4100 + 3211 i210 + 10211 ).
The pictures below depict the probability distributions in space for the Hydrogen wavefunctions.

301

15. Hydrogen

TOC

The graphs below show the radial wave functions. Again, for a given n the maximum
` state has no radial excitation, and hence no nodes in the radial wavefunction. As ell
gets smaller for a fixed n, we see more radial excitation.

302

15. Hydrogen

TOC

303

15. Hydrogen

TOC

A useful integral for Hydrogen atom calculations is.

Useful Hydrogen Radial Integral

dx xn eax =

n!
an+1

This can be easily derived by integrating by parts n times. All the finite terms give
zero due to x 0 at r = 0 and eax 0 at infinity.
304

15. Hydrogen

Example: What is the expectation value of

TOC

1
r

in the state 100 ?

Example: What is the expectation value of r in the state 100 ?

Example: What is the expectation value of the radial component of velocity squared vr2 in the state 100 ?

15.2

The Hydrogen Spectrum

The figure shows the transitions between Hydrogen atom states.

The ground state of Hydrogen has n = 1 and ` = 0. This is conventionally called the
1s state. The convention is to name ` = 0 states s, ` = 1 states p, ` = 2 states
d, and ` = 3 states f. From there on follow the alphabet with g, h, i, ...
305

15. Hydrogen

TOC

The first excited state of Hydrogen has n = 2. There are actually four degenerate
states (not counting different spin states) for n = 2. In terms of n`m , these are 200 ,
211 , 210 , and 211 . These would be called the 2s and 2p states. Remember, all
values of ` < n are allowed.
The second excited state has n = 3 with the 3s, 3p and 3d states being degenerate.
This totals 9 states with the different allowed m values.
In general there are n2 degenerate states, again not counting different spin states.
The Hydrogen spectrum was primarily investigated by measuring the energy of photons
emitted in transitions between the states, as depicted in the figures above and below.

Transitions which change ` by one unit are strongly preferred, as we will later learn.

306

15. Hydrogen

15.3
15.3.1

TOC

Derivations and Calculations

k+`+1
(k + ` + 1)(k + `) + 2(k + ` + 1) `(` + 1)
k+`+1
k+`+1
=
k(k + 2` + 1) + 2(k + ` + 1)
k(k + 2` + 2) + (k + 2` + 2)
k+`+1
1

(k + 1)(k + 2` + 2)
k

This is then the power series for

G() ` e
308

15. Hydrogen

TOC

unless it somehow terminates. We can terminate the series if for some value of
k = nr ,
= nr + ` + 1 n.
The number of nodes in G will be nr . We will call n the principal quantum number,
since the energy will depend only on n.
Plugging in for we get the energy eigenvalues.
r
c2
= n.
Z
2E
1
E = 2 Z 2 2 c2
2n

The solutions are

Rn` () = `

ak k e/2 .

k=0

The recursion relation is

ak+1 =

k+`+1n
k nr
ak =
ak .
(k + 1)(k + 2` + 2)
(k + 1)(k + 2` + 2)

We can rewrite , substituting the energy eigenvalue.

r
r
8E
42 c2 Z 2 2
2cZ
2Z
=
r=
r=
r=
r
2
2
2
~
~ n
~n
na0

15.3.2

Computing the Radial Wavefunctions *

The radial wavefunctions are given by

R() = `

n`1
X

ak k e/2

k=0

where
=

2Z
r
na0

and the coefficients come from the recursion relation

ak+1 =

k+`+1n
ak .
(k + 1)(k + 2` + 2)
309

15. Hydrogen

TOC

The series terminates for k = n ` 1.

Lets start with R10 .
R10 (r) = 0

0
X

ak k e/2

k=0

R10 (r) = CeZr/a0

We determine C from the normalization condition.

r2 Rn`
Rn` dr = 1
0

r2 e2Zr/a0 dr = 1

|C|2
0

This can be integrated by parts twice.

a 2
0

|C|

e2Zr/a0 dr = 1

2
0

a 3
0

|C|2 = 1
3
1 2Z
2
C =
2 a0
32
1
2Z
C=
2 a0
32
Z
R10 (r) = 2
eZr/a0
a0
2

R21 can be computed in a similar way. No recursion is needed.

310

15. Hydrogen

TOC

Lets try R20 .

R20 (r) = 0

1
X

ak k e/2

k=0

R20 (r) = (a0 + a1 )e/2

k+`+1n
ak
ak+1 =
(k + 1)(k + 2` + 2)
0+0+12
1
a1 =
a0 =
a0
(0 + 1)(0 + 2(0) + 2)
2

Zr
eZr/2a0
R20 (r) = C 1
2a0
We again normalize to determine the constant.

15.4

Examples

15.4.1

Expectation Values in Hydrogen States

An electron in the Coulomb field of a proton

is in the state described by the

wave function = 16 (4100 + 3211 i210 + 10211 ). Find the expected value of
the Energy, L2 , Lz , and Ly .
First check the normalization.

36
|4|2 + |3|2 + | i|2 + | 10|2
=
=1
36
36
The terms are eigenstates of E, L2 , and Lz , so we can easily compute expectation
values of those operators.
1
1
2 c2 2
2
n
16 12 + 9 212 + 1 212 + 10 212
21
7
1
1
1
hEi = 2 c2 1
= 2 c2
= 2 c2
2
36
2
36
2
12
En

Similarly, we can just square probability amplitudes to compute the expectation value
of L2 . The eigenvalues are `(` + 1)~2 .
hL2 i = ~2

16(0) + 9(2) + 1(2) + 10(2)

10 2
=
~
36
9

311

15. Hydrogen

TOC

The Eigenvalues of Lz are m~.

hLz i = ~

16(0) + 9(1) + 1(0) + 10(1)

1
=
~
36
36

Computing the expectation value of Ly is harder because the states are not eigenstates
of Ly . We must write Ly = (L+ L )/2i and compute.
hLy i = h|Ly i

1
= 72i
h4100 + 3211 i210 + 10211 |L+ L |4100 + 3211 i210 + 10211 i
1
h4100 + 3211 i210 + 10211 | 3L
= 72i
211 i(L
+ L )
210 + 10L+ 211
i
~
=
h4
+
3

i
+
10
|

3
2

i
2
+
i
2
+
10
2210 i
100
211
210
211
210
211
211
72i

2~
= 72i
h4100 + 3211 i210 + 10211| 3

i
+
i
+
10
i
211
210
210
211

2)~
2~
= 72i
(3i 3i + 10i + 10i) = (6+2 72i10)i 2~ = (2 53
36

15.4.2

The Expectation of

R10

1
h100 | |100 i =
r

1
r

in the Ground State

Z
a0

eZr/a0

Y00
Y00

1
r2 R10
R10 dr
r

rR10
R10

Z
a0

2Zr/a0

dr = 4

Z
a0

a0 2
1!
2Z

We can compute the expectation value of the potential energy.

h100 |

15.4.3

Z~c
Z 2 ~c
mc
|100 i =
= Z 2 ~c
= Z 2 2 mc2 = 2E100
r
a0
~

The Expectation Value of r in the Ground State

h100 |r|100 i =

r
0

R10
R10

dr = 4

Z
a0

r3 e2Zr/a0 dr = 3!

1 a0
3 a0
=
4Z
2Z

312

15. Hydrogen

15.4.4

TOC

The Expectation Value of vr2 in the Ground State

Clearly the average of the radial velocity must be zero, or the electron would migrate
to higher or lower radius in this stationary state. We can compute the square of the
radial velocity, and hence the RMS.
For ` = 0, there is no angular dependence to the wavefunction so no velocity except in
the radial direction. So it makes sense to compute the radial component of the velocity
which is the full velocity.
We can find the term for

p2r
2m

in the radial equation.

h100 |(vr ) |100 i =

2
~
R10
2

2 d
d2
+
2
dr
r dr

R10 dr

~2
4
m2

Z
a0

r e

Zr
a0

=
=

Since a0 =

~
mc ,

Z2
2Z

a20
a0 r

Zr
a0

3 2

~2
Z
Z
a0 3 2Z a0 2
4
2

m2
a0
a20
2Z
a0 2Z

2
~2 Z
m2 a0

we get
h100 |(vr )2 |100 i = Z 2 2 c2

For Z = 1, the RMS velocity is c or

1
137

We can compute the expected value of the kinetic energy.

K.E. =

1
~2 Z 2
1
mv 2 =
= Z 2 2 mc2 = E100
2
2m a20
2

This is what we expect from the Virial theorem.

15.5

Homework

1. Tritium is a unstable isotope of Hydrogen with a proton and two neutrons in

the nucleus. Assume an atom of Tritium starts out in the ground state. The
313

15. Hydrogen

TOC

nucleus (beta) decays suddenly into that of He3 . Calculate the probability that
the electron remains in the ground state.
important fundamental easy
Zr
3
radial wfn for ground state
R10 = 2( aZ0 ) 2 e a0

(Z=2) (Z=1) 2
Z changes, nuclear mass almost
Pg.s. = h100 |100 i

2
2r
r
3
3

Pg.s. = 2( a20 ) 2 e a0 Y00

plug in wfns
2( a10 ) 2 e a0 Y00 r2 dr d

2
2 23 1 32 3r

Pg.s. = 2( a0 ) 2( a0 )
e a0 r2 dr
plug in wfns

9 6
7 2!a3 2
9
2 a
Pg.s. = a26 330 = 36 a06 = 326 = 0.702
plug in wfns
0

2. A hydrogen atom is in the state = 16 4100 + 3211 210 + 10211 .
What are the possible energies that can be measured and what are the probabilities of each? What is the expectation value of L2 ? What is the expectation
value of Lz ? What is the expectation value of Lx ?
important fundamental easy
E1 = 13.6 eV, with prob. 49
E2 = 13.6/4 eV, with prob. 59
hL2 i = 49 0 + 59 2 ~2 = 10
9

9
1
10
1
hLz i = 16
36 0 + 36 1 + 36 0 +36 (1) ~ = 36 ~

1
hLx i = 72
h3
10
10211 i
211 210 +
211 |L+ + L |3211 210 +

1
hLx i = 72
2~h3211 210 + 10211 | 211 + 10210 + 3210 211 i

1
hLx i = 72
2~ 3 10 3 10 = 362 (3 + 10)~ = 0.242~
3. What is P (pz ), the probability distribution of pz for the Hydrogen energy eigenstate 210 ? You may find the expansion of eikz in terms of Bessel functions useful.
extension difficult
4. An electron in the Hydrogen potential V (r) = ~c
r) = Cer .
r is in the state (~
Find the value of C that properly normalizes the state. What is the probability
that the electron be found in the ground state of Hydrogen?
important fundamental medium
2
r dr d = 1
0

2
4|C|2
r dr e2r = 4|C|2
q0
3
C =

2
83

Pg.s. = |h100 |i|

q
r

3
3 r 2

h100 |i = 2( a10 ) 2 e a0 Y00

r dr d
e

q q
1

3
3
1

2 a0 + r
h100 |i = 2( a10 ) 2 4
4
r
e
dr

h100 |i = 4( a0 ) 2

1
a0

2!
3
+

1
a0 + a0

314

15. Hydrogen

This goes to 1 if =

TOC

1
a0 ,

making this state equal to the ground state.

5. An electron is in the 210 state of hydrogen. Find its wave function in momentum
space.

15.6

Sample Test Problems

1. A Hydrogen atom is in its 4D state (n = 4, ` = 2). The atom decays to a

lower state by emitting a photon. Find the possible photon energies that may be
observed. Give your answers in eV.
Answer
The n = 4 state can decay into states with n = 1, 2, 3. (Really the n = 1 state will
be suppressed due to selection rules but this is supposed to be a simple question.)
The energies of the states are
En =

13.6
eV.
n2

The photon energy is given by the energy difference between the states.

1
1
E = 13.6

n2
42
For the n = 1 final state, E =
For the n = 2 final state, E =
For the n = 3 final state, E =

15
16 13.6 = 12.8 eV.
3
16 13.6 = 2.6 eV.
7
144 13.6 = 0.7 eV.

2. Using the n`m notation, list all the n = 1, 2, 3 hydrogen states. (Neglect the
existence of spin.)
Answer
The states are, 100 , 200 , 211 , 210 , 211 , 300 , 311 , 310 , 311 , 322 , 321 ,
320 , 321 , 322 .
3. Find the difference in wavelength between light emitted from the 3P 2S transition in Hydrogen and light from the same transition in Deuterium. (Deuterium
is an isotope of Hydrogen with a proton and a neutron in the nucleus.) Get a
numerical answer.
4. An electron in the Coulomb field of a
proton is in the state described by
the wave function 16 (4100 + 3211 210 + 10211 ). Find the expected value
of the Energy, L2 and Lz . Now find the expected value of Ly .
5. * Write out the (normalized) hydrogen energy eigenstate 311 (r, , ).
6. Calculate the expected value of r in the Hydrogen state 200 .
315

15. Hydrogen

TOC

7. Write down the wave function of the hydrogen atom state 321 (r).
8. A Hydrogen atom is in its 4D state (n = 4, l = 2). The atom decays to a
lower state by emitting a photon. Find the possible photon energies that may be
observed. Give your answers in eV .
9. A Hydrogen atom is in the state:
1
(r) = (100 + 2211 322 2i310 + 2i300 4433 )
30
For the Hydrogen eigenstates, hnlm | 1r |nlm i = a0Zn2 . Find the expected value of
the potential energy for this state. Find the expected value of Lx .
10. A Hydrogen atom is in its 3D state (n = 3, l = 2). The atom decays to a
lower state by emitting a photon. Find the possible photon energies that may be
observed. Give your answers in eV .
11. The hydrogen atom is made up of a proton and an electron bound together by
the Coulomb force. The electron has a mass of 0.51 MeV/c2 . It is possible to
make a hydrogen-like atom from a proton and a muon. The force binding the
muon to the proton is identical to that for the electron but the muon has a mass
of 106 MeV/c2 .
a) What is the ground state energy of muonic hydrogen (in eV).
b) What is theBohr Radius of the ground state of muonic hydrogen.
12. A hydrogen atom is in the state: (r) = 110 (322 + 2221 + 2i220 + 111 ) Find
the possible measured energies and the probabilities of each. Find the expected
value of Lz .
13. Find the difference in frequency between light emitted from the 2P 1S transition in Hydrogen and light from the same transition in Deuterium. (Deuterium
is an isotope of Hydrogen with a proton and a neutron in the nucleus.)
14. Tritium is an isotope of hydrogen having 1 proton and 2 neutrons in the nucleus.
The nucleus is unstable and decays by changing one of the neutrons into a proton
with the emission of a positron and a neutrino. The atomic electron is undisturbed by this decay process and therefore finds itself in exactly the same state
immediately after the decay as before it. If the electron started off in the 200
(n = 2, l = 0) state of tritium, compute the probability to find the electron in
the ground state of the new atom with Z=2.
15. At t = 0 a hydrogen atom is in the state (t = 0) =
the expected value of r as a function of time.
Answer

1 (100
2

200 ). Calculate

1
1
(t) = (100 eiE1 t/~ 200 eiE2 t/~ ) = eiE1 t/~ (100 200 ei(E1 E2 )t/~ )
2
2
316

15. Hydrogen

TOC

1
h100 200 ei(E1 E2 )t/~ |r|100 200 ei(E1 E2 )t/~ i
2
The angular part of the integral can be done. All the terms of the wavefunction
contain a Y00 and r does not depend on angles, so the angular integral just gives
1.
h|r|i =

1
h|r|i =
2

(R10 R20 ei(E2 E1 )t/~ ) r(R10 R20 ei(E2 E1 )t/~ )r2 dr

The cross terms are not zero because of the r.

1
h|r|i =
2

2
2
R10
+ R20
R10 R20 ei(E2 E1 )t/~ + ei(E2 E1 )t/~ r3 dr

1
h|r|i =
2

2
2
R10
+ R20
2R10 R20 cos

E2 E1
t
~

r3 dr

317

15. Hydrogen

TOC

Now we will need to put in the actual radial wavefunctions.

R10
R20

=
=

h|r|i =

+
=
+

1
a0

er/a0

3

1 2
r
er/2a0
1
a0
2A0

1
r
1
r2
2r/a0
4e
+
1

er/a0
+
2a30
2
a0
4a20
0

r
E2 E1
3r/2a0
t
r3 dr
2 2 1
e
cos
2a0
~

r
2r
r
1
1
1 4 r
1
4r3 e a0 + r3 e a0
r e a0 + 2 r5 e a0
3
2a0
2
2a0
8a0
0
!

!
3 3r
3r
E

E
2
2
1
4
2 2r e 2a0 +
t
dr
r e 2a0 cos
a0
~

1
1
1
a0 4
+ 3a40
24a50 + 2 120a50
24
3
2a0
2
2a0
8a0

4
5 !

#

2a0
2
E2 E1
2a0
2 26
+
cos
t
24
3
a0
3
~
"
!

#
16
2 32
a0 3
E2 E1
+ 3 12 + 15 + 12 2 +
24
cos
t
2 2
81
a0 243
~
"
!

#
64 256 2
a0 3
E2 E1
+ 3 12 + 15 + 2 +
cos
t
2 2
27
81
~
"

#
15 32 2
E2 E1
a0
+
cos
t
4
81
~
1

318

16. 3D Symmetric HO in Spherical Coordinates *

TOC

3D Symmetric HO in Spherical Coordinates *

Since Hydrogen has this unusual accidental degeneracy, it is instructive to study

another problem with spherical symmetry, and the harmonic oscillator is an ideal candidate to illustrate the normal behaviour of a central potential problem. We have
already solved the problem of a 3D harmonic oscillator by separation of variables in
Cartesian coordinates. It is interesting to solve the same problem in spherical
coordinates and compare the results. The potential is
V (r) =

1 2 2
r .
2

Our radial equation is

2

d
2 d
2
`(` + 1)~2
+
R
(r)
+
E

V
(r)

RE` (r) = 0
E`
dr2
r dr
~2
2r2
d2 R 2 dR 2 2 2
`(` + 1)
2E
+
2 r R
R+ 2 R = 0
dr2
r dr
~
r2
~
Write the equation in terms of the dimensionless variable
y

r
d
dr
d2
dr2

r
.

s
~

= y
dy d
1 d
=
=
dr dy
dy
1 d
=
2 dy 2

Plugging these into the radial equation, we get

1
1 d2 R
1 2 dR
1 `(` + 1)
2E
4 2 y 2 R 2
+ 2
R+ 2 R
2
2
2
dy
y dy

y
~
d2 R 2 dR
`(`
+
1)
2E
+
y2 R
R+
R
dy 2
y dy
y2
~

Now find the behavior for large y.

d2 R
y2 R = 0
dy 2
R ey

319

16. 3D Symmetric HO in Spherical Coordinates *

TOC

Also, find the behavior for small y.

d2 R 2 dR `(` + 1)
+
R=0

dy 2
y dy
y2
R ys
s(s 1)y s2 + 2sy s2 = `(` + 1)y s2
s(s + 1) = `(` + 1)
R y`

Explicitly put in this behavior and use a polynomial to solve the full equation.
R = y`

ak y k ey

k=0

ak y `+k ey

k=0

Well need to compute the derivatives.

2
dR X
=
ak [(` + k)y `+k1 y `+k+1 ]ey /2
dy

k=0

d2 R X
ak [(` + k)(` + k 1)y `+k2 (` + k)y `+k
=
dy 2
k=0

(` + k + 1)y `+k + y `+k+2 ]ey /2

d2 R X
=
ak [(` + k)(` + k 1)y `+k2
dy 2
k=0

(2` + 2k + 1)y `+k + y `+k+2 ]ey

We can now plug these into the radial equation.

d2 R 2 dR
`(` + 1)
2E
+
y2 R
R+
R=0
dy 2
y dy
y2
~
Each term will contain the exponential ey
run a single sum over all the terms.

, so we can factor that out. We can also

ak (` + k)(` + k 1)y `+k2 (2` + 2k + 1)y `+k + y `+k+2

k=0

+2(` + k)y

`+k2

`+k

`+k+2

`(` + 1)y

`+k2

2E `+k
+
y
=0
~
320

16. 3D Symmetric HO in Spherical Coordinates *

TOC

The terms for large y which go like y `+k+2 and some of the terms for small y which go
like y `+k2 should cancel if we did our job right.

ak [(` + k)(` + k 1) `(` + 1) + 2(` + k)]y `+k2

k=0

2E
2 (2` + 2k + 1) y `+k = 0
~

ak [`(` 1) + k(2` + k 1) `(` + 1) + 2` + 2k]y `+k2

k=0

2E
2 (2` + 2k + 1) y `+k = 0
~

X
2E
`+k
`+k2
ak [k(2` + k + 1)]y
(2` + 2k + 3) y
=0
+
~
+

k=0

Now as usual, the coefficient for each power of y must be zero for this sum to be zero
for all y. Before shifting terms, we must examine the first few terms of this sum to
learn about conditions on a0 and a1 . The first term in the sum runs the risk of giving
us a power of y which cannot be canceled by the second term if k < 2. For k = 0, there
is no problem because the term is zero. For k = 1 the term is (2` + 2)y `1 which
cannot be made zero unless
a1 = 0.
This indicates that all the odd terms in the sum will be zero, as we will see from the
recursion relation.
Now we will do the usual shift of the first term of the sum so that everything has
a y `+k in it.
k k+2

X
2E
`+k
`+k
ak+2 (k + 2)(2` + k + 3)y
+ ak
(2` + 2k + 3) y
=0
~
k=0

2E
ak+2 (k + 2)(2` + k + 3) + ak
(2` + 2k + 3) = 0
~

2E
ak+2 (k + 2)(2` + k + 3) = ak
(2` + 2k + 3)
~
2E
(2` + 2k + 3)
ak
ak+2 = ~
(k + 2)(2` + k + 3)

For large k,
ak+2

2
ak ,
k

321

16. 3D Symmetric HO in Spherical Coordinates *

TOC

Which will cause the wave function to diverge. We must terminate the series for
some k = nr = 0, 2, 4..., by requiring
2E
(2` + 2nr + 3) = 0
~

3
E = nr + ` +
~
2
These are the same energies as we found in Cartesian coordinates. Lets plug this back
into the recursion relation.
ak+2 =

(2` + 2nr + 3) (2` + 2k + 3)

ak
(k + 2)(2` + k + 3)
2(k nr )
ak+2 =
ak
(k + 2)(2` + k + 3)

To rewrite the series in terms of y 2 and let k take on every integer value, we make
the substitutions nr 2nr and k 2k in the recursion relation for ak+1 in terms of
ak .
Solution to Harmonic Oscillator in Spherical Coordinates
r
y=

Rnr ` =

r
~

ak y `+2k ey

k=0

(k nr )
ak
(k + 1)(` + k + 3/2)

3
E = 2nr + ` +
~
2

ak+1 =

The table shows the quantum numbers for the states of each energy for our separation
in spherical coordinates, and for separation in Cartesian coordinates. Remember that
there are 2` + 1 states with different z components of angular momentum for the
spherical coordinate states.

322

16. 3D Symmetric HO in Spherical Coordinates *

E
3
2 ~
5
2 ~
7
2 ~
9
2 ~
11
2 ~

nr `
00
01
10, 02
11, 03
20, 12, 04

nx ny nz
000
001(3 perm)
002(3 perm), 011(3 perm)
003(3 perm), 210(6 perm), 111
004(3), 310(6), 220(3), 211(3)

TOC

NSpherical
1
3
6
10
15

NCartesian
1
3
6
10
15

The number of states at each energy matches exactly. The parities of the states
also match. Remember that the parity is (1)` for the angular momentum states and
that it is (1)nx +ny +nz for the Cartesian states. If we were more industrious, we could
verify that the wavefunctions in spherical coordinates are just linear combinations of
the solutions in Cartesian coordinates.

16.1

Homework
2

p
1. The differential equation for the 3D harmonic oscillator H = 2m
+ 12 m 2 r2 has
been solved in the notes, using the same techniques as we used for Hydrogen. Use
the recursion relations derived there to write out the wave functions n`m (r, , )
for the three lowest energies. You may write them in terms of the standard Y`m
but please write out the radial parts of the wavefunction completely. Note that
there is a good deal of degeneracy in this problem so the three lowest energies
actually means 4 radial wavefunctions and 10 total states. Try to write the
solutions 000 and 010 in terms of the solutions in cartesian coordinates with
the same energy nx,ny,nz .
extension medium
The energy eigenvalues for the 3D HO depend on both n and `. (Remember
Hydrogen is a special and unusual case.)

3
En` = 2nr + ` +
~
2

Clearly the lowest energy state is n`m = 000 with energy E000 = 32 ~. The
next possible energy is E01m = 25 ~. There are two ways to get the third
energy E100 = E02m = 72 ~. Three of these radial wave functions are relatively
simple, having just one term with the coefficient a0 , since nr = 0.
Rnr ` =

ak y `+2k ey

a0 y ` ey

k=0

The state with n = 1 will have two terms in the polynomial. So we have:
2
000 = a0 ey /2 Y00
2
010 = a0 yey /2 Y1m
2
020 = a0 y 2 ey /2 Y2m
2
100 = (a0 + a1 y 2 )ey /2 Y00
323

16. 3D Symmetric HO in Spherical Coordinates *

TOC

We can use the recursion relation for n = 1, ` = 0 and k = 0 to compute a1 in

the last case.
ak+1 =

(k nr )
a0
2
ak a1 =
= a0
(k + 1)(` + k + 3/2)
3/2
3

So we have the last wavefunction:

100 = a0

3 2 y2 /2
Y00
1 y e
2

For all of these, we need to set a0 to normalize the wave function. The normalization integral is:

2
r2 Rn`
(r)dr = 1
0

We can chage the variable of integration from r to y =

HO problem and save toner.

p
~

r as defined for this

2
y 2 Rn`
(y)dy = 1

0
2

Since the exponential in this integral becomes er there is substantial simplification in constants in all the terms. Lets do the hardest example for normalization,
R100 .
3
2

2
3 2
~ 2 2
2
a0 y 1 y
ey dy = 1

2
0

a20

2
9
y 2 3y 4 + y 6 ey dy = 1
4

a20

9
2 3(24) + (1320) = 1
4
3
2
a20 =

2900
3
4

a0 = ~
2900

324

17. Operators Matrices and Spin

TOC

Operators Matrices and Spin

We have already solved many problems in Quantum Mechanics using wavefunctions and
differential operators. Since the eigenfunctions of Hermitian operators are orthogonal
(and we normalize them) we can now use the standard linear algebra to solve quantum
problems with vectors and matrices. To include the spin of electrons and nuclei in our
discussion of atomic energy levels, we will need the matrix representation.
These topics are covered at very different levels in Gasiorowicz Chapter 14, Griffiths Chapters 3, 4 and, more rigorously, in Cohen-Tannoudji et al. Chapters
II, IV and IX.

17.1

The Matrix Representation of Operators and Wavefunctions

We will define our vectors and matrices using a complete set of, orthonormal basis
states ui , usually the set of eigenfunctions of a Hermitian operator. These basis states
are analogous to the orthonormal unit vectors in Euclidean space x
i .
hui |uj i = ij
Define the components of a state vector (analogous to xi ).
The State Vector
i hui |i
|i =

i |ui i

The wavefunctions are therefore represented as vectors. Define the matrix element.
The Operator Matrix
Oij hui |O|uj i

We know that an operator acting on a wavefunction gives a wavefunction.

X
X
|Oi = O|i = O
j |uj i =
j O|uj i
j

325

17. Operators Matrices and Spin

TOC

If we dot hui | into this equation from the left, we get

X
X
(O)i = hui |Oi =
j hui |O|uj i =
Oij j
j

This is exactly the formula for a state vector equals a matrix operator times a state
vector.

(O)1
O11 O12 ... O1j ...
1
(O)2 O21 O22 ... O2j ... 2

... = ...

... ... ... ...

...
(O)i Oi1 Oi2 ... Oij ... j
...
...
... ... ... ...
...
Similarly, we can look at the product of two operators (using the identity

|uk ihuk | =

1).
(OP )ij = hui |OP |uj i =

X
X
hui |O|uk ihuk |P |uj i =
Oik Pkj
k

This is exactly the formula for the product of two matrices.

(OP )11
(OP )21

...

(OP )i1
...

O11
O21

...

Oi1
...

O12
O22
...
Oi2
...

(OP )12
(OP )22
...
(OP )i2
...

... O1j
... O2j
... ...
... Oij
... ...

... (OP )1j

... (OP )2j
...
...
... (OP )ij
...
...

...
P11 P12

...
P21 P22

...
...
...

... Pi1 Pi2

...
...
...

...
...

...
=
...
...
... P1j
... P2j
... ...
... Pij
... ...

...
...

...

...
...

So, wave functions are represented by vectors and operators by matrices,

all in the space of orthonormal functions.
Example: The Harmonic Oscillator Hamiltonian Matrix.
Example: The harmonic oscillator raising operator.
Example: The harmonic oscillator lowering operator.

Now compute the matrix for the Hermitian Conjugate of an operator.

(O )ij = hui |O |uj i = hOui |uj i = huj |Oui i = Oji

The Hermitian Conjugate matrix is the (complex) conjugate transpose.

326

17. Operators Matrices and Spin

TOC

Check that this is true for A and A .

We know that there is a difference between a bra vector

and a ket P
vector. This
P
becomes explicit in the matrix representation. If =
j uj and =
k uk then,
j

the dot product is

h|i =

X
j,k

j k huj |uk i =

j k jk =

j,k

k k .

We can write this in dot product in matrix notation as

1
2

h|i = 1 2 3 ...
3
...
The bra vector is the conjugate transpose of the ket vector. They both represent the
same state but are different mathematical objects.

17.2

The Angular Momentum Matrices*

An important case of the use of the matrix form of operators is that of Angular Momentum. Assume we have an atomic state with ` = 1 (fixed) but m free. This is very
reasonable physically since the Energy only depends on n and ` leaving the different
m states degenerate. So any linear combination of them has the same energy. We can
apply a small external magnetic field and study these states. We may use the eigenstates of Lz as a basis for our states and operators. Ignoring the (fixed) radial part of
the wavefunction, our state vectors for ` = 1 must be a linear combination of the Y1m

X+
= Rn1 (r) (X+ Y11 + X0 Y10 + X Y11 ) = Rn1 (r)X = Rn1 (r) X0
X
where X+ , for example, is just the numerical coefficient of the eigenstate.
We will write our 3 component vectors like

X+
X = X0 .
X
The angular momentum operators are therefore 3X3 matrices. We can easily derive
the matrices representing the angular momentum operators for ` = 1.
327

17. Operators Matrices and Spin

TOC

The ` = 1 Matrix Operators

0
~
1
Lx =
2 0

0
~
1
Ly =
2i
0

1
Lz = ~ 0
0

0
1
0

1
0
1

0
0
0

1
0
1

0
1
0

0
0
1

The matrices must satisfy the same commutation relations as the differential operators.
[Lx , Ly ] = i~Lz
We verify this with an explicit computation of the commutator.
Since these matrices represent physical variables, we expect them to be Hermitian.
That is, they are equal to their conjugate transpose. Note that they are also traceless.
As an example of the use of these matrices, lets compute an expectation value of
Lx in the matrix representation for the general state X .

0 1 0
X1

~
1 0 1 X2
hX |Lx |X i = X1 X2 X3
2 0 1 0
X3

X2

~
= X1 X2 X3 X1 + X3
2
X
2

17.3

~
(X1 X2 + X2 (X1 + X3 ) + X3 X2 )
2

Eigenvalue Problems with Matrices

It is often convenient to solve eigenvalue problems like A = a using matrices.

Many problems in Quantum Mechanics are solved by limiting the calculation to a
finite, manageable, number of states, then finding the linear combinations which are
the energy eigenstates. The calculation is simple in principle but large dimension
328

17. Operators Matrices and Spin

TOC

matrices are difficult to work with by hand. Standard computer utilities are readily
available to help solve this problem.

A11 A12 A13 ...
1
1
A21 A22 A23 ... 2
2

A31 A32 A33 ... 3 = a 3
...
...
... ...
...
...
Subtracting the right hand side of the equation, we have

1
A11 a
A12
A13
...
2
A21
A

a
A
...
22
23
= 0.

A31
A32
A33 a ... 3
...
...
...
...
...
For the product to be zero, the determinant of the
this equation to get the eigenvalues.

A11 a
A12
A13

A21
A

a
A
22
23

A31
A
A
a
32
33

...
...
...

matrix must be zero. We solve

...
...
=0
...

...

Example: Eigenvectors of Lx .

The eigenvectors computed in the above example show that the x axis is not really any
different than the z axis. The eigenvalues are +~, 0, and ~, the same as for z. The
normalized eigenvectors of Lx are
The Eigenstates of Lx
1
2
1
2
1
2

(x)

X+~ =

(x)

X0~ =
0
1
2

(x)
X~ = 12

1
2

2
1
2

These vectors, and any ` = 1 vectors, can be written in terms of the eigenvectors of
Lz .
We can check whether the eigenvectors are orthogonal, as they must be.
1

2

0 12 12 = 0
hX0~ |X+~ i = 12
1
2

329

17. Operators Matrices and Spin

TOC

The others will also prove orthogonal.

(x)

(z)

Should X+~ and X~ be orthogonal?

NO. They are eigenvectors of different hermitian operators.
The eigenvectors may be used to compute the probability or amplitude of a particular
measurement. For example, if a particle is in a angular momentum state and the
angular momentum in the x direction is measured, the probability to measure +~ is

(x) 2
P+~ = hX+~ |i

17.4

An ` = 1 System in a Magnetic Field*

We will derive the Hamiltonian terms added when an atom is put in a magnetic field
in section 19. For now, we can be satisfied with the classical explanation that the circulating current associated with nonzero angular momentum generates a magnetic
moment, as does a classical current loop. This magnetic moment has the same interaction as in classical EM.
A Magnetic Moment in a B Field
~
H = ~
B

For the orbital angular momentum in a normal atom, the magnetic moment is

e ~
L.
2mc

For the electron mass, in normal atoms, the magnitude of

~ is one Bohr magneton,
B =

e~
.
2me c

If we choose the direction of B to be the z direction, then the magnetic moment

term in the Hamiltonian becomes:
Hamiltonian for Orbital Angular Momentum in B Field

1
B B
H=
Lz = B B 0
~
0

0
0
0

0
0
1
330

17. Operators Matrices and Spin

TOC

So the eigenstates of this magnetic interaction are the eigenstates of Lz and the energy
eigenvalues are +B B, 0, and B B.
Example: The energy eigenstates of an ` = 1 system in a B-field.
Example: Time development of a state in a B field.

17.5

Splitting the Eigenstates with Stern-Gerlach

A beam of atoms can be split into the eigenstates of Lz with a Stern-Gerlach

apparatus. A magnetic moment is associated with angular momentum.

~
e ~
L
L = B
2mc
~

This magnetic moment interacts with an external field, adding a term to the Hamiltonian.
~
H = ~
B
If the magnetic field has a gradient in the z direction, there is a force exerted (classically).
U
B
F =
= z
z
z
A magnet with a strong gradient to the field is shown below.

Lets assume the field gradient is in the z direction.

In the Stern-Gerlach experiment, a beam of atoms (assume ` = 1) is sent into a
magnet with a strong field gradient. The atoms come from an oven through some
collimator to form a beam. The beam is said to be unpolarized since the three m
states are equally likely no particular state has been prepared. An unpolarized, ` = 1
beam of atoms will be split into the three beams (of equal intensity) corresponding to
the different eigenvalues of Lz .
331

17. Operators Matrices and Spin

TOC

The atoms in the top beam are in the m = 1 state. If we put them through another
Stern-Gerlach apparatus, they will all go into the top beam again. Similarly for the
middle beam in the m = 0 state and the lower beam in the m = 1 state.
We can make a fancy Stern-Gerlach apparatus which puts the beam back together
as shown below.

+
0

z
We can represent the apparatus by the symbol to the right.
We can use this apparatus to prepare an eigenstate. The apparatus below picks
out the m = 1 state

+
0|

| z
again representing the apparatus by the symbol at the right. We could also represent
our apparatus plus blocking by an operator
O = |+i h+|
where we are writing the states according to the m value, either +, -, or 0. This is a
projection operator onto the + state.
An apparatus which blocks both the + and - beams
332

17. Operators Matrices and Spin

TOC

+|
0

| z
would be represented by the projection operator
O = |0i h0|

Similarly, an apparatus with no blocking could be written as the sum of the three
projection operators.

+1
+
X
0
|zm i hzm | = 1
= |+i h+| + |0i h0| + |i h| =

m=1
z
If we block only the m = 1 beam, the apparatus would be represented by

+|
0
= |0i h0| + |i h|.

z
Example: A series of Stern-Gerlachs.

17.6

Rotation operators for ` = 1 *

We have chosen the z axis arbitrarily. We could choose any other direction to define
our basis states. We wish to know how to transform from one coordinate system
to another. Experience has shown that knowing how an object transforms under
rotations is important in classifying the object: scalars, vectors, tensors...
We can derive the operator for rotations about the z-axis. This operator
transforms an angular momentum state vector into an angular momentum state vector
in the rotated system.
333

17. Operators Matrices and Spin

TOC

z-Axis Rotation Operator

Rz (z ) = eiz Lz /~
X 0 = Rz (z )X

Since there is nothing special about the z-axis, rotations about the other axes follow
the same form.
x and y Axis Rotation Operators
Rx (x ) = eix Lx /~
Ry (y ) = eiy Ly /~

The above formulas for the rotation operators must apply in both the matrix
representation and in the differential operator representation.
Redefining the coordinate axes cannot change any scalars, like dot products of state
vectors. Operators which preserve dot products are called unitary. We proved that
operators of the above form, (with hermitian matrices in the exponent) are unitary.
A computation of the operator for rotations about the z-axis gives :
z-Axis Rotation Matrix
i
e z
Rz (z ) = 0
0

0
1
0

0
0

eiz

A computation of the operator for rotations about the y-axis yields :

y-Axis Rotation Matrix
1
Ry (y ) =

2 (1 +
1

2
1
(1

cos(y ))
sin(y )
cos(y ))

1
2

sin(y )
cos(y )
12 sin(y )

1
2 (1 cos(y ))
1 sin(y )

2
1
(1
+
cos(
))
y
2

334

17. Operators Matrices and Spin

TOC

Try calculating the rotation operator for the x-axis yourself.

Note that both of the above rotation matrices reduce to the identity matrix for
rotations of 2 radians. For a rotation of radians, Ry interchanges the plus and
minus components (and changes the sign of the zero component), which is consistent
with what we expect. Note also that the above rotation matrices are quite different
than the ones used to transform vectors and tensors in normal Euclidean space. Hence,
the states here are of a new type and are referred to as spinors.
Example: A 90 degree rotation about the z axis.

17.7

A Rotated Stern-Gerlach Apparatus*

Imagine a Stern-Gerlach apparatus that first separates an ` = 1 atomic beam with a

strong B-field gradient in the z-direction. Lets assume the beam has atoms moving in
the y-direction. The apparatus blocks two separated beams, leaving only the eigenstate
of Lz with eigenvalue +~. We follow this with an apparatus which separates in the
u-direction, which is at an angle from the z-direction, but still perpendicular to the
direction of travel of the beam, y. What fraction of the (remaining) beam will
go into each of the three beams which are split in the u-direction?
We could represent this problem with the following diagram.

+
+ D+
Oven 0|
0 D0

| z
D u
We put a detector in each of the beams split in u to determine the intensity.
To solve this with the rotation
matrices,
we first determine the state after the first

1
(z)
apparatus. It is just X+ = 0 with the usual basis. Now we rotate to a new
0
(primed) set of basis states with the z 0 along the u direction. This means a rotation
through an angle about the y direction. The problem didnt clearly define whether
it is + or , but, if we only need to know the intensities, it doesnt matter. So the

335

17. Operators Matrices and Spin

TOC

state coming out of the second apparatus is

1
1 sin(y )
2 (1 + cos(y ))
2

(z)
cos(y )
Ry ()X+
= 12 sin(y )
1
1
2 (1 cos(y )) 2 sin(y )
1

2 (1 + cos(y ))
= 12 sin(y )
1
2 (1 cos(y ))

1
1
2 (1 cos(y ))

1 sin(y ) 0
2
1
0
2 (1 + cos(y ))

The 3 amplitudes in this vector just need to be (absolute) squared to get the 3
intensities.
1
1
1
I+ = (1 + cos(y ))2
I0 = sin2 (y )
I = (1 cos(y ))2
4
2
4
These add up to 1.
~ = cos Lz + sin Lx operator.
An alternate solution would be to use the Lu = u
L
(u)
Find the eigenvectors of this operator, like X+ . The intensity in the + beam is then
(z)
(u)
I+ = |hX+ |X+ i|2 .

17.8

Spin

1
2

Earlier, we showed that both integer and half integer angular momentum could satisfy
the commutation relations for angular momentum operators but that there is no single
valued functional representation for the half integer type.
Some particles, like electrons, neutrinos, and quarks have half integer internal angular momentum, also called spin. We will now develop a spinor representation for
spin 21 . There are no coordinates and associated with internal angular momentum
so the only thing we have is our spinor representation.
Electrons, for example, have total spin one half. There are no spin 3/2 electrons so
there are only two possible spin states for an electron. The usual basis states are the
eigenstates of Sz . We know from our study of angular momentum, that the eigenvalues
of Sz are + 12 ~ and 21 ~. We will simply represent the + 12 ~ eigenstate as the upper
component of a 2-component vector. The 12 ~ eigenstate amplitude is in the lower
component. So the pure eigenstates are.
Sz Eigenvectors
(z)
+

1
=
0

(z)

0
=
1
336

17. Operators Matrices and Spin

TOC

An arbitrary spin one half state can be represented by a spinor.

a
=
b
with the normalization condition that |a|2 + |b|2 = 1.
It is easy to derive the matrix operators for spin.

~
Sx =
2

0
1

1
0

The Spin

1
2

~
Sy =
2

Operators
0
i

i
0

~
Sz =
2

1
0

0
1

These satisfy the usual commutation relations from which we derived the properties of
angular momentum operators. For example lets calculate the basic commutator.

~2
0 1
0 i
0 i
0 1
[Sx , Sy ] =

i 0
1 0
4 1 0 i 0

2
2
~
~
~ 1 0
i 0
i 0
i 0
=

=
= i~
= i~Sz
0 i
0 i
4
2 0 i
2 0 1
The spin operators are an (axial) vector of matrices. To form the spin operator for
an arbitrary direction u
, we simply dot the unit vector into the vector of matrices.

Spin Operator for Any Axis

~
Su = u
S

The Pauli Spin Matrices, i , are simply defined and have the following properties.

337

17. Operators Matrices and Spin

TOC

Yes, The Pauli Spin Matrices Form a Vector

~ =

0
1

1
0
,
0
i
Si

i
1
,
0
0

0
1

~
i
2

~ = ~ ~
S
2
[i , j ] = 2iijk k
i2 = 1

They also anti-commute.

The Pauli Matrices AntiCommute

x y = y x

x z = z x

z y = y z

{i , j } i j + j i = 2ij

The matrices are the Hermitian, Traceless matrices of dimension 2. Any 2 by

2 Hermitian matrix can be written as a linear combination of the matrices and the
identity.
Example: The expectation value of Sx .
Example: The eigenvectors of Sx .
Example: The eigenvectors of Sy .

The (passive) rotation operators, for rotations of the coordinate axes can be computed from the formula Ri (i ) = eiSi i /~ .

338

17. Operators Matrices and Spin

TOC

Rotation Operators for Spin

Rz () =

Rx () =

Ry () =

1
2

ei/2

cos 2

i sin 2

cos 2

sin 2

cos 2

Note that the operator for a rotation through 2 radians is minus the identity matrix
for any of the axes (because 2 appears everywhere). The surprising result is that
the sign of the wave function of all fermions is changed if we rotate through 360
degrees.
Example: The eigenvectors of Su .
~ there is also a magnetic moment associated
As for orbital angular momentum (L),
~
with internal angular momentum (S).
The Magnetic Moment Due to Spin

~ spin =

eg ~
S
2mc

This formula has an additional factor of g, the gyromagnetic ratio, compared to

the formula for orbital angular momenta. For point-like particles, like the electron, g
has been computed in Quantum ElectroDynamics to be a bit over 2, g = 2 +
+ ....
For particles with structure, like the proton or neutron, g is hard to compute, but has
been measured. Because the factor of 2 from g cancels the factor of 2 from s = 21 , the
magnetic moment due to the spin of an electron is almost exactly equal to the magnetic
moment due to the orbital angular momentum in an ` = 1 state. Both are 1 Bohr
e~
Magneton, B = 2mc
.
eg~
~ = B ~ B
~
~ B
4mc
If we choose the z axis to be in the direction of B, then this reduces to
~ =
H = ~
B

339

17. Operators Matrices and Spin

The Hamiltonian for Spin

TOC

1
2

in a B Field

H = B Bz

Example: The time development of an arbitrary electron state in a magnetic field.

Example: Nuclear Magnetic Resonance (NMR and MRI).

A beam of spin one-half particles can also be separated by a Stern-Gerlach apparatus

which uses a large gradient in the magnetic field to exert a force on particles proprtional
to the component of spin along the field gradient. Thus, we can measure the component
of spin along a direction we choose. A field gradient will separate a beam of spin onehalf particles into two beams. The particles in each of those beams will be in a definite
spin state, the eigenstate with the component of spin along the field gradient direction
either up or down, depending on which beam the particle is in.
We may represent a Stern-Gerlach appartatus which blocks the lower beam by the
symbol below.

+

| z
This apparatus is equivalent to the operator that projects out the + ~2 eigenstate.

1
1 0
1 0 =
|+i h+| =
0
0 0
We can perform several thought experiments. The appartus below starts with an
unpolarized beam. In such a beam we dont know the state of any of the particles. For
a really unpolarized beam, half of the particles will go into each of the separated beams.
(Note that an unpolarized beam cannot be simply represented by a state vector.) In
the apparatus below, we block the upper beam so that only half of the particles come
out of the first part of the apparatus and all of those particles are in the definite state
having spin down along the z axis. The second part of the apparatus blocks the lower
separated beam. All of the particles are in the lower beam so nothing is left coming
out of the apparatus.

N 0
+
+|

0
Unpolarized Beam (N particles)

1
|
z
2
z
The result is unaffected if we insert an additional apparatus that separates in the x
direction in the middle of the apparatus above. While the apparatus separates, neither
340

17. Operators Matrices and Spin

TOC

beam is blocked (and we assume we cannot observe which particles go into which
beam). This apparatus does not change the state of the beam!

N 0
N 0
+
+
+|

0
(N particles)

x
| z
z
2 1
2 1
Now if we block one of the beams separated according to the x direction, particles can
get through the whole apparatus. The middle part of the apparatus projects the state
onto the positive eigenstate of Sx . This state has equal amplitudes to have spin up
and spin down along the z direction. So now, 1/8 of the particles come out of the
apparatus. By blocking one beam, the number of particles coming out increased from
0 to N/8. This seems a bit strange but the simple explanation is that the upper and
lower beams of the middle part of the apparatus were interfering to give zero particles.
With one beam blocked, the inteference is gone.
!

1
N 0
N 2
N 1
+|
+
+

(N )

z
| x
| z
2 1
4 12
8 0
Note that we can compute the number of particles coming out of the second (and third)
part by squaring the amplitude to go from the input state to the output state

2
N 1
N
0
1
=

2
2
1
2
4
!

1 1
1
1
1
2

or we can just use the projection operator 1

= 21 21 .
2
2
2

1
2
1
2

17.9
17.9.1

1
2
1
2

N
2

N
0
=
1
4

1
2
1
2

Other Two State Systems*

The Ammonia Molecule (Maser)

The Feynman Lectures (Volume III, chapters 8 and 9) makes a complete study of the
two ground states of the Ammonia Molecule. Feynman starts with two states, one
with the Nitrogen atom above the plane defined by the three Hydrogen atoms, and
the other with the Nitrogen below the plane. We can ignore high energy excitations
and factor out vibrations and rotations from our analysis, except for a rotation about
the symmetry axis that helps define above. There is clearly symmetry between the
two states which we can call up and down. They have identical properties. This is
341

17. Operators Matrices and Spin

TOC

an example of an SU(2) symmetry, like that in angular momentum (and the weak
interactions). We just have two states which are different but completely symmetric.

Figure 47: The Ammonia Molecule has symmetric ground states with the Nitrogen
atom above and below the plane of Hydrogens. It is possible for the Nitrogen to tunnel
through the plane and transition to the other state.

Since the Nitrogen atom can tunnel from one side of the molecule to the other, there
are cross terms in the Hamiltonian (limiting ourselves to the two symmetric ground
states).
habove |H|above i = hbelow |H|below i = E0
habove |H|below i = A

E0 A
H=
= E0 1 Ax
A E0
We can adjust the phases of the above and below states to make A real.
The energy eigenvalues can be found from the usual equation.

E0 E
A

= 0
A
E0 E
(E0 E)2

E E0

E0 A

Now find the eigenvectors.

H = E

E0 A
a
a
= (E0 A)
A E0
b
b

E0 a Ab
(E0 A)a
=
E0 b Aa
(E0 A)b
342

17. Operators Matrices and Spin

TOC

These are solved if b = a. Substituting auspiciously, we get.

E0 a Aa
(E0 A)a
=
E0 b Ab
(E0 A)b
So the eigenstates are
!

E = E0 A

1
2
1
2

E = E0 + A

1
2
12

The states are split by the interaction term. If the molecule starts out in the above
state, we can write that in terms of the energy eigenstates and find the time dependence.
This is essentially the same as starting with a spin up electron in a B field in the x
direction. The spin will precess around the field and the thus oscillate between up and
down along the z axis. Try this time dependence of Ammonia yourself. We could also
solve the two coupled time dependent Schr
odinger equations, but that is the hard way
to find the time dependence.
The molecule has an electric dipole moment d~ with the Hydrogens more positive than
the Nitrogen. If we put the molecule in an electric field, there is an additional contribution to the Hamiltonian. Feynman goes on to further split the states by putting
the molecules in an electric field E = d. This makes the diagonal terms of the
Hamiltonian slightly different, like a magnetic field does in the case of spin.

E0 + d
A
H=
= E0 1 Ax + dz
A
E0 d
We can find the energy eigenvalues fairly simply.

(E0 E) + d

A

=

A
(E0 E) d

(E0 E)2 d2 2

(E0 E)2

E E0

A2 + d 2 2
p
A2 + d2 2
p
E0 A2 + d2 2

In principle either term could dominate depending on the strength of the -field, but
usually the field is too weak to domiate A.
Finally, Feynman studies the effect of Ammonia in an oscillating Electric field, and
looks at the Ammonia Maser. The Hamiltonian itself becomes time dependent so
we now must solve the coupled differential equations.

E0 + deit
A
H=
A
E0 deit
343

17. Operators Matrices and Spin

17.9.2

TOC

The Neutral Kaon System*

The neutral Kaons, K 0 and K0 form a very interesting two state system. As in
the Ammonia molecule, there is a small amplitude to make a transition form one
to the other. The Energy (mass) eigenstates are similar to those in the example above,
but the CP (Charge conjugation times Parity) eigenstates are important because they
determine how the particles can decay. A violation of CP symmetry is seen in the
decays of these particles.

17.10

Examples

17.10.1

Harmonic Oscillator Hamiltonian Matrix

We wish to find the matrix form of the Hamiltonian for a 1D harmonic oscillator.
The basis states are the harmonic oscillator energy eigenstates. We know the eigenvalues of H.
Huj = Ej uj

1
~ij
hi|H|ji = Ej ij = j +
2
The Kronecker delta gives us a diagonal matrix.
1
0 0 0
2
0 3 0 0
2

5
H = ~
0 0 2 07
0 0 0
2
... ... ... ...

17.10.2

...
...

...

...
...

Harmonic Oscillator Raising Operator

We wish to find the matrix representing the 1D harmonic oscillator raising operator.
We use the raising operator equation for an energy eigenstate.

A un = n + 1un+1
Now simply compute the matrix element.
Aij = hi|A |ji =

p
j + 1i(j+1)
344

17. Operators Matrices and Spin

TOC

Now this Kronecker delta puts us one off the diagonal. As we have it set up, i gives
the row and j gives the column. Remember that in the Harmonic Oscillator we start
counting at 0. For i=0, there is no allowed value of j so the first row is all 0. For i=1,
j=0, so we have an entry for A10 in the second row and first column. All he entries
will be on a diagonal from that one.

0
1

A =
0

0
...

17.10.3

0
0
2
0
0
...

0
0
0
3
0
...

0
0
0
0
4
...

...
...

...

...
...

Harmonic Oscillator Lowering Operator

We wish to find the matrix representing the 1D harmonic oscillator lowering operator.
This is similar to the last section.
The lowering operator equation is.
Aun =

nun1

Now we compute the matrix element from the definition.

p
Aij = hi|A|ji = ji(j1)

0
0

A=
0
0
...

1
0
0
0
...

0
2
0
0
...

0
0
3
0
...

0
0
0
4
...

...
...

...

...
...

This should be the Hermitian conjugate of A .

345

17. Operators Matrices and Spin

17.10.4

TOC

Eigenvectors of Lx

We will do it as if we dont already know that the eigenvalues are m~.

0
1
0

where a =

1
0
1

L X = aX
x

0
X1
X1
X1

1 X2 = ~2a X2 b X2
0
X3
X3
X3

b 1
0

1 b 1 = 0

0
1 b

b.
2

b(b2 1) 1(b 0) = 0
b(b2 2) = 0

There are three solutions to this equation: b = 0, b = + 2, and b = 2 or a = 0,

a = +~, and a = ~. These are the eigenvalues we expected for ` = 1. For each of
these three eigenvalues, we should go back and find the corresponding eigenvector by
using the matrix equation.

0 1 0
X1
X1
1 0 1 X2 = b X2
0 1 0
X3
X3

X2
X1
X1 + X3 = b X2
X2
X3
Up to a normalization constant, the solutions are:
1

1
2
X+~ = c 1
X0~ = c 0
1
1
2

1
2

X~ = c 1 .
1

We should normalize these eigenvectors to represent one particle. For example:

|c|2

1
2

hX+~ |X+~ i =
1
2

1 =
1
2

1
2|c|2 = 1

1
2

c =

1
.
2

Try calculating the eigenvectors of Ly .

You already know what the eigenvalues are.
346

17. Operators Matrices and Spin

17.10.5

TOC

A 90 degree rotation about the z axis.

If we rotate our coordinate system by 90 degrees about the z axis, the old x axis
becomes the new -y axis. So we would expect that the state with angular momentum
(y)
(x)
+~ in the x direction, X+ , will rotate into X within a phase factor. Lets do the
rotation.

i
e z 0
0
1
0 .
Rz (z ) = 0
iz
0
0 e

i 0 0
Rz (z = 90) = 0 1 0 .
0 0 i
Before rotation the state is

1
2
1
2
1
2

(x)

X+~ =

The rotated state is.

i
X 0 = 0
0

0
1
0

1 i
0
2
2
0 12 = 12
1
i
i
2
2
(y)

Now, what remains is to check whether this state is the one we expect. What is X ?
We find that state by solving the eigenvalue problem.
(y)

(y)

Ly X = ~X

0
1 0
a
a
~

1 0 1 b = ~ b
2i
0 1 0
c
c
ib

a
2
i(ca)

2 = b
ib
c

Setting b = 1, we get the unnormalized result.

i
2

(y)
X = C 1
i

347

17. Operators Matrices and Spin

TOC

Normalizing, we get.
i
2
1

2
i
2

(y)
X

This is exactly the same as the rotated state. A 90 degree rotation about the z axis
(y)
(x)
changes X+ into X .

17.10.6

Energy Eigenstates of an ` = 1 System in a B-field

Recall that the Hamiltonian for a magnetic moment in an external B-field is

B B
Lz .
~

As usual, we find the eigenstates (eigenvectors) and eigenvalues of a system by solving

the time-independent Schr
odinger equation HX = EX . We see that everything in the
Hamiltonian above is a (scalar) constant except the operator Lz , so that
HX =

B B
Lz X = constant (Lz X ).
~

Now if Xm is an eigenstate of Lz , then Lz Xm = m~Xm , thus

HXm =

B B
(m~Xm ) = (mB B)Xm
~

Hence the normalized eigenstates must be just those of the operator Lz itself, i.e., for
the three values of m (eigenvalues of Lz ), we have

Xm=+1

1
= 0
0

Xm=0

0
= 1
0

Xm=1

0
= 0 .
1

and the energy eigenvalues are just the values that E = mB B takes on for the three
values of m i.e.,

Em=+1 = +B B

Em=0 = 0

Em=1 = B B.

348

17. Operators Matrices and Spin

17.10.7

TOC

A series of Stern-Gerlachs

Now that we have the shorthand notation for a Stern-Gerlach apparatus, we can
put some together and think about what happens. The following is a simple example in
which three successive apparati separate the atomic beam using a field gradient along
the z direction.

+|
+
+|
(I3 )
(I2 ) 0
(I1 ) 0|
Oven(I0 ) 0

| z
z
z
If the intensity coming out of the oven is I0 , what are the intensities at
positions 1, 2, and 3? We assume an unpolarized beam coming out of the oven so
that 1/3 of the atoms will go into each initial beam in apparatus 1. This is essentially
a classical calculation since we dont know the exact state of the particles coming from
the oven. Now apparatus 1 removes the m = 1 component of the beam, leaving a state
with a mixture of m = 0 and m = 1.
I1 =

2
I0
3

We still dont know the relative phase of those two components and, in fact, different
atoms in the beam will have different phases.
The beam will split into only two parts in the second apparatus since there is no m = 1
component left. Apparatus 2 blocks the m = 0 part, now leaving us with a state that
we can write.
1
I2 = I0
3
All the particles in the beam are in the same state.
(z)

X = X

The beam in apparatus 3 all goes along the same path, the lower one. Apparatus 3
blocks that path.
I3 = 0
The following is a more complex example using a field gradients in the z and x directions
(assuming the beam is moving in y).

+
+|
+|
Oven(I0 ) 0|
(I1 ) 0|
(I2 ) 0
(I )

3
| z
x
z
If the intensity coming out of the oven is I0 , what are the intensities at
positions 1, 2, and 3?
349

17. Operators Matrices and Spin

TOC

Now we have a Quantum Mechanics problem. After the first apparatus, we have an
intensity as before
1
I1 = I0
3
and all the particles are in the state

1
(z)
X+ = 0 .
0
The second apparatus is oriented to separate the beam in the x direction. The beam
separates into 3 parts. We can compute the intensity of each but lets concentrate on
the bottom one because we block the other two.

(x) (z) 2
I2 = hX |X+ i I1
(z)

We have written the probability that one particle, initially in the the state X+ , goes
(x)
into the state X when measured in the x direction (times the intensity coming
into the apparatus). Lets compute that probability.

1
1
(x)
(z)
hX |X+ i = 12 12 12 0 =
2
0
So the probability is 14 .
I2 =

1
1
I1 =
I0
4
12

The third apparatus goes back to a separation in z and blocks the m = 1 component.
The incoming state is
1
2
(x)
X = 12
12
Remember that the components of this vector are just the amplitudes to be in the
different m states (using the z axis). The probability to get through this apparatus is
just the probability to be in the m = 0 beam plus the probability to be in the m = 1
beam.

1 2 1 2
3

P = + =
2
4
2
I3 =

3
1
I2 =
I0
4
16

350

17. Operators Matrices and Spin

TOC

Now lets see what happens if we remove the blocking in apparatus 2.

+
+
+|
Oven(I0 ) 0|
(I1 ) 0
(I2 ) 0
(I )

3
| z
x
z
Assuming there are no bright lights in apparatus 2, the beam splits into 3 parts then
(z)
recombines yielding the same state as was coming in, X+ . The intensity coming out
(z)
of apparatus 2 is I2 = I1 . Now with the pure state X+ going into apparatus 3 and
the top beam being blocked there, no particles come out of apparatus 3.
I3 = 0
By removing the blocking in apparatus 2, the intensity dropped from
to zero. How could this happen?

1
16 I0

What would happen if there were bright lights in apparatus 2?

17.10.8

Time Development of an ` = 1 System in a B-field: Version I

We wish to determine how an angular momentum 1 state develops with time, develops
with time, in an applied B field. In particular, if an atom is in the state with x
(x)
component of angular momentum equal to +~, X+ , what is the state at a later time
t? What is the expected value of Lx as a function of time?
We will choose the z axis so that the B field is in the z direction. Then we know the
energy eigenstates are the eigenstates of Lz and are the basis states for our vector
representation of the wave function. Assume that we start with a general state which
is known at t = 0.

X+
X (t = 0) = X0 .
X
But we know how each of the energy eigenfunctions develops with time so its easy to
write

X+ eiE+ t/~
X+ eiB Bt/~
.
X0
X (t) = X0 eiE0 t/~ =
iB Bt/~
iE t/~
X e
X e
As a concrete example, lets assume we start out in the eigenstate of Lx with eigenvalue

351

17. Operators Matrices and Spin

TOC

+~.
1
2
1
2
1
2

X (t = 0)

= Xx+ =

eiB Bt/~
2

2
eiB Bt/~
2

X (t)

hX (t)|Lx |X (t)i =

hX (t)|Lx |X (t)i =
=

Xx+ =

0

~
e+iB Bt/~
eiB Bt/~
1

1
2
2
2
2 0
+iB Bt/~
iB Bt/~
1
~
e
e
1

+
2
2
2
2
2
~
B Bt
B Bt
(4 cos(
)) = ~ cos(
)
4
~
~

eiB Bt/~
0
2

1
1

2
iB Bt/~
e
0
2

eiB Bt/~
eiB Bt/~ 1

+
+
2
2
2
1
0
1

Note that this agrees with what we expect at t = 0 and is consistent with the angular
momentum precessing about the z axis. If we checked hX |Ly |X i, we would see a sine
instead of a cosine, confirming the precession.

17.10.9

Expectation of Sx in General Spin

Let =

+
, be some arbitrary spin

1
2

State

state. Then the expectation value of the

operator
hSx i =
=
=

h|Sx |i

2 +

~ 0 1
+

2 1 0

~

+ +
+ .
=
+
2

352

17. Operators Matrices and Spin

17.10.10

TOC

Eigenvectors of Sx for Spin

1
2

First the quick solution. Since there is no difference between x and z, we know the
eigenvalues of Sx must be ~2 . So, factoring out the constant, we have

0 1
a
a
=
1 0
b
b

b
a
=
a
b
a = b

Sx Eigenvectors
(x)
+

1
2
1
2

!
(x)

1
2
1

These are the eigenvectors of Sx . We see that if we are in an eigenstate of Sx the

spin measured in the z direction is equally likely to be up and down since the absolute
square of either amplitude is 12 .
The remainder of this section goes into more detail on this calculation but is currently
notationally challenged.
Recall the standard method of finding eigenvectors and eigenvalues:
A =
(A ) = 0
For spin

1
2

system we have, in matrix notation,

a1 a2
1
1
=
a3 a4
2
0

a1
a3

a2
a4

0
1
1
2

1
=0
2

For a matrix times a nonzero vector to give zero, the determinant of the matrix must
be zero. This gives the characteristic equation which for spin 21 systems will be a
quadratic equation in the eigenvalue :

a1
a2

= (a1 )(a4 ) a2 a3 = 0
a3
a4
353

17. Operators Matrices and Spin

TOC

2 (a1 + a4 ) + (a1 a4 a2 a3 ) = 0
whose solution is
s
=

(a1 + a4 )

(a1 + a4 )
(a1 a4 a2 a3 )
2

To find the eigenvectors, we simply replace (one at a time) each of the eigenvalues
above into the equation

a1
a2
1
=0
a3
a4
2
and solve for 1 and 2 .

Now specifically, for the operator A = Sx =

~
2

) = 0 becomes, in matrix notation,

~ 0 1
1

2
0
2 1 0

~/2

0
1

1
, the eigenvalue equation (Sx
0

1
=0
2

~/2
1
=0

The characteristic equation is det|Sx | = 0, or

~2
=0
4

~
2

These are the two eigenvalues (we knew this, of course). Now, substituting + back
into the eigenvalue equation, we obtain

~ 1 1
1
+ ~/2
1
~/2 ~/2
1
=0
=
=
2
~/2 +
2
~/2 ~/2
2
2 1 1
The last equality is satisfied only if 1 = 2 (just write out the two component equations
to see this). Hence the normalized eigenvector corresponding to the eigenvalue =
+~/2 is

1 1
(x)

.
+ =
2 1
Similarly, we find for the eigenvalue = ~/2,

1
1
(x)
=
.
2 1
354

17. Operators Matrices and Spin

17.10.11

TOC

Eigenvectors of Sy for Spin

1
2

To find the eigenvectors of the operator Sy we follow precisely the same procedure as
we did for Sx (see previous example for details). The steps are:
1. Write the eigenvalue equation (Sy ) = 0
2. Solve the characteristic equation for the eigenvalues
3. Substitute the eigenvalues back into the original equation
4. Solve this equation for the eigenvectors

0 i
, so that, in matrix notation the eigenvalue
Here we go! The operator Sy = ~2
i 0
equation becomes

i~/2
1
=0
i~/2

2
The characteristic equation is det|Sy | = 0, or
2

~2
=0
4

~
2

These are the same eigenvalues we found for Sx (no surprise!) Plugging + back into
the equation, we obtain

~ 1 i
+ i~/2
1
1
=
=0
i~/2 +
2
i 1
2
2
Writing this out in components gives the pair of equations
1 i2 = 0

and

i1 2 = 0

which are both equivalent to 2 = i1 . Repeating the process for , we find that
2 = i1 . Hence the two eigenvalues and their corresponding normalized eigenvectors
are

1 1
(y)
+ = +~/2
+ =
2 i

1
1
(y)
= ~/2
=
i
2
Sy Eigenvectors
(y)
+

1
=
2

1
i

(y)

1
=
2

1
i
355

17. Operators Matrices and Spin

17.10.12

TOC

Eigenvectors of Su

As an example, lets take the u direction to be in the xz plane, between the positive x and
zq
axes, 30 degrees from the x axis. The unit vector is then u
= (cos(30), 0, sin(30)) =
3
1
~
( 4 , 0, 2 ). We may simply calculate the matrix Su = u
S.
r
Su =

1
~
3
Sx + Sz =
4
2
2

1
q2

3
4

We expect the eigenvalues to be ~2 as for all axes.

Factoring out the

~
2,

the equation for the eigenvectors is.

1
q2

3
4 a

a
b

b
q

1
a + 34 b
2
= a
q
b
1
3
4a 2b
3
4

For the positive eigenvalue, we have a =

For the negative eigenvalue, we have a =

q !
3b, giving the eigenvector

1
3 b,

(u)
+

(u)

giving the eigenvector =

3
4

1
2 !
21
q
3 .
4

Of course each of these could be multiplied by an arbitrary phase factor.

There is an alternate way to solve the problem using rotation matrices. We take the
(z)
states and rotate the axes so that the u axis is where the z axis was. We must
think carefully about exacty what rotation to do. Clearly we need a rotation about the
y axis. Thinking about the signs carefully, we see that a rotation of -60 degrees moves

356

17. Operators Matrices and Spin

TOC

the u axis to the old z axis.

Ry =

3
sin(30)
4

=
cos(30)
1
2

q

3
1
4
q2 1 =
=
0
1
3
2
4
q

3
1

4
2 0

q
=
=
1
1
3

cos(30)
Ry (60) =
sin(30)
(u)
+

(u)

cos 2
sin 2
q

sin 2
cos 2

12
q
3
4

q !
3
4

1
2

21
q

3
4

This gives the same answer. By using the rotation operator, the phase of the eigen(z)
vectors is consistent with the choice made for . For most problems, this is not
important but it is for some.

17.10.13

Time Development of a Spin

1
2

State in a B field

a
and we have chosen the z
b
axis to be in the field direction. The upper component of the vector (a) is the amplitude
to have spin up along the z direction, and the lower component (b) is the amplitude to
have spin down. Because of our choice of axes, the spin up and spin down states are
also the energy eigenstates with energy eigenvalues of B B and B B respectively. We
know that the energy eigenstates evolve with time quite simply (recall the separation
of the Schr
odinger equation where T (t) = eiEt/~ ). So its simple to write down the
time evolved state vector.
i Bt/~ it
ae B
ae
(t) =
=
beit
beiB Bt/~
Assume that we are in an arbitrary spin state (t = 0) =

where =

B B
~ .

So lets say we start out in the state with spin up along the x axis, (0) =

1
2
1
2

!
. We

357

17. Operators Matrices and Spin

TOC

then have
(t)

1 eit
2
1 eit
2

h(t)|Sx |(t)i =

1 e+it
2

.
1 eit
2

~ 0
2

1
0

1 e+it
2
1 eit
2

1 eit
2
1 eit
2

~ 1 +it
e
2
2

~
~ 1 +2it
e
+ e2it = cos(2B Bt/~)
22
2

1 eit
2

So again the spin precesses around the magnetic field. Because g = 2 the rate is twice
as high as for ` = 1.

17.10.14

Nuclear Magnetic Resonance (NMR and MRI)

Nuclear Magnetic Resonance is an important tool in chemical analysis. As the

name implies, it uses the spin magnetic moments of nuclei (particularly hydrogen) and
resonant excitation. Magnetic Resonance Imaging uses the same principle to get
an image (of the inside of the body for example).
In basic NMR, a strong static B field is applied. A spin 12 proton in a hydrogen nucleus
then has two energy eigenstates. After some time, most of the protons fall into the
lower of the two states. We now use an electromagnetic wave (RF pulse) to excite some
of the protons back into the higher energy state. Surprisingly, we can calculate this
process already. The protons magnetic moment interacts with the oscillating B field
of the EM wave.

358

17. Operators Matrices and Spin

TOC

As we derived, the Hamiltonian is

~ = gp e S
~ B
~ = gp e~ ~ B
~ = gp N ~ B
~
H = ~
B
2mp c
4mp c
2
Note that the gyromagnetic ratio of the proton is about +5.6. The magnetic moment
is 2.79 N ( Nuclear Magnetons). Different nuclei will have different gyromagnetic
ratios, giving us more tools to work with. Lets choose our strong static B field to be
in the z direction and the polarization on our oscillating EM wave so that the B field
points in the x direction. The EM wave has (angular) frequency .

gp
gp
Bz
Bx cos t
H = N (Bz z + Bx cos(t)x ) = N
Bx cos t
Bz
2
2
Now we apply the time dependent Schr
odinger equation.
d
i~
= H
dt

gp
a
Bz
Bx cos t
a
i~
= N
Bx cos t
Bz
b
b
2

gp N
a
Bz
Bx cos t
a
= i
Bx cos t
Bz
b
b
2~

0
1 cos t
a
= i
1 cos t
0
b
The solution of these equations represents and early example of time dependent
perturbation theory.
d
i1 i(+20 )t
(bei0 t ) =
(e
+ ei(20 t) )
dt
2

359

17. Operators Matrices and Spin

TOC

Terms that oscillate rapidly will average to zero. The first term oscillates very rapidly.
The second term will only cause significant transitions if 20 . Note that this is
exactly the condition that requires the energy of the photons in the EM field E = ~ to
be equal to the energy difference between the two spin states E = 2~0 . The conservation of energy condition must be satisfied well enough to get a significant transition
rate. Actually we will find later that for rapid transitions, energy conservation does
not have to be exact.
So we have proven that we should set the frequency of our EM wave according to
the energy difference between the two spin states. This allows us to cause transitions
to the higher energy state. In NMR, we observe the transitions back to the lower
energy state. These emit EM radiation at the same frequency and we can detect it
after the stronger input pulse ends (or by more complex methods). We dont yet know
why the higher energy state will spontaneously decay to the lower energy state. To
calculate this, we will have to quantize the field. But we already see that the energy
terms eiEt/~ of standard wave mechanics will require energy conservation with photon
energies of E = ~.
NMR is a powerful tool in chemical analysis because the molecular field adds to the
external B field so that the resonant frequency depends on the molecule as well as the
nucleus. We can learn about molecular fields or just use NMR to see what molecules
are present in a sample.
In MRI, we typically concentrate on one nucleus like hydrogen. We can put a gradient
in Bz so that only a thin slice of the material has tuned to the resonant frequency.
Therefore we can excite transitions to the higher energy state in only a slice of the
sample. If we vary (in the orthogonal direction!) the B field during the decay of the
excited state, we can get a two dimensional picture. If we vary B as a function of
time during the decay, we can get to 3D. While there are more complex methods used
in MRI, we now understand the basis of the technique. MRIs are a very safe way
to examine the inside of the body. All the field variation takes some time though.
Ultimately, a very powerful tool for scanning materials (a la Star Trek) is possible.

17.11

Derivations and Computations

17.11.1

The ` = 1 Angular Momentum Operators*

We will use states of definite Lz , the Y1m .

h`m0 |Lz |`mi = m~m0 m

1 0 0
Lz = ~ 0 0 0
0 0 1
360

17. Operators Matrices and Spin

h`m0 |L |`mi =
L+

TOC

`(` + 1) m(m 1)~m0 (m1)

0
2 0
= ~ 0 0
2
0 0
0

0 0
0
= ~ 2 0 0
2 0
0

0 1 0
1
~
1 0 1
(L+ + L ) =
2
2 0 1 0

0
1 0
1
~
1 0 1
(L+ L ) =
2i
2i
0 1 0

What is the dimension of the matrices for ` = 2?

Dimension 5. Derive the matrix operators for ` = 2.
Just do it.

17.11.2

Compute [Lx , Ly ] Using Matrices *

0 1 0
0
1 0
0
1 0
0 1 0
~
1 0 1 1 0 1 1 0 1 1 0 1
2i
0 1 0
0 1 0
0 1 0
0 1 0

1
1 0 1
2 0 0
1 0 0
2
~
0 0 0 = i~~ 0 0 0 = i~Lz
0 0 0 0 =
2i
1
1 0 1
0 0 2
0 0 1
2

[Lx , Ly ] =

1
~2
0
2i
1

0
0
0

The other relations will prove to be correct too, as they must. Its a reassuring check
and a calculational example.

17.11.3

Derive the Expression for Rotation Operator Rz *

The laws of physics do not depend on what axes we choose for our coordinate systemThere is rotational symmetry. If we make an infinitesimal rotation (through and angle
d) about the z-axis, we get the transformed coordinates
x0

x + dy

y dx.

361

17. Operators Matrices and Spin

TOC

We can Taylor expand any function f ,

f (x0 , y 0 ) = f (x, y) +

f
f
i
dy
dx = (1 dLz )f (x, y).
x
y
~

So, by changing the coordinates, the value of the function changes. We want to modify
the function so the rotation does not cause a change. What we want is the operator
Rz (d) such that we can get the value back to f (x, y).
f (x, y) = Rz (d)f (x0 , y 0 ) f 0 (x0 , y 0 ).
That is we want to transform the wavefunction so that a coordinate transformation
leaves the physical state unchanged. This is the problem of a passive rotation. Clearly
for this infinitesimal rotation, if we choose
Rz (d) = (1 +

i
dLz )
~

then
i
i
dLz )(1 dLz )f (x, y) = f (x, y)
~
~
satisfies that condition since terms proportional to d2 go to zero.
f (x, y) = Rz (d)f (x0 , y 0 ) = (1 +

As a simpler example, lets consider a translation of the origin. The laws of physics are
independent of this translation. Note that for an infinitesimal positive translation of
df
dx = (1 ~i p dx)f (x).
the origin in the x direction, x0 = x dx, and f (x0 ) = f (x) dx
If we start off in a definite state of x, (p) = eipx0 /~ and we translate the origin a
distance dx then the wavefunction would change to (p) = eip(x0 dx)/~ . However,
the translation operator Tx (dx) = eipdx/~ (1 + ipdx) would transform the state to
0 (p) = Tx (dx)(p) = eipdx/~ eip(x0 dx)/~ = eipx0 /~ , which takes it back to the original
state, even though the coordinates have changed.
So the rotation operator for the function is
i
dLz )
~
A finite rotation can be made by applying the operator for an infinitesimal rotation
over and over. Let z = nd. Then
Rz (d) = (1 +

Rz (z ) = lim (1 +
n

i z
Lz )n = eiz Lz /~ .
~ n

The last step, converting the limit to an exponential is a known identity. We can
verify it by using the log of the quantity. First we expand ln(x) about x = 1: ln(x) =
ln(1) + x1 x=1 (x 1) = (x 1).
i z
i z
i
Lz )n = n(
Lz ) = z Lz
~ n
~ n
~
So exponentiating, we get the identity. Also one can easily see that the exponential
gives the correct infinitesimal rotation in the limit of d 0.
362
lim ln(1 +

17. Operators Matrices and Spin

17.11.4

TOC

Compute the ` = 1 Rotation Operator Rz (z ) *

iLz /~

iLz n
~

n=0

0
1
Lz
= 0
~
0

1
1
Lz
= 0
~
0

2
1
Lz
= 0
~
0

3
1 0
Lz
= 0 0
~
0 0
...

0
0
1

0 0
0 0
0 1

0
0 = Lz /~
1
0
1
0

All the odd powers are the same. All the nonzero even powers are the same. The ~s
all cancel out. We now must look at the sums for each term in the matrix and identify
the function it represents. If we look at the sum for the upper left term of the matrix,
n
i
we get a 1 times (i)
n! . This is just e . There is only one contribution to the middle
term, that is a one from n = 0. The lower right term is like the upper left except the
n
term n. This is just ei . The rest of the
odd terms have a minus sign. We get (i)
n!
terms are zero.
i

e z 0
0
1
0 .
Rz (z ) = 0
iz
0
0 e

17.11.5

Compute the ` = 1 Rotation Operator Ry (y ) *

eiLy /~ =

iLy
~

n=0

0
1
Ly
= 0
~
0

0
1
0

0
0
1
363

17. Operators Matrices and Spin

TOC

0
1 0
1
1 0 1
=
2i
0 1 0

2
1 0 1
1
Ly
0 2 0
=
~
2
1 0 1

3

0
2 0
Ly
Ly
1 1
2 0 2 =
=
~
~
2i 2
0 2 0
...

Ly
~

All the odd powers are the same. All the nonzero even powers are the same. The ~s
all cancel out. We now must look at the sums for each term in the matrix and identify
the function it represents.
The n = 0 term contributes 1 on the diagonals.

iL
The n = 1, 3, 5, ... terms sum to sin() ~y .
The n = 2, 4, 6, ... terms (with a -1 in the matrix) are nearly the series for 12 cos().
The n = 0 term is is missing so subtract 1. The middle matrix element is twice
the other even terms.

eiLy /~

1
= 0
0

0
1
0

0
0
1
0 + sin() 1
2
1
0

1
0
1

0
1
1
1 + (cos() 1) 0
2
1
0

Putting this all together, we get

1
1 sin(y )
2 (1 + cos(y ))
2
1 sin( )
cos(y )
Ry (y ) =
y
2
1
1
2 (1 cos(y )) 2 sin(y )

0 1
2 0
0 1

1
2 (1 cos(y ))
1 sin(y )
.
2
1
(1
+
cos(
))
y
2

364

17. Operators Matrices and Spin

17.11.6

Derive Spin

1
2

TOC

Operators

We will again use eigenstates of Sz , as the basis states.

1
+ =
0

0
=
1
~
Sz =
2

~ 1 0
Sz =
2 0 1
Its easy to see that this is the only matrix that works. It must be diagonal since
the basis states are eigenvectors of the matrix. The correct eigenvalues appear on the
diagonal.
Now we do the raising and lowering operators.
S+ +

S+

S +

0
p

s(s + 1) m(m + 1)~+ = ~+

0 1
~
0 0
0
p
s(s + 1) m(m 1)~ = ~

0 0
~
1 0

We can now calculate Sx and Sy .

1
(S+ + S )
2
1
Sy = (S+ S )
2i
Sx =

=
=

~ 0
2 1

~ 0
2 i

1
0

i
0

These are again Hermitian, Traceless matrices.

17.11.7

Derive Spin

1
2

Rotation Matrices *

In section 17.11.3, we derived the expression for the rotation operator for orbital angular
momentum vectors. The rotation operators for internal angular momentum will follow
365

17. Operators Matrices and Spin

TOC

the same formula.

iSz
~

= ei 2 z

Rz ()

= e

Rx ()

= ei 2 x

Ry ()

= ei 2 y

i n
X
2
jn
=
n!
n=0

ei 2 j

We now can compute the series by looking at the behavior of jn .

1 0
0 1

0 i
i 0

0 1
1 0

1 0
0 1

1 0
y2 =
0 1

1 0
x2 =
0 1
z2 =

Doing the sums

Rz ()

= e

i 2 z

= n=0

( i2 )

Ry ()

( )
i
2

n=0,2,4... n!
=
n

( i2 )
i
n!

n=1,3,5...

( i2 )

n=0,2,4... n!
Rx () =
n

P
( i2 )
n=1,3,5...

ei 2
0

n =

P
( i
2 )
i

( i2 )

n=1,3,5...

i
P

n!
n

(2)

n=0,2,4...

i
P

(2)

n=1,3,5...

P
n=0,2,4...

ei 2

n=0

n!
n

sin 2
cos 2

=
n
( i2 )
n!

cos 2
sin 2

cos 2
i sin 2

i sin 2
cos 2

Note that all of these rotation matrices become the identity matrix for rotations through
720 degrees and are minus the identity for rotations through 360 degrees.

366

17. Operators Matrices and Spin

17.11.8

TOC

NMR Transition Rate in a Oscillating B Field

We have the time dependent Schr

odinger equation for a proton in a static field in the
z direction plus an oscillating field in the x direction.
d
= H
i~
dt

gp
a
Bz
Bx cos t
a
i~
= N
Bx cos t
Bz
b
b
2

gp N
a
Bz
Bx cos t
a
0
1 cos t
a
=
i
=
i
Bx cos t
Bz
b
1 cos t
0
b
b
2~
So far all we have done is plugged things into the Schr
odinger equation. Now we have
to solve this system of two equations. This could be hard but we will do it only near
t = 0, when the EM wave starts. Assume that at t = 0, a = 1 and b = 0, that is, the
nucleus is in the lower energy state. Then we have
a

= i0 a

a
b

= 1ei0 t
= i1 cos ta i0 b = i1 cos tei0 t i0 b
i1 i(+0 )t
(e
+ ei(0 )t ) i0 b
=
2

Now comes the one tricky part of the calculation. The diagonal terms in the Hamiltonian cause a very rapid time dependence to the amplitudes. To get b to grow, we need
to keep adding b in phase with b. To see that clearly, lets compute the time derivative
of bei0 t .
d
(bei0 t )
dt

=
=

i1 i(+20 )t
(e
+ ei(20 )t ) i0 bei0 t + i0 bei0 t
2
i1 i(+20 )t
(e
+ ei(20 )t )
2

Terms that oscillate rapidly will average to zero. To get a net change in bei0 t , we
need to have 20 . Then the first term is important and we can neglect the second
which oscillates with a frequency of the order of 1011 . Note that this is exactly the
condition that requires the energy of the photons in the EM field E = ~ to be equal
to the energy difference between the two spin states E = 2~0 .

367

17. Operators Matrices and Spin

TOC

d
(bei0 t )
dt

bei0 t

i1
2
i1
t
2

It appears that the amplitude grows linearly with time and hence the probability would
grow like t2 . Actually, once we do the calculation (only a bit) more carefully, we will
see that the probability increases linearly with time and there is a delta function of
energy conservation. We will do this more generally in the section on time dependent
perturbation theory.
In any case, we can only cause transitions if the EM field is tuned so that 20
which means the photons in the EM wave have an energy equal to the difference in
energy between the spin down state and the spin up state. The transition rate increases
as we increase the strength of the oscillating B field.

17.12

Homework Problems

1
1. An angular momentum 1 system is in the state = 126 3. What is the
4
probability that a measurement of Lx yields a value of 0?
important
fundamental
easy
1
0
0
2

1
0
Lx = ~ 12
2
1

0
0
2

1
0
0
a
2
1
1

0
b =0
2
2
1

c
0
0
2

b
(a + c) = 0
b
1
(x)

(x)

= 0
1

2

= 126 12

2
1
3 =

4

9
52

2. A spin 12 particle is in an eigenstate of Sy with eigenvalue + ~2 at time t = 0. At

that time it is placed in a constant magnetic field B in the z direction. The spin is
allowed to precess for a time T . At that instant, the magnetic field is very quickly
368

17. Operators Matrices and Spin

TOC

switched to the x direction. After another time interval T , a measurement of the

y component of the spin is made. What is the probability that the value ~2 will
be found?
importantfundamental
medium

0 i
a
a
(y)
y =
=
first eigenstates of Sy
i
0
b
b

ib
a
=
ia
!b
(y)

1
2
i

ge~
ge ~ ~
~ = B ~
B
2mc S B = !
4mc ~
!
1 eit
1 eiB Bt/~
2
2

i eit
i eiB Bt/~
2
2 !
iT

~ =
H = ~
B
(t T ) =
(T ) =

1 e
2
i eiT
2

H
B B
x = ~
=

x
0 1
a
a
=
1
0
b
b

b
a
=
a
b!
(x)

~ = B BZ
B

Hamiltonian

time dep. from energy in B field

evaluate at time T
Hamiltonian for t > T
find eigenstates of x

1
2
1

Write the state as a vector in terms of the E eigenstates for B in the x direction.
(x)
(x)
(x)
(x)
(T ) = h+ |(T )i+ + h!|(T )i
!
!
!
1
1
iT

e

1 eiT

1
1
1
1
1
2
2
2
2
(T ) = 2 2
+ 2 2
1
1

i eiT
i eiT
2
2
2

2

1
1
1
iT
iT
iT
iT
+ ie
+ e
ie
(T ) = 22 e
1
1

iT 1
1
1
iT
iT
iT
iT
iT
e
e
(2T ) = 22 e
+ ie
+ e
ie
1
1

1
1
1
(2T ) = 2
e2iT + i
+ 1 ie2iT
2
1
1
2iT

2iT
e

ie
+
i
+
1
1
(2T ) = 2
2 e2iT + ie2iT + i 1

e2iT ie2iT + i + 12
1 1
(y)
(y)
i

P = 2 2
dot into
2
2
e2iT + ie2iT + i 1

2
(y)
P = 41 e2iT ie2iT + i + 1 + i(e2iT + ie2iT + i 1)

2
(y)
P = 41 e2iT ie2iT + i + 1 + ie2iT e2iT 1 i
1

2
(y)
P = 4 (1 + i)e2iT (1 + i)e2iT

2
(y)

P = 41 (1 + i) (2i sin(2T ))

1i
2
(y)
2 2B BT

P = 2 sin(2T ) = sin
~
369

17. Operators Matrices and Spin

TOC

3. Consider a system of spin 12 . What are the eigenstates and eigenvalues of the
operator Sx + Sy ? Suppose a measurement of this quantity is made, and the
system is found to be in the eigenstate with the larger eigenvalue. What is the
probability that a subsequent measurement of Sy yields ~2 ?
important
fundamental
easy
0
1i
x + y =
1 + i
0

1 i

1 + i = 0
2 2= 0
= 2
Eigenvalues for Sx + Sy are ~2 = ~2 .

~ a
0
1i
a
(1 i)b
~
~
= 2
= 22
find upper eigenvalue state
2 1+i
0
b
(1 + i)a
b
1

(x+y)
2
+
= 1+i
2

1 2
(y) (x+y) 2 1
(y)
i
2

i = 2
P+ = h+ |+
1+i
2
2

2
2

(y)
i
= 2+1
= 3+2 2 + 1
P+ = 12 + 21i
+
8
8
2
2 2
2 2
(y)

P+ =

2+ 2
4

4. The Hamiltonian matrix is given to be

8
H = ~ 4
6

4 6
10 4 .
4 8

What are the eigen-energies and corresponding eigenstates of the system? (This
isnt too messy.)
important
calculational

medium
8

4
6

4
14
4 = 0 eigenvalue eq.

6
4
8

2
0
2

4
14
4 = 0 subtract third row from first

6
4
8
(2 )[(14 )(8 ) 16] + ( 2)[16 6(14 )] = 0
( 2)[2 22 + 112 16] ( 2)[6 + 16 84] = 0
( 2)[2 22 + 112 16 6 16 + 84] = 0
( 2)[2 28+ 164] = 0

= 2, = 282 128 = 14 4 2

The eigen-energies are E = 2~, and E = (14 4 2)~.

370

17. Operators Matrices and Spin

TOC

8 4 6
a
a
4 14 4 b = 2 b
6 4 8 c c
4a + 2b + 3c
a
2a + 7b + 2c = b
c
3a + 2b + 4c
3a + 2b + 3c
2a + 6b + 2c = 0
3a + 2b + 3c
c = a 32 b
2a + 6b 2a 34 b = 0
b = 0,
c = a
1
2

= 0

for eigenvalue 2

4a + 2b + 3c
a

2a + 7b + 2c = (7 + 2 2) b
3a + 2b
c
+ 4c

(3 2 2)a + 2b + 3c = 0
2a 2 2b + 2c = 0

2a 2 2b + 2 (1 + 2 3 2 )a 23 b = 0
c = (1 + 2 3 2 )a 23 b

4 + 4 3 2 a = 2 2 + 34 b

b
2

c = (1 + 2 3 2 ) b2 32 b =
1
2

= 12

b
2

for eigenvalue 7 + 2 2

2
1
2

= 12

for eigenvalue 7 2 2

1
2

5. What are the eigenfunctions and eigenvalues of the operator Lx Ly + Ly Lx for a

spin 1 system?
fundamental medium

371

17. Operators Matrices and Spin

TOC

1 0
0 1
1 0
i 0
0 i
i
0
i 0 i
2
Lx Ly = ~2 0 0 0
i 0 i
i 0 i
2
Ly Lx = ~2 0 0 0
i 0 i

0 0 i
Lx Ly + Ly Lx = ~2 0 0 0
i 0 0

0

i

0 0 = 0

i
0
3 i(i) = 0
3 = ( + 1)( 1) = 0
2
2
The
are
0, i~
, and
i~ .
eigenvalues
a
ic
0 0 i
0 0 0 b = 0 = 0
c
ia
i 0 0
0
0 = 1

0

ic
a
0 = 1 b
ia
c
1
0
Lx = ~2 1
0
0
Ly = ~2 i
0

+1 = 0
i
2

ic
a
0 = 1 b
ia
c
1
2

1 = 0
i

6. Calculate the ` = 1 operator for arbitrary rotations about the x-axis. Use the
usual Lz eigenstates as a basis.
fundamental medium

372

17. Operators Matrices and Spin

x n
P
( iL
~ )
Rx () = eiLx /~ =
n!
n=0

1 0 0

Lx 0
= 0 1 0
~
0 0 1
0 1 0

Lx 1
= 12 1 0 1
~
0
0 1
1
0
1

Lx 2
= 12 0 2 0
~
1 0 1
0 2 0

Lx 3
1 1 2 0 2 = Lx
=
~
2
~
2
0 2 0
All the odd powers are the same. All
The ~s all
cancel out.

1 0 0
0
eiLx /~ = 0 1 0 + sin() 12 1
0
10 0 1
1 sin()
(1
+
cos())
2
2

cos()
Rx () = 12 sin()
12 (1 cos()) 12 sin()

TOC

the nonzero even powers are the same.

1 0
1
0 1 + 12 (cos() 1) 0
1 0
1

12 (1 cos())

1 sin()
.
2
1
(1
+
cos())
2

0
2
0

1
0
1

7. An electron is in an eigenstate of Sx with eigenvalue ~2 . What are the amplitudes

to find the electron with a) Sz = + ~2 , b) Sz = ~2 , c) Sy = + ~2 , d) Su = + ~2 ,
where the u-axis is assumed to be in the x y plane rotated by and angle from
the x-axis.
important fundamental easy

373

17. Operators Matrices and Spin

electron =
(y)

(x)
+

1
2
i

1
2
1
2

TOC

a) just take the top component of spinor, amplitude= 12

b) take bottom component, amplitude= 12
!

1
(y) (x)
1
i
2

= 1i
c) amplitude= h+ |+ i =
2
2
2
1
2

0
cos i sin
0
d) u = cos x + sin y =
= i
cos

+
i
sin

0
e

i
0 ei
a
a
e b
=
=
b
ei
0 ! b
ei a
1
(u)

+ =

ei
0

2
ei

amplitude=

(u) (x)
h+ |+ i

1
2

i
e
2

1
2
1
2

!
=

1+ei
2

8. Particles with angular momentum 1 are passed through a Stern-Gerlach apparatus which separates them according to the z-component of their angular momentum. Only the m = 1 component is allowed to pass through the apparatus. A
second apparatus separates the beam according to its angular momentum component along the u-axis. The u-axis and the z-axis are both perpendicular to the
beam direction but have an angle between them. Find the relative intensities
of the three beams separated in the second apparatus.
important fundamental easy
Choose the y axis as the beam axis, so the u axis is in the x-z plane, an angle
from the z axis.

+|
+ D+
Oven(I0 ) 0|
0 D0

z
D z

0
The state coming out of the first apparatus is 0.
1
To get the z axis to be where the u axis is, we rotate the coordinate system by
around the y axis.
(z)
0 =
Ry () =

1
1
1
1 sin(y )
(1

cos(
))
y
0
2 (1 + cos(y ))
2
2
2 (1 cos(y ))

1
1 sin(y )
cos(y )
0 = 12 sin(y )
0 = 2 sin(y )
2
1
1
1
1

1
2 (1 cos(y )) 2 sin(y )
2 (1 + cos(y ))
2 (1 + cos(y ))
The three relative intensities are proportional to the squares of the three amplitudes in the 0 vector. The intensities are the same for rotations through .
374

17. Operators Matrices and Spin

TOC

9. Find the eigenstates of the harmonic oscillator lowering operator A. They should
satisfy the equation A|i = |i. Do this by finding the coefficients hn|i where
|ni is the nth energy eigenstate. Make sure that the states |i are normalized so
that h|i = 1. Suppose |0 i is another such state with a different eigenvalue.
Compute h0 |i. Would you expect these states to be orthogonal?
extension medium
P
Lets write the solution as |i =
an un .
n
P
P
A an un = an un
n P
P n
nan un1 = an un
n
n P
P
n + 1an+1 un = an un
nn+1
n

an+1 = n+1
an
n

an = n! a0
P n
un
|i = a0
n!
n
P 2n
2 2
2
h|i = a0
n! = a0 e
nP
2
n

|i = e /2
u
n! n
n
(0 )2
02 2 P
02 +20 2
0n n
2
2
h0 |i = e 2
=e
=e
n!
n

These need not be orthogonal since A is not Hermitian.

10. Find the matrix which represents the p2 operator for a 1D harmonic oscillator.
Write out the upper left 5 5 part of the matrix.

m~
AA AA A A + A A
2
p

p
m~
(p2 )ij =
j(j 1)i,j2 + (j + 1)i,j + ji,j + (j + 1)(j + 2)i,j+2
2

p
m~ p
(p2 )ij =
j(j 1)i,j2 + (2j + 1)i,j + (j + 1)(j + 2)i,j+2
2

1
0
2 0
0
0
3
0
6 0

m~
2

p =
2 0
5
0
12

2
6 0
7
0
0
0
12 0
9
p2 =

11. Lets define the u axis to be in the x-z plane, between the positive x and z axes
and at an angle of 30 degrees to the x axis. Given an unpolarized spin 21 beam of
intensity I going into the following Stern-Gerlach apparati, what intensity comes
out?

+
+
a) I

?
| z
| x
375

17. Operators Matrices and Spin

TOC

+
+|

?
| z

u
+
+|
+|
c) I

?
| z
u

z

+|
+
+
?

d) I

u
| z
z

+|
+|
+
?

e) I
x
u
| z
b) I

17.13

Sample Test Problems

~ is
1. * We have shown that the Hermitian conjugate of a rotation operator R()
~ Use this to prove that if the i form an orthonormal complete set, then
R().
~ i are also orthonormal and complete.
the set 0i = R()
2. Given that un is the nth one dimensional harmonic oscillator energy eigenstate:
a) Evaluate the matrix element hum |p2 |un i. b) Write the upper left 5 by 5 part
of the p2 matrix.

2
3. A spin 1 system is in the following state in the usual Lz basis: = 15 1 + i.
i
What is the probability that a measurement of the x component of spin yields
zero? What is the probability that a measurement of the y component of spin
yields +~?
4. In a three state system, the matrix elements are given as h1 |H|1 i = E1 ,
h2 |H|2 i = h3 |H|3 i = E2 , h1 |H|2 i = 0, h1 |H|3 i = 0, and h 2 |H|3 i =
. Assume all of the matrix elements are real. What are the energy eigenvalues
and eigenstates of the system? At t = 0 the system is in the state 2 . What is
(t)?
5. Find the (normalized) eigenvectors and eigenvalues of the Sx (matrix) operator
for s = 1 in the usual (Sz ) basis.
6. * A spin 21 particle is in a magnetic field in the x direction giving a Hamiltonian
H = B Bx . Find the
time development (matrix) operator eiHt/~ in the usual

1
basis. If (t = 0) =
, find (t).
0

3
7. A spin 12 system is in the following state in the usual Sz basis: = 15
.
1+i
What is the probability that a measurement of the x component of spin yields
+ 21 ?
376

17. Operators Matrices and Spin

TOC

i
(in the usual Sz eigenstate ba2
sis). What is the probability that a measurement of Sx yields ~
2 ? What is the
?
probability that a measurement of Sy yields ~
2

8. A spin

1
2

system is in the state =

1
5

9. A spin 12 object is in an eigenstate of Sy with eigenvalue ~2 at t=0. The particle

is in a magnetic field B = (0, 0, B) which makes the Hamiltonian for the system
H = B Bz . Find the probability to measure Sy = ~2 as a function of time.
10. Two degenerate eigenfunctions of the Hamiltonian are properly normalized and
have the following properties.
H1 = E0 1

H2 = E0 2

P 1 = 2

P 2 = 1

What are the properly normalized states that are eigenfunctions of H and P?
What are their energies?
11. What are the eigenvectors and eigenvalues for the spin

1
2

operator Sx + Sz ?

12. A spin 12 object is in an eigenstate of Sy with eigenvalue ~2 at t=0. The particle

is in a magnetic field B = (0, 0, B) which makes the Hamiltonian for the system
H = B Bz . Find the probability to measure Sy = ~2 as a function of time.
13. A spin 1 system is in the following state, (in the usual Lz eigenstate basis):

i
1
=
2 .
5 1+i
What is the probability that a measurement of Lx yields 0? What is the probability that a measurement of Ly yields ~?
14. A spin 12 object is in an eigenstate of Sz with eigenvalue ~2 at t=0. The particle
is in a magnetic field B = (0, B, 0) which makes the Hamiltonian for the system
H = B By . Find the probability to measure Sz = ~2 as a function of time.
15. A spin 1 particle is placed in an external field in the u direction such that the
Hamiltonian is given by
!

3
1
H=
Sx + Sy
2
2
Find the energy eigenstates and eigenvalues.
16. A (spin 12 ) electron is in an eigenstate of Sy with eigenvalue ~2 at t = 0. The
~ = (0, 0, B) which makes the Hamiltonian for the
particle is in a magnetic field B
system H = B Bz . Find the spin state of the particle as a function of time.
Find the probability to measure Sy = + ~2 as a function of time.
377

18. Homework Problems 130A

18
18.1

TOC

Homework Problems 130A

HOMEWORK 1

1. A polished Aluminum plate is hit by beams of photons of known energy. It is

measured that the maximum electron energy is 2.3 0.1 eV for 2000 Angstrom
light and 0.90 0.04 eV for 2580 Angstrom light. Determine Plancks constant
and its error based on these measurements.
2. A 200 keV photon collides with an electron initially at rest. The photon is
observed to scatter at 90 degrees in the electron rest frame. What are the kinetic
energies of the electron and photon after the scattering?
3. Use the energy density in a cavity as a function of frequency and T
u(, T ) =

8h
3
c3 eh/kT 1

to calculate the emissive power of a black body E(, T ) as a function of wavelength and temperature.
4. What is the DeBroglie wavelength for each of the following particles? The energies
given are the kinetic energies.
a 1 eV electron
a 104 MeV proton
a 1 gram lead ball moving with a velocity of 100 cm/sec.
5. The Dirac delta function has the property that

f (x)(x x0 ) dx = f (x0 )

Find the momentum space wave function (p) if (x) = (x x0 ).

6. Use the calculation of a spreading Gaussian wave packet to find the fractional
change in size of a wave packet between t = 0 and t = 1 second for an electron
localized to 1 Angstrom. Now find the fraction change for a 1 gram weight
localized to 1 nanometer.
7. Use the uncertainty principle to estimate the energy of the ground state of a
p2
+ 12 kx2 .
harmonic oscillator with the Hamiltonian H = 2m
8. Estimate the kinetic energy of an electron confined to be inside a nucleus of
radius 5 Fermis. Estimate the kinetic energy of a neutron confined inside the
same nucleus.

378

18. Homework Problems 130A

18.2

TOC

Homework 2

1. Show that

(x)x(x)dx =

(p) i~
p

(p)dp.

Remember that the wave functions go to zero at infinity.

2. Directly calculate the the RMS uncertainty in x for the state (x) =
by computing
p
x = h|(x hxi)2 |i.

ax2
2

3. Calculate hpn i for the state in the previous problem. Use this to calculate p in
a similar way to the x calculation.
4. Calculate the commutator [p2 , x2 ].
5. Consider the functions of one angle () with and () = ().
d
Show that the angular momentum operator L = ~i d
has real expectation values.
6. A particle is in the first excited state of a box of length L. What is that state?
Now one wall of the box is suddenly moved outward so that the new box has
length D. What is the probability for the particle to be in the ground state of
the new box? What is the probability for the particle to be in the first excited
state of the new box? You may find it useful to know that

sin ((A B)x) sin ((A + B)x)

sin(Ax) sin(Bx)dx =

.
2(A B)
2(A + B)
7. A particle is initially in the nth eigenstate of a box of length L. Suddenly the
walls of the box are completely removed. Calculate the probability to find that
the particle has momentum between p and p + dp. Is energy conserved?
8. A particle is in a box with solid walls at x = a2 . The state at t = 0 is constant
q
(x, 0) = a2 for a2 < x < 0 and the (x, 0) = 0 everywhere else. Write this
state as a sum of energy eigenstates of the particle in a box. Write (x, t) in
terms of the energy eigenstates. Write the state at t = 0 as (p). Would it be
correct (and why) to use (p) to compute (x, t)?
9. The wave function for a particle is initially (x) = Aeikx + Beikx . What is the
probability flux j(x)?
10. Prove that the parity operator defined by P (x) = (x) is a hermitian operator
and find its possible eigenvalues.

379

18. Homework Problems 130A

18.3

TOC

Homework 3

1. A general one dimensional scattering problem could be characterized by an (arbitrary) potential V (x) which is localized by the requirement that V (x) = 0 for
|x| > a. Assume that the wave-function is
ikx
Ae + Beikx
x < a
(x) =
Ceikx + Deikx
x>a
Relating the outgoing waves to the incoming waves by the matrix equation

C
S11 S12
A
=
B
S21 S22
D
show that
|S11 |2 + |S21 |2 = 1
|S12 |2 + |S22 |2 = 1

S11 S12
+ S21 S22
=0

Use this to show that the S matrix is unitary.

2. Calculate the S matrix for the potential

V0
V (x) =
0

|x| < a
|x| > a

and show that the above conditions are satisfied.

3. The odd bound state solution to the potential well problem bears many similarities to the zero angular momentum solution to the 3D spherical potential well.
Assume the range of the potential is 2.3 1013 cm, the binding energy is -2.9
MeV, and the mass of the particle is 940 MeV. Find the depth of the potential
in MeV. (The equation to solve is transcendental.)
4. Find the three lowest energy wave-functions for the harmonic oscillator.
5. Assume the potential for particle bound inside a nucleus is given by
(
V0
x<R
V (x) = ~2 `(`+1)
x>R
2mx2
and that the particle has mass m and energy e > 0. Estimate the lifetime of the
particle inside this potential.

380

18. Homework Problems 130A

18.4

TOC

Homework 4

1. The 1D model of a crystal puts the following constraint on the wave number k.
cos() = cos(ka) +

ma2 V0 sin(ka)
~2
ka

Assume that ma~2V0 = 3

2 and plot the constraint as a function of ka. Plot the
allowed energy bands on an energy axis assuming V0 = 2 eV and the spacing
between atoms is 5 Angstroms.
2. In a 1D square well, there is always at least one bound state. Assume the width
of the square well is a. By the uncertainty principle, the kinetic energy of an
~2
electron localized to that width is 2ma
2 . How can there be a bound state even
for small values of V0 ?
3. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =
0) = 12 (u0 + u1 ). Is correctly normalized? Compute the expected value of x
as a function of time by doing the integrals in the x representation.
2

4. Prove the Schwartz inequality |h|i| h|ih|i. (Start from the fact that
h + C| + Ci 0 for any C.
5. The hyper-parity operator H has the property that H 4 = for any state .
Find the eigenvalues of H for the case that it is not Hermitian and the case that
it is Hermitian.
6. Find the correctly normalized energy eigenfunction u5 (x) for the 1D harmonic
oscillator.
7. A beam of particles of energy E > 0 coming from is incident upon a double
delta function potential in one dimension. That is V (x) = (x + a) (x a).
a) Find the solution to the Schr
odinger equation for this problem.
b) Determine the coefficients needed to satisfy the boundary conditions.
c) Calculate the probability for a particle in the beam to be reflected by the
potential and the probability to be transmitted.

381

18. Homework Problems 130A

18.5

TOC

Homework 5

1. An operator is Unitary if U U = U U = 1. Prove that a unitary operator

6. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =

0) = 12 (u0 + u1 ).
Compute the expected value of x as a function of time using A and A in

the Schrodinger
picture.
Now do the same in the Heisenberg picture.

382

18. Homework Problems 130A

18.6

TOC

Homework 6

1. The energy spectrum of hydrogen can be written in terms of the principal quan2
2
tum number n to be E = 2nc2 . What are the energies (in eV) of the photons
from the n = 2 n = 1 transition in hydrogen and deuterium? What is the
difference in photon energy between the two isotopes of hydrogen?
2. Prove that the operator that exchanges two identical particles is Hermitian.
3. Two identical, non-interacting spin 21 particles are in a box. Write down the full
lowest energy wave function for both particles with spin up and for one with spin
up and the other spin down. Be sure your answer has the correct symmetry under
the interchange of identical particles.
4. At t = 0 a particle is in the one dimensional Harmonic Oscillator state (t =
0) = 12 (u1 + u3 ). Compute the expected value of x2 as a function of time.
5. Calculate the Fermi energy of a gas of massless fermions with n particles per unit
volume.
6. The number density of conduction electrons in copper is 8.5 1022 per cubic
centimeter. What is the Fermi energy in electron volts?
1

7. The volume of a nucleus is approximately 1.1A 3 Fermis, where A = N + Z, N is

the number of neutrons, and Z is the number of protons. A Lead nucleus consists
of 82 protons and 126 neutrons. Estimate the Fermi energy of the protons and
neutrons separately.

8. The momentum operator conjugate to any cooridinate xi is ~i x

. Calculate the
i
commutators of the center of mass coordinates and momenta [Pi , Rj ] and of the
internal coordinates and momenta [pi , rj ]. Calculate the commutators [Pi , rj ] and
[pi , Rj ].

383

18. Homework Problems 130A

18.7

TOC

Homework 7

1. A particle is in the state = R(r)

1
3 Y21

1
3 Y20

1
3 Y22

. Find the

expected values of L , Lz , Lx , and Ly .

q
q

1
2
Y
+
i
Y
2. A particle is in the state = R(r)
. If a measurement of the
11
10
3
3
x component of angular momentum is made, what are the possilbe outcomes and
what are the probabilities of each?
3. Calculate the matrix elements hY`m1 |Lx |Y`m2 i and hY`m1 |L2x |Y`m2 i
L2 +L2

4. The Hamiltonian for a rotor with axial symmetry is H = x2I1 y + 2Iz2 where the
I are constant moments of inertia. Determine and plot the eigenvalues of H for
dumbbell-like case that I1 >> I2 .
5. Prove that hL2x i = hL2y i = 0 is only possible for ` = 0.
6. Write the spherical harmonics for ` 2 in terms of the Cartesian coordinates x,
y, and z.
7. A particle in a spherically symmetric potential has the wave-function (x, y, z) =
2
C(xy+yz+zx)er . A measurement of L2 is made. What are the possible results
and the probabilities of each? If the measurement of L2 yields 6~2 , what are the
possible measured values of Lz and what are the corresponding probabilities?
8. The deuteron, a bound state of a proton and neutron with ` = 0, has a binding
energy of -2.18 MeV. Assume that the potential is a spherical well with potential
of V0 for r < 2.8 Fermis and zero potential outside. Find the approximate value
of V0 using numerical techniques.

384

18. Homework Problems 130A

18.8

TOC

Homework 8

1. Calculate the ` = 0 phase shift for the spherical potential well for both and
attractive and repulsive potential.
2. Calculate the ` = 0 phase shift for a hard sphere V = for r < a and V = 0 for
r > a. What are the limits for ka large and small?
3. Show that at large r, the radial flux is large compared to the angular components
ikr
of the flux for wave-functions of the form C e r Y`m (, ).
4. Calculate the difference in wavelengths of the 2p to 1s transition in Hydrogen and
Deuterium. Calculate the wavelength of the 2p to 1s transition in positronium.
5. Tritium is a unstable isotope of Hydrogen with a proton and two neutrons in
the nucleus. Assume an atom of Tritium starts out in the ground state. The
nucleus (beta) decays suddenly into that of He3 . Calculate the probability that
the electron remains in the ground state.

6. A hydrogen atom is in the state = 61 4100 + 3211 210 + 10211 .
What are the possible energies that can be measured and what are the probabilities of each? What is the expectation value of L2 ? What is the expectation
value of Lz ? What is the expectation value of Lx ?
7. What is P (pz ), the probability distribution of pz for the Hydrogen energy eigenstate 210 ? You may find the expansion of eikz in terms of Bessel functions
useful.
2

p
8. The differential equation for the 3D harmonic oscillator H = 2m
+ 12 m 2 r2 has
been solved in the notes, using the same techniques as we used for Hydrogen. Use
the recursion relations derived there to write out the wave functions n`m (r, , )
for the three lowest energies. You may write them in terms of the standard Y`m
but please write out the radial parts of the wavefunction completely. Note that
there is a good deal of degeneracy in this problem so the three lowest energies
actually means 4 radial wavefunctions and 10 total states. Try to write the
solutions 000 and 010 in terms of the solutions in cartesian coordinates with
the same energy nx,ny,nz .

385

18. Homework Problems 130A

18.9

TOC

Homework 9
2

1. An electron in the Hydrogen potential V (r) = er is in the state (~r) = Cer .

Find the value of C that properly normalizes the state. What is the probability
that the electron be found in the ground state of Hydrogen?
2. An electron is in the 210 state of hydrogen. Find its wave function in momentum
space.
3. A spin 12 particle is in an eigenstate of Sy with eigenvalue + ~2 at time t = 0. At
that time it is placed in a constant magnetic field B in the z direction. The spin is
allowed to precess for a time T . At that instant, the magnetic field is very quickly
switched to the x direction. After another time interval T , a measurement of the
y component of the spin is made. What is the probability that the value ~2 will
be found?
4. Consider a system of spin 21 . What are the eigenstates and eigenvalues of the
operator Sx + Sy ? Suppose a measurement of this quantity is made, and the
system is found to be in the eigenstate with the larger eigenvalue. What is the
probability that a subsequent measurement of Sy yields ~2 ?
5. Lets define the u axis to be in the x-z plane, between the positive x and z axes
and at an angle of 30 degrees to the x axis. Given an unpolarized spin 12 beam of
intensity I going into the following Stern-Gerlach apparati, what intensity comes
out?

+
+
?

I
| x
| z

+
+|
I

?
| z
u

+
+|
+|
I

?
| z
u
z

+
+
+|
I

?
| z
u
z

+
+|
+|
?
I

| z
u
x

386

19. Electrons in an Electromagnetic Field

TOC

Electrons in an Electromagnetic Field

In this section, we will study the interactions of electrons in an electromagnetic field.

We will compute the additions to the Hamiltonian for magnetic fields. The gauge
symmetry exhibited in electromagnetism will be examined in quantum mechanics. We
will show that a symmetry allowing us to change the phase of the electron wave function
requires the existence of EM interactions (with the gauge symmetry).
These topics are covered in Gasiorowicz Chapter 13, and in Cohen-Tannoudji
et al. Complements EV I , DV II and HIII .

19.1

Review of the Classical Equations of Electricity and Magnetism in CGS Units

You may be most familiar with Maxwells equations and the Lorentz force law in SI
units as given below.
~ B
~

~ E
~ + B

t
~
~
E

~ B
~ 1 E = 0 J~

c2 t
~ + ~v B).
~
F~ = e(E
These equations have needless extra constants (not) of nature in them so we dont like
to work in these units. Since the Lorentz force law depends on the product of the
charge and the field, there is the freedom to, for example, increase the field by a factor
of 2 but decrease the charge by a factor of 2 at the same time. This will put a factor
of 4 into Maxwells equations but not change physics. Similar tradeoffs can be made
with the magnetic field strength and the constant on the Lorentz force law.
The choices made in CGS units are more physical (but still not perfect). There are no
extra constants other than . Our textbook and many other advanced texts use CGS
units and so will we in this chapter. Maxwells Equations in CGS units are
~ B
~

~ E
~ + 1 B

c t
~ E
~

~ B
~ 1 E

c t

4
4 ~
=
J.
c

387

19. Electrons in an Electromagnetic Field

The Lorentz Force is

TOC

~ + 1 ~v B).
~
F~ = e(E
c

In fact, an even better definition (rationalized Heaviside-Lorentz units) of the charges

and fields can be made as shown in the introduction to field theory in chapter 31. For
now we will stick with the more standard CGS version of Maxwells equations.
If we derive the fields from potentials,
~
B

~ A
~
=

~
E

~
=

1 A
c t

then the first two Maxwell equations are automatically satisfied. Applying the second
two equations we get wave equations in the potentials.
1 ~ ~
2
( A) =
c t

2~
~ A
~ + 1
~+ 1 A +
~
2 A
=
c2 t2
c t

4
4 ~
J
c

These derivations are fairly simple using Einstein notation.

The two results we want to use as inputs for our study of Quantum Physics are
the classical gauge symmetry and
the classical Hamiltonian.
The Maxwell equations are invariant under a gauge transformation of the potentials.
~
A

~ f
~ (~r, t)
A
1 f (~r, t)
+
c t

Note that when we quantize the field, the potentials will play the role that wave functions do for the electron, so this gauge symmetry will be important in quantum mechanics. We can use the gauge symmetry to simplify our equations. For time independent
~ A
~ = 0, is often used. For
charge and current distributions, the coulomb gauge,
~ A
~ + 1 = 0, is often convenient.
time dependent conditions, the Lorentz gauge,
c t
These greatly simplify the above wave equations in an obvious way.
Finally, the classical Hamiltonian for electrons in an electromagnetic field
becomes
p2
1
e ~ 2
H=

p~ + A
e
2m
2m
c
388

19. Electrons in an Electromagnetic Field

TOC

The magnetic force is not a conservative one so we cannot just add a scalar potential.
We know that there is momentum contained in the field so the additional momentum
term, as well as the usual force due to an electric field, makes sense. The electron
~ B
~ gives rise to momentum
generates an E-field and if there is a B-field present, E
density in the field. The evidence that this is the correct classical Hamiltonian is that
we can derive the Lorentz Force from it.

19.2

The Quantum Hamiltonian Including a B-field

We will quantize the Hamiltonian

e ~ 2
1
e
p~ + A
2m
c

in the usual way, by replacing the momentum by the momentum operator, for the case
of a constant magnetic field.
Note that the momentum operator will now include momentum in the field, not just
the particles momentum. As this Hamiltonian is written, p~ is the variable conjugate
to ~r and is related to the velocity by
e~
p~ = m~v A
c
as seen in our derivation of the Lorentz force.
The computation yields

~2 2
e ~ ~
e2 2 2
~ 2 = (E + e).
+
B L +
r
B

(~
r

B)
2m
2mc
8mc2
The usual kinetic energy term, the first term on the left side, has been recovered. The
standard potential energy of an electron in an Electric field is visible on the right side.
We see two additional terms due to the magnetic field. An estimate of the size of
the two B field terms for atoms shows that, for realizable magnetic fields, the first
B
term is fairly small (down by a factor of 2.410
9 gauss compared to hydrogen binding
energy), and the second can be neglected. The second term may be important in very
high magnetic fields like those produced near neutron stars or if distance scales are
larger than in atoms like in a plasma (see example below).
So, for atoms, the dominant additional term is the one we anticipated classically in
section 17.4,
e ~ ~
~
HB =
B L = ~
B,
2mc
e ~
where
~ = 2mc
L. This is, effectively, the magnetic moment due to the electrons
orbital angular momentum. In atoms, this term gives rise to the Zeeman effect:
389

19. Electrons in an Electromagnetic Field

TOC

otherwise degenerate atomic states split in energy when a magnetic field is applied.
Note that the electron spin which is not included here also contributes to the splitting
and will be studied later.
The Zeeman effect, neglecting electron spin, is particularly simple to calculate because the the hydrogen energy eigenstates are also eigenstates of the additional term
in the Hamiltonian. Hence, the correction can be calculated exactly and easily.
Example: Splitting of orbital angular momentum states in a B field.

The result is that the shifts in the eigen-energies are

E = B Bm`
where m` is the usual quantum number for the z component of orbital angular momentum. The Zeeman splitting of Hydrogen states, with spin included, was a powerful
tool in understanding Quantum Physics and we will discuss it in detail in chapter 22.
The additional magnetic field terms are important in a plasma because the typical
radii can be much bigger than in an atom. A plasma is composed of ions and
electrons, together to make a (usually) electrically neutral mix. The charged particles
are essentially free to move in the plasma. If we apply an external magnetic field, we
have a quantum mechanics problem to solve. On earth, we use plasmas in magnetic
fields for many things, including nuclear fusion reactors. Most regions of space contain
plasmas and magnetic fields.
In the example below, we will solve the Quantum Mechanics problem two ways: one
using our new Hamiltonian with B field terms, and the other writing the Hamiltonian
in terms of A. The first one will exploit both rotational symmetry about the B
field direction and translational symmetry along the B field direction. We
will turn the radial equation into the equation we solved for Hydrogen. In the
second solution, we will use translational symmetry along the B field direction
as well as translational symmetry transverse to the B field. We will now turn
the remaining 1D part of the Schr
odinger equation into the 1D harmonic oscillator
equation, showing that the two problems we have solved analytically are actually
related to each other!
Example: A neutral plasma in a constant magnetic field.

The result in either solution for the eigen-energies can be written as

1
~2 k 2
eB~
n+
+
.
En =
me c
2
2me
which depends on 2 quantum numbers. ~k is the conserved momentum along the field
direction which can take on any value. n is an integer dealing with the state in x
390

19. Electrons in an Electromagnetic Field

TOC

and y. In the first solution we understand n in terms of the radial wavefunction in

cylindrical coordinates and the angular momentum about the field direction. In the
second solution, the physical meaning is less clear.

19.3

Gauge Symmetry in Quantum Mechanics

Gauge symmetry in Electromagnetism was recognized before the advent of quantum

mechanics. We have seen that symmetries play a very important role in the quantum
theory. Indeed, in quantum mechanics, gauge symmetry can be seen as the basis for
electromagnetism and conservation of charge.
We know that the all observables are unchanged if we make a global change of the
phase of the wavefunction, ei . We could call this global phase symmetry.
All relative phases (say for amplitudes to go through different slits in a diffraction
experiment) remain the same and no physical observable changes. This is a symmetry
in the theory which we already know about. Lets postulate that there is a bigger
symmetry and see what the consequences are.
(~r, t) ei(~r,t) (~r, t)
That is, we can change the phase by a different amount at each point in spacetime and
the physics will remain unchanged. This local phase symmetry is bigger than the
global one.
Its clear that this transformation leaves the absolute square of the wavefunction
the same, but what about the Schr
odinger equation? It must also be unchanged. The
derivatives in the Schr
odinger equation will act on (~r, t) changing the equation
unless we do something else to cancel the changes.
1
e ~ 2
p~ + A
= (E + e)
2m
c
A little calculation shows that the equation remains unchanged if we also transform
the potentials
~
A

~ f
~ (~r, t)
A
1 f (~r, t)
+
c t
~c
f (~r, t) =
(~r, t).
e
This is just the standard gauge transformation of electromagnetism, but, we
now see that local phase symmetry of the wavefunction requires gauge symmetry for
the fields and indeed even requires the existence of the EM fields to cancel terms in the
391

19. Electrons in an Electromagnetic Field

TOC

Schr
odinger equation. Electromagnetism is called a gauge theory because the gauge
symmetry actually defines the theory. It turns out that the weak and the strong
interactions are also gauge theories and, in some sense, have the next simplest
possible gauge symmetries after the one in Electromagnetism.
We will write our standard gauge transformation in the traditional way to conform
a bit better to the textbooks.
~
A

(~r, t)

~ f
~ (~r, t)
A
1 f (~r, t)
+
c t
e
i ~c
f (~
r ,t)
e
(~r, t)

There are measurable quantum physics consequences of this symmetry. We can

understand a number of them by looking at the vector potential in a field free
~ can be written as the gradient of a function f (~r). To
regions. If B = 0 then A
be specific, take our gauge transformation of the vector potential. Make a gauge
~ 0 = 0. This of course is still consistent with B
~ = 0.
transformation such that A
~0 = A
~ f
~ (~r) = 0
A
Then the old vector potential is then given by
~ = f
~ (~r).
A
~ r).
Integrating this equation, we can write the function f (~r) in terms of A(~
~r

~r
~=
d~r A

~
r0

~ = f (~r) f (r~0 )
d~r f
~
r0

If we choose f so that f (r~0 ) = 0, then we have a very useful relation between the
gauge function and the vector potential in a field free region.
~r
~
d~r A.

f (~r) =
~
r0

We can derive the quantization of magnetic flux by calculating the line integral
~ around a closed loop in a field free region.
of A
=

2n~c
e

A good example of a B = 0 region is a superconductor. Magnetic flux is excluded

from the superconducting region. If we have a superconducting ring, we have a B=0
region surrounding some flux. We have shown then, that the flux going through a ring
of superconductor is quantized.
392

19. Electrons in an Electromagnetic Field

TOC

Flux is observed to be quantized but the charge of the particle seen is 2e.
=

2n~c
2e

This is due to the pairing of electrons inside a superconductor.

The Aharanov B
ohm Effect brings us back to the two slit diffraction experiment
but adds magnetic fields.

393

19. Electrons in an Electromagnetic Field

TOC

The electron beams travel through two slits in field free regions but we have the ability
to vary a magnetic field enclosed by the path of the electrons. At the screen, the
amplitudes from the two slits interfere = 1 + 2 . Lets start with B = 0 and A = 0
everywhere. When we change the B field, the wavefunctions must change.
1

1 e

e
i ~c
e
i ~c

~
d~
r A

2
2 e

e

i ~c
~
d~
r A
e
2
1 ei ~c + 2 e

The relative phase from the two slits depends on the flux between the slits. By varying
the B field, we will shift the diffraction pattern even though B = 0 along the whole
path of the electrons. While this may at first seem amazing, we have seen similar effects
in classical E&M with an EMF induced in a loop by a changing B field which does not
touch the actual loop.

394

19. Electrons in an Electromagnetic Field

19.4
19.4.1

TOC

Examples
The Naive Zeeman Splitting

e ~ ~
The additional term we wish to consider in the Hamiltonian is 2c
B L. Choosing the
z axis so that the constant field points in the z direction, we have

HZeeman =

eBz
Lz .
2c

In general, the addition of a new term to the Hamiltonian will require us to use an
approximation to solve the problem. In this case, however, the energy eigenstates we
derived in the Hydrogen problem are still eigenstates of the full Hamiltonian
H = Hhydrogen + HZeeman . Remember, our hydrogen states are eigenstates of H, L2
and Lz .
(Hhydrogen + HZeeman )n`m = (En + mB B)n`m
This would be a really nice tool to study the number of degenerate states in each
hydrogen level. When the experiment was done, things did not work our according
to plan at all. The magnetic moment of the electron s spin greatly complicates the
problem. We will solve this later.

19.4.2

A Plasma in a Magnetic Field

An important place where both magnetic terms come into play is in a plasma. There,
many electrons are not bound to atoms and external Electric fields are screened out.
Lets assume there is a constant (enough) B field in the z direction. We then have
cylindrical symmetry and will work in the coordinates, , , and z.
~2 2
eB
e2 B 2 2
+
Lz +
(x + y 2 ) = (E + e)
2me
2me c
8me c2
The problem clearly has translational symmetry along the z direction and
rotational symmetry around the z axis. Given the symmetry, we know that Lz
and pz commute with the Hamiltonian and will give constants of the motion. We
therefore will be able to separate variables in the usual way.
(~r) = unmk ()eim eikz
In solving the equation in we may reuse the Hydrogen solution ultimately get the
energies

1 + m + |m|
~2 k 2
eB~
n+
+
E=
me c
2
2m
and associated LaGuerre polynomials (as in Hydrogen) in 2 (instead of r).
395

19. Electrons in an Electromagnetic Field

TOC

The solution turns out to be simpler using the Hamiltonian written in terms
~ if we choose the right gauge by setting A
~ = Bx
of
A
y.
!
2

e ~ 2
1
1
eB
2
2
p~ + A =
x + pz
H =
px + py +
2me
c
2me
c
!

2
1
eB
2eB
2
2
2
2
=
xpy +
x + pz
px + py +
2me
c
c
This Hamiltonian does not depend on y or z and therefore has translational symmetry in both x and y so their conjugate momenta are conserved. We can use this
symmetry to write the solution and reduce to a 1D equation in v(x).
= v(x)eiky y eikz z
Then we actually can use our harmonic oscillator solution instead of hydrogen! The
energies come out to be

eB~
1
~2 k 2
En =
n+
+
.
me c
2
2me
Neglecting the free particle behavior in z, these are called the Landau Levels. This
is an example of the equivalence of the two real problems we know how to solve.

19.5
19.5.1

Derivations and Computations

Deriving Maxwells Equations for the Potentials

We take Maxwells equations and the fields written in terms of the potentials as input.
In the left column the equations are given in the standard form while the right column
gives the equivalent equation in terms of indexed components. The right column uses
the totally antisymmetric tensor in 3D ijk and assumes summation over repeated
indices (Einstein notaton). So in this notation, dot products can be simply written as
~a ~b = ai bi and any component of a cross product is written (~a ~b)k = ai bj ijk .
~ B
~ =0

~ E
~ + 1 B = 0

c t
~ E
~ = 4

~ B
~ 1 E = 4 J~

c t
c
~ =
~ A
~
B
~ =
~ 1 A
E
c t

Bi = 0
xi

1 Bk
Ej ijk +
=0
xi
c t

Ek = 4
xk

1 Ek
4
Bj ijk
=
Jk
xi
c t
c

Bj =
An mnj
xm

1 Ak
Ek =

xk
c t

396

19. Electrons in an Electromagnetic Field

TOC

If the fields are written in terms of potentials, then the first two Maxwell equations
are automatically satisfied. Lets verify the first equation by plugging in the B field in
terms of the potential and noticing that we can interchange the order of differentiation.
~ B
~ = Bi = An mni = An mni

xi
xi xm
xm xi
We could also just interchange the index names i and m then switch those indices
around in the antisymmetric tensor.
~ B
~ =

An inm =
An mni
xm xi
xm xi

~ B
~ = 0.
We have the same expression except for a minus sign which means that
For the second equation, we write it out in terms of the potentials and notice that the

first term x
ijk = 0 for the same reason as above.
i xj

1 Bk
Ej ijk +
xi
c t

=
=
=

1 Aj
1

+
ijk +
An mnk
xi xj
c t
c t xm

An
1
Aj
ijk +
mnk

c
xi t
xm t

Aj
1
Aj
ijk +
ijk = 0

c
xi t
xi t

The last step was simply done by renaming dummy indices (that are summed over) so
the two terms cancel.
Similarly we may work with the Gausss law equation

1 Ak

~
~
Ek =
+
= 4
E =
xk
xk xk
c t
1 ~ ~
2
( A) = 4
c t

For the fourth equation we have.

1 Ek
Bj ijk
=
xi
c t

1

1 Ak
An mnj ijk +
+
=
xi xm
c t xk
c t

4
Jk
c
4
Jk
c

Its easy to derive an identity for the product of two totally antisymmetric tensors
mnj ijk as occurs above. All the indices of any tensor have to be different in order to
397

19. Electrons in an Electromagnetic Field

TOC

get a nonzero result. Since the j occurs in both tensors (and is summed over) we can
simplify things. Take the case that i = 1 and k = 2. We only have a nonzero term
if j = 3 so the other 2 terms in the sum are zero. But if j = 3, we must have either
m = 1 and n = 2 or vice versa. We also must not have i = k since all the indices have
to be different on each epsilon. So we can write.
mnj ijk = mnj kij = (km in kn im )
Applying this identity to the Maxwell equation above, we get.
1 2 Ak
4

1

+ 2
=
Jk
Ai
Ak +
xi xk
xi xi
c xk t
c t2
c
~ ~
1
1 2 Ak
4
A 2 Ak +
+ 2
=
Jk
2
xk
c xk t
c t
c

1 2 Ak
~ A
~ + 1 = 4 Jk
+

2 Ak + 2
2
c t
xk
c t
c

2~
~
~ A
~ + 1 = 4 J~
~+ 1 A +
2 A
c2 t2
c t
c

The last two equations derived are wave equations with source terms obeyed by the
potentials. As discussed in the opening section of this chapter, they can be simplified
with a choice of gauge.

19.5.2

The Lorentz Force from the Classical Hamiltonian

In this section, we wish to verify that the Hamiltonian

1
e ~ 2
p~ + A
e
2m
c

gives the correct Lorentz Force law in classical physics. We will then proceed to use
this Hamiltonian in Quantum Mechanics.
Hamiltons equations are
q
p

H
p
H
=
q
=

where ~q ~r and the conjugate momentum is already identified correctly p~ p~. Remember that these are applied assuming q and p are independent variables.
398

19. Electrons in an Electromagnetic Field

TOC

Beginning with q = H/p, we have

e ~
1
p~ + A
m
c
e~
m~v = p~ + A
c
e~
p~ = m~v A
c
d~r
dt

Note that p~ 6= m~v . The momentum conjugate to ~r includes momentum in the field.
We now time differentiate this equation and write it in terms of the components of a
vector.
dpi
dvi
e dAi
=m

dt
dt
c dt.
H
Similarly for the other Hamilton equation (in each vector component) pi = x
, we
i
have
~
~
e
dpi
e ~ A

= pi =

p~ + A
+e
.
dt
mc
c
xi
xi
i
We now have two equations for dp
dt derived from the two Hamilton equations. We
equate the two right hand sides yielding

mai = m

~
e
e ~ A
dvi

e dAi
=
p~ + A

.
+e
+
dt
mc
c
xi
xi
c dt

~
e
A

e dAi
(m~v )
+e
+
.
mc
xi
xi
c dt
The total time derivative of A has one part from A changing with time and another
from the particle moving and A changing in space.
mai =

~
~
A
dA
~ A
~
=
+ ~v
dt
t
so that

~
e
A

e Ai
e ~
Fi = mai = ~v
+e
+
+
~v Ai .
c xi
xi
c t
c
We notice the electric field term in this equation.

e Ai
+
= eEi
xi
c t
"
#

~
e
A
~
Fi = mai = eEi +
~v
+ ~v Ai .
c
xi
e

Lets work with the other two terms to see if they give us the rest of the Lorentz Force.
"
#

~
e ~
A
e

Aj
e
Ai
Aj
~v Ai ~v
=
vj
Ai v j
= vj

c
xi
c
xj
xi
c
xj
xi
399

19. Electrons in an Electromagnetic Field

TOC

We need only prove that

~ = vj Aj Ai .
~v B
xi
xj
i
To prove this, we will expand the expression using the totally antisymmetric tensor.

An
An
~
~
~v B = ~v A
= vj
mnk jki = vj
(mnk jki )
xm
xm
i
i

Ai
An
An
Aj
= vj
(mnk jik ) = vj
(mj ni mi nj ) = +vj

.
xm
xm
xi
xj
Q.E.D.
So we have

e
~
~v B
c
i
which is the Lorentz force law. So this is the right Hamiltonian for an electron in a
electromagnetic field. We now need to quantize it.
Fi = eEi

19.5.3

The Hamiltonian in terms of B

Start with the Hamiltonian

1
e ~ 2
p~ + A
e
2
c

Now write the Schr

odinger equation.

1 ~~
e~
~~
e~
+ A

+ A
= (E + e)
2 i
c
i
c
ie~ ~ ~ ie~ ~ ~
e2
~2 2

A
A +
A2 = (E + e)
2
2c
2c
2mc2
ie~ ~ ~
ie~ ~ ~
e2
~2 2

A
A +
A2 = (E + e)
2
2c
c
2mc2
~ A
~ = 0, so
The second term vanishes in the Coulomb gauge i.e.,

ie~ ~ ~
e2
~2 2

A +
A2 = (E + e)
2
c
2mc2

Now for constant Bz , we choose the vector potential

~ = 1 ~r B
~
A
2
400

19. Electrons in an Electromagnetic Field

TOC

since

~ A
~

1
Aj ijk =
(xm Bn mnj ) ijk
xi
2 xi
1
1
= im Bn mnj ijk = Bn inj ijk
2
2

XX
XX
1
1
Bn
ijn ijk = Bk
2ijk = Bk
2
2
i
j
i
j

it gives the right field and satisfies the Coulomb gauge condition.
Substituting back, we obtain
2
2
ie~
~2 2
~
~ + e
~ = (E + e)
+
~r B
~r B
2
2
2c
8mc
Now lets work on the vector arithmetic.

~
~
~ L
~
~ ~r
~ =iB
~r B
= ri Bj ijk
= Bj ri
ikj = B
xk
xk
~

2
~
~r B

=
=

ri Bj ijk rm Bn mnk = (ri Bj ri Bj ri Bj rj Bi )

2
~ 0
r2 B 2 ~r B

So, plugging these two equations in, we get

2
e ~ ~
e2
~2 2
2 2
~
+
B L +
r B ~r B
= (E + e) .
2
2c
8mc2
We see that there are two new terms due to the magnetic field. The first one is the
magnetic moment term we have already used and the second will be negligible in atoms.

19.5.4

The Size of the B field Terms in Atoms

In the equation

2
~2 2
e ~ ~
e2
2 2
~
+
B L +
r B ~r B
= (E + e) .
2
2c
8mc2
the second term divided by (e2 /a0 )
e ~ ~ 2
e
B L/(e /a0 )
B (m~) /(e2 /a0 )
2c
2c

eBa0
a2
= m
/ e2 /a0 = m 0 B
2
2e
2
8
0.5 10 cm
B
= mB
=m
(2)(137) (4.8 1010 )
5 109 gauss

401

19. Electrons in an Electromagnetic Field

e2
~c

a0 =

TOC

~
mc

Divide the third term by the second:

2
e2
0.5 108
B 2 a20 8mc
B
a20
2
= 10
= B=
e
10
4e
(4)(137) (4.8 10 )
10 gauss
2c B~

19.5.5

Energy States of Electrons in a Plasma I

~2 2
e2 B 2
eB
Lz +
x2 + y 2 = E
+
2
2me
2me c
8me c

~ field, cylindrical symmetry apply cylindrical coordinates , , z.

For uniform B
Then
2
2
1
1 2
2 = 2 + 2 +
+ 2 2
z

From the symmetry of the problem, we can guess (and verify) that [H, pz ] = [H, Lz ] =
0. These variables will be constants of the motion and we therefore choose
= umk () eim eikz .
~
= m~
Lz =
i
~
pz =
= ~k
i z
2u
1 u im ikz
m2
e
e
2 = k 2 2 + 2 eim eikz +

d2 u 1 du m2
e2 B 2 2
2me E
eBm
2
+

u
+

k
u=0
d2
d
2
4~2 c2
~2
~c
q

2 2
eB
ec
Let x = 2~c
(dummy variable, not the coordinate) and = 4m
E ~2mke 2m.
eB~
Then
d2 u 1 du m2
+
2 u x2 u + = 0
dx2
x dx
x
In the limit x ,
(~r)

d2 u
x2 u = 0
dx2

u ex

while in the other limit x 0,

d2 u 1 du m2
+
2u=0
dx2
x dx
x

402

19. Electrons in an Electromagnetic Field

TOC

Try a solution of the form xs . Then

s(s 1)xs2 + sxs2 m2 xs2 = 0

s2 = m2

A well behaved function s 0 s = |m|

u(x) = x|m| ex

G(x)

Plugging this in, we have

2|m| + 1
dG
d2 G
+
2x
+ ( 2 2|m|) G = 0
dx2
x
dx
We can turn this into the hydrogen equation for
y = x2
and hence
dy = 2x dx
d
1 d
=
.
dy
2x dx
Transforming the equation we get

|m| + 1
dG 2 2|m|
d2 G
+
1
+
G = 0.
dy 2
y
dy
4y
Compare this to the equation we had for hydrogen

dH
1`
d2 H
2` + 2

1
+
H=0
+
d2

with = nr +`+1. The equations are the

same if WE
set our
4me c
eB~

= nr + 1+|m|
2

where nr =

2 2

~ k
2me

0, 1, 2, . . .. Recall that our =

E
2m. This gives us the energy
eigenvalues

~2 k 2
eB~
1 + |m| + m
nr +
.

E
=
2me
me c
2
As in Hydrogen, the eigenfunctions are
G(y) = Ln|m|
(y).
r
We can localize electrons in classical orbits for large E and nr 0. This is the classical
limit.
2
nr = 0

L0 = const

||2 ex x2|m|
Max when

2
2
d||2
= 0 = 2xex x2|m| + 2|m|ex x2|m|1
dx

403

19. Electrons in an Electromagnetic Field

|m| = x2

TOC

1/2
2c
~m
eB

Now lets put in some numbers: Let B 20 kGauss = 2 104 Gauss. Then
s
=

cm (1.05 1027 erg sec)

2 3 1010 sec
m 2.5 106 m cm
10
4
(4.8 10
esu) (2 10 g)

This can be compared to the purely

q classical calculation for an electron with angular
momentum m~ which gives = m~c
Be . This simple calculation neglects to count the
angular momentum stored in the field.

19.5.6

Energy States of Electrons in a Plasma II

We are going to solve the same plasma in a constant B field in a different gauge. If
~ = (0, Bx, 0), then
A
~ =
~ A
~ = Ay z = B z.
B
x
This A gives us the same B field. We can then compute H for a constant B field in
the z direction.
!

2
1
e ~ 2
1
eB
2
2
H =
p~ + A =
px + py +
x + pz
2me
c
2me
c
!

2
1
2eB
eB
2
2
2
2
=
px + py +
xpy +
x + pz
2me
c
c
With this version of the same problem, we have
[H, py ] = [H, pz ] = 0.
We can treat pz and py as constants of the motion and solve the problem in Cartesian
coordinates! The terms in x and py are actually a perfect square.
= v(x)eiky y eikz z

2
d2
eB
~
+
x+
dx2
c

2
~2 d2
1
eB
+
me
x+
2me dx2
2
me c
1
2me

~cky
eB
~cky
eB

2 !
v(x)

v(x)

2 !

~2 kz2
E
v(x)
2me

~2 kz2
E
v(x)
2me
404

19. Electrons in an Electromagnetic Field

This is the same as the 1D harmonic oscillator equation with =

x0 = ~ck
eB .

E=

1
2

~ =

~eB
me c

TOC

eB
me c

and

1
~2 kz2
n+
+
2
2me

So we get the same energies with a much simpler calculation. The resulting
states are somewhat strange and are not analogous to the classical solutions. (Note
that an electron could be circulating about any field line so there are many possible
states, just in case you are worrying about the choice of ky and x0 and counting states.)

19.5.7

A Hamiltonian Invariant Under Wavefunction Phase (or Gauge)

Transformations

We want to investigate what it takes for the Hamiltonian to be invariant under

a local phase transformation of the wave function.
(~r, t) ei(~r,t) (~r, t)
That is, we can change the phase by a different amount at each point in spacetime
and the physics will remain unchanged. We know that the absolute square of the
wavefunction is the same. The Schr
odinger must also be unchanged.

e ~ 2
p~ + A
= (E + e)
c
So lets postulate the following transformation then see what we need to keep
the equation invariant.
(~r, t) ei(~r,t) (~r, t)
~ A
~ + A
~
A

We now need to apply this transformation to the Schr

odinger equation.

2

~~
e~ e ~

+ A
+ A ei(~r,t) = i~ + e + e ei(~r,t)
i
c
c
t

Now we will apply the differential operator to the exponential to identify the
~ i(~r,t) = ei(~r,t) i(~
~ r, t).
new terms. Note that e

2
~~
e~ e ~

(~r, t)
~ r, t)
= ei(~r,t) i~ + e + e ~

ei(~r,t)
+ A
+ A + ~ (~
i
c
c
t
t

2
~~
e~ e ~

(~r, t)
~
+ A + A + ~ (~r, t)
=
i~ + e + e ~

i
c
c
t
t
405

19. Electrons in an Electromagnetic Field

~ ~j = 0.
+

~ =
4. Consider the problem of a charged particle in an external magnetic field B
~
(0, 0, B) with the gauge chosen so that A = (yB, 0, 0). What are the constants
of the motion? Go as far as you can in solving the equations of motion and
obtain the energy spectrum. Can you explain why the same problem in the
~ = (yB/2, xB/2, 0) and A
~ = (0, xB, 0) can represent the same physical
gauges A
situation? Why do the solutions look so different?
5. Calculate the top left 4 4 corner of the matrix representation of x4 for the
harmonic oscillator. Use the energy eigenstates as the basis states.
6. The Hamiltonian for an electron in a electromagnetic field can be written as
1
~ , t)
~ r, t)]2 e(~r, t) + e~ ~ B(~
H = 2m
[~
p + ec A(~
r. Show that this can be written
2mc
as the Pauli Hamiltonian
2
1
e~
H=
~ [~
p + A(~
r, t)] e(~r, t).
2m
c

19.7

Sample Test Problems

1. A charged particle is in an external magnetic field. The vector potential is given

by A = (yB, 0, 0). What are the constants of the motion? Prove that these are
constants by evaluating their commutator with the Hamiltonian.
2. A charged particle is in an external magnetic field. The vector potential is given
by A = (0, xB, 0). What are the constants of the motion? Prove that these are
constants by evaluating their commutator with the Hamiltonian.
3. Gauge symmetry was noticed in electromagnetism before the advent of Quantum
Mechanics. What is the symmetry transformation for the wave function of an
electron from which the gauge symmetry for EM can be derived?

p
408

20. Addition of Angular Momentum

TOC

Addition of Angular Momentum

Since total angular momentum is conserved in nature, we will find that eigenstates of
the total angular momentum operator are usually energy eigenstates. The exceptions
will be when we apply external Fields which break the rotational symmetry. We must
therefore learn how to add different components of angular momentum together. One
of our first uses of this will be to add the orbital angular momentum in Hydrogen to
the spin angular momentum of the electron.
~ +S
~
J~ = L

audio
Our results can be applied to the addition of all types of angular momentum.
This material is covered in Gasiorowicz Chapter 15, in Cohen-Tannoudji et al.
Chapter X and very briefly in Griffiths Chapter 6.

20.1

Adding the Spins of Two Electrons

The coordinates of two particles commute with each other: [p(1)i , x(2)j ] = 0. They
are independent variables except that the overall wave functions for identical particles
must satisfy the (anti)symmetrization requirements. This will also be the case for the
spin coordinates.
[S(1)i , S(2)j ] = 0
We define the total spin operators
~=S
~(1) + S
~(2) .
S
Its easy to show the total spin operators obey the same commutation relations as individual spin operators
[Si , Sj ] = i~ijk Sk .
audio
This is a very important result since we derived everything about angular momentum
from the commutators. The sum of angular momentum will be quantized in the same
way as orbital angular momentum.
409

20. Addition of Angular Momentum

TOC

As with the combination of independent spatial coordinates, we can make product

states to describe the spins of two particles. These products just mean, for example,
the spin of particle 1 is up and the spin of particle 2 is down. There are four possible
(product) spin states when we combine two spin 12 particles. These product states
are eigenstates of total Sz but not necessarily of total S 2 . The states and their Sz
eigenvalues are given below.
Product State
(1) (2)
+ +
(1) (2)
+
(1) (2)
+
(1) (2)

Total Sz eigenvalue
~
0
0
~

audio
Verify the quoted eigenvalues by calculation using the operator Sz = S(1)z +
S(2)z .
We expect to be able to form eigenstates of S 2 from linear combinations of these
four states. From pure counting of the number of states for each Sz eigenvalue, we can
guess that we can make one s = 1 multiplet plus one s = 0 multiplet. The s = 1
multiplet has three component states, two of which are obvious from the list above.
We can use the lowering operator to derive the other eigenstates of S 2 .
s=1,m=1
s=1,m=0
s=1,m=1
s=0,m=0

(1) (2)

= + +

1 (1) (2)
(1) (2)
= + + +
2
(1) (2)

=

1 (1) (2)
(1) (2)
= + +
2

audio
As a necessary check, we operate on these states with S 2 and verify that they are
indeed the correct eigenstates.
Note that by deciding to add the spins together, we could not change the nature of
2
the electrons. They are still spin 21 and hence, these are all still eigenstates of S(1)
and
2
S(2) , however, (some of) the above states are not eigenstates of S(1)z and S(2)z . This
will prove to be a general feature of adding angular momenta. Our states of definite
total angular momentum and z component of total angular momentum will still also
be eigenstates of the individual angular momenta squared.
410

20. Addition of Angular Momentum

20.2

TOC

Total Angular Momentum and The Spin Orbit Interaction

The spin-orbit interaction (between magnetic dipoles) will play a role in the fine
structure of Hydrogen as well as in other problems. It is a good example of the need
for states of total angular momentum. The additional term in the Hamiltonian is
HSO =

~ S
~
Ze2 L
2m2 c2 r3

~ S
~
If we define the total angular momentum J~ in the obvious way we can write L
in terms of quantum numbers.
audio
~ +S
~
J~ = L
J2
~ S
~
L

~ S
~ + S2
= L2 + 2L
1 2
~2
=
(J L2 S 2 )
(j(j + 1) `(` + 1) s(s + 1))
2
2

Since our eigenstates of J 2 and Jz are also eigenstates of L2 and S 2 (but not Lz or
Sz ), these are ideal for computing the spin orbit interaction. In fact, they are going to
be the true energy eigenstates, as rotational symmetry tells us they must.

20.3

Adding Spin

1
2

to Integer Orbital Angular Momentum

Our goal is to add orbital angular momentum with quantum number ` to spin 21 . We
can show in several ways that, for ` 6= 0, that the total angular momentum quantum
number has two possible values j = ` + 12 or j = ` 12 . For ` = 0, only j = 12 is
allowed. First lets argue that this makes sense when we are adding two vectors. For
example if we add a vector of length 3 to a vector of length 0.5, the resulting vector
could take on a length between 2.5 and 3.5 For quantized angular momentum, we will
only have the half integers allowed, rather than a continuous range. Also we know that
the quantum numbers like ` are not exactly the length of the vector but are close. So
these two values make sense physically.
We can also count states for each eigenvalue of Jz as in the following examples.
Example: Counting states for ` = 3 plus spin 12 .
Example: Counting states for any ` plus spin 12 .
As in the last section, we could start with the highest Jz state, Y`` + , and apply the
lowering operator to find the rest of the multiplet with j = ` + 12 . This works well
for some specific ` but is hard to generalize.
411

20. Addition of Angular Momentum

TOC

We can work the problem in general. We know that each eigenstate of J 2 and Jz will
be a linear combination of the two product states with the right m.
j(m+ 12 ) = Y`m + + Y`(m+1)
audio
The coefficients and must be determined by operating with J 2 .
r
r
`+m+1
`m
Y`m + +
Y`(m+1)
(`+ 12 )(m+ 12 ) =
2` + 1
2` + 1
r
r
`m
`+m+1
(` 12 )(m+ 12 ) =
Y`m +
Y`(m+1)
2` + 1
2` + 1
We have made a choice in how to write these equations: m must be the same throughout. The negative m states are symmetric with the positive ones. These equations will
be applied when we calculate the fine structure of Hydrogen and when we study
the anomalous Zeeman effect.

20.4

Spectroscopic Notation

A common way to name states in atomic physics is to use spectroscopic notation.

It is essentially a standard way to write down the angular momementum quantum
numbers of a state. The general form is N 2s+1 Lj , where N is the principal quantum
number and will often be omitted, s is the total spin quantum number ((2s + 1) is the
number of spin states), L refers to the orbital angular momentum quantum number
` but is written as S, P, D, F, . . . for ` = 0, 1, 2, 3, . . . , and j is the total angular
momentum quantum number.
A quick example is the single electron states, as we find in Hydrogen. These are:
1 2 S 21 2 2 S 12 2 2 P 32 2 2 P 21 3 2 S 12 3 2 P 23 3 2 P 12 3 2 D 25
3 2 D 32 4 2 S 21 4 2 P 23 4 2 P 12 4 2 D 52 4 2 D 32 4 2 F 72 4 2 F 52

...

All of these have the pre-superscript 2 because they are all spin one-half. There are
two j values for each `.
For atoms with more than one electron, the total spin state has more possibilities and
perhaps several ways to make a state with the same quantum numbers.

412

20. Addition of Angular Momentum

20.5

TOC

General Addition of Angular Momentum: The ClebschGordan Series

We have already worked several examples of addition of angular momentum. Lets work
one more.
Example: Adding ` = 4 to ` = 2.

The result, in agreement with our classical vector model, is multiplets with j =
2, 3, 4, 5, 6.
The vector model qualitatively explains the limits.
audio

In general, j takes on every value between the maximum an minimum in

integer steps.
|`1 `2 | j `1 + `2
The maximum and minimum lengths of the sum of the vectors makes sense physically.
Quantum Mechanics tells up that the result is quantized and that, because of the
uncertainty principle, the two vectors can never quite achieve the maximum allowed
classically. Just like the z component of one vector can never be as great as the full
vector length in QM.
We can check that the number of states agrees with the number of product
states.
We have been expanding the states of definite total angular momentum j in terms of the
product states for several cases. The general expansion is called the Clebsch-Gordan
series:
X
jm =
h`1 m1 `2 m2 |jm`1 `2 iY`1 m1 Y`2 m2
m1 m2

413

20. Addition of Angular Momentum

TOC

or in terms of the ket vectors

|jm`1 `2 i =

h`1 m1 `2 m2 |jm`1 `2 i|`1 m1 `2 m2 i

m1 m2

The Clebsch-Gordan coefficients are tabulated. We have computed some of them here
by using the lowering operator and some by making eigenstates of J 2 .

20.6

Interchange Symmetry for States with Identical Particles

If we are combining the angular momentum from two identical particles, like two electrons in an atom, we will be interested in the symmetry under interchange of the
angular momentum state. Lets use the combination of two spin 12 particles as an example. We know that we get total spin states of s = 1 and s = 0. The s = 1 state is
called a triplet because there are three states with different m values. The s = 0 state
is called a singlet. The triplet state is symmetric under interchange. The highest
total angular momentum state, s = s1 + s2 , will always be symmetric under
interchange. We can see this by looking at the highest m state, m = s. To get the
maximum m, both spins to have the maximum z component. So the product state has
just one term and it is symmetric under interchange, in this case,
(1) (2)

11 = + + .
When we lower this state with the (symmetric) lowering operator S = S(1) + S(2) ,
the result remains symmetric under interchange. To make the next highest state,
with two terms, we must choose a state orthogonal to the symmetric state and this will
always be antisymmetric.
In fact, for identical particles, the symmetry of the angular momentum wave
function will alternate, beginning with a symmetric state for the maximum total
angular momentum. For example, if we add two spin 2 states together, the resulting
states are: 4S , 3A , 2S , 1A and 0S . In the language of group theory, when we take
the direct product of two representations of the the SU(2) group we get:
5 5 = 9S 7A 5S 3A 1S
where the numbers are the number of states in the multiplet.
Example: Two electrons in a P state.
Example: The parity of the pion from d nn.

414

20. Addition of Angular Momentum

20.7
20.7.1

TOC

Examples
Counting states for ` = 3 Plus spin

1
2

For ` = 3 there are 2` + 1 = 7 different eigenstates of Lz . There are two different

eigenstates of Sz for spin 12 . We can have any combination of these states, implying
2 7 = 14 possible product states like Y31 + .
We will argue based on adding vectors... that there will be two total angular momentum
states that can be made up from the 14 product states, j = ` 12 , in this case j = 25
and j = 72 . Each of these has 2j + 1 states, that is 6 and 8 states respectively. Since
6 + 8 = 14 this gives us the right number of states.

20.7.2

Counting states for Arbitrary ` Plus spin

1
2

For angular momentum quantum number `, there are (2` + 1) different m states, while
for spin we have 2 states . Hence the composite system has 2(2` + 1) states total.
Max jz = ` + 21 so we have a state with j = ` + 12 . This makes up (2j + 1) = (2` + 2)
states, leaving

1
(4` + 2) (2` + 2) = 2` = 2 `
+1
2
Thus we have a state with j = `

20.7.3

1
2

and thats all.

Adding ` = 4 to ` = 2

As an example, we count the states for each value of total m (z component quantum
number) if we add `1 = 4 to `2 = 2.
audio

415

20. Addition of Angular Momentum

Total m
6
5
4
3
2
1
0
-1
-2
-3
-4
-5
-6

TOC

(m1 , m2 )
(4,2)
(3,2) (4,1)
(2,2) (3,1) (4,0)
(1,2) (2,1) (3,0) (4,-1)
(0,2) (1,1) (2,0) (3,-1) (4,-2)
(-1,2) (0,1) (1,0) (2,-1) (3,-2)
(-2,2) (-1,1) (0,0) (1,-1) (2,-2)
(1,-2) (0,-1) (-1,0) (-2,1) (-3,2)
(0,-2) (-1,-1) (-2,0) (-3,1) (-4,2)
(-1,-2) (-2,-1) (-3,0) (-4,1)
(-2,-2) (-3,-1) (-4,0)
(-3,-2) (-4,-1)
(-4,-2)

Since the highest m value is 6, we expect to have a j = 6 state which uses up one state
for each m value from -6 to +6. Now the highest m value left is 5, so a j = 5 states
uses up a state at each m value between -5 and +5. Similarly we find a j = 4, j = 3,
and j = 2 state. This uses up all the states, and uses up the states at each value of m.
So we find in this case,
|`1 `2 | j |`1 + `2 |
and that j takes on every integer value between the limits. This makes sense in the
vector model.

20.7.4

Two electrons in an atomic P state

If we have two atomic electrons in a P state with no external fields applied, states
of definite total angular momentum will be the energy eigenstates. We will
learn later that closed shells in atoms (or nuclei) have a total angular momentum of
zero, allowing us to treat only the valence electrons. Examples of atoms like this would
be Carbon, Silicon, and Germanium.
Our two electrons each have ell = 1 (P state) and s =
four angluar momenta together to get the total.

1
2

(electrons). We need to add

J~ = L~1 + L~2 + S~1 + S~2

We will find it useful to do this addition in two steps. For low Z atoms, it is most
~ and S~1 + S~2 = S
~ then to add these results L
~ +S
~ = J.
~
useful to add L~1 + L~2 = L
Since the electrons are identical particles and they are in the same radial state, the
angular momentum part of the wavefunction must be antisymmetric under interchange.
416

20. Addition of Angular Momentum

TOC

This will limit the allowed states. So lets do the spinor arithmetic.
|`1 `2 |

`1 + `2

` =

0, 1, 2

s =

0, 1

These states have a definite symmetry under interchange. Before going on to make the
total angular momentum states, lets note the symmetry of each of the above states.
The maximum allowed state will always need to be symmetric in order to achieve the
maximum. The symmetry will alternate as we go down in the quantum number. So,
for example, the ` = 2 and ` = 0 states are symmetric, while the ` = 1 state is
antisymmetric. The s = 1 state is symmetric and the s = 0 state is antisymmetric.
The overall symmetry of a state will be a product of the these two symmetries (since
when we add ` and s to give j we are not adding identical things anymore). The overall
state must be antisymmetic so we can use:
` =

1 s = 1 j = 0, 1, 2

P0 , 3 P1 , 3 P2

` =

2 s=0 j=2

` =

0 s=0 j=0

Each atomic state will have the angular momentum quantum numbers
`1 , `2 , s1 , s2 , `, s, j, m.
Normally we will not bother to include that the spins are one half since thats always
true for electrons. We will (and must) keep track of the intermediate ` and s quantum
numbers. As can be seen above, we need them to identify the states.
In the atomic physics section, we will even deal with more than two electrons outside
a closed shell.

20.7.5

The parity of the pion from d nn.

audio
We can determine the internal parity of the pion by studying pion capture by a
deuteron, + d n + n. The pion is known to have spin 0, the deuteron spin 1, and
the neutron spin 21 . The internal parity of the deuteron is +1. The pion is captured by
the deuteron from a 1S states, implying ` = 0 in the initial state. So the total angular
momentum quantum number of the initial state is j = 1.
417

20. Addition of Angular Momentum

TOC

So the parity of the initial state is

(1)` P Pd = (1)0 P Pd = P
The parity of the final state is
Pn Pn (1)` = (1)`
Therefore,
P = (1)` .
Because the neutrons are identical fermions, the allowed states of two neutrons are 1 S0 ,
P0,1,2 , 1 D2 , 3 F2,3,4 ... The only state with j = 1 is the 3 P1 state, so ` = 1

TOC

(1) (2)

Now we want to identify 11 = + + . Lets operate on this equation with S . First

the RHS gives

(1) (2)
(1) (1)
(2)
(1)
(2) (2)
(1) (2)
(1) (2)
S + + = S + + + + S + = ~ + + + .
Operating on the LHS gives
p

S 11 = ~ (1)(2) (1)(0)10 = 2~10 .

So equating the two we have

(1) (2)
(1) (2)
2~10 = ~ + + + .

1 (1) (2)
(1) (2)
10 = + + + .
2
Now we can lower this state. Lowering the LHS, we get
p

S 10 = ~ (1)(2) (0)(1)1(1) = 2~1,1 .

Lowering the RHS, gives

1 (1) (2)
1 (1) (2)
(1) (2)
(1) (2)
(1) (2)
S + + + = ~ + = 2~
2
2

(1) (2)

1,1 =

Therefore we have found 3 s=1 states that work together. They are all symmetric
under interchange of the two particles.
There is one state left over which is orthogonal to the three states we identified. Orthogonal state:

1 (1) (2)
(1) (2)
00 = + +
2
We have guessed that this is an s = 0 state since there is only one state and it has
m=0. We could verify this by using the S 2 operator.

20.8.3

Applying the S 2 Operator to 1m and 00 .

We wish to verify that the states we have deduced are really eigenstates of the S 2
operator. We will really compute this in the most brute force.

2
~1 + S
~2 = S12 + S22 + 2S
~1 S
~2
S
S2 =
(1) (2)

S 2 + +

=
=

(1) (2)
(1) (2)
~1 (1) S
~2 (2)
s1 (s1 + 1)~2 + + + s2 (s2 + 1)~2 + + + 2S
+
+

3 2 (1) (2)
(1) (2)
(1) (2)
(1) (2)
(1) (2)
~ + + + 2 Sx Sx + Sy Sy + Sz Sz
+ +
2

419

20. Addition of Angular Momentum

(1) (2)
S 2 + +

Sx +

Sy +

=
=
=

~ 0
2 1

~ 0
2 i

~ 0
2 1

~ 0
2 i

TOC

~
~ 0
1
1
=
=
0
0
2 1
2

~
~
i
1
0
=
= i
0
0
2 i
2

~ 1
~
1
0
=
= +
0
1
2 0
2

~ i
~
i
0
=
= i +
0
1
0
2
2

3 2 (1) (2) ~2
~ + + +
2
2

1
1
0
0
0
0
+
+
0 1 0 2
i 1 i 2
1 1 1 2

3 2 (1) (2) ~2
1
1
0
0
0
0
+

~ + + +
0
0 2
1
1
1
1
2
2
1
1
2
1
2
3 2 (1) (2) ~2 (1) (2)
(1) (2)
~ + + + + + = 2~2 + +
2
2

Note that s(s + 1) = 2, so that the 2~ is what we expected to get. This confirms that
we have an s=1 state.
Now lets do the 00 state.

~1 S
~2 00
S 2 00 =
S12 + S22 + 2S

=
S12 + S22 + 2 Sx(1) Sx(2) + Sy(1) Sy(2) + Sz(1) Sz(2) 00

=
S12 + S22 + 2Sz(1) Sz(2) + 2 Sx(1) Sx(2) + Sy(1) Sy(2) 00

3 3
1
=
+ 2
~2 00 + 2 Sx(1) Sx(2) + Sy(1) Sy(2) 00
4 4
4

1
(1) (2)
(1) (2)
= ~2 00 + 2 Sx(1) Sx(2) + Sy(1) Sy(2) + +
2

2
~
1
(2)
(1) (2)
(1) (2)
(1) (2)
(1)
= ~2 00 +

+
i(i)

(i)i

+
+

2 2

1
(1) (2)
(1) (2)
(1) (2)
(1) (2)
= ~2 00 + + + + +
2 2
= ~2 (1 1) 00 = 0~2 00

420

20. Addition of Angular Momentum

20.8.4

Adding any ` plus spin

TOC

1
2.

We wish to write the states of total angular momentum j in terms of the product states
Y`m . We will do this by operating with the J 2 operator and setting the coefficients
so that we have eigenstates.
J 2 jmj = j(j + 1)~2 jmj
We choose to write the the quantum number mj as m + 21 . This is really just the
defintion of the dummy variable m. (Other choices would have been possible.)
The z component of the total angular momentum is just the sum of the z components
from the orbital and the spin.
mj = ml + ms
There are only two product states which have the right mj = m + 12 . If the spin is up
we need Y`m and if the spin is down, Y`(m+1) .
j(m+ 12 ) = Y`m + + Y`(m+1)
audio
We will find the coefficients and so that will be an eigenstate of
~ + S)
~ 2 = L2 + S 2 + 2Lz Sz + L+ S + L S+ .
J 2 = (L
So operate on the right hand side with J 2 .

1
3
J 2 j,m+ 12 = ~2 `(` + 1)Ylm + + Ylm + + 2m Ylm +
4
2
i
p

+ `(` + 1) m(m + 1) 1Yl(m+1)

3
1
2
+ ~ `(` + 1)Y`,m+1 + Y`,m+1 + 2(m + 1)
Y`,m+1
4
2
i
p
p
+ `(` + 1) (m + 1)m (1)Ylm +
And operate on the left hand side.
J 2 j,m+ 12 = j(j + 1)~2 j,m+ 12 = j(j + 1)~2 Ylm + + Y`,(m+1)

Since the two terms are orthogonal, we can equate the coefficients for each term, giving
us two equations. The Y`m + term gives

p
3
j(j + 1) = `(` + 1) + + m + `(` + 1) m(m + 1).
4
421

20. Addition of Angular Momentum

TOC

The Y`(m+1) term gives

p
3
j(j + 1) = `(` + 1) + (m + 1) + `(` + 1) m(m + 1).
4
Collecting terms on the LHS and terms on the RHS, we get two equations.

p
3
j(j + 1) `(` + 1) m =
(` m)(` + m + 1)
4

p
3
(` m)(` + m + 1) =
j(j + 1) `(` + 1) + (m + 1)
4
Now we just cross multiply so we have one equation with a common factor of .

3
3
(`m)(`+m+1) = j(j + 1) `(` + 1) m j(j + 1) `(` + 1) + (m + 1)
4
4
While this equation looks like a mess to solve, if we notice the similarity between the
LHS and RHS, we can solve it if
3
` = j(j + 1) `(` + 1) .
4
If we look a little more carefully at the LHS, we can see that another solution (which
just interchanges the two terms in parentheses) is to replace ` by ` 1.
3
` 1 = j(j + 1) `(` + 1) .
4
These are now simple to solve
j(j + 1) = `(` + 1) + ` +

j =`+

1
2

3
1

j =`
4
2
So these are (again) the two possible values for j. We now need to go ahead and find
and .
j(j + 1) = `(` + 1) ` 1 +

Plugging j = ` +

1
2

into our first equation,

p
(` m) = (` m)(` + m + 1)

we get the ratio between and . We will normalize the wave function by setting
2 + 2 = 1. So lets get the squares.
2 =

(` m)2
(` m)
2 =
2
(` m)(` + m + 1)
(` + m + 1)

422

20. Addition of Angular Momentum

TOC

`+m+1+`m 2
=1
`+m+1
r
`+m+1
=
2` + 1
r
r
r
`m
`+m+1
`m
=
=
`+m+1
2` + 1
2` + 1

2 + 2 = 1

So we have completed the calculation of the coefficients. We will make use of these in
the hydrogen atom, particularly for the anomalous Zeeman effect.
Writing this in the notation of matrix elements or Clebsch-Gordan coefficients
of the form,
hjmj `s|`m` sms i
we get.
audio

1
1
1 11
`+
m+
` `m
2
2
2
22

1 1
1
1
1
`+
m+
` `(m + 1)
2
2
2
2 2

1
1
1 1 1
`+
m+
` `m
2
2
2
2 2

1
1
1
11
m+
` `(m + 1)
`+
2
2
2
22

`+m+1
2` + 1

`m
2` + 1

= =
= =
=

Similarly

1
1
1 11
`
m+
` `m
2
2
2
22

1
1
1
1 1
`
m+
` `(m + 1)
2
2
2
2 2

20.8.5

`m
2` + 1
r
`+m+1
=
2` + 1
=

Counting the States for |`1 `2 | j `1 + `2 .

If we add `1 to `2 there are (2`1 + 1)(2`2 + 1) product states. Lets add up the number
of states of total `. To keep things simple we assume we ordered things so `1 `2 .
`X
1 +`2
`=`1 `2

(2` + 1) =

2`2
X

(2(`1 `2 + n) + 1) = (2`2 + 1)(2`1 2`2 + 1) + 2

n=0

2`2
X
n=0

n
423

20. Addition of Angular Momentum

TOC

= (2`2 + 1)(2`1 2`2 + 1) + (2`2 + 1)(2`2 ) = (2`2 + 1)(2`1 + 1)

This is what we expect.

20.9

Homework Problems

1. Find the allowed total spin states of two spin 1 particles. Explicitly write out the
9 states which are eigenfunctions of S 2 and Sz .
~

2. The Hamiltonian of a spin system is given by H = A + B S~12S2 + C(S1z~+S2z ) . Find

the eigenvalues and eigenfunctions of the system of two particles (a) when both
particles have spin 12 , (b) when one particle has spin 12 and the other spin 1.
What happens in (a) when the two particles are identical?
3. Consider a system of two spinless identical particles. Show that the orbital
angular momentum of their relative motion can only be even. (l = 0, 2, 4, ...)
Show by direct calculation that, for the triplet spin states of two spin 21 particles, ~1 ~2 1m = 1m for all allowed m. Show that for the singlet state
~1 ~2 00 = 300 .
4. A deuteron has spin 1. What are the possible spin and total angular momentum
states of two deuterons. Include orbital angular momentum and assume the two
particles are identical.
q
q
5. The state of an electron is given by = R(r)[ 13 Y10 (, )+ + 23 Y11 (, ) ].
Find the possible values and the probabilities of the z component of the electrons
total angular momentum. Do the same for the total angular momentum squared.
What is the probability density for finding an electron with spin up at r, , ?
What is it for spin down? What is the probability density independent of spin?
(Do not leave your answer in terms of spherical harmonics.)
6. The n = 2 states of hydrogen have an 8-fold degeneracy due to the various l
and m states allowed and the two spin states of the electron. The spin orbit
interaction partially breaks the degeneracy by adding a term to the Hamiltonian
2
~ ~
H1 = 2mAe
2 c2 r 3 L S. Use first order perturbation theory to find how the degeneracy
is broken under the full Hamiltonian and write the approximate energy eigenstates
in terms of Rnl , Ylm , and .
7. The nucleus of a deuterium (A=2 isotope of H) atom is found to have spin 1.
With a neutral atom, we have three angular momenta to add, the nuclear spin,
~ +S
~ in the
the electron spin, and the orbital angular momentum. Define J~ = L
~
~
~
usual way and F = J + I where I denotes the nuclear spin operator. What are
the possible quantum numbers j and f for an atom in the ground state? What
are the possible quantum numbers for an atom in the 2p state?

424

X
= N n +
cnk k
k6=n

cnk

(1)
cnk

(2)
cnk

+ ...
427

21. Time Independent Perturbation Theory

TOC

En(2)

= hn |H1 |n i
hk |H1 |n i
=
(0)
(0)
En Ek
X |hk |H1 |n i|2
=
(0)
(0)
k6=n En Ek

audio
(1)

So the first order correction to the energy of the nth eigenstate, En , is just
the expectation value of the perturbation in the unperturbed state. The first order
(1)
admixture of k in n , cnk , depends on a matrix element and the energy difference
(2)
between states. The second order correction to the energy, En , has a similar
dependence. Note that the higher order corrections may not be small if states are
nearby in energy.
The application of the first order perturbation equations is quite simple in principal.
The actual calculation of the matrix elements depends greatly on the problem being
solved.
Example: H.O. with anharmonic perturbation (ax4 ).

Sometimes the first order correction to the energy is zero. Then we will need to use
(2)
the second order term En to estimate the correction. This is true when we apply an
electric field to a hydrogen atom.
Example: Hydrogen Atom in a E-field, the Stark Effect.

We will exercise the use of perturbation theory in section 22 when we compute the fine
structure, and other effects in Hydrogen.

21.2

Degenerate State Perturbation Theory

The perturbation expansion has a problem for states very close in energy. The energy
difference in the denominators goes to zero and the corrections are no longer small.
428

21. Time Independent Perturbation Theory

TOC

The series does not converge. We can very effectively solve this problem by treating
all the (nearly) degenerate states like we did n in the regular perturbation
expansion. That is, the zeroth order state will be allowed to be an arbitrary linear
combination of the degenerate states and the eigenvalue problem will be solved.
Assume that two or more states are (nearly) degenerate. Define N to be the set of
those nearly degenerate states. Choose a set of basis state in N which are orthonormal
h(j) |(i) i = ji
where i and j are in the set N . We will use the indices i and j to label the states in
N.
By looking at the zeroth and first order terms in the Schrodinger equation and dotting
it into one of the degenerate states (j) , we derive the energy equation for first
order (nearly) degenerate state perturbation theory
X
h(j) |H0 + H1 |(i) ii = Ej ,
iN

This is an eigenvalue equation with as many solutions as there are degnerate states in
our set. audio
We recognize this as simply the (matrix) energy eigenvalue equation limited the list
of degenerate states. We solve the equation to get the energy eigenvalues and energy
eigenstates, correct to first order.
Written as a matrix, the equation

H11 H12
H21 H22

...
...
Hn1 Hn2

is
...
...
...
...

H1n
1
1
2
2
H2n
= E
...
... ...
Hnn
n
n

where Hji = h(j) |H0 + H1 |(i) i is the matrix element of the full Hamiltonian. If there
are n nearly degenerate states, there are n solutions to this equation.
The Stark effect for the (principle quantum number) n=2 states of hydrogen requires
the use of degenerate state perturbation theory since there are four states with (nearly)
the same energies. For our first calculation, we will ignore the hydrogen fine structure
and assume that the four states are exactly degenerate, each with unperturbed energy
of E0 . That is H0 2`m = E0 2`m . The degenerate states 200 , 211 , 210 , and 21(1) .
Example: The Stark Effect for n=2 States.

The perturbation due to an electric field in the z direction is H1 = +eEz. The linear
combinations that are found to diagonalize the full Hamiltonian in the subspace of
429

21. Time Independent Perturbation Theory

degenerate states are: 211 , 21(1) and

E2 3eEa0 .

21.3
21.3.1

1
2

TOC

(200 210 ) with energies of E2 , E2 , and

Examples
H.O. with anharmonic perturbation (ax4 ).

We add an anharmonic perturbation to the Harmonic Oscillator problem.

H1 = ax4
Since this is a symmetric perturbation we expect that it will give a nonzero result in
first order perturbation theory. First, write x in terms of A and A and compute the
expectation value as we have done before.
En(1)

=
=
=
=

21.3.2

ahn|x4 |ni =

a~2
hn|(A + A )4 |ni
4m2 2

a~2
hn|(AAA A + AA AA + AA A A + A AAA + A AA A + A A AA)|n
4m2 2

a~2
(n + 1)(n + 2) + (n + 1)2 + n(n + 1) + n(n + 1) + n2 + n(n 1)
2
2
4m
3a~2
(2n2 + 2n + 1)
4m2 2

Hydrogen Atom Ground State in a E-field, the Stark Effect.

We have solved the Hydrogen problem with the following Hamiltonian.

H0 =

Ze2
p2

2
r

Now we want to find the correction to that solution if an Electric field is applied
to the atom. We choose the axes so that the Electric field is in the z direction. The
perturbtion is then.
H1 = eEz
It is typically a small perturbation. For non-degenerate states, the first order
correction to the energy is zero because the expectation value of z is an odd function.
(1)

Enlm = eEhnlm |z|nlm i = 0

430

21. Time Independent Perturbation Theory

TOC

We therefore need to calculate the second order correction. This involves a sum
over all the other states.
2

(2)

E100 = e2 E 2

|hnlm |z|100 i|

nlm6=100

E1 En

(0)

We need to compute all the matrix elements of z between the ground state and the
other Hydrogen states.

hnlm |z|100 i = d3 rRnl

(r cos )R10 Ylm
Y00

We can do the angular integral by converting the cos term into a spherical harmonic.
r
1
4
Y00 cos =
Y10
3
4
The we can just use the orthonormality of the spherical harmonics to do the angular
integral, leaving us with a radial integral to do.

hnlm |z|100 i =
r3 drRnl
R10 dYlm
Y10
3

`1 m0

=
r3 Rnl
R10 dr
3
The radial part of the integral can be done with some work, yielding.
2n5

|hnlm |z|100 i| =

1 28 n7 (n 1)
3 (n + 1)2n+5

a20 `0 m0 f (n)a20 `0 m0

We put this back into the sum. The Kronecker deltas eliminate the sums over ` and
m. We write the energy denominators in terms of the Bohr radius.
(2)

E100

= e2 E 2
= a30 E 2

f (n)a20
e2
2a0 + 2a0 n2

e2
n=2

2f (n)
1 + n12
n=2

= 2a30 E 2

X
n2 f (n)
n2 1
n=2

431

21. Time Independent Perturbation Theory

TOC

This is all a little dissatisfying because we had to insert the general formula for the
radial integral and it just goes into a nasty sum. In fact, we could just start with the
first few states to get a good idea of the size of the effect. The result comes out to be.
(2)

E100 = 2a30 E 2 (0.74 + 0.10 + . . . ) = 2.25a30 E 2

The first two terms of the sum get us pretty close to the right answer. We could have
just done those radial integrals.
Now we compute d, the electric dipole moment of the atom which is induced by
the electric field.
E
= 4(1.125)a30 E
d=
E
The dipole moment is proportional to the Electric field, indicating that it is induced.
The E field induces the dipole moment, then the dipole moment interacts with the E
field causing a energy shift. This indicates why the energy shift is second order.

21.3.3

The Stark Effect for n=2 Hydrogen.

The Stark effect for the n=2 states of hydrogen requires the use of degenerate state
perturbation theory since there are four states with (nearly) the same energies. For our
first calculation, we will ignore the hydrogen fine structure and assume that the four
states are exactly degenerate, each with unperturbed energy of E0 . That is H0 2`m =
E0 2`m . The degenerate states 200 , 211 , 210 , and 21(1) .
The perturbation due to an electric field in the z direction is H1 = +eEz. So our first
order degenerate state perturbation theory equation is
E
X D
i (j) |H0 + eEz| (i) = (E0 + E (1) )j .
i

This is esentially a 4X4 matrix eigenvalue equation. There are 4 eigenvalues (E0 +E (1) ),
distinguished by the index n.
Because of the exact degeneracy (H0 (j) = E0 (j) ), H0 and E0 can be eliminated from
the equation.
E
D
X
i (E0 ij + (j) |eEz| (i) ) = (E0 + E (1) )j
i

E0 j +

D
E
i (j) |eEz| (i)

= E0 j + E (1) j

D
E
i (j) |eEz| (i)

= E (1) j

X
i

432

21. Time Independent Perturbation Theory

TOC

This is just the eigenvalue equation for H1 which we can write out in (pseudo)matrix
form

1
1
2
2
(1)
H1 = E
3
3
4
4
Now, in fact, most of the matrix elements of H1 are zero. We will show that because
[Lz , z] = 0, that all the matrix elements between states of unequal m are zero. Another
way of saying this is that the operator z doesnt change m. Here is a little proof.
hYlm |[Lz , z]| Yl0 m0 i = 0 = (m m0 ) hYlm |z| Yl0 m0 i
This implies that hYlm |z| Yl0 m0 i = 0 unless m = m0 .
Lets define the one remaining nonzero (real) matrix element to be .
= eE h200 |z| 210 i
The equation (labeled with the basis states to define the order) is.

200
1
0 0 0
1

211
0 0 0 0 2 = E (1) 2
3
210 0 0 0 3
211 0 0 0 0
4
4
We can see by inspection that the eigenfunctions of this operator are 211 , 211 , and
1 (200 210 ) with eigenvalues (of H1 ) of 0, 0, and .
2
What remains is to compute . Recall Y00 =

=
=
=
=
=

and Y10 =

3
4

cos .

r
r
3/2 1

2 1
er/2a0 Y00 z (2a0 )
er/2a0 Y10 d3 r
2a0
3 a0

r
r
1
3 1
cos Y10 d
2eE (2a0 )
r3 d3 r 1
er/a0
2a0
a0
3
4

4
5
1 1
r
r
2eE(2)3
5 er/a0 dr
a40
2a0
3 3 0

a0 eE
1 5 x
x4 ex dx
x e dx
12
2 0
0

a0 eE
54321
4321
12
2
a0 eE
(36)
12
3eEa0

E (1) = 3eEa0

= eE
=

1
4

3/2

(2a0 )

433

21. Time Independent Perturbation Theory

TOC

This is first order in the electric field, as we would expect in first order (degenerate)
perturbation theory.
If the states are not exactly degenerate, we have to leave in the diagonal terms of H0 .
Assume that the energies of the two (mixed) states are E0 , where comes from
some other perturbation, like the hydrogen fine structure. (The 211 and 21(1) are
still not mixed by the electric field.)

E0

1
1
=E

E0 +
2
2
E = E0

2 + 2

This is OK in both limits, , and . It is also correct when the two

corrections are of the same order.

21.4
21.4.1

Derivations and Computations

Derivation of 1st and 2nd Order Perturbation Equations

To keep track of powers of the perturbation in this derivation we will make the substitution H1 H1 where is assumed to be a small parameter in which we are making
the series expansion of our energy eigenvalues and eigenstates. It is there to do the
book-keeping correctly and can go away at the end of the derivations.
To solve the problem using a perturbation series, we will expand both our energy
eigenvalues and eigenstates in powers of .
En
n

= En(0) + En(1) + 2 En(2) + ...

X
cnk ()k
= N () n +
k6=n

cnk ()

(1)
cnk

(2)
2 cnk

+ ...

The full Schr

odinger equation is

X
X
(H0 +H1 ) n +
cnk ()k = (En(0) +En(1) +2 En(2) +...) n +
cnk ()k
k6=n

k6=n

where the N () has been factored out on both sides. For this equation to hold as we
vary , it must hold for each power of . We will investigate the first three terms.
434

21. Time Independent Perturbation Theory

0
1
2

TOC

(0)

H0 n = En n
P (1)
(1)
(0) P (1)
H1 n + H0
cnk k = En n + En
cnk k
k6=n
k6=n
P 2 (2)
P (1)
(0) P 2 (2)
(1) P
(1)
(2)
H0
cnk k + H1
cnk k = En
cnk k + En
cnk k + 2 En n
k6=n

k6=n

The zero order term is just the solution to the unperturbed problem so there is no new
information there. The other two terms contain linear combinations of the orthonormal
functions i . This means we can dot the equations into each of the i to get information,
much like getting the components of a vector individually. Since n is treated separately
in this analysis, we will dot the equation into n and separately into all the other
functions k .
The first order equation dotted into n yields
hn |H1 |n i = En(1)
and dotted into k yields
(0)

(1)

hk |H1 |n i + Ek cnk = En(0) cnk .

From these it is simple to derive the first order corrections
En(1) = hn |H1 |n i
hk |H1 |n i
(1)
cnk = (0)
(0)
En Ek

The second order equation projected on n yields

X (1)
(2)
cnk hn |H1 |k i = 2 EN .
k6=n

We will not need the projection on k but could proceed with it to get the second order
correction to the wave function, if that were needed. Solving for the second order
(1)
correction to the energy and substituting for cnk , we have
2 En(2) =

X |hk |H1 |n i|2

(0)

En Ek

k6=n

The normalization factor N () played no role in the solutions to the Schrodinger

equation since that equation is independent of normalization. We do need to go back
and check whether the first order corrected wavefunction needs normalization.
P (1)
P (1)
P 2 (1) 2
1
cnk k |n +
cnk k i = 1 +
|cnk |
N ()2 = hn +
k6=n

k6=n

N () 1

1
2

P
k6=n

k6=n

(1)
|cnk |2

435

21. Time Independent Perturbation Theory

TOC

The correction is of order 2 and can be neglected at this level of approximation.

These results are rewritten with all the removed in section 21.1.

21.4.2

Derivation of 1st Order Degenerate Perturbation Equations

To deal with the problem of degenerate states, we will allow an arbitrary linear combination of those states at zeroth order. In the following equation, the sum over i is the
sum over all the states degenerate with n and the sum over k runs over all the other
states.

X
X (1)
n = N ()
i (i) +
cnk k + ...
iN

k6N

where N is the set of zeroth order states which are (nearly) degenerate with n . We will
only go to first order in this derivation and we will use as in the previous derivation
to keep track of the order in the perturbation.
The full Schr
odinger equation is.

X
X
X
X
(H0 +H1 )
i (i) +
cnk ()k = (En(0) +E (1) +...)
i (i) +
cnk ()k
k6N

k6N

If we keep the zeroth and first order terms, we have

X
X (1)
X
X (1)
(H0 + H1 )
i (i) + H0
cnk k = (En(0) + E (1) )
i (i) + En(0)
cnk k .
k6N

k6N

Projecting this onto one of the degenerate states (j) , we get

X
h(j) |H0 + H1 |(i) ii = (En(0) + E (1) )j .
iN

By putting both terms together, our calculation gives us the full energy to first order,
not just the correction. It is useful both for degenerate states and for nearly degenerate
states. The result may be simplified to
X
h(j) |H|(i) ii = Ej .
iN

This is just the standard eigenvalue problem for the full Hamiltonian in the subspace
of (nearly) degenerate states.
436

21. Time Independent Perturbation Theory

21.5

TOC

Homework Problems

1. An electron is bound in a harmonic oscillator potential V0 = 21 m 2 x2 . Small

electric fields in the x direction are applied to the system. Find the lowest order
nonzero shifts in the energies of the ground state and the first excited state if a
constant field E1 is applied. Find the same shifts if a field E1 x3 is applied.
2. A particle is in a box from a to a in one dimension. A small additional potential
V1 = cos( x
2b ) is applied. Calculate the energies of the first and second excited
states in this new potential.
3. The proton in the hydrogen nucleus is not really a point particle like the electron
is. It has a complicated structure, but, a good approximation to its charge distribution is a uniform charge density over a sphere of radius 0.5 fermis. Calculate
the effect of this potential change for the energy of the ground state of hydrogen.
Calculate the effect for the n = 2 states.
4. Consider a two dimensional harmonic oscillator problem described by the Hamilp2 +p2

tonian H0 = x2m y + 12 m 2 (x2 + y 2 ). Calculate the energy shifts of the ground

state and the degenerate first excited states, to first order, if the additional potential V = 2xy is applied. Now solve the problem exactly. Compare the exact
result for the ground state to that from second order perturbation theory.
P
~2
5. Prove that (En Ea )|hn|x|ai|2 = 2m
by starting from the expectation value
n

of the commutator [p, x] in the state a and summing over all energy eigenstates.
dx
Assume p = m dx
dt and write dt in terms of the commutator [H, x] to get the
result.
~
6. If the general form of the spin-orbit coupling for a particle of mass m and spin S
dV
(r)
1
1
~ S
~
moving in a potential V (r) is HSO = 2m2 c2 L
r dr , what is the effect of that
coupling on the spectrum of a three dimensional harmonic oscillator? Compute
the relativistic correction for the ground state of the three dimensional harmonic
oscillator.

21.6

Sample Test Problems

1. Assume an electron is bound to a heavy positive particle with a harmonic potential V (x) = 21 m 2 x2 . Calculate the energy shifts to all the energy eigenstates in
an electric field E (in the x direction).
2. Find the energies of the n = 2 hydrogen states in a strong uniform electric field
in the z-direction. (Note, since spin plays no role here there are just 4 degenerate
states. Ignore the fine structure corrections to the energy since the E-field is
strong. Remember to use the fact that [Lz , z] = 0. If you are pressed for time,
dont bother to evaluate the radial integrals.)
437

21. Time Independent Perturbation Theory

TOC

3. An electron is in a three dimensional harmonic oscillator potential V (r) = 12 m 2 r2 .

A small electric field, of strength Ez , is applied in the z direction. Calculate the
lowest order nonzero correction to the ground state energy.
4. Hydrogen atoms in the n = 2 state are put in a strong Electric field. Assume that
the 2s and 2p states of Hydrogen are degenerate and spin is not important. Under
these assumptions, there are 4 states: the 2s and three 2p states. Calculate the
shifts in energy due to the E-field and give the states that have those energies.
Please work out the problem in principle before attempting any integrals.

438

22. Fine Structure in Hydrogen

TOC

Fine Structure in Hydrogen

In this section, we will calculate the fine structure corrections to the Hydrogen spectrum. Some of the degeneracy will be broken. Since the Hydrogen problem still has
spherical symmetry, states of definite total angular momentum will be the energy eigenstates.
We will break the spherical symmetry by applying a weak magnetic field, further breaking the degeneracy of the energy eigenstates. The effect of a weak magnetic field is
known as the anomalous Zeeman effect, because it was hard to understand at the time
it was first measured. It will not be anomalous for us.
We will use many of the tools of the last three sections to make our calculations.
Nevertheless, a few of the correction terms we use will not be fully derived here.
This material is covered in Gasiorowicz Chapter 17, in Cohen-Tannoudji et al.
Chapter XII, and in Griffiths 6.3 and 6.4.

22.1

Hydrogen Fine Structure

The basic hydrogen problem we have solved has the following Hamiltonian.
H0 =

p2
Ze2

2
r

To this simple Coulomb problem, we will add several corrections:

1. The relativistic correction to the electrons kinetic energy.
2. The Spin-Orbit correction.
3. The Darwin Term correction to s states from Dirac eq.
4. The ((anomalouus) Zeeman) effect of an external magnetic field.
Correction (1) comes from relativity. The electrons velocity in hydrogen is of order
c. It is not very relativistic but a small correction is in order. By calculating the
next order relativistic correction to the kinetic energy we find the additional
term in the Hamiltonian
1 p4e
H1 =
.
8 m3 c2
439

22. Fine Structure in Hydrogen

TOC

Our energy eigenstates are not eigenfunctions of this operator so we will have to treat
it as a perturbation.
We can estimate the size of this correction compared to the Hydrogen binding
energy by taking the ratio
D to
E the Hydrogen kinetic energy. (Remember that, in the
hydrogen ground state,

p2
2m

= E = 12 2 mc2 .)

p4
p2
p2
(p2 /2m)
1

=
=
= 2
3
2
2
2
2
8m c
2m
4m c
2mc
4
Like all the fine structure corrections, this is down by a factor of order 2 from the
Hydrogen binding energy.
The second term, due to Spin-Orbit interactions, is harder to derive correctly.
We understand the basis for this term. The magnetic moment from the electrons spin
interacts with the B field produced by the current seen in the electrons rest frame from
the circulating proton.
~
H2 = ~
e B
We can derive B from a Lorentz transformation of the E field of a static
proton
(We must also add in the Thomas Precession which we will not try to
understand here).
1 ge2 ~ ~
H2 =
LS
2 2m2 c2 r3
This will be of the same order as the relativistic correction.
Now we compute the relativity correction in first order perturbation theory
.

(0) 2
En
4n
hnlm |H1 | nlm i = +
3

2mc2
` + 21
The result depends on ` and n, but not on m` or j. This means that we could use
either the njmj `s or the n`m` sms to calculate the effect of H1 . We will need to use
the njmj `s to add in the spin-orbit.
The first order perturbation energy shift from the spin orbit correction is
calculated for the states of definite j.

ge2 ~2 1
1
[j(j + 1) `(` + 1) s(s + 1)]
hnlm |H2 | nlm i =
2
2
4m c 2
r3 nlm
=

g E (0) 2
n

2mc

2
2

n
(`+ 21 )(`+1)
n
`(`+ 21 )

j =`+
j =`

1
2
1
2

440

22. Fine Structure in Hydrogen

TOC

~ S
~ term should give 0 for ` = 0! In the above calculation there is
Actually, the L
`
an ` factor which makes the result for ` = 0 undefined. There is an additional Dirac
Equation contribution called the Darwin term
which is important for ` = 0
and surprisingly makes the above calculation right, even for ` = 0!
We will now add these three fine structure corrections together for states of definite j.
We start with a formula which has slightly different forms for j = ` 21 .
(0) 2
En
2mc2

Enjmj `s = En(0) +

Enjmj `s =

En(0)

2
j+ 12
+ 12

4
`

1
2

3 4n +
` + 21

)(+)

(`+ 12 )(`+1)
2n
`(`+
1
)
2

()

"
(+) #

(0) 2
2
n
En
4 `+1
3
+
4 + 2` ()
2mc2
(` + 12 )

Enjmj `s =

We can write (` + 12 ) as (j +

En(0)

"
#
(0) 2
4 j+2 1
En
2
3n
+
2mc2
` + 12

12 ), so that

(j + 21 12 )
4j + 2 2
=
4
(j + 21 )(j + 12 12 )
(j + 12 21 )(j + 12 )

and we get a nice cancellation giving us a simple formula.

(0)

Enlm = En +

(0)
En
2mc2

h
3

4n
j+ 12

This is independent of ` so the states of different total angular momentum split in

energy but there is still a good deal of degeneracy.

441

22. Fine Structure in Hydrogen

TOC

We have calculated the fine structure effects in Hydrogen. There are, of course, other,
smaller corrections to the energies. A correction from field theory, the Lamb Shift,
causes states of different ` to shift apart slightly. Nevertheless, the states of definite
total angular momentum are the energy eigenstates until we somehow break spherical
symmetry.

22.2

Hydrogen Atom in a Weak Magnetic Field

One way to break the spherical symmetry is to apply an external B field. Lets assume
that the field is weak enough that the energy shifts due to it are smaller than the fine
structure corrections. Our Hamiltonian can now be written as H = H0 +(H1 +H2 )+H3 ,
2
p2
where H0 = 2
Zer is the normal Hydrogen problem, H1 + H2 is the fine structure
correction, and
~
eB
~ + 2S)
~ = eB (Lz + 2Sz )
(L
H3 =
2mc
2mc
is the term due to the weak magnetic field.
442

22. Fine Structure in Hydrogen

TOC

We now run into a problem because H1 + H2 picks eigenstates of J 2 and Jz while H3

picks eigenstates of Lz and Sz . In the weak field limit, we can do perturbation theory
using the states of definite j. A direct calculation of the Anomalous Zeeman
Effect gives the energy shifts in a weak B field.

E = n`jmj 2mc
(Lz + 2Sz ) n`jmj =

e~B
2mc mj

1
2`+1

This is the correction, due to a weak magnetic field, which we should add to the fine
structure energies.

1 2 2 1
2
1
3
Enjmj `s = mc
+ 3

2
n2
n j + 21
4n
Thus, in a weak field, the the degeneracy is completely broken for the states
njmj `s . All the states can be detected spectroscopically.

443

22. Fine Structure in Hydrogen

TOC

1
is known as the Lande g Factor because the state splits as if
The factor 1 2`+1
it had this gyromagnetic ratio. We know that it is in fact a combination of the orbital
and spin g factors in a state of definite j.
In the strong field limit we could use states of definite m` and ms and calculate the
effects of the fine structure, H1 + H2 , as a correction. We will not calculate this here. If
the field is very strong, we can neglect the fine structure entirely. Then the calculation
is easy.
eB~
E = En0 +
(m` + 2ms )
2mc
In this limit, the field has partially removed the degeneracy in m` and ms , but not `.
For example, the energies of all these n = 3 states are the same.
` = 2 m` = 0 ms = 12
` = 1 m` = 0 ms = 12
` = 2 m` = 2 ms = 21

22.3

Examples

22.4

Derivations and Computations

22.4.1

The Relativistic Correction

Moving from the non-relativistic formula for the energy of an electron to the relativistic
formula we make the change

1/2
1/2
p2 c2
p2
= mc2 1 + 2 4
.
mc2 + e p2 c2 + m2 c4
2m
m c
Taylor expanding the square root around p2 = 0, we find
p2 c2 + m2 c4

1/2

= mc2 +

1 p2 c2
p4
1 p4 c4
p2
2

mc
+
2 mc2
8 m3 c6
2m 8m3 c2

So we have our next order correction term. Notice that

correction to mc2 .

p2
2m

was just the lowest order

What about the reduced mass problem? The proton is very non-relativistic so only
the electron term is important and the reduced mass is very close to the electron mass.
We can therefore neglect the small correction to the small correction and use

H1 =

1 p4e
.
8 m3 c2
444

22. Fine Structure in Hydrogen

22.4.2

TOC

The Spin-Orbit Correction

We calculate the classical Hamiltonian for the spin-orbit interaction which we will later
apply as a perturbation. The B field from the proton in the electrons rest frame is
~ = ~v E.
~
B
c
Therefore the correction is
H2 =

ge ~ ~
ge ~
~
SB =
S ~v E
2mc
2mc2
=

ge ~
~
S p~ .
2m2 c2

only depends on r = r d
dr =
H2 =

~
r d
r dr

1 d
ge ~ ~ 1 d
ge ~
S p~ ~r
=
SL
2m2 c2
r dr
2m2 c2
r dr

e
r
H2 =

d
e
= 2
dr
r

1 ge2 ~ ~
LS
2 2m2 c2 r3

Note that this was just a classical calculation which we will apply to quantum states
later. It is correct for the EM forces, but, the electron is actually in a rotating system
~ S
~ term (not from the B field!). This term is 1/2 the size
which gives an additional L
and of opposite sign. We have already included this factor of 2 in the answer given
above.
Recall that

1 2
J L2 S 2
2
and we will therefore want to work with states of definite j, `, and s.
~ S
~=
H2 L

22.4.3

Perturbation Calculation for Relativistic Energy Shift

2 2
p4
p
1
Rewriting H1 = 81 m3ec2 as H1 = 2mc
we calculate the energy shift for a
2
2m
state njmj `s . While there is no spin involved here, we will need to use these states for

445

22. Fine Structure in Hydrogen

TOC

the spin-orbit interaction

njmj `s |H1 | njmj `s

=
=
=

+
p 2 2

njmj `s

2m njmj `s

*
+
2

1
e2

njmj `s H0 +
njmj `s

2mc2
r

2 e2
e4
1
e2

H
+
H
+
H
+

njmj `s 0
0
0
njmj `s
2mc2
r
r
r2
2

e
1

En2 + njmj `s H0 njmj `s

2
2mc
r
2

e
+ H0 njmj `s njmj `s
r
4

e

+ njmj `s 2 njmj `s
r

1
1
1
2
2
4

En + 2En e
+e
2mc2
r n
r2 nl
1

2mc2

where we can use some of our previous results.

En =

1
=
r n

1
=
r2

njmj `s |H1 | njmj `s

1
e2
2 mc2 /n2 =
2
2a0 n2

1
a0 n2

1
a20 n3 (` + 12 )

#
2
1 2 2 2
12 2 mc2
2 mc
e
e4
+2
+ 2 3
n2
n2
a0 n2
a0 n (` + 21 )

2
1
4n
=
E (0) 1 4 +
2mc2 n
` + 21

(0) 2
En
4n
= +
3

2mc2
` + 12

1
=
2mc2

Since this does not depend on either m` or j, total j states and the product states give
the same answer. We will choose to use the total j states, njmj `s , so that we can
combine this correction with the spin-orbit correction.

446

22. Fine Structure in Hydrogen

22.4.4

TOC

Perturbation Calculation for H2 Energy Shift

We now calculate the expectation value of H2 . We will immediately use the fact that
j = ` 12 .

1
ge2 ~2 1
[j(j
+
1)

`(`
+
1)

s(s
+
1)]
njmj `s |H2 | njmj `s
=
4m2 c2 2
r3 nl

ge2 ~2
1
1
3 1
1
=
(`

)(`
+
1

`(`
+
1)

8m2 c2
2
2
4 a30 n3 `(` + 21 )(` + 1

3
1
g~2
1 1
2
2
= En
` +`` + ` `
1
2
2
2
4m c a0
2 4
4
n`(` + 2 )(` +
(+)
g E ~2
n
`
n
=
1
2
2
2
(`
+
1)
2
2mc
ma0 n
() `(` + 2 )(` + 1)
(+)
g E ~2 2 m2 c2
n
`
n
=
1
2
2
2
(`
+
1)
2
2mc
m~ n
() `(` + 2 )(` + 1)
#
"
2
n
g E (0)
j = ` + 21
n
(`+ 21 )(`+1)
=
2
n
j = ` 21
2 2mc2
`(`+ 1 )
2

Note that in the above equation, we have canceled a term

` = 0. We will return to this later.

22.4.5

`
`

which is not defined for

The Darwin Term

We get a correction at the origin from the Dirac equation.

HD =

e2 ~2 3
(~r)
2m2e c2

When we take the expectation value of this, we get the probability for the electron and
proton to be at the same point.
h |HD | i =

Now, (0) = 0 for ` > 0 and (0) =

hHD in00 =

1 2
4

e2 ~2
2
|(0)|
2m2e c2

z
na0

3/2

for ` = 0, so

4e2 ~2
e2 ~2 2 m2 c2
2nEn2
=
=
8n3 a30 m2 c2
2n3 a0 m2 c2 ~2
mc2

This is the same as ` = 0 term that we got for the spin orbit correction. This actually
replaces the ` = 0 term in the spin-orbit correction (which should be zero) making the
formula correct!
447

22. Fine Structure in Hydrogen

22.4.6

TOC

The Anomalous Zeeman Effect

We compute the energy change due to a weak magnetic field using first order Perturbation Theory.

eB

n`jmj
(Lz + 2Sz ) n`jmj
2mc
(Lz + 2Sz ) = Jz + Sz

The Jz part is easy since we are in eigenstates of that operator.

eB
eB
Jz n`jmj =
~mj
n`jmj
2mc
2mc
eB

n`jmj ,
The Sz is harder since we are not in eigenstates of that one. We need
S
z
j
2mc
n`jm

but we dont know how Sz acts on these. So, we must write njmj `s in terms of
|n`m` sms i.

eB
(Jz + Sz ) nj`mj
En(1) =
nj`mj
2mc

eB
=
mj ~ + nj`mj |Sz | nj`mj
2mc
We already know how to write in terms of these states of definite m` and ms .
n(`+ 12 )`(m+ 12 )

= Y`m + + Y`(m+1)

n(` 21 )`(m+ 12 )

= Y`m + Y`(m+1)
r
`+m+1
=
2` + 1
r
`m
=
2` + 1

Lets do the j = ` +

nj`mj |Sz | nj`mj

1
2

state first.
D
E
=
Y`(mj 12 ) + + Y`(mj + 21 ) |Sz | Y`(mj 12 ) + + Y`(mj + 12 )
=

For j = ` 12 ,

1
~ 2 2 m=m 1
j
2
2

1

nj`mj |Sz | nj`mj = ~ 2 2 m=mj 1
2
2

448

22. Fine Structure in Hydrogen

TOC

We can combine the two formulas for j = ` 12 .

nj`mj |Sz | nj`mj

~ 2
~`+m+1`+m
2 =
2
2
2` + 1
mj ~
~ 2(mj 12 ) + 1
=
=
2
2` + 1
2` + 1
=

So adding this to the (easier) part above, we have

mj ~
e~B
eB
1
(1)
mj ~
=
mj 1
En =
2mc
2` + 1
2mc
2` + 1
for j = ` 12 .
In summary then, we rewrite the fine structure shift.

1 2
3
1
4 1
E = mc (Z) 3

.
2
n j + 12
4n
To this we add the anomalous Zeeman effect

e~B
1
E =
mj 1
.
2mc
2` + 1

22.5

Homework Problems

1. Consider the fine structure of the n = 2 states of the hydrogen atom. What is the
spectrum in the absence of a magnetic field? How is the spectrum changed when
the atom is placed in a magnetic field of 25,000 gauss? Give numerical values for
the energy shifts in each of the above cases. Now, try to estimate the binding
energy for the lowest energy n = 2 state including the relativistic, spin-orbit, and
magnetic field.
2. Verify the relations used for 1r ,
and for any n if l = n 1.

1
r2 ,

and

1
r3

for hydrogen atom states up to n = 3

3. Calculate the fine structure of hydrogen atoms for spin 1 electrons for n = 1 and
n = 2. Compute the energy shifts in eV.

22.6

Sample Test Problems

1. The relativistic correction to the Hydrogen Hamiltonian is H1 = 8mp3 c2 . Assume

that electrons have spin zero and that there is therefore no spin orbit correction.
Calculate the energy shifts and draw an energy diagram for the n=3 states of
Hydrogen. You may use hnlm | 1r |nlm i = n21a0 and hnlm | r12 |nlm i = n3 a2 1(l+ 1 ) .
0
2
449

22. Fine Structure in Hydrogen

TOC

2. Calculate the fine structure energy shifts (in eV!) for the n = 1, n = 2, and
n = 3 states of Hydrogen. Include the effects of relativistic corrections, the spinorbit interaction, and the so-called Darwin term (due to Dirac equation). Do not
include hyperfine splitting or the effects of an external magnetic field. (Note: I
am not asking you to derive the equations.) Clearly list the states in spectroscopic
notation and make a diagram showing the allowed electric dipole decays of these
states.
3. Calculate and show the splitting of the n = 3 states (as in the previous problem)
in a weak magnetic field B. Draw a diagram showing the states before and after
the field is applied
~
4. If the general form of the spin-orbit coupling for a particle of mass m and spin S
1 dV
1
~
~
moving in a potential V (r) is HSO = 2m2 c2 L S r dr , what is the effect of that
coupling on the spectrum of an electron bound in a 3D harmonic oscillator? Give
the energy shifts and and draw a diagram for the 0s and 1p states.
V = 12 m 2 r2
dV
dr

= m 2 r

~2 1
2
2m2 c2 2 [j(j + 1) l(l + 1) s(s + 1)]m
~2 2
hHSO i = 4mc
2 [j(j + 1) l(l + 1) s(s + 1)]

hHSO i =

for the 0S 21 , E = 0,

~
for the 1P 21 , E = 2 4mc
2,
~
for the 1P 23 , E = +1 4mc
2.

5. We computed that the energies after the fine structure corrections to the hydrogen
2
4
mc2
4n
mc2
+ 8n
). Now a weak magnetic field B
spectrum are Enlj = 2n
2
4 (3
j+ 12
is applied to hydrogen atoms in the 3d state. Calculate the energies of all the 3d
states (ignoring hyperfine effects). Draw an energy level diagram, showing the
quantum numbers of the states and the energy splittings.
6. In Hydrogen, the n = 3 state is split by fine structure corrections into states
of definite j,mj ,`, and s. According to our calculations of the fine structure,
the energy only depends on j. We label these states in spectroscopic notation:
N 2s+1 Lj . Draw an energy diagram for the n = 3 states, labeling each state in
spectroscopic notation. Give the energy shift due to the fine structure corrections
in units of 4 mc2 .
7. The energies of photons emitted in the Hydrogen atom transition between the
3S and the 2P states are measured, first with no external field, then, with the
atoms in a uniform magnetic field B. Explain in detail the spectrum of photons
before and after the field is applied. Be sure to give an expression for any relevant
energy differences.
450

23. Hyperfine Structure

TOC

Hyperfine Structure

The interaction between the magnetic moment, due to the spin of the nucleus, and
the larger magnetic moment, due to the electrons spin, results in energy shifts which
me 2
En and are
are much smaller than those of the fine structure. They are of order m
p
hence called hyperfine.
The hyperfine corrections may be difficult to measure in transitions between states
of different n, however, they are quite measurable and important because they split
the ground state. The different hyperfine levels of the ground state are populated
thermally. Hyperfine transitions, which emit radio frequency waves, can be used to
detect interstellar gas.
This material is covered in Gasiorowicz Chapter 17, in Cohen-Tannoudji et al.
Chapter XII, and briefly in Griffiths 6.5.

23.1

Hyperfine Splitting

~ This particle is actually

We can think of the nucleus as a single particle with spin I.
1
made up of protons and neutrons which are both spin 2 particles. The protons and
neutrons in turn are made of spin 21 quarks. The magnetic dipole moment due to the
nuclear spin is much smaller than that of the electron because the mass appears in the
denominator. The magnetic moment of the nucleus is

~N =

ZegN ~
I
2MN c

where I~ is the nuclear spin vector. Because the nucleus has internal structure, the
nuclear gyromagnetic ratio is not just 2. For the proton, it is gp 5.56. This is
the nucleus of hydrogen upon which we will concentrate. Even though the neutron is
neutral, the gyromagnetic ratio is about -3.83. (The quarks have gyromagnetic ratios
of 2 (plus corrections) like the electron but the problem is complicated by the strong
interactions which make it hard to define a quarks mass.) We can compute (to some
accuracy) the gyromagnetic ratio of nuclei from that of protons and neutrons as we
can compute the protons gyromagnetic ratio from its quark constituents.
In any case, the nuclear dipole moment is about 1000 times smaller than that for e-spin
~ We will calculate E for ` = 0 states (see Condon and Shortley for more details).
or L.
This is particularly important because it will break the degeneracy of the Hydrogen
ground state.
~ from
To get the perturbation, we should find B
~ (see Gasiorowicz page
D 287) then
E
~ .
calculate the energy change in first order perturbation theory E = ~
e B
451

23. Hyperfine Structure

TOC

Calculating the energy shift for ` = 0 states.

D e
E 4
~ I~
m
1 S
4
~
~
E =
S B = (Z)
(mc2 )gN 3 2
mc
3
MN
n ~
~ S,
~ spin-orbit interaction, we will define the total
Now, just as in the case of the L
angular momentum
~ + I.
~
F~ = S
It is in the states of definite f and mf that the hyperfine perturbation will be diagonal.
In essence, we are doing degenerate state perturbation theory. We could diagonalize
the 4 by 4 matrix for the perturbation to solve the problem or we can use what we
know to pick the right states to start with. Again like the spin orbit interaction, the
total angular momentum states will be the right states because we can write
the perturbation in terms of quantum numbers of those states.

1 2
1
3 3
2
2
2
~
~
SI =
F S I = ~ f (f + 1)
2
2
4 4

E = 23 (Z)4

m
MN

(mc2 )gN n13 f (f + 1)

3
2

For the hydrogen ground state we are just adding two spin
values are f = 0, 1.

A
2

1
2

f (f + 1)

3
2

particles so the possible

Example: The Hyperfine Splitting of the Hydrogen Ground State.

The transition between the two states gives rise to EM waves with = 21 cm.

23.2

Hyperfine Splitting in a B Field

If we apply a B-field the states will split further. As usual, we choose our coordinates
so that the field is in z direction. The perturbation then is
Wz

~ (~
= B
L +
~S +
~I)
gN
B B
(Lz + 2Sz ) +
BIz
=
~
~

where the magnetic moments from orbital motion, electron spin, and nuclear spin are
considered for now. Since we have already specialized to s states, we can drop the
452

23. Hyperfine Structure

TOC

orbital term. For fields achievable in the laboratory, we can neglect the nuclear
magnetic moment in the perturbation. Then we have
Wz = 2B B

Sz
.
~

As an examples of perturbation theory, we will work this problem for weak fields, for
strong fields, and also work the general case for intermediate fields. Just as in the
Zeeman effect, if one perturbation is much bigger than another, we choose
the set of states in which the larger perturbation is diagonal. In this case,
the hyperfine splitting is diagonal in states of definite f while the above perturbation
due to the B field is diagonal in states of definite ms . For a weak field, the hyperfine
dominates and we use the states of definite f . For a strong field, we use the ms , mf
states. If the two perturbations are of the same order, we must diagonalize the full
perturbation matrix. This calculation will always be correct but more time consuming.
We can estimate the field at which the perturbations are the same size by comparing
me
mc2 gN = 2.9 106 . The weak field limit is achieved if B 500 gauss.
B B to 23 4 m
p
Example: The Hyperfine Splitting in a Weak B Field.

3
The result of this is example is quite simple E = En00 + A
2 f (f + 1) 2 +B Bmf . It
has the hyperfine term we computed before and adds a term proportional to B which
depends on mf .
In the strong field limit we use states |ms mi i and treat the hyperfine interaction as
a perturbation. The unperturbed energies of these states are E = En00 + 2B Bms +
gN BmI . We kept the small term due to the nuclear moment in the B field without
extra effort.
Example: The Hyperfine Splitting in a Strong B Field.

The result in this case is

E = En00 + 2B Bms + gn BmI + Ams mI .
Finally, we do the full calculation.
Example: The Hyperfine Splitting in an Intermediate B Field.
The general result consists of four energies which depend on the strength of the B field.
Two of the energy eigenstates mix in a way that also depends on B. The four energies
are
A
E = En00 + B B
4
453

23. Hyperfine Structure

and
A
E = En00
4

TOC

A
2

+ (B B) .

These should agree with the previous calculations in the two limits: B small, or B
large. The figure shows how the eigenenergies depend on B.

We can make a more general calculation, in which the interaction of the nuclear magnetic moment is of the same order as the electron. This occurs in muonic hydrogen or
positronium. Example: The Hyperfine Splitting in an Intermediate B Field.

23.3
23.3.1

Examples
Splitting of the Hydrogen Ground State

The ground state of Hydrogen has a spin 12 electron coupled to a spin 21 proton, giving
total angular momentum state of f = 0, 1. We have computed in first order perturbation theory that

2
m
1
3
E = (Z)4
(mc2 )gN 3 f (f + 1)
.
3
MN
n
2
454

23. Hyperfine Structure

TOC

The energy difference between the two hyperfine levels determines the wave length of
the radiation emitted in hyperfine transitions.

Ef =1 Ef =0 =

4
(Z)4
3

m
MN

(mc2 )gN

1
n3

For n = 1 Hydrogen, this gives

Ef =1 Ef =0 =

4
3

1
137

.51
938

(.51 106 )(5.56) = 5.84 106 eV

1
Recall that at room temperature, kB t is about 40
eV, so the states have about equal
population at room temperature. Even at a few degrees Kelvin, the upper state is
populated so that transitions are possible. The wavelength is well known.

~c
1973

= 2
A = 2 109
A = 21.2 cm
E
5.84 106
This transition is seen in interstellar gas. The f = 1 state is excited by collisions.
Electromagnetic transitions are slow because of the selection rule ` = 1 we will
learn later, and because of the small energy difference. The f = 1 state does emit a
photon to de-excite and those photons have a long mean free path in the gas.
= 2

23.3.2

Hyperfine Splitting in a Weak B Field

Since the field is weak we work in the states |f mf i in which the hyperfine perturbation
is diagonal and compute the matrix elements for Wz = B Bz . But to do the computation, we will have to write those states in terms of |ms mi i which we will abbreviate
like | + i, which means the electrons spin is up and the protons spin is down.

z |11i = z |++i = |11i

z |1 1i = z |i = |1 1i
1

z |10i = z 2 (|+i + |+i) = 12 (|+i |+i) = |00i

z |00i = z 12 (|+i |+i) = 12 (|+i + |+i) = |10i
Now since the three (f = 1) states are degenerate, we have to make sure all the matrix
elements between those states are zero, otherwise we should bite the bullet and do the
455

23. Hyperfine Structure

TOC

full problem as in the intermediate field case.

could have guessed.

1 0
B B 0 0
0 0

The f = 1 matrix is diagonal, as we

0
0
1

The only nonzero connection between states is between f = 1 and f = 0 and we are
assuming the hyperfine splitting between these states is large compared to the matrix
element.
So the full answer is
Ez(1) = B Bmf
which is correct for both f states.

23.3.3

Hydrogen in a Strong B Field

We need to compute the matrix elements of the hyperfine perturbation using |ms mi i
as a basis with energies E = En00 + 2B Bms . The perturbation is
Hhf = A
where A = 43 (Z)4

me
MN

~ I~
S
~2

me c2 gN n13 .

Recalling that we can write

~ I~ = Iz Sz + 1 I+ S + 1 I S+ ,
S
2
2
the matrix elements can be easily computed. Note that the terms like I S+ which
change the state will give zero.

E

1
AD
A
1
~ ~

+ I S + = 2 + Iz Sz + I+ S + I S+ +
~2
~
2
2
=

A
A
h+ |Iz Sz | + i =
~2
4

h + |Hhf | +i =

h+ + |Hhf | + +i =

A
.
4

A
4

456

23. Hyperfine Structure

TOC

h |Hhf | i =

A
4

We can write all of these in one simple formula that only depends on relative sign of
ms and mi .
E = En00 + 2B Bms

23.3.4

A
= En00 + 2B Bms + A(ms mI )
4

Intermediate Field

Now we will work the full problem with no assumptions about which perturbation
is stronger. This is really not that hard so if we were just doing this problem on
the homework, this assumption free method would be the one to use. The reason we
work the problem all three ways is as an example of how to apply degenerate state
perturbation theory to other problems.
We continue on as in the last section but work in the states of |f mf i. The matrix for
hf mf |Hhf + HB |f 0 m0f i is
1
1
1
0

A
1
4 + B B
1
0

0
0
0
0

A
4

0
B B
0
0

0
0
A
4
B B

0
0
.
B B
3A
4

The top part is already diagonal so we only need to work in bottom right 2 by 2 matrix,
solving the eigenvalue problem.

A
B

B
3A

a
a
=E
b
b

where

A A
4
B B B

Setting the determinant equal to zero, we get

(A E)(3A E) B 2 = 0.
E 2 + 2AE 3A2 B 2 = 0
E=

p
4A2 + 4(3A2 + B 2 )
= A A2 + (3A2 + B 2 )
2
p
= A 4A2 + B 2

457

23. Hyperfine Structure

TOC

The eigenvalues for the mf = 0 states, which mix differently as a function of the field
strength, are
s
2
A
A
2
E=
+ (B B) .
4
2
The eigenvalues for the other two states which remain eigenstates independent of the
field strength are
A
+ B B
4
and
A
B B.
4

23.3.5

Positronium

Positronium, the Hydrogen-like bound state of an electron and a positron, has a hyperfine correction which is as large as the fine structure corrections since the magnetic
moment of the positron is the same size as that of the electron. It is also an interesting
laboratory for the study of Quantum Physics. The two particles bound together are
symmetric in mass and all other properties. Positronium can decay by anihilation into
two or more photons.
In analyzing positronium, we must take some care to correctly handle the relativistic
correction in the case of a reduced mass much different from the electron mass and to
correctly handle the large magnetic moment of the positron.
The zero order energy of positronium states is
En =
where the reduced mass is given by =

1 2 2 1
c 2
2
n
me
2 .

The relativistic correction must take account of both the motion of the electron and the

~
r2
. Since the electron and positron
positron. We use ~r ~r1 ~r2 and p~ = ~r = m~r1 m
2
are of equal mass, they are always exactly oposite each other in the center of mass and
so the momentum vector we use is easily related to an individual momentum.
p~ =

p~1 p~2
= p~1
2

We will add the relativistic correction for both the electron and the positron.
Hrel

1 p4 + p4
1 p4
1 p4
1
= 1 3 22 =
=
=
8 m c
4 m 3 c2
32 3 c2
8c2

p2
2

2
458

23. Hyperfine Structure

TOC

This is just half the correction we had in Hydrogen (with me essentially replaced by
).
ge ~
~ as the
The spin-orbit correction should be checked also. We had HSO = 2mc
v
2 S ~
interaction between the spin and the B field producded by the orbital motion. Since
p~ = ~v , we have
ge ~
~
S p~
HSO =
2mc2
for the electron. We just need to add the positron. A little thinking about signs shows
that we just at the positron spin. Lets assume the Thomas precession is also the same.
We have the same fomula as in the fine structure section except that we have m in
the denominator. The final formula then is

2
1 ge2 ~ ~
~2 = 1 e
~ S
~1 + S
~2
HSO =
L

S
+
S
L
1
2 2mc2 r3
2 22 c2 r3

again just one-half of the Hydrogen result if we write everything in terms of for the
electron spin, but, we add the interaction with the positron spin.
The calculation of the spin-spin (or hyperfine) term also needs some attention. We had
3

2 Ze2 gN ~ ~ 4 Zme c
ESS =
SI 3
3 2me MN c2
n
~
where the masses in the deonominator of the first term come from the magnetic moments and thus are correctly the mass of the particle and the mas in the last term
comes from the wavefunction and should be replaced by . For positronium, the result
is
2 e2 2 ~ ~ 4 c 3
ESS =
S1 S2 3
3 2m2e c2
n
~

2
2 e 8 ~ ~ 4 c 3
=
S1 S2 3
3 22 c2
n
~
~
~
32 4 2 1 S1 S2
=
c 3
3
n
~2
23.3.6

Hyperfine and Zeeman for H, muonium, positronium

We are able to set up the full hyperfine (plus B field) problem in a general way so that
different hydrogen-like systems can be handled. We know that as the masses become
more equal, the hyperfine interaction becomes more important.
Lets define our perturbation W as
W

A~ ~
S1 S2 + w1 S1z + w2 S2z
~2

459

23. Hyperfine Structure

TOC

Here, we have three constants that are determined by the strength of the interactions.
We include the interaction of the nuclear magnetic moment with the field, which
we have so far neglected. This is required because the positron, for example, has a
magnetic moment equal to the electron so that it could not be neglected.

1
1
1
0

A ~
1
4 + 2 (w1 + w2 )
0
1

0
0
0
0

A
4

A
E3 = +
4

A
E4 =
4

0
~2 (w1 + w2 )
0
0

A
2

A
4
~
2 (w1

w2 )

2
~2
(w1 w2 )
2

0
0

2 (w1 w2 )
3A
4

~ S
~ term into account.
Like previous hf except now we take (proton) other B

23.4

Derivations and Computations

23.4.1

Hyperfine Correction in Hydrogen

We start from the magnetic moment of the nucleus

ZegN ~
I.
2MN c

Now we use the classical vector potential from a point dipole (see (green) Jackson page
147)
~ r) = (~
~ 1.
A(~
)
r
We compute the field from this.
~ =
~ A
~
B

1

1
Bk =
Aj ijk =
m
mnj ijk = m
(mnj ikj )
xi
xi
xn
r
xi xn
r

1

1
= m
(km in kn im ) = k
i
xi xn
r
xn xn
xi xk r

1 ~
~ =
~ 1
B
~ 2 (~
)
r
r
460

23. Hyperfine Structure

TOC

Then we compute the energy shift in first order perturbation theory for s

e ~ ~
Ze2 gN
~ I~ 2 1 Si Ij
E =
S
SB =
me c
2me MN c2
r
xi xj

states.

1
r

The second term can be simplified because of the spherical symmetry of s states.
(Basically the derivative with respect to x is odd in x so when the integral is done, only
the terms where i = j are nonzero).

ij
1
1
=
d3 r |n00 (~r)|2 2
d3 r |n00 (~r)|2
xi xj r
3
r
So we have

2 Ze2 gN ~ ~
21
SI
E =
.
3 2me MN c2
r

Now working out the 2 term in spherical coordinates,

2
2 1
2 1
+
=0
= 3+
r2
r r r
r
r r2
we find that it is zero everywhere but we must be careful at r = 0.
To find the effect at r = 0 we will integrate.

~ 2 1 d3 r =

~ (
~ 1 )d3 r =

~ =
~ 1 ) dS
(
r

1
dS
r r

r=0

1
1
dS = (42 )( 2 ) = 4
r2

r=0

So the integral is nonzero for any region including the origin, which implies

21
= 4 3 (~r).

r
We can now evaluate the expectation value.
2 Ze2 gN ~ ~
S I(4|n00 (0)|2 )
3 2me MN c2

3
4 Zme c
2
2
4|n00 (0)| = |Rn0 (0)| = 3
n
~
E =

461

23. Hyperfine Structure

TOC

2 Ze2 gN ~ ~ 4
SI 3
E =
3 2me MN c2
n

Zme c
~

Simply writing the e2 in terms of and regrouping, we get

~ I~
4
me
1 S
E = (Z)4
(me c2 )gN 3 2 .
3
MN
n ~
We will sometimes group the constants such that
E

A~ ~
S I.
~2

(The textbook has numerous mistakes in this section.)

23.5

Homework Problems

1. Calculate the shifts in the hydrogen ground states due to a 1 kilogauss magnetic
field.
2. Consider positronium, a hydrogen-like atom consisting of an electron and a
positron (anti-electron). Calculate the fine structure of positronium for n = 1
and n = 2. Determine the hyperfine structure for the ground state. Compute
the energy shifts in eV.
3. List the spectroscopic states allowed that arise from combining (s =
(s = 2 with l = 1), and (s1 = 12 , s2 = 1 and l = 4).

23.6

1
2

with l = 3),

Sample Test Problems

1. Calculate the energy shifts to the four hyperfine ground states of hydrogen in a
weak magnetic field. (The field is weak enough so that the perturbation is smaller
than the hyperfine splitting.)
2. Calculate the splitting for the ground state of positronium due to the spin-spin
interaction between the electron and the positron. Try to correctly use the reduced mass where required but dont let this detail keep you from working the
problem.
3. A muonic hydrogen atom (proton plus muon) is in a relative 1s state in an external
magnetic field. Assume that the perturbation due to the hyperfine interaction
and the magnetic field is given by W = AS~1 S~2 + 1 S1z + 2 S2z . Calculate the
energies of the four nearly degenerate ground states. Do not assume that any
terms in the Hamiltonian are small.
462

23. Hyperfine Structure

TOC

4. A hydrogen atom in the ground state is put in a magnetic field. Assume that the
energy shift due to the B field is of the same order as the hyperfine splitting of
the ground state. Find the eigenenergies of the (four) ground states as a function
of the B field strength. Make sure you define any constants (like A) you use in
terms of fundamental constants.

463

24. The Helium Atom

TOC

The Helium Atom

Hydrogen has been a great laboratory for Quantum Mechanics. After Hydrogen, Helium is the simplest atom we can use to begin to study atomic physics. Helium has
two protons in the nucleus (Z = 2), usually two neutrons (A = 4), and two electrons
bound to the nucleus.
This material is covered in Gasiorowicz Chapters 18, in Cohen-Tannoudji et
al. Complement BXIV , and briefly in Griffiths Chapter 7.

24.1

General Features of Helium States

We can use the hydrogenic states to begin to understand Helium. The Hamiltonian has
the same terms as Hydrogen but has a large perturbation due to the repulsion between
the two electrons.
p2
p2
Ze2
Ze2
e2
H= 1 + 2

+
2m 2m
r1
r2
|~r1 ~r2 |
We can write this in terms of the (Z = 2) Hydrogen Hamiltonian for each electron plus
a perturbation,
H = H1 + H2 + V
2

where V (~r1 , ~r2 ) = |~r1e~r2 | . Note that V is about the same size as the the rest of the
Hamiltonian so first order perturbation theory is unlikely to be accurate.
For our zeroth order energy eigenstates, we will use product states of Hydrogen
wavefunctions.
u(~r1 , ~r2 ) = n1 `1 m1 (~r1 )n2 `2 m2 (~r2 )
These are not eigenfunctions of H because of V , the electron coulomb repulsion term.
Ignoring V , the problem separates into the energy for electron 1 and the energy for
electron 2 and we can solve the problem exactly.
(H1 + H2 )u = Eu
We can write these zeroth order energies in terms of the principal quantum numbers
of the two electrons, n1 and n2 . Recalling that there is a factor of Z 2 = 4 in these
energies compared to hydrogen, we get

1 2 2
1
1
1
1
2
E = En1 + En2 = Z me c
+ 2 = 54.4 eV
+ 2 .
2
n21
n2
n21
n2

E11 = Egs = 108.8 eV

464

24. The Helium Atom

TOC

E12 = E1st = 68.0 eV

E1 = Eionization = 54.4 eV
E22 = 27.2eV
Note that E22 is above ionization energy, so the state can decay rapidly by ejecting an
electron.

Now lets look at the (anti) symmetry of the states of two identical electrons. For
the ground state, the spatial state is symmetric, so the spin state must be antisymmetric
s = 0.
1
u0 = 100 100 (+ + )
2
465

24. The Helium Atom

TOC

For excited states, we can make either symmetric or antisymmetric space states.
1
1
(s)
u1 = (100 2`m + 2`m 100 ) (+ + )
2
2
1
(t)
u1 = (100 2`m 2`m 100 )+ +
2
The first state is s = 0 or spin singlet. The second state is s = 1 or spin triplet and
has three ms states. Only the +1 state is shown. Because the large correction
due to electron repulsion is much larger for symmetric space states, the
spin of the state determines the energy.
We label the states according to the spin quantum numbers, singlet or triplet. We will
treat V as a perturbation. It is very large, so first order perturbation theory will be
quite inaccurate.

24.2

The Helium Ground State

Calculating the first order correction to the ground state is simple in principle.

100 (~r1 ) 100

+ 2
+
h|H|i = d3 r1 d3 r2 100 (~r1 ) 100 (~r2 )
2m
r1
2m
r2
|~r1 ~r2 |
We can recycle our previous work to do these integrals. First, replace the Z in H1 with
a Z and put in a correction term. This makes the H1 part just a hydrogen energy.
The correction term is just a constant over r so we can also write that in terms of the
hydrogen ground state energy.
2

p1
Ze2
3

x =
d r1 100

100
2m
r1

2

Z e2
(Z Z) e2
p1
3

+
100
=
d r1 100
2m
r1
r1

1
= Z 2 (13.6 eV ) + (Z Z)e2 d3 r1 |100 |2
r1
Z
= Z 2 (13.6 eV ) + (Z Z)e2
a0
2

2 1 2
= Z
mc + Z (Z Z)2 mc2
2

1 2
2
2

= mc Z (Z Z) Z
2
Then we reuse the perturbation theory calculation to get the V term.

5 1 2 2
h|H|i = 2[x] + Z
mc
4
2

1 2 2
5
2

= mc 2Z 4Z (Z Z) Z
2
4

1
5
= 2 mc2 2Z 2 + 4ZZ Z
2
4

472

24. The Helium Atom

TOC

Use the variational principle to determine the best Z .

h|H|i
=0
Z

4Z + 4Z

5
=0
4

5
16
estimate of the ground state energy.

5
1
2 mc2 Z 2Z + 4Z
2
4

1 2 2
5
5
5
mc (Z ) 2Z + + 4Z
2
16
8
4
"
2 #
5
1
= 77.38 eV
2 mc2 2 Z
2
16
Z = Z

Putting these together we get our

h|H|i =
=
=
(really 78.975eV ).

Now we are within a few percent. We could use more parameters for better results.

24.6
24.6.1

Examples
1D Harmonic Oscillator

Use
= a2 x2

|x| a

and = 0 otherwise as a trial wave function. Recall the actual wave function is
2
2
emx /2~ . The energy estimate is
D
E
2
2
d2
1
2 2
2
2 2
+
m
x
|
a

x
a2 x2 | ~
2m dx2
2
E
D
.
E0 =
2
2
(a2 x2 ) | (a2 x2 )

We need to do some integrals of polynomials to compute

E0 =

1
3 ~2
+ m 2 a2 .
2
2 ma
22

Now we optimize the parameter.

dE 0
3 ~2
1
=0=
+ m 2 a2 =
2
4
da
2 ma
22

r
33

~2
~
= 33
2
m
m

473

24. The Helium Atom

TOC

!
3
33
1
+
~ = ~
22
2
2 33

1
3 ~
~
E = + m 2 33
=
2 33 22
m
0

!
33 + 33
11

1
1
43
12
= ~
= ~
2
2
11
11
This is close to the right answer. As always, it is treated as an upper limit on the
ground state energy.

24.6.2

1-D H.O. with exponential wavefunction

As a check of the procedure, take trial function eax

ground state energy.

E0 =

. This should give us the actual

h 2 2
i
~
1
2 2
2m
x2 + 2 m x dx

(
~2
2m

ax2

)
2 2

2 ax2
1
2
a x a dx + 2 m
x e
dx

ax2

2 2

1
a ~
+ m 2
2m
2

x2 eax dx

eax2 dx

~2 a
2m

eax dx =

1/2
= a
a

x2 eax dx =

1

a3/2
2

474

24. The Helium Atom

TOC

2 ax2

x e

1
dx =
2

1
=
3
a
2a

1
~2 a
1
~2
a~2
2
+ m +
=
m 2 +
a
E =
4m
4a
2m
4a
4m

E 0
m 2
~2
=
+
=0
2
a
4a
4m
4a2 ~2 = 4m2 2
m
a=
~
m

= e 2~ x
E0 =

~2 m
1
1
m 2 ~
+
= ~ + ~
4 m 4m ~
4
4

OK.

24.7

Derivations and Computations

24.7.1

Calculation of the ground state energy shift

To calculate the first order correction to the He ground state energy, we gotta do this
integral.

e2
Egs = hu0 |V |u0 i = d3 r1 d3 r2 |100 (~r1 )|2 |100 (~r2 )|2
|~r1 ~r2 |
First, plug in the Hydrogen ground state wave function (twice).
"

Egs

1
=
4
4

Z
a0

3 # 2

r12 dr1 e2Zr1 /a0

r22 dr1 e2Zr2 /a0

1
|~r1 ~r2 |

1
1
=p 2
2
|~r1 ~r2 |
r1 + r2 2r1 r2 cos
Do the d1 integral and prepare the other.
Egs

4
= 2 e2

Z
a0

6
0

r12 dr1 e2Zr1 /a0

r22 dr2 e2Zr2 /a0

d2 d cos 2 p

1
r12 + r22 2r1 r2 cos
475

24. The Helium Atom

TOC

The angular integrals are not hard to do.

Egs

4 2
e
2

Z
a0

r12 dr1 e2Zr1 /a0

Egs

2
2r1 r2

r12 + r22 2r1 r2 cos

2
4 2 Z
e
r12 dr1 e2Zr1 /a0 r22 dr2 e2Zr2 /a0
2
a0
r1 r2
0
0
q

q
2
2
2
2
r1 + r2 2r1 r2 + r1 + r2 + 2r1 r2
4 2
e
2

Z
a0

r12 dr1 e2Zr1 /a0

Egs

r22 dr2 e2Zr2 /a0 2

8e2

Z
a0

6
0

r22 dr2 e2Zr2 /a0

2
[|r1 r2 | + (r1 + r2 )]
r1 r2

r1 dr1 e2Zr1 /a0

r2 dr2 e2Zr2 /a0 (r1 + r2 |r1 r2 |)

We can do the integral for r2 < r1 and simplify the expression. Because of the symmetry

476

24. The Helium Atom

TOC

between r1 and r2 the rest of the integral just doubles the result.
Egs

16e

Z
a0

6
r1 dr1 e

Egs

Z
a0

x1
x1 dx1 ex1

0
2

Ze
a0

Ze2
a0

r2 dr2 e2Zr2 /a0 (2r2 )

2Zr1 /a0

x1 dx1 ex1
0

x1 dx1 ex1

x22 dx2 ex2

x1
x21 ex1 +

2x2 dx2 ex2

x1
x21 ex1 2x1 ex1 + 2

ex2 dx2
0

x1 dx1 ex1 x21 ex1 2x1 ex1 2 ex1 1
0

Ze2

x31 + 2x21 + 2x1 e2x1 2x1 ex1 dx1

=
Egs

=
=

24.8

211
11
11
Ze2 3 2 1 1
+2
+2
2

a0 2 2 2 2
222
22
11

2
2
Ze 3 4 4 16
5 Ze

+ +
=+
a0 8 8 8
8
8 a0
5
Z(13.6 eV )
34 eV for Z=2
4

Homework Problems

1. Calculate the lowest order energy shift for the (0th order degenerate first) excited
(s,t)
states of Helium E2,l where ` = 0, 1. This problem is set up in the The
following formulas will aid you in the computation. First, we can expand the
formula for the inverse distance between the two electrons as follows.

X r`
1
<
=
P (cos 12 )
`+1 `
|~r1 ~r2 |
r>
`=0

Here r< is the smaller of the two radii and r> is the larger. As in the ground state
calculation, we can use the symmetry of the problem to specify which radius is the
larger. Then we can use a version of the addition theorem to write the Legendre
477

24. The Helium Atom

TOC

Polynomial P` (cos 12 ) in terms of the spherical hamonics for each electron.

`
X
4
P` (cos 12 ) =
(1)m Y`m (1 , 1 )Y`(m) (2 , 2 )
2` + 1
m=`

Using the equation Y`(m) = (1)` Y`m

, this sets us up to do our integrals nicely.

2. Consider the lowest state of ortho-helium. What is the magnetic moment? That
is what is the interaction with an external magnetic field?
3. A proton and neutron are bound together into a deuteron, the nucleus of an
isotope of hydrogen. The binding energy is found to be -2.23 MeV for the nuclear
r/r0
ground state, an ` = 0 state. Assuming a potential of the form V (r) = V0 e r/r0 ,
with r0 = 2.8 Fermis, use the variational principle to estimate the strength of the
potential.
4. Use the variational principle with a gaussian trial wave function to prove that a
one dimensional attractive potential will always have a bound state.
5. Use the variational principle to estimate the ground state energy of the anharp2
monic oscillator, H = 2m
+ x4 .

24.9

Sample Test Problems

1. We wish to get a good upper limit on the Helium ground state energy. Use as a
trial wave function the 1s hydrogen state with the parameter a screened nuclear
charge Z to get this limit. Determine the value of Z which gives the best limit.
2
The integral h(1s)2 | |~r1e~r2 | |(1s)2 i = 58 Z 2 mc2 for a nucleus of charge Z e.
2. A Helium atom has two electrons bound to a Z = 2 nucleus. We have to add the
coulomb repulsion term (between the two electrons) to the zeroth order Hamiltonian.
H=

p21
Ze2
p2
Ze2
e2

+ 2
+
= H1 + H2 + V
2m
r1
2m
r2
|~r1 ~r2 |

The first excited state of Helium has one electron in the 1S state and the other in
the 2S state. Calculate the energy of this state to zeroth order in the perturbation
V. Give the answer in eV. The spins of the two electrons can be added to give
~ The possible total spin states are s = 0 and s = 1. Write
states of total spin S.
out the full first excited Helium state which has s = 1 and ms = 1. Include the
spatial wave function and dont forget the Pauli principle. Use bra-ket notation
to calculate the energy shift to this state in first order perturbation theory. Dont
do any integrals.
478

25. Atomic Physics

TOC

Atomic Physics

This material is covered in Gasiorowicz Chapter 19, and in Cohen-Tannoudji

et al. Complement AXIV .

25.1

Atomic Shell Model

The Hamiltonian for an atom with Z electrons and protons is

X
Z 2
2
2
X
Ze
p
e
i
= E.

+
2m
ri
|~
ri r~j |
i=1
i>j
We have seen that the coulomb repulsion between electrons is a very large correction
in Helium and that the three body problem in quantum mechanics is only solved by
approximation. The states we have from hydrogen are modified significantly. What
hope do we have to understand even more complicated atoms?
The physics of closed shells and angular momentum enable us to make sense of even
the most complex atoms. Because of the Pauli principle, we can put only one electron
into each state. When we have enough electrons to fill a shell, say the 1s or 2p, The
resulting electron distribution is spherically symmetric because
`
X

|Y`m (, )| =

m=`

2` + 1
.
4

With all the states filled and the relative phases determined by the antisymmetry
required by Pauli, the quantum numbers of the closed shell are determined. There
is only one possible state representing a closed shell.
As in Helium, the two electrons in the same spatial state, n`m , must by symmetric
in space and hence antisymmetric in spin. This implies each pair of electrons has a
total spin of 0. Adding these together gives a total spin state with s = 0, which is
antisymmetric under interchange. The spatial state must be totally symmetric under
interchange and, since all the states in the shell have the same n and `, it is the different
m states which are symmetrized. This can be shown to give us a total ` = 0 state.
So the closed shell contributes a spherically symmetric charge and spin
distribution with the quantum numbers
s=0
`=0
j=0
479

25. Atomic Physics

TOC

The closed shell screens the nuclear charge. Because of the screening, the potential
no longer has a pure 1r behavior. Electrons which are far away from the nucleus see less
of the nuclear charge and shift up in energy. This is a large effect and single electron
states with larger ` have larger energy. From lowest to highest energy, the atomic shells
have the order
1s, 2s, 2p, 3s, 3p, 4s, 3d, 4p, 5s, 4d, 5p, 6s, 4f, 5d, 6p.
The effect of screening not only breaks the degeneracy between states with the same n
but different `, it even moves the 6s state, for example, to have lower energy than the
4f or 5d states. The 4s and 3d states have about the same energy in atoms because of
screening.

25.2

The Hartree Equations

The Hartree method allows us to to change the 3Z dimensional Schrodinger equation

(Z electrons in 3 dimensions) into a 3 dimensional equation for each electron. This
equation depends on the wavefunctions of the other electrons but can be solved in a
self consistent way using the variational principle and iterating.
= 1 (r~1 ) 2 (r~2 ) . . . Z (r~Z )

2
2
2
X
~
Ze
|
(
r
~
)|
j
j

i (~
ri ) = i i (~
ri )
2
+ e2
d3 rj
2m i
ri
|~
ri r~j |
j6=i

In the Hartree equation above, i represents the energy contribution of electron i.

P 3 |j (r~j )|2
The term e2
d rj |r~i r~j | represents the potential due to the other electrons in
j6=i

which electron i moves. In this equation we can formally see the effect of screening
by the other electrons. The equation is derived (see Gasiorowicz pp 309-311) from the
Schr
odinger equation using = 1 2 . . . Z . Since we will not apply these equations
to solve problems, we will not go into the derivation, however, it is useful to know how
one might proceed to solve more difficult problems.
An improved formalism known as the Hartree-Fock equations, accounts for the required
antisymmetry and gives slightly different results.

25.3

Hunds Rules

A set of guidelines, known as Hunds rules, help us determine the quantum numbers
for the ground states of atoms. The hydrogenic shells fill up giving well defined j = 0
states for the closed shells. As we add valence electrons we follow Hunds rules to
480

25. Atomic Physics

TOC

determine the ground state. We get a great simplification by treating nearly closed
shells as a closed shell plus positively charged, spin 12 holes. For example, if an atom
is two electrons short of a closed shell, we treat it as a closed shell plus two positive
holes.)
1. Couple the valence electrons (or holes) to give maximum total spin.
2. Now choose the state of maximum ` (subject to the Pauli principle. The Pauli
principle rather than the rule, often determines everything here.)
3. If the shell is more than half full, pick the highest total angular momentum state
j = ` + s otherwise pick the lowest j = |` s|.
This method of adding up all the spins and all the Ls, is called LS or Russel-Saunders
coupling. This method and these rule are quite good until the electrons become
relativistic in heavy atoms and spin-orbit effects become comparable to the electron
repulsion (arond Z=40). We choose the states in which the total s and the total ` are
good quantum numbers are best for minimizing the overlap of electrons, and hence the
positive contribution to the energy.
For very heavy atoms, we add the total angular momentum from each electron first
then add up the Js. This is called j-j coupling. For heavy atoms, electrons are
relativistic and the spin-orbit interaction becomes more important than the effect of
electron repulsion. Thus we need to use states in which the total angular momentum
of each electron is a good quantum number.
We can understand Hunds rules to some extent. The maximum spin state is symmetric
under interchange, requiring an antisymmetric spatial wavefunction which has a lower
energy as we showed for Helium. We have not demonstated it, but, the larger the total
` the more lobes there are in the overall electron wavefunction and the lower the effect
of electron repulsion. Now the spin orbit interaction comes into play. For electrons
with their negative charge, larger j increases the energy. The reverse is true for holes
which have an effective postive charge.
A simpler set of rules has been developed for chemists, who cant understand addition
of angular momentum. It is based on the same principles. The only way to have a
totally antisymmetric state is to have no two electrons in the same state. We use the
same kind of trick we used to get a feel for addition of angular momentum; that is,
we look at the maximum z component we can get consistent with the Pauli principle.
Make a table with space for each of the different m` states in the outer shell. We can
put two electrons into each space, one with spin up and one with spin down. Fill the
table with the number of valence electrons according to the following rules.
1. Make as many spins as possible parallel, then compute ms and call that s.
481

25. Atomic Physics

TOC

2. Now set the orbital states to make maximum m` , and call this `, but dont allow
any two electrons to be in the same state (of ms and m` ).
3. Couple to get j as before.
This method is rather easy to use compared to the other where addition of more than
two angular momenta can make the symmetry hard to determine.
Example:
Example:
Example:
Example:

25.4

The
The
The
The

Boron ground State.

Carbon ground State.
Nitrogen ground State.
Oxygen ground State.

The Periodic Table

The following table gives the electron configurations for the ground states of light
atoms.

482

25. Atomic Physics

Z
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
36
54
86

El.
H
He
Li
Be
B
C
N
O
F
Ne
Na
Mg
Al
Si
P
S
Cl
Ar
K
Ca
Sc
Ti
V
Cr
Mn
Fe
Kr
Xe
Rn

TOC

Electron Configuration
(1s)
(1s)2
He (2s)
He (2s)2
He (2s)2 (2p)
He (2s)2 (2p)2
He (2s)2 (2p)3
He (2s)2 (2p)4
He (2s)2 (2p)5
He (2s)2 (2p)6
Ne (3s)
Ne (3s)2
Ne (3s)2 (3p)
Ne (3s)2 (3p)2
Ne (3s)2 (3p)3
Ne (3s)2 (3p)4
Ne (3s)2 (3p)5
Ne (3s)2 (3p)6
Ar (4s)
Ar (4s)2
Ar (4s)2 (3d)
Ar (4s)2 (3d)2
Ar (4s)2 (3d)3
Ar (4s)(3d)5
Ar (4s)2 (3d)5
Ar (4s)2 (3d)6
(Ar) (4s)2 (3d)10 (4p)6
(Kr) (5s)2 (4d)10 (5p)6
(Xe) (6s)2 (4f )14 (5d)10 (6p)6

2s+1

Lj
S1/2
1
S0
2
S1/2
1
S0
2
P1/2
3
P0
4
S3/2
3
P2
2
P3/2
1
S0
2
S1/2
1
S0
2
P1/2
3
P0
4
S3/2
3
P2
2
P3/2
1
S0
2
S1/2
1
S0
2
D3/2
3
F2
4
F3/2
7
S3
6
S3/2
5
D4
1
s0
1
s0
1
s0
2

Ioniz. Pot.
13.6
24.6
5.4
9.3
8.3
11.3
14.5
13.6
17.4
21.6
5.1
7.6
6.0
8.1
11.0
10.4
13.0
15.8
4.3
6.1
6.5
6.8
6.7
6.7
7.4
7.9
14.0
12.1
10.7

We see that the atomic shells fill up in the order 1s, 2s, 2p, 3s, 3p, 4s, 3d, 4p, 5s, 4d,
5p, 6s, 4f, 5d, 6p. The effect of screening increasing the energy of higher ` states is
clear. Its no wonder that the periodic table is not completely periodic.
The Ionization Potential column gives the energy in eV needed to remove one electron from the atom, essentially the Binding energy of the last electron. The Ionization
Potential peaks for atoms with closed shells, as the elctron gains binding energy from
more positive charge in the the nucleus without much penalty from repulsion of the
other electrons in the shell. As charge is added to the nucleus, the atom shrinks in
size and becomes more tightly bound. A single electron outside a closed shell often has
the lowest Ionization Potential because it is well screened by the inner electrons. The
figure below shows a plot of ionization potential versus Z.
483

25. Atomic Physics

TOC

484

25. Atomic Physics

TOC

The perodic table of elements is based on the fact that atoms with the same number of
electrons outside a closed shell have similar properties. The rows of the periodic table
contain the following states.
1. 1s
2. 2s, 2p
485

Oxygen Ground State

Oxygen, with Z = 8 has the 1S and 2S levels filled giving j = 0 as a base. It has four
valence 2P electrons which we will treat as two valence 2P holes. Hunds first rule ,
maximum total s, tells us to couple the two hole spins to s = 1. This is the symmetric
spin state so well need to make the space state antisymmetric. Hunds second rule,
maximum `, doesnt play a role because only the ` = 1 state is antisymmetric. Since
the shell is more than half full we couple to the the highest j = ` + s = 2. So the
ground state is 3 P2 .
491

25. Atomic Physics

TOC

m`
e
1

-1 P
s = P ms = 1
`=
m` = 1

25.7

Homework Problems

1. List the possible spectroscopic states that can arise in the following electronic
configurations: (1s)2 , (2p)2 , (2p)3 , (2p)4 , and (3d)4 . Take the exclusion principle
into account. Which should be the ground state?
2. Use Hunds rules to find the spectroscopic description of the ground states of the
following atoms: N(Z=7), K(Z=19), Sc(Z=21), Co(Z=27). Also determine the
electronic configuration.
3. Use Hunds rules to check the (S, L, J) quantum numbers of the elements with
Z =14, 15, 24, 30, 34.

25.8

Sample Test Problems

1. Write down the electron configuration and ground state for the elements from
Z = 1 to Z = 10. Use the standard 2s+1 Lj notation.
2. Write down the ground state (in spectroscopic notation) for the element Oxygen
(Z = 8).

492

26. Molecular Physics

TOC

Molecular Physics

In this section, we will study the binding and excitation of simple molecules. Atoms
bind into molecules by sharing electrons, thus reducing the kinetic energy. Molecules
can be excited in three ways.
Excitation of electrons to higher states. E 4 eV
Vibrational modes (Harmonic Oscillator). Nuclei move slowly in background of
electrons. E 0.1 eV
Rotational modes (L = n~). Entire molecule rotates. E 0.001 eV
Why dont atoms have rotational states?
The atomic state already accounts for electrons angular momentum around
the nucleus.
About which axes can a molecule rotate?

Do you think identical atoms will make a difference?

This material is covered in Gasiorowicz Chapter 20, and in Cohen-Tannoudji

et al. Complements CV I, EV II , CXI .

26.1

The H+
2 Ion

The simplest molecule we can work with is the H+

2 ion. It has two nuclei (A and B)
sharing one electron (1).
H0 =

p2e
e2
e2
e2

+
2m r1A
r1B
RAB

RAB is the distance between the two nuclei.

The lowest energy wavefunction can be thought of as a (anti)symmetric linear combination of an electron in the ground state near nucleus A and the ground state near
nucleus B

~ = C (R) [A B ]
~r, R

493

26. Molecular Physics

where A =

1 r1A /a0
e
a30

TOC

is g.s. around nucleus A. A and B are not orthogonal;

there is overlap. We must compute the normalization constant to estimate the energy.
1
2 = hA B |A B i = 2 2hA |B i 2 2S(R)
C
where

R2
R
+ 2 eR/a0
S(R) hA |B i = 1 +
a0
3a0

These calculations are straightforward but tedious (Gasiorowicz).

We can now compute the energy of these states.
hH0 i

=
=
=

1
hA B |H0 |A B i
2[1 S(R)]
1
[hA |H0 |A i + hB |H0 |B i hA |H0 |B i hB |H0 |A i]
2[1 S(R)]
hA |H0 |A i hA |H0 |B i
1 S(R)

We can compute the integrals needed.

hA |H0 |A i =
hA |H0 |B i =

e2
R
E1 +
1+
e2R/a0
R
a0

e2
e2
R
E1 +
S(R)
1+
eR/a0
R
a0
a0

We have reused the calculation of S(R) in the above. Now, we plug these in and
rewrite things in terms of y = R/a0 , the distance between the atoms in units of the
Bohr radius.

2
2
2
E1 + eR (1 + R/a0 ) e2R/a0 E1 + eR S(R) ae0 (1 + R/a0 ) eR/a0
hH0 i =
1 S(R)

2y
1 (2/y)(1 + y)e
(1 2/y)(1 + y + y 2 /3)ey 2(1 + y)ey
hH0 i = E1
1 (1 + y + y 2 /3)ey
The symmetric (bonding) state has a large probability for the electron to be found
between nuclei. The antisymmetric (antibonding) state has a small probability there,
and hence, a much larger energy.
The graph below shows the energies from our calculation for the space symmetric (Eg )
and antisymmetric (Eu ) states as well as the result of a more complete calculation
(Exact Eg ) as a function of the distance between the protons R. Our calculation for
the symmetric state shows a minimum arount 1.3 Angstroms between the nuclei and
494

26. Molecular Physics

TOC

a Binding Energy of 1.76 eV. We could get a better estimate by introduction some
parameters in our trial wave function and using the variational method.
The antisymmetric state shows no minimum and never goes below -13.6 eV so there is
no binding in this state.

By setting

dhHi
dy

= 0, we can get the distance between atoms and the energy.

Calculated
Actual

Distance
1.3
A
1.06
A

Energy
-1.76 eV
-2.8 eV

Its clear we would need to introduce some wfn. parameters to get good precision.

26.2

The H2 Molecule

The H2 molecule consists of four particles bound together: e1 , e2 , protonA , and

protonB . The Hamiltonian can be written in terms of the H+
2 Hamiltonian, the repulsion between electrons, plus a correction term for double counting the repulsion
495

26. Molecular Physics

TOC

between protons.
H = H1 + H2 +
H1 =

e2
e2

r12
RAB

p21
e2
e2
e2

+
2m rA1
rB1
RAB

We wish to compute variational upper bound on RAB and the energy.

We will again use symmetric electron wavefunctions,
(r1 , r2 ) =

1
[A (r~1 ) + B (r~1 )] [A (r~2 ) + B (r~2 )] s
2[1 + S(RAB )]

where the spin singlet is required because the spatial wfn is symmetric under interchange.
The space symmetric state will be the ground state as before.
2
e
e2

h|H|i = 2EH + (RAB )
+
2
RAB
r12
From this point, we can do the calculation to obtain

Calculated
Actual

Distance
0.85
A
0.74
A

Energy
-2.68 eV
-4.75 eV.

wIth a multiterm wavefunction, we could get good agreement.

26.3

Importance of Unpaired Valence Electrons

Inner (closed shell) electrons stick close to nucleus so they do not get near to other
atoms. The outer (valence) electrons may participate in bonding either by sharing or
migrating to the other atom. Electrons which are paired into spin singlets
dont bond. If we try to share one of the paired electrons, in a bonding state, with
another atom, the electron from the other atom is not antisymmetric with the (other)
paired electron. Therefore only the antibonding (or some excited state) will work and
binding is unlikely. Unpaired electrons dont have this problem.
. . . first four dont bond!
The strongest bonds come from s and p orbitals (not d,f).
496

26. Molecular Physics

26.4

TOC

Molecular Orbitals

Even with additional parameters, parity symmetry in diatomic molecules implies we

will have symmetric and antisymmetric wavefunctions for single electrons. The symmetric or bonding state has a larger probability to be between the two
nuclei, sees more positive charge, and is therefore lower energy. As in our simple
model of a molecule, the kinetic energy can be lowered by sharing an electron.
There is an axis of symmetry for diatomic molecules. This means Lz commutes
with H and m` is a good quantum number. The different m` states, we have seen,
have quite different shapes therefore bond differently. Imagine that a valence electron
is in a d state. The m` = 0, 1, 2 are called molecular orbitals , , respectively.
Each has a bonding and an antibonding state.
Pictures of molecular orbitals are shown for s and p states in the following figure. Both bonding and antibonding orbitals are shown first as atomic states then as
molecular. The antibonding states are denoted by a *.

497

26. Molecular Physics

TOC

498

26. Molecular Physics

26.5

TOC

Vibrational States

We have seen that the energy of a molecule has a minimum for some particular separation between atoms. This looks just like a harmonic oscillator potential
for small variations from the minimum. The molecule can vibrate in this potential
giving rise to a harmonic oscillator energy spectrum.
We can estimate the energy of the vibrational levels. If Ee ~ = ~

E m
then crudely the proton has the same spring constant k e ~ e .
r
Evib ~

k
=
M

Recalling that room temperature is about

infrared. The energy levels are simply

k
me ,

1
m
Ee
eV
M
10

1
40

eV, this is approximately thermal energy,

1
E = (n + )~vib
2
Complex molecules can have many different modes of vibration. Diatomic molecules
have just one.
The graph below shows the energy spectrum of electrons knocked out of molecular
hydrogen by UV photons (photoelectric effect). The different peaks correspond
to the vibrational state of the final H+
2 ion.

499

26. Molecular Physics

TOC

Can you calculate the number of vibrational modes for a molecule compose
of N > 3 atoms.

26.6

Rotational States

Molecules can rotate like classical rigid bodies subject to the constraint that angular
momentum is quantized in units of ~. We can estimate the energy of these rotations
to be
Erot =

`(` + 1)~2
~2
m 2 mc2
m
1
1 L2
=

E
eV
2
2 I
2I
2M a0
M
2
M
1000

where we have used a0 =

~
mc .

These states are strongly excited at room temperature.

Lets look at the energy changes between states as we might get in a radiative transition
with ` = 1..
`(` + 1)~2
E=
2I
2
~
~2
~2 `
E =
[`(` + 1) (` 1)`] =
(2`) =
2I
2I
I
500

26. Molecular Physics

TOC

These also have equal energy steps in emitted photon energy.

With identical nuclei, ` is required to be even for (nuclear) spin singlet and odd for
triplet. This means steps will be larger.
A complex molecule will have three principle axes, and hence, three moments of inertia
to use in our quantized formula.
Counting degrees of freedom, which should be equal to the number of quantum numbers
needed to describe the state, we have 3 coordinates to give the position of the center
of mass, 3 for the rotational state, and 3N-6 for vibrational. This formula should be
modified if the molecule is too simple to have three principle axes.
The graph below shows the absorption coefficient of water for light of various energies.
For low energies, rotational and vibrational states cause the absorption of light. At
higher energies, electronic excitation and photoelectric effect take over. It is only in
the region around the visible spectrum that water transmits light well. Can you think
of a reason for that?

501

26. Molecular Physics

26.7

Examples

26.8

Derivations and Computations

26.9

Homework Problems

TOC

1. In HCl, absorption lines with wave numbers in inverse centimeters of 83.03,

103.73, 124.30, 145.05, 165.51 and 185.86 have been observed. Are these rotational or vibrational transitions? Estimate some physical parameters of the
molecule from these data.
2. What is the ratio of the number of HCl molecules in the j = 10 rotational state
to that in the j = 0 state if the gas is at room temperature?
502

26. Molecular Physics

26.10

TOC

Sample Test Problems

503

27. Time Dependent Perturbation Theory

TOC

Time Dependent Perturbation Theory

We have used time independent perturbation theory to find the energy shifts of states
and to find the change in energy eigenstates in the presence of a small perturbation. We
will now consider the case of a perturbation that is time dependent. Such a perturbation
can cause transitions between energy eigenstates. We will calculate the rate of those
transitions.
This material is covered in Gasiorowicz Chapter 21, in Cohen-Tannoudji et al.
Chapter XIII, and briefly in Griffiths Chapter 9.

27.1

General Time Dependent Perturbations

Assume that we solve the unperturbed energy eigenvalue problem exactly: H0 n =

En n . Now we add a perturbation that depends on time, V(t). Our problem is now
inherently time dependent so we go back to the time dependent Schr
odinger
equation.
(t)
(H0 + V(t)) (t) = i~
t
P
We will expand in terms of the eigenfunctions: (t) =
ck (t)k eiEk t/~ with
k

ck (t)eiEk t/~ = hk |(t)i. The time dependent Schr

odinger equations is
X

(H0 + V(t)) ck (t)eiEk t/~ k

= i~

X ck (t)eiEk t/~
t

ck (t)eiEk t/~ (Ek + V(t)) k

X
k

V(t)ck (t)eiEk t/~ k

= i~

ck (t)
+ Ek ck (t) eiEk t/~ k
t

X ck (t)
k

eiEk t/~ k

Now dot hn | into this equation to get the time dependence of one coefficient.
X

Vnk (t)ck (t)eiEk t/~

cn (t)
t

cn (t) iEn t/~

e
t
1 X
Vnk (t)ck (t)ei(En Ek )t/~
i~

= i~
=

504

27. Time Dependent Perturbation Theory

TOC

Assume that at t = 0, we are in an initial state (t = 0) = i and hence all the

other ck are equal to zero: ck = ki .

X
1
cn (t)
Vnk (t)ck (t)eink t
=
Vni (t)eini t +
t
i~
k6=i

Now we want to calculate transition rates. To first order, all the ck (t) are small
compared to ci (t) = 1, so the sum can be neglected.
(1)

cn (t)
t

c(1)
n (t)

1
=
i~

1
Vni (t)eini t
i~

eini t Vni (t0 )dt0

This is the equation to use to compute transition probabilities for a general

time dependent perturbation. We will also use it as a basis to compute transition
rates for the specific problem of harmonic potentials. Again we are assuming t is small
enough that ci has not changed much. This is not a limitation. We can deal with the
decrease of the population of the initial state later.
Note that, if there is a large energy difference between the initial and final states, a
slowly varying perturbation can average to zero. We will find that the perturbation
will need frequency components compatible with ni to cause transitions.
If the first order term is zero or higher accuracy is required, the second order term can
be computed. In second order, a transition can be made to an intermediate state k ,
(1)
then a transition to n . We just put the first order ck (t) into the sum.

X
cn (t)
1
(1)
=
Vni (t)eini t +
Vnk (t)ck (t)eink t
t
i~
k6=i

t
X
0
cn (t)
1
1
=
Vni (t)eini t +
Vnk (t) eink t eiki t Vki (t0 )dt0
t
i~
i~
k6=i

c(2)
n (t)

1
~2

t
X
k6=i 0

dt00 Vnk (t00 )eink t

dt0 eiki t Vki (t0 )

505

27. Time Dependent Perturbation Theory

1 X
c(2)
n (t) =
~2

TOC

t
dt00 Vnk (t00 )eink t

00
0

dt0 eiki t Vki (t0 )

k6=i 0

Example: Transitions of a 1D harmonic oscillator in a transient E field.

27.2

Sinusoidal Perturbations

An important case is a pure sinusoidal oscillating (harmonic) perturbation. We can

make up any time dependence from a linear combination of sine and cosine
waves. We define our perturbation carefully.

V(~r, t) = 2V (~r) cos(t) 2V cos(t) = V eit + eit
We have introduced the factor of 2 for later convenience. With that factor, we have
V times a positive exponential plus a negative exponential. As before, V depends on
position but we dont bother to write that for most of our calculations.
Putting this perturbation into the expression for cn (t), we get
cn (t)

1
i~

eini t Vni (t0 )dt0

1
Vni
i~

t
dt0 eini t

eit + eit

1
Vni
i~

t
dt0

ei(ni +)t + ei(ni )t

Note that the terms in the time integral will average to zero unless one of the
exponents is nearly zero. If one of the exponents is zero, the amplitude to be in
the state n will increase with time. To make an exponent zero we must have one of
two conditions satisfied.

= ni
En Ei
=
~
= Ei En

= En + ~

506

27. Time Dependent Perturbation Theory

TOC

This is energy conservation for the emission of a quantum of energy ~.

ni
En Ei
~
En Ei

En ~

This is energy conservation for the absorption of a quantum of energy ~. We can see
the possibility of absorption of radiation or of stimulated emission.
For t , the time integral of the exponential gives (some kind of) delta function
of energy conservation. We will expend some effort to determine exactly what delta
function it is.
Lets take the case of radiation of an energy quantum ~. If the initial and final
states have energies such that this transition goes, the absorption term is completely
negligible. (We can just use one of the exponentials at a time to make our formulas
simpler.)
The amplitude to be in state n as a function of time is
cn (t)

1
Vni
i~

dt0 ei(ni +)t

=
=
=
=
Pn (t)

Vni
i~

ei(ni +)t
i(ni + )

i(ni +)t

#t0 =t
t0 =0

Vni e
1
i~
i(ni + )

Vni i(ni +)t/2 ei(ni +)t/2 ei(ni +)t/2
e
i~
i(ni + )
Vni i(ni +)t/2 2 sin ((ni + )t/2)
e
i~
i(ni + )

2
Vni
4 sin2 ((ni + )t/2)
~2
(ni + )2

In the last line above we have squared the amplitude to get the probability to be in the
final state. The last formula is appropriate to use, as is, for short times. For long times
(compared to ni1+ which can be a VERY short time), the term in square brackets
looks like some kind of delta function.
We will show that the quantity in square brackets in the last equation is 2t (ni + ).
The probability to be in state n then is
Pn (t) =

2
2
2
Vni
2Vni
2Vni
2t (ni + ) =
(ni + )t =
(En Ei + ~)t
2
2
~
~
~

507

27. Time Dependent Perturbation Theory

TOC

The probability to be in the final state n increases linearly with time. There is a delta
function expressing energy conservation. The frequency of the harmonic perturbation
must be set so that ~ is the energy difference between initial and final states. This is
true both for the (stimulated) emission of a quantum of energy and for the absorption
of a quantum.
Since the probability to be in the final state increases linearly with time, it is reasonable
to describe this in terms of a transition rate. The transition rate is then given by

2
dPn
2Vni
=
(En Ei + ~)
dt
~

We would get a similar result for increasing E (absorbing energy) from the other
exponential.
2
2Vni
in =
(En Ei ~)
~
It does not make a lot of sense to use this equation with a delta function to calculate
the transition rate from a discrete state to a discrete state. If we tune the frequency
just right we get infinity otherwise we get zero. This formula is what we need if either
the initial or final state is a continuum state. If there is a free particle in the initial
state or the final state, we have a continuum state. So, the absorption or emission of
a particle, satisfies this condition.
The above results are very close to a transition rate formula known as Fermis Golden
Rule. Imagine that instead of one final state n there are a continuum of final
states. The total rate to that continuum would be obtained by integrating over final
state energy, an integral done simply with the delta function. We then have

if =

2
2Vni
f (E)
~

where f (E) is the density of final states. When particles (like photons or electrons)
are emitted, the final state will be a continuum due to the continuum of states available
to a free particle. We will need to carefully compute the density of those states, often
known as phase space.

508

27. Time Dependent Perturbation Theory

27.3
27.3.1

TOC

Examples
Harmonic Oscillator in a Transient E Field

Assume we have an electron in a standard one dimensional harmonic oscillator of

frequency in its ground state. An weak electric field is applied for a time interval T .
Calculate the probability to make a transition to the first (and second) excited state.
The perturbation is eEx for 0 < t < T and zero for other times. We can write this in
terms of the raising an lowering operators.
r
~
(A + A )
V = eE
2m
We now use our time dependent perturbation result to compute the transition probability to the first excited state.
cn (t)

1
i~

eini t Vni (t0 )dt0

1
eE
i~

~
2m

eit h1|A + A |0idt0

eE
i~

~
2m

eit dt0
0

=
=
=
=
P1

#T
0
~
eit
2m i
0
r

eE
~

eiT 1
~ 2m
r
i
eE
~ iT /2 h iT /2
e
e
eiT /2

~ 2m
r
eE
~ iT /2

e
2i sin(T /2)
~ 2m
e2 E 2 ~
4 sin2 (T /2)
~2 2 2m
2e2 E 2
sin2 (T /2)
m~ 3
eE
i~

As long as the E field is weak, the initial state will not be significantly depleted and
the assumption we have made concerning that is valid. We do see that the transition
509

27. Time Dependent Perturbation Theory

TOC

probability oscillates with the time during which the E field is applied. We would get
a (much) larger transition probability if we applied an oscillating E field tuned to have
the right frequency to drive the transition.
Clearly the probability to make a transition to the second excited state is zero in first
order. If we really want to compute this, we can use our first order result for c1 and
calculate the transition probability to the n = 2 state from that. This is a second order
calculation. Its not too bad to do since there is only one intermediate state.

27.4
27.4.1

Derivations and Computations

The Delta Function of Energy Conservation

For harmonic perturbations, we have derived a probability to be in the final state n

proportional to the following.

4 sin2 ((ni + )t/2)
Pn
(ni + )2
For simplicity of analysis lets consider the characteristics of the function

4 sin2 ((ni + )t/2)
4 sin2 (t/2)
g( ni + ) =

2
2
(ni + ) t
2 t2
1
for values of t >>
. (Note that we have divided our function to be investigated by
2
t . For = 0, g() = 1 while for all other values for , g() approaches zero for
large t. This is clearly some form of a delta function.

510

27. Time Dependent Perturbation Theory

TOC

To find out exactly what delta function it is, we need to integrate over .

d f ()g()

f ( = 0)

d g()

f ( = 0)

4 sin2 (t/2)
2 t 2

4 sin2 (y)
4y 2

f ( = 0)

2
t
2
t

sin2 (y)
y2

We have made the substitution that y = t

2 . The definite integral over y just gives
(consult your table of integrals), so the result is simple.

d f ()g()

f ( = 0)

2
t

2
g() =
()
t

4 sin2 ((ni + )t/2)
= 2t (ni + )
(ni + )2
Q.E.D.

27.5

Homework Problems

1. A hydrogen atom is placed in an electric field which is uniform in space and

~
turns on at t = 0 then decays exponentially. That is, E(t)
= 0 for t < 0 and
t
~
~
E(t) = E0 e
for t > 0. What is the probability that, as t , the hydrogen
atom has made a transition to the 2p state?
2. A one dimensional harmonic oscillator is in its ground state. It is subjected to the
additional potential W = ex for a a time interval . Calculate the probability
to make a transition to the first excited state (in first order). Now calculate the
probability to make a transition to the second excited state. You will need to
calculate to second order.
511

27. Time Dependent Perturbation Theory

27.6

TOC

Sample Test Problems

1. A hydrogen atom is in a uniform electric field in the z direction which turns on

abruptly at t = 0 and decays exponentially as a function of time, E(t) = E0 et/ .
The atom is initially in its ground state. Find the probability for the atom to
have made a transition to the 2P state as t . You need not evaluate the
radial part of the integral. What z components of orbital angular momentum are
allowed in the 2P states generated by this transition?

512

28. Radiation in Atoms

TOC

Radiation in Atoms

Now we will go all the way back to Planck who proposed that the emission of radiation
be in quanta with E = ~ to solve the problem of Black Body Radiation. So far, in
our treatment of atoms, we have not included the possibility to emit or absorb real
photons nor have we worried about the fact that Electric and Magnetic fields are
made up of virtual photons. This is really the realm of Quantum Electrodynamics, but
we do have the tools to understand what happens as we quantize the EM field.
We now have the solution of the Harmonic Oscillator problem using operator methods.
Notice that the emission of a quantum of radiation with energy of ~ is like the
raising of a Harmonic Oscillator state. Similarly the absorption of a quantum of
radiation is like the lowering of a HO state. Planck was already integrating over an
infinite number of photon (like HO) states, the same integral we would do if we had
an infinite number of Harmonic Oscillator states. Planck was also correctly counting
this infinite number of states to get the correct Black Body formula. He did it by
considering a cavity with some volume, setting the boundary conditions, then letting
the volume go to infinity.
This material is covered in Gasiorowicz Chapter 22, in Cohen-Tannoudji et al.
Chapter XIII, and briefly in Griffiths Chapter 9.

28.1

The Photon Field in the Quantum Hamiltonian

The Hamiltonian for a charged particle in an ElectroMagnetic field is given by

1
e ~ 2
H=
p~ + A
+ V (r).
2m
c
Lets assume that there is some ElectroMagnetic field around the atom. The
field is not extremely strong so that the A2 term can be neglected (for our purposes) and
~ = ~
~ A
~ = 0. The Hamiltonian
we will work in the Coulomb gauge for which p~ A
i
then becomes
p2
e ~
H
+
A p~ + V (r).
2m mc
Now we have a potentially time dependent perturbation that may drive transitions between the atomic states.
e ~
V=
A p~
mc
Lets also assume that the field has some frequency and corresponding wave vector
~k. (In fact, and arbitrary field would be a linear combination of many frequencies,
513

28. Radiation in Atoms

TOC

directions of propagation, and polarizations.)

~ r, t) 2A
~ 0 cos(~k ~r t)
A(~
~ 0 is a real vector and we have again introduced the factor of 2 for convenience
where A
of splitting the cosine into two exponentials.
We need to quantize the EM field into photons satisfying Plancks original hypothesis,
E = ~. Lets start by writing A in terms of the number of photons in the field
(at frequency and wave vector ~k). Using classical E&M to compute the energy in
a fieldclassical E&M to compute the energy in a field (See Section represented by a
~ r, t) = 2A
~ 0 cos(~k ~r t), we find that the energy inside a volume
vector potential A(~
V is
2
V |A0 |2 = N ~.
Energy =
2c2
~ in terms of the number of photons
We may then turn this around and write A
N.
|A0 |2

~ r, t)
A(~

2c2
2~c2 N
=
2 V
V

12

2
2~c N
2 cos(~k ~r t)
V

1

2~c2 N 2 i(~k~rt)
~
e
+ ei(k~rt)
V

N ~

We have introduced the unit vector to give the direction (or polarization) of the
vector potential. We now have a perturbation that may induce radiative transitions.
There are terms with both negative and positive so that we expect to see both
stimulated emission of quanta and absorption of quanta in the the presence of
a time dependent EM field.
But what about decays of atoms with no applied field? Here we need to go beyond our
classical E&M calculation and quantize the field. Since the terms in the perturbation
above emit or absorb a photon, and the photon has energy ~, lets assume the
number of photons in the field is the n of a harmonic oscillator. It has the
right steps in energy. Essentially, we are postulating that the vacuum contains an
infinite number of harmonic oscillators, one for each wave vector (or frequency...) of
light.
We now want to go from a classical harmonic oscillator to a quantum oscillator, in
which the ground state energy is not zero, and the hence the perturbing field is never
really zero. We do this by changing N to N +1 in the term that creates a photon
in analogy to the raising operator A in the HO. With this change, our perturbation
becomes
1

2~c2 2 i(~k~rt)
~
~
N e
+ N + 1ei(k~rt)
A(~r, t) =
V
514

28. Radiation in Atoms

TOC

Remember that one exponential corresponds to the emission of a photon and the other
~ as an operator which either
corresponds to the the absorption of a photon. We view A
creates or absorbs a photon, raising or lowering the harmonic oscillator in the vacuum.
Now there is a perturbation even with no applied field (N = 0).

VN =0 = VN =0 eit =

1
e 2~c2 2 i(~k~rt)
e ~
A p~ =
e
p~
mc
mc V

We can plug this right into our expression for the decay rate (removing the eit into
the delta function as was done when we considered a general sinusoidal time dependent
perturbation). Of course we have this for all frequencies, not just the one we have been
assuming without justification. Also note that our perturbation still depends on
the volume we assume. This factor will be canceled when we correctly compute the
density of final states.
We have taken a step toward quantization of the EM field, at least when we emit
or absorb a photon. With this step, we can correctly compute the EM transition
rates in atoms. Note that we have postulated that the vacuum has an infinite number
of oscillators corresponding to the different possible modes of EM waves. When we
quantize these oscillators, the vacuum has a ground state energy density in the EM field
(equivalent to half a photon of each type). That vacuum EM field is then responsible for
the spontaneous decay of excited states of atoms through the emission of a photon. We
have not yet written the quantum equations that the EM field must satisfy, although
they are closely related to Maxwells equations.

28.2

Decay Rates for the Emission of Photons

Our expression for the decay rate of an initial state i into some particular final
state n is
2
2Vni
in =
(En Ei + ~).
~
The delta function reminds us that we will have to integrate over final states to get
a sensible answer. Nevertheless, we proceed to include the matrix element of the
perturbing potential.
Taking out the harmonic time dependence (to the delta function) as before, we have
the matrix element of the perturbing potential.

1
e ~
e 2~c2 2
~
Vni = hn |
A p~|i i =
hn |eik~r p~|i i
mc
mc V
515

28. Radiation in Atoms

TOC

We just put these together to get

2 e2
2~c2
~
in =
|hn |eik~r p~|i i|2 (En Ei + ~)
~ m2 c2
V
(2)2 e2
~
|hn |eik~r p~|i i|2 (En Ei + ~)
in =
m2 V

We must sum (or integrate) over final states. The states are distinguishable so we
add the decay rates, not the amplitudes. We will integrate over photon energies and
directions, with the aid of the delta function. We will sum over photon polarizations.
We will sum over the final atomic states when that is applicable. All of this is quite
doable. Our first step is to understand the number of states of photons as Planck (and
even Rayleigh) did to get the Black Body formulas.

28.3

Phase Space: The Density of Final States

We have some experience with calculating the number of states for fermions in a 3D
box. For the box we had boundary conditions that the wavefunction go to zero at
the wall of the box. Now we wish to know how many photon states are in a region
of phase space centered on the wave vector ~k with (small) volume in k-space of
d3~k. (Remember = |~k|c for light.) We will assume for the sake of calculation that
the photons are confined to a cubic volume in position space of V = L3 and impose
periodic boundary conditions on our fields. (Really we could require the fields to
be zero on the boundaries of the box by choosing a sine wave. The PBC are equivalent
to this but allow us to deal with single exponentials instead of real functions.) Our
final result, the decay rate, will be independent of volume so we can let the volume go
to infinity.
kx L = 2nx

dnx =

ky L = 2ny

dny =

kz L = 2nz

dnz =
3

d n=

L
2 dkx
L
2 dky
L
2 dkz

L3
3
(2)3 d k

V
3
(2)3 d k

That was easy. We will use this phase space formula for decays of atoms emitting
a photon. A more general phase space formula based on our calculation can be used
with more than one free particle in the final state. (In fact, even our simple case,
the atom recoils in the final state, however, its momentum is fixed due to momentum
conservation.)

516

28. Radiation in Atoms

28.4

TOC

Total Decay Rate Using Phase Space

Now we are ready to sum over final (photon) states to get the total transition rate.
Since both the momentum of the photon and the electron show up in this equation, we
will label the electrons momentum to avoid confusion.
X
X V d3 k
X V d3 p
tot =
in
in =
in
(2)3
(2~)3
pol.
pol.
~
k,pol
X V d3 p (2)2 e2
~
|hn |eik~r () p~e |i i|2 (En Ei + ~)
=
(2~)3 m2 V

X d3 p
e2
~
=
|hn |eik~r () p~e |i i|2 (En Ei + ~)
2~3 m2

X p2 d(~)d ~
e2
~
=
|hn |eik~r () p~e |i i|2 (En Ei + ~)
2~3 m2
pc
c

X
e2
~
pd(~)d |hn |eik~r () p~e |i i|2 (En Ei + ~)
=
2~2 m2 c2

2
X
Ei En
e
~
=
d |hn |eik~r () p~e |i i|2
2~2 m2 c2
c

tot =

e2 (Ei En ) X
2~2 m2 c3

d |hn |eik~r () p~e |i i|2

This is the general formula for the decay rate emitting one photon. Depending on the
problem, we may also need to sum over final states of the atom. The two polarizations
are transverse to the photon direction, so they must vary inside the integral.
A quick estimate of the decay rate of an atom gives
50 psec.

28.5

Electric Dipole Approximation and Selection Rules

~
We can now expand the eik~r 1 i~k ~r + ... term to allow us to compute matrix
~
elements more easily. Since k ~r 2 and the matrix element is squared, our expansion

517

28. Radiation in Atoms

TOC

will be in powers of 2 which is a small number. The dominant decays will be those
from the zeroth order approximation which is
~

eik~r 1.
This is called the Electric dipole approximation.
In this Electric Dipole approximation, we can make general progress on computation
p2
+ V and [V, ~r] = 0,
of the matrix element. If the Hamiltonian is of the form H = 2m
then
~ p
[H, ~r] =
im
and we can write p~ =

im
r]
~ [H, ~

in terms of the commutator.

hn |eik~r p~e |i i

hn |~
pe |i i
im
hn |[H, ~r]|i i
=
~
im
=
(En Ei )
hn |~r|i i
~
im(En Ei )
=
hn |
~r|i i
~

This equation indicates the origin of the name Electric Dipole: the matrix element is
of the vector ~r which is a dipole.
We can proceed further, with the angular part of the (matrix element) integral.

hn |
~r|i i =

drRn n `n Rni `i

dY`n mn ~rY`i mi

dY`n mn rY`i mi

r = x sin cos + y sin sin + z cos

r

4
x + iy
x + iy

z Y10 +
Y11 +
Y11
=
3
2
2
r

4
x + iy
x + iy

hn |
~r|i i =
Y11 +
Y11
r3 drRn n `n Rni `i dY`n mn z Y10 +
3
2
2
0

At this point, lets bring all the terms in the formula back together so we know what
518

28. Radiation in Atoms

TOC

we are doing.
tot

e2 (Ei En ) X
2~2 m2 c3

d |hn |eik~r () p~e |i i|2

im(En Ei )
e (Ei En ) X

d
hn |
~r|i i
2~2 m2 c3
~

3 X
in
d |hn |
~r|i i|2
2c2
2

=
=

This is a useful version of the total decay rate formula to remember.

tot

3 X
in
=
d |hn |
~r|i i|2
2c2

We proceed with the calculation to find the E1 selection rules.

r
2
4
3 X

+
i
in

+
i

x
y
x
y

d
n z Y10 +
=
Y11 +
Y11 i
2

2c
3
2
2

3 X
4
x + iy
x + i
in
3

r drRnn `n Rni `i dY`n mn z Y10 +

Y11 +
d
=
2c2
3
2
2

We will attempt to clearly separate the terms due to hn |

~r|i i for the sake of modularity of the calculation.

The integral with three spherical harmonics in each term looks a bit difficult,
but, we can use a Clebsch-Gordan series like the one in addition of angular
momentum to help us solve the problem. We will write the product of two spherical
harmonics in terms of a sum of spherical harmonics. Its very similar to adding the
angular momentum from the two Y s. Its the same series as we had for addition
of angular momentum (up to a constant). (Note that things will be very simple
if either the initial or the final state have ` = 0, a case we will work out below for
transitions to s states.) The general formula for rewriting the product of two spherical
harmonics (which are functions of the same coordinates) is
s
`X
1 +`2
(2`1 + 1)(2`2 + 1)
Y`1 m1 (, )Y`2 m2 (, ) =
h`0|`1 `2 00ih`(m1 +m2 )|`1 `2 m1 m2 iY`(m1
4(2` + 1)
`=|`1 `2 |

519

28. Radiation in Atoms

TOC

The square root and h`0|`1 `2 00i can be thought of as a normalization constant in an
otherwise normal Clebsch-Gordan series. (Note that the normal addition of the orbital
angular momenta of two particles would have product states of two spherical harmonics
in different coordinates, the coordinates of particle one and of particle two.) (The
derivation of the above equation involves a somewhat detailed study of the properties
of rotation matrices and would take us pretty far off the current track (See Merzbacher
page 396).)
First add the angular momentum from the initial state (Yì mi ) and the photon (Y1m )
using the Clebsch-Gordan series, with the usual notation for the Clebsch-Gordan
coefficients h`n mn |ì 1mi mi.
s
`X
i +1
3(2ì + 1)
h`0|ì 100ih`(m + mi )|ì 1mi miY`(mi +m) (, )
Y1m (, )Yì mi (, ) =
4(2` + 1)
`=|ì 1|
s

3(2`i + 1)

dY`n mn Y1m Y`i mi =

h`n 0|`i 100ih`n mn |`i 1mi mi
4(2`n + 1)

x + iy
x + iy

Y11 +
Y11 Y`i mi
z Y10 +
2
2
s

3(2`i + 1)
x + iy
x + iy

h`n mn |`i 1mi 1i +

h`n
=
h`n 0|`i 100i z h`n mn |`i 1mi 0i +
4(2`n + 1)
2
2

dY`n mn

I remind you that the Clebsch-Gordan coefficients in these equations are just numbers
which are less than one. They can often be shown to be zero if the angular momentum doesnt add up. The equation we derive can be used to give us a great deal of
information.
s

(2ì + 1)
hn |
~r|i i
=
h`n 0|ì 100i r3 drRn n `n Rni ì
(2`n + 1)
0

x + iy
x + iy

h`n mn |`i 1mi 1i +

h`n mn |ì 1mi
z h`n mn |ì 1mi 0i +
2
2
We know, from the addition of angular momentum, that adding angular momentum
1 to `1 can only give answers in the range |`1 1| < `n < `1 + 1 so the change in in
` between the initial and final state can only be ` = 0, 1. For other values, all the
Clebsch-Gordan coefficients above will be zero.
We also know that the Y1m are odd under parity so the other two spherical harmonics
must have opposite parity to each other implying that `n 6= ì , therefore
` = 1.
520

28. Radiation in Atoms

TOC

We also know from the addition of angular momentum that the z components just add
like integers, so the three Clebsch-Gordan coefficients allow
m = 0, 1.

We can also easily note that we have no operators which can change the spin here. So
certainly
s = 0.
We actually havent yet included the interaction between the spin and the field in our
calculation, but, it is a small effect compared to the Electric Dipole term.
The above selection rules apply only for the Electric Dipole (E1) approximation. Higher
order terms in the expansion, like the Electric Quadrupole (E2) or the Magnetic Dipole
(M1), allow other decays but the rates are down by a factor of 2 or more. There is one
absolute selection rule coming from angular momentum conservation, since the photon
is spin 1. No j = 0 to j = 0 transitions in any order of approximation.
As a summary of our calculations in the Electric Dipole approximation, lets write out
the decay rate formula.

28.6

Explicit 2p to 1s Decay Rate

Starting from the summary equation for electric dipole transitions, above,
r

3 X
4
x + iy
x + i
in
3

d
r
drR
R
dY
Y11 +
tot =

n
`
nn `n
`n mn z Y10 +
i i

2
2c
2
2
3

we specialize to the 2p to 1s decay,

3 X
4
in
x + iy
x + iy
3

tot =
d
r
drR
R
dY
Y11 +
Y11

10 21
00 z Y10 +
2c2
3
2
2

521

28. Radiation in Atoms

TOC

perform the radial integration,

r3 drR10
R21

" 3
#"
#
52
2
1
1
1

r3 dr 2
er/a0
rer/2a0
a0
24 a0

1
a0

r4 dre3r/2a0

1
a0

2a0
3

x4 dxex

5
2
1
=
a0 (4!)
6 3
5

2
a0
= 4 6
3
and perform the angular integration.

x + iy
x + iy

d Y00
z Y10 +
Y11 +
Y11 Y1mi
2
2

x + iy
x + iy
1

d z Y10 +
Y11 +
Y11 Y1mi
=
4
2
2

x + iy
x + iy
1

z mi 0 +
mi (1) +
mi 1
=
4
2
2
2

x + iy
x + iy

d Y00

Y11 +
Y11 Y1mi
z Y10 +

2
2

1
1
2
2
2
=
m 0 + (x + y )(mi (1) + mi 1 )
4 z i
2

Lets assume the initial state is unpolarized, so we will sum over mi and divide by 3,

522

28. Radiation in Atoms

TOC

the number of different mi allowed.

2

x + iy
1 X
x + iy

d Y`n mn
z Y10 +
Y11 +
Y11 Y`i mi
3 m
2
2
i

X
1 1
1 2
2
2
=
z mi 0 + (x + y )(mi (1) + mi 1 )
4 3 m
2
i

1
1 2
2
2
+ ( + y )(1 + 1)
=
12 z 2 x

1
=
2z + 2x + 2y
12
1
=
12

Our result is independent of photon polarization since we assumed the initial state was
unpolarized, but, we must still sum over photon polarization. Lets assume that we
are not interested in measuring the photons polarization. The polarization vector is
constrained to be perpendicular to the photons direction
~kp = 0
so there are two linearly independent polarizations to sum over. This just introduces
a factor of two as we sum over final polarization states.
The integral over photon direction clearly just gives a factor of 4 since there is no
direction dependence left in the integrand (due to our assumption of an unpolarized
initial state).
tot

28.7

2 5 2 1
3
3
2in
4in

=
(2)(4)
6
a
=
4

0
12

3c2
3
9c2

2 5 2

a0
4 6

3

General Unpolarized Initial State

If we are just interested in the total decay rate, we can go further. The decay rate should
not depend on the polarization of the initial state, based on the rotational symmetry
of our theory. Usually we only want the total decay rate to some final state so we sum
over polarizations of the photon, integrate over photon directions, and (eventually)
sum over the different mn of the final state atoms. We begin with a simple version of

523

28. Radiation in Atoms

TOC

the total decay rate formula in the E1 approximation.

3 X
in
tot =
d |hn |
~r|i i|2
2c2

3 X
in
tot =
d |hn |~r|i i |2
2c2

3 X
in
tot =
d |~rni |2
2c2

3 X
in
tot =
d |~rni |2 cos2
2c2

Where is the angle between the matrix element of the position vector ~rni and the
polarization vector . It is far easier to understand the sum over polarizations in terms
of familiar vectors in 3-space than by using sums of Clebsch-Gordan coefficients.
Lets pick two transverse polarization vectors (to sum over) that form a right handed
system with the direction of photon propagation.
(1) (2) = k
The figure below shows the angles, basically picking the photon direction as the polar
axis, and the (1) direction as what is usually called the x-axis.

524

28. Radiation in Atoms

TOC

The projection of the vector ~rni into the transverse plan gives a factor of sin . It is
then easy to see that
cos 1

sin cos

cos 2

sin sin

The sum of cos2 over the two polarizations then just gives sin2 . Therefore the decay

525

28. Radiation in Atoms

TOC

rate becomes
tot

tot

3 X
in
d |~rni |2 cos2
2c2

3
in
2
|~rni |
d sin2
2c2

3
in
2
|~rni | 2 d(cos ) sin2
2c2
1
3
in
2
|~rni | 2 d(cos )(1 cos2 )
2c2
1

3
in
|~rni |2 2
2c2

tot

3
in
|~rni |2 2
2c2

tot

1
dx(1 x2 )

1
x3
x
3 1

3
in
2
2
|~
r
|
2
2

ni
2c2
3
3
in
8
|~rni |2
2c2
3
3
4in
|~rni |2
3c2

This is now a very nice and simple result for the total decay rate of a state, summed
over photon polarizations and integrated over photon direction.

tot =

3
4in
|~rni |2
2
3c

We still need to sum over the final atomic states as necessary. For the case of a transition
in a single electron atom n`m n0 `0 m0 + , summed over m0 , the properties of the
Clebsch-Gordan coefficients can be used to show (See Merzbacher, second edition, page
467).

526

28. Radiation in Atoms

tot

3
4in
=
2
3c

TOC

`+1
2`+1
`

2`+1
0

2

Rn 0 `0 Rn` r3 dr

for

`0 =

`+1
`1

The result is independent of m as we would expect from rotational symmetry.

As a simple check, lets recompute the 2p to 1s decay rate for hydrogen. We must
choose the `0 = ` 1 case and ` = 1.

2

2

3
3

4in
`
4
in

3

tot =
R10 R21 r dr =
R10 R21 r dr

2
2
3c 2` + 1
9c

0

This is the same result we got in the explicit calculation.

28.8

Angular Distributions

We may also deduce the angular distribution of photons from our calculation. Lets
take the 2p to 1s calculation as an example. We had the equation for the decay rate.
r

3 X
4
x + iy
in
x + iy
3

r
drR
R
dY
tot =
d
Y11 +
Y11

10 21
00 z Y10 +
2c2
3
2
2

We have performed that radial integration which will be unchanged. Assume that we
start in a polarized state with mi = 1. We then look at our result for the angular
integration in the matrix element

2

x + iy
x + iy

dY00

z Y10 +
Y11 +
Y11 Y1mi

2
2

1
1
2
2
2
=
m 0 + (x + y )(mi (1) + mi 1 )
4 z i
2

1 1 2
2
( + y )
=
4 2 x
where we have set mi = 1 eliminating two terms.
Lets study the rate as a function of the angle of the photon from the z axis, . The
rate will be independent of the azimuthal angle. We see that the rate is proportional
to 2x + 2y . We still must sum over the two independent transverse polarizations.
527

28. Radiation in Atoms

TOC

For clarity, assume that = 0 and the photon is therefore emitted in the x-z plane.
One transverse polarization can be in the y direction. The other is in the x-z plane
perpendicular to the direction of the photon. The x component is proportional to
cos . So the rate is proportional to 2x + 2y = 1 + cos2 .
If we assume that mi = 0 then only the z term remains and the rate is proportional
to 2z . The angular distribution then goes like sin2 .

28.9

Vector Operators and the Wigner Eckart Theorem

There are some general features that we can derive about operators which are vectors,
that is, operators that transform like a vector under rotations. We have seen in the
sections on the Electric Dipole approximation and subsequent calculations that the
vector operator ~r could be written as its magnitude r and the spherical harmonics
Y1m . We found that the Y1m could change the orbital angular momentum (from initial
to final state) by zero or one unit. This will be true for any vector operator.
In fact, because the vector operator is very much like adding an additional ` = 1 to the
initial state angular momentum, Wigner and Eckart proved that all matrix elements
of vector operators can be written as a reduced matrix element which does not
depend on any of the m, and Clebsch-Gordan coefficients. The basic reason for this is
that all vectors transform the same way under rotations, so all have the same angular
properties, being written in terms of the Y1m .
~ in terms of the spherical harmonics using
Note that it makes sense to write a vector V
V =

Vx iVy

and
V0 = Vz .
We have already done this for angular momentum operators.
Lets consider our vector V q where the integer q runs from -1 to +1. The Wigner-Eckart
theorem says
h0 j 0 m0 |V q |jmi = hj 0 m0 |j1mqih0 j 0 ||V ||ji
Here represents all the (other) quantum numbers of the state, not the angular momentum quantum numbers. jm represent the usual angular momentum quantum numbers
of the states. h0 j 0 ||V ||ji is a reduced matrix element. Its the same for all values
of m and q. (Its easy to understand that if we take a matrix element of 10r it will
be 10 times the matrix element of r. Nevertheless, all the angular part is the same.
This theorem states that all vectors have essentially the same angular behavior. This
theorem again allows us to deduce that ` = 1, 0. + 1.
528

28. Radiation in Atoms

TOC

The theorem can be generalized for spherical tensors of higher (or even lower) rank
than a vector.

28.10

Exponential Decay

We have computed transition rates using our theory of radiation. In doing this, we
have assumed that our calculations need only be valid near t = 0. More specifically, we
have assumed that we start out in some initial state i and that the amplitude to be in
that initial state is one. The probability to be in the initial state will become depleted
for times on the order of the lifetime of the state. We can account for this in terms of
the probability to remain in the initial state.
Assume we have computed the total transition rate.
X
tot =
in
n

This transition rate is the probability per unit time to make a transition away from
the initial state evaluated at t = 0. Writing this as an equation we have.

dPi
= tot
dt t=0
For larger times we can assume that the probability to make a transition away from
the initial state is proportional to the probability to be in the initial state.
dPi (t)
= tot Pi (t)
dt
The solution to this simple first order differential equation is
Pi (t) = Pi (t = 0)etot t
If you are having any trouble buying this calculation, think of a large ensemble of
hydrogen atoms prepared to be in the 2p state at t = 0. Clearly the number of atoms
remaining in the 2p state will obey the equation
dN2p (t)
= tot N2p (t)
dt
and we will have our exponential time distribution.
We may define the lifetime of a state to the the time after which only
state remains.
1
=
tot

1
e

of the decaying

529

28. Radiation in Atoms

28.11

TOC

Lifetime and Line Width

Now we have computed the lifetime of a state. For some atomic, nuclear, or particle
states, this lifetime can be very short. We know that energy conservation can be
violated for short times according to the uncertainty principle
Et

~
.
2

This means that a unstable state can have an energy width on the order of
E

~tot
.
2

We may be more quantitative. If the probability to be in the initial state is proportional

to et , then we have
|i (t)|2 = et
i (t) et/2
i (t) eiEi t/~ et/2
We may take the Fourier transform of this time function to the the amplitude as a
function of frequency.

i (t)eit dt

i ()
0

eiEi t/~ et/2 eit dt

ei0 t et/2 eit dt

=
0

ei(0 +i 2 )t dt

=
0

"
=
=

1
ei(0 +i 2 )t
i( 0 + i 2 )

#
0

i
( 0 + i 2 )

We may square this to get the probability or intensity as a function of (and hence
E = ~).
1
Ii () = |i ()|2 =
2
( 0 )2 + 4
530

28. Radiation in Atoms

TOC

This gives the energy distribution of an unstable state. It is called the Breit-Wigner
line shape. It can be characterized by its Full Width at Half Maximum (FWHM) of
.
The Breit-Wigner will be the observed line shape as long as the density of final states
is nearly constant over the width of the line.
As 0 this line shape approaches a delta function, ( 0 ).
For the 2p to 1s transition in hydrogen, weve calculated a decay rate of 0.6 109 per
second. We can compute the FWHM of the width of the photon line.
E = ~ =

(1.05 1027 erg sec)(0.6 109 sec1 )

0.4 106 eV
1.602 1012 erg/eV

Since the energy of the photon is about 10 eV, the width is about 107 of the photon
energy. Its narrow but not enough for example make an atomic clock. Weaker transitions, like those from E2 or M1 will be relatively narrower, allowing use in precision
systems.

28.11.1

Other Phenomena Influencing Line Width

We have calculated the line shape due to the finite lifetime of a state. If we attempt to
measure line widths, other phenomena, both of a quantum and non-quantum nature,
can play a role in the observed line width. These are:
Collision broadening,
Doppler broadening, and
Recoil.
Collision broadening occurs when excited atoms or molecules have a large probability
to change state when they collide with other atoms or molecules. If this is true, and
it usually is, the mean time to collision is an important consideration when we are
assessing the lifetime of a state. If the mean time between collisions is less than the
lifetime, then the line-width will be dominated by collision broadening.
An atom or molecule moving through a gas sweeps through a volume per second proportional to its cross section and velocity. The number of collisions it will have per
second is then
c = Ncollision/sec = nv
531

28. Radiation in Atoms

TOC

where n is the number density of molecules to collide with per unit volume. We can
estimate the velocity from the temperature.
1
3
mv 2 = kT
2
r2
3kT
vRM S =
m
r
3kT
c = n

m
The width due to collision broadening increases with he pressure of the gas. It also
depends on temperature. This is basically a quantum mechanical effect broadening a
state because the state only exists for a short period of time.
Doppler broadening is a simple non-quantum effect. We know that the frequency of
photons is shifted if the source is moving shifted higher if the source is moving toward
the detector, and shifted lower if it is moving away.
vk
=
r c
p
kT /m

kT
=
=

c
mc2
This becomes important when the temperature is high.
Finally, we should be aware of the effect of recoil. When an atom emits a photon, the
atom must recoil to conserve momentum. Because the atom is heavy, it can carry a
great deal of momentum while taking little energy, still the energy shift due to recoil
can be bigger than the natural line width of a state. The photon energy is shifted
downward compared to the energy difference between initial and final atomic states.
This has the consequence that a photon emitted by an atom will not have the right
energy to be absorbed by another atom, raising it up to the same excited state that
decayed. The same recoil effect shifts the energy need to excite a state upward. Lets
do the calculation for Hydrogen.

p~H = p~
E
p
c
p2
E2
=
=
2mp
2mp c2
E
E
=
E
2mp c2

For our 2p to 1s decay in Hydrogen, this is about 10 eV over 1860 MeV, or less than
one part in 108 . One can see that the effect of recoil becomes more important as the
energy radiated increases. The energy shift due to recoil is more significant for nuclear
decays.
532

28. Radiation in Atoms

TOC

28.12

Phenomena of Radiation Theory

28.12.1

The M
ossbauer Effect

In the case of the emission of x-rays from atoms, the recoil of the atom will shift the
energy of the x-ray so that it is not reabsorbed. For some experiments it is useful to be
able to measure the energy of the x-ray by reabsorbing it. One could move the detector
at different velocities to find out when re-absorption was maximum and thus make a
very accurate measurement of energy shifts. One example of this would be to measure
the gravitational red (blue) shift of x-rays.
Mossbauer discovered that atoms in a crystal need not recoil significantly. In fact,
the whole crystal, or at least a large part of it may recoil, making the energy shift
very small. Basically, the atom emitting an x-ray is in a harmonic oscillator (ground)
state bound to the rest of the crystal. When the x-ray is emitted, there is a good
chance the HO remains in the ground state. An analysis shows that the probability is
approximately
P0 = eErecoil /~HO
Thus a large fraction of the radiation is emitted (and reabsorbed) without a large
energy shift. (Remember that the crystal may have 1023 atoms in it and that is a large
number.
The M
ossbauer effect has be used to measure the gravitational red shift on earth. The
red shift was compensated by moving a detector, made from the same material as the
emitter, at a velocity (should be equal to the free fall velocity). The blue shift was
measured to be

= (5.13 0.51) 1015

when 4.92 1015 was expected based upon the general principle of equivalence.

28.12.2

LASERs

Light Amplification through Stimulated Emission of Radiation is the phenomenon with

the acronym LASER. As the name would indicate, the LASER uses stimulated emission
to genrate an intense pulse of light. Our equations show that the decay rate of a state
by emission of a photon is proportional to the number (plus one) of photons in the field
(with the same wave-number as the photon to be emitted).
~ r, t)
A(~

2~c2
V

~
~
N ei(k~rt) + N + 1ei(k~rt)

Here plus one is not really important since the number of photons is very large.
533

28. Radiation in Atoms

TOC

Lets assume the material we wish to use is in a cavity. Assume this material has an
excited state that can decay by the emission of a photon to the ground state. In normal
equilibrium, there will be many more atoms in the ground state and transitions from
one state to the other will be in equilibrium and black body radiation will exist in the
cavity. We need to circumvent equilibrium to make the LASER work. To cause many
more photons to be emitted than are reabsorbed a LASER is designed to produce a
poplation inversion. That is, we find a way to put many more atoms in the excited
state than would be the case in equilibrium.
If this population inversion is achieved, the emission from one atom will increase the
emission rate from the other atoms and that emission will stimulate more. In a pulsed
laser, the population of the excited state will become depleted and the light pulse will
end until the inversion can be achieved again. If the population of the excited state
can be continuously pumped up, then the LASER can run continously.
This optical pumping to achieve a population inversion can be done in a number of
ways. For example, a Helium-Neon LASER has a mixture of the two gasses. If a high
voltage is applied and an electric current flows through the gasses, both atoms can be
excited. It turns out that the first and second excited states of Helium have almost
the same excitation energy as the 4s and 5s excitations of Neon. The Helium states
cant make an E1 transition so they are likely to excite a Neon atom instead. An
excited Helium atom can de-excite in a collision with a Neon atom, putting the Neon
in a highly excited state. Now there is a population inversion in the Neon. The Neon
decays more quickly so its de-excitation is dominated by photon emission.

534

28. Radiation in Atoms

TOC

Another way to get the population inversion is just the use of a metastable state as in
a ruby laser. A normal light sorce can excite a higher excited state which decays to
a metastable excited state. The metastable state will have a much larger population
than in equilibrium.
A laser with a beam coming out if it would be made in a cavity with a half silvered
mirror so that the radiation can build up inside the cavity, but some of the radiation
leaks out to make the beam.

535

28. Radiation in Atoms

TOC

28.13

Examples

28.13.1

The 2P to 1S Decay Rate in Hydrogen

28.14

Derivations and Computations

28.14.1

Energy in Field for a Given Vector Potential

We have the vector potential

~ r, t) 2A
~ 0 cos(~k ~r t).
A(~
First find the fields.
~
1 A
~
~ r t)
=2 A
0 sin(k ~
c t
c
~ A
~ = 2~k A
~ 0 sin(~k ~r t)
2

~
E

~
B

Note that, for an EM wave, the vector potential is transverse to the wave vector. The
energy density in the field is
2

1
1

2 2 2 2 ~
2
U=
E2 + B2 =
4
+
k
A20 sin2 (~k ~r t) =
A sin (k ~r t)
2
8
8
c
2c2 0
Averaging the sine square gives one half, so, the energy in a volume V is
Energy =

2 A20 V
2c2
536

28. Radiation in Atoms

28.14.2

TOC

General Phase Space Formula

If there are N particles in the final state, we must consider the number of states
available for each one. Our phase space calculation for photons was correct even for
particles with masses.
V d3 p
d3 n =
(2~)3
Using Fermis Golden Rule as a basis, we include the general phase space formula into
our formula for transition rates.

if =

Y
N
V d3 pk
k=1

(2~)3

!
2

|Mf i | Ei Ef

p~i p~f

k
X

!
p~k

In our case, for example, of an atom decaying by the emission of one photon, we have
two particles in the final state and the delta function of momentum conservation will
do one of the 3D integrals getting us back to the same result. We have not bothered
to deal with the free particle wave function of the recoiling atom, which will give the
factor of V1 to cancel the V in the phase space for the atom.

28.14.3

Estimate of Atomic Decay Rate

We have the formula

tot =

e2 (Ei En )
2~2 m2 c3

d |hn |ei(k~r) p~e |i i|2

537

28. Radiation in Atoms

TOC

Lets make some approximations.

p~
~k ~r

ei(k~r)

tot

=
=
=
=
=
=

|p| = m|v| mc = mc
1 2
mc2
~
2 mc2 ~

ka0 =
a0 2
a0 =
=
~c
~c
2~c mc
2
i
i
e 2 1+
1
2
2
e (Ei En )
(4)|mc|2
2~2 m2 c3
(Ei En )
(4)|mc|2
2~m2 c2
( 12 2 mc2 )
(4)|mc|2
2~m2 c2
5 mc2
~
5 mc2 c
~c
(0.51 MeV )3 1010 cm/sec
(1013 F/cm) 2 1010 sec1
(1375 )(197 MeV F )

This gives a life time of about 50 psec.

28.15

Homework Problems

1. The interaction term for Electric Quadrupole transitions correspond to a linear

combination of spherical harmonics, Y2m , and are parity even. Find the selection
rules for E2 transitions.
2. Magnetic dipole transitions are due to an axial vector operator and hence are
proportional to the Y1m but do not change parity (unlike a vector operator).
What are the M1 selection rules?
3. Draw the energy level diagram for hydrogen up to n = 3. Show the allowed E1
transitions. Use another color to show the allowed E2 and M1 transitions.
4. Calculate the decay rate for the 3p 1s transition.
5. Calculate the decay rate for the 3d 2p transition in hydrogen.
6. Assume that we prepare Hydrogen atoms in the n`m = 211 state. We set up an
experiment with the atoms at the origin and detectors sensitive to the polariztion
along each of the 3 coordinate axes. What is the probability that a photon with
its wave vector pointing along the axis will be Left Circularly Polarized?
538

28. Radiation in Atoms

TOC

7. Photons from the 3p 1s transition are observed coming from the sun. Quantitatively compare the natural line width to the widths from Doppler broadening
and collision broadening expected for radiation from the suns surface.

28.16

Sample Test Problems

1. A hydrogen atom is in the n = 5, 3 D 25 state. To which states is it allowed to

decay via electric dipole transitions? What will be the polarization for a photon
emitted along the z-axis if ml decreases by one unit in the decay?
2. Derive the selection rules for radiative transitions between hydrogen atom states
in the electric dipole approximation. These are rules for the change in l, m, and
s.
3. State the selection rules for radiative transitions between hydrogen atom states
in the electric dipole approximation. These are rules for the allowed changes in
l, m, s, and parity. They can be easily derived from the matrix element given on
the front of the test. Draw an energy level diagram (up to n = 3) for hydrogen
atoms in a weak B field. Show the allowed E1 transitions from n = 3 to n = 1
on that diagram.
d
4. Calculate the differential cross section, d
, for high energy scattering of particles
of momentum p, from a spherical shell delta function

V (r) = (r r0 )
Assume that the potential is weak so that perturbation theory can be used. Be
sure to write your answer in terms of the scattering angles.
5. Assume that a heavy nucleus attracts K0 mesons with a weak Yakawa potential
d
, for scattering high
V (r) = Vr0 er . Calculate the differential cross section, d
energy K0 mesons (mass mK ) from that nucleus. Give your answer in terms of
the scattering angle .

539

29. Scattering

TOC

Scattering

This material is covered in Gasiorowicz Chapter 23.

Scattering of one object from another is perhaps our best way of observing and learning
about the microscopic world. Indeed it is the scattering of light from objects and
the subsequent detection of the scattered light with our eyes that gives us the best
information about the macroscopic world. We can learn the shapes of objects as well
as some color properties simply by observing scattered light.
There is a limit to what we can learn with visible light. In Quantum Mechanics we know
that we cannot discern details of microscopic systems (like atoms) that are smaller than
the wavelength of the particle we are scattering. Since the minimum wavelength of
visible light is about 0.25 microns, we cannot see atoms or anything smaller even
with the use of optical microscopes. The physics of atoms, nuclei, subatomic particles,
and the fundamental particles and interactions in nature must be studied by scattering
particles of higher energy than the photons of visible light.
Scattering is also something that we are familiar with from our every day experience.
For example, billiard balls scatter from each other in a predictable way. We can fairly
easily calculate how billiard balls would scatter if the collisions were elastic but with
some energy loss and the possibility of transfer of energy to spin, the calculation becomes more difficult.
Let us take the macroscopic example of BBs scattering from billiard balls as
an example to study. We will motivate some of the terminology used in scattering
macroscopically. Assume we fire a BB at a billiard ball. If we miss the BB does not
scatter. If we hit, the BB bounces off the ball and goes off in a direction different from
the original direction. Assume our aim is bad and that the BB has a uniform probability
distribution over the area around the billiard ball. The area of the projection of the
billiard ball into two dimensions is just R2 if R is the radius of the billiard ball.
Assume the BB is much smaller so that its radius can be neglected for now. We might
say that R2 is the cross section of the billard ball. The cross sectional area is just
the area of the projection of the ball into 2D.
We can then say something about the probability for a scattering to occur if we know
the area of the projection of the billiard ball and number of BBs per unit area that we
shot.
N
Nscat = R2
A
Where N is the number of BBs we shot, A is the area over which they are spread, and
R is the radius of the billiard ball.

540

29. Scattering

TOC

Figure 48: Classical scattering from a Billard Ball. The total scattering rate depends
on the cross sectional area of the ball.
The incident flux is the number of particles per unit area per unit time in the beam.
This is a well defined quantity in quantum mechanics, |~j|. The cross section is the
projected area of the billiard ball in this case. It may be more complicated in other
cases. For example, if we do not neglect the radius r of the BB, the cross section for
scattering is
= (R + r)2 .
Clearly there is more information available from scattering than whether a particle scatters or not. For example, Rutherford discovered that atomic nucleus by seeing
that high energy alpha particles sometimes backscatter from a foil containing atoms.
The atomic model of the time did not allow this since the positive charge was spread
over a large volume. We measure the probability to scatter into different directions.
In the case of Rutherford, the scattering angle can be computed classically from the
impact parameter b. There will be azimuthal symmetry for Rutherford scattering
but in general, the azimuthal scattering distribution can also contain information.

541

29. Scattering

TOC

Figure 49: Classical scattering with the polar angle depending on the impact parameter
of the incoming particle. We may divide the cross sectional area of the beam that
scatters into (infinitesimal) parts scattering into elements of solid angle d.
This will also happen in the case of the BB and the billiard ball. The polar angle
of scattering will depend on the impact parameter of the incoming BB. We can
measure the scattering into some small solid angle d. The part of the cross section
d
that scatters into that solid angle can be called the differential cross section d
.
The integral over solid angle will give us back the total cross section.

d
d =
d
The idea of cross sections and incident fluxes translates well to the quantum mechanics
we are using. If the incoming beam is a plane wave, that is a beam of particles of
definite momentum or wave number, we can describe it simply in terms of the number
or particles per unit area per second, the incident flux. The scattered particle is
also a plane wave going in the direction defined by d. What is left is the interaction
between the target particle and the beam particle which causes the transition from the
initial plane wave state to the final plane wave state.
We will study scattering in quantum mechanics using two methods of calculation:
The Born Approximation: Use time dependent perturbation theory to calculate the
~0
~
transition rate from an incoming plane wave eiki ~x into an outgoing plane wave eikf ~x .
Since this is based on perturbation theory, it will be accurate for weak potentials. It
will also be good if the wavelength of the beam is small compared to the range over
which the potential varies, that is at high energy.
Partial Wave Analysis: For strong potentials with short range or at low energy, we
need a more exact solution. In principle we can solve the problem exactly in spherical
coordiantes by dividing the incoming beam up into different angular momentum terms
which are called partial waves. We solve the scattering problem for each wave. In
practice, only the low angular momentum waves have a substantial interaction with a
542

29. Scattering

TOC

short range potential. Usually, only the ` = 0 wave is important, but we could also
included ` = 1 if needed. Often, given the long wavelength of the beam and short range
of the potential, the scattering is diffractive.

29.1

The Born Approximation

For high energies relative to the inverse range of the potential, a partial wave analysis is
not helpful and it is far better to use perturbation theory. The Born approximation
is valid for high energy and weak potentials. If the potential is weak, only one or two
terms in the perturbation series need be calculated.
If we work in the usual center of mass system, we have a problem with one particle
scattering in a potential. The incoming plane wave can be written as
1 ~
i (~r) = eiki ~x .
V
The scattered plane wave is
1 ~
f (~r) = eikf ~x .
V
We can use Fermis golden rule to calculate the transition rate to first order in perturbation theory.

2
V d3~kf
2
Rif =
|hf |V (~r)|i i| (Ef Ei )
~
(2)3
The delta function expresses energy conservation for elastic scattering which we are
assuming at this point. If inelastic scattering is to be calculated, the energy of the
atomic state changes and that change should be included in the delta function and the
change in the atomic state should be included in the matrix element.
The elastic scattering matrix element is

1
1
1
~
~
~
~
d3~reikf ~x V (~r)eiki ~x =
d3~rei~x V (~r) = V ()
hf |V (~r)|i i =
V
V
V
~ = ~kf ~ki . We notice that this is just proportional to the Fourier Transform
where
of the potential.
Assuming for now non-relativistic final state particles we calculate
!

2
V df kf2 dkf 1
~2 kf2
2

~
Rif =
Ei
V ()
~
(2)3
V2
2

2
2
1
2 ~
=
d
k
V
(
)

2
f f
~ (2)3 V
~ kf

2
1
~
=
df kf V ()

4 2 ~3 V
543

29. Scattering

TOC

We now need to convert this transition rate to a cross section. Our wave functions are
normalize to one particle per unit volume and we should modify that so that there is a
flux of one particle per square centimeter per second to get a cross section. To do this
we set the volume to be V = (1 cm2 )(vrel )(1 second). The relative velocity is just
the momentum divided by the reduced mass.

1
~ 2
V
(
)
d
k
=

f
f
4 2 ~3 vrel

2

1
2 ~
=
V
(
)
d

f
4 2 ~4
d
2 ~ 2
=
V ()
d
4 2 ~4

This is a very useful formula for scattering from a weak potential or for scattering
at high energy for problems in which the cross section gets small because the Fourier
Transform of the potential diminishes for large values of k. It is not good for scattering
due to the strong interaction since cross sections are large and do not typically decrease
at high energy. Note that the matrix elements and hence the scattering amplitudes
calculated in the Born approximation are real and therefore do not satisfy the Optical
Theorem. This is a shortcoming of the approximation.

29.1.1

Scattering from a Screened Coulomb Potential

A standard Born approximation example is Rutherford Scattering, that is, Coulomb

scattering of a particle of charge Z1 e in a screened Coulomb potential (r) = Zr2 e er/a .
The exponential represents the screening of the nuclear charge by atomic electrons.
Without screening, the total Coulomb scattering cross section is infinite because the
range of the force in infinite.
The potential energy then is
V (r) =

Z1 Z2 e2 r/a
e
r

We need to calculate its Fourier Transform.

r/a
~ e
~ = Z1 Z2 e2
V ()
d3 rei~r
r
Since the potential has spherical symmetry, we can choose to be in the z direction

544

29. Scattering

TOC

and proceed with the integral.

~
V ()

1
d(cos )eir cos

Z1 Z2 e 2

r dr
1

eirx
r dr
ir
2

= Z1 Z2 e 2

= Z1 Z2 e2

2
i

er/a
r

x=1
x=1

er/a
r

dr eir eir er/a
0

2
= Z1 Z2 e
i

h
i
1
1
dr e( a +i)r e( a i)r

e( a i)r
2
e( a +i)r
+ 1
= Z1 Z2 e2
1
i
a + i
a i

1
1
2 2
= Z1 Z2 e
1
i a1 + i
a i
1

1
2
a i a i
= Z1 Z2 e2
1
2
i
a2 +

2i
2
= Z1 Z2 e2
i a12 + 2
=

#
0

4Z1 Z2 e2
1
2
a2 +

~ = ~kf ~ki , we have 2 = k 2 + k 2 2kf ki cos . For elastic scattering, 2 =

Since
i
f
2
2k (1 cos ). The differential cross section is
d
d

2 ~ 2
V ()
4 2 ~4
2

2
4Z1 Z2 e2

=
1

2
4
2
4 ~ a2 + 2k (1 cos )

2

Z1 Z2 e2

= ~2

2 + p2 (1 cos )
2a

2

Z1 Z2 e2

= ~2
2

+
4E
sin
2
=

In the last step we have used the non-relativistic formula for energy and 1 cos =
2
1
2 sin 2 .
545

29. Scattering

TOC

The screened Coulomb potential gives a finite total cross section. It corresponds well
with the experiment Rutherford did in which particles were scattered from atoms in
a foil. If we scatter from a bare charge where there is no screening, we can take the
limit in which a .

Z Z e2 2
1 2

4E sin2 2
The total cross section diverges in due to the region around zero scattering angle.

29.2

The Radial Equation and Constant Potentials *

After separation of variables, the radial equation depends on `.

"
#
2
~2 1

1
`(` + 1)
r
+

RE` (r) + V (r)RE` (r) = ERE` (r)

2 r2
r
r r
r2
It can be simplified a bit.

2
`(` + 1)
~2 2
+

RE` (r) + V (r)RE` (r) = ERE` (r)

2 r2
r r
r2
The term due to angular momentum is often included with the potential.

~2 2
2
`(` + 1)~2
R
(r)
+
V
(r)
+
+
Rn` (r) = ERn` (r)
n`
2 r2
r r
2r2

This pseudo-potential repels the particle from the origin.

29.3

Behavior at the Origin *

The pseudo-potential dominates the behavior of the wavefunction at the

origin if the potential is less singular than r12 .

~2 2
2
`(` + 1)~2
+
R
(r)
+
V
(r)
+
Rn` (r) = ERn` (r)
n`
2 r2
r r
2r2
For small r, the equation becomes
2

2
`(` + 1)
+
Rn` (r)
Rn` (r) = 0
r2
r r
r2

546

29. Scattering

TOC

The dominant term at the origin will be given by some power of r

R(r) = rs .
Higher powers of r are OK, but are not dominant. Plugging this into the equation we
get

s(s 1)rs2 + 2srs2 `(` + 1)rs2 = 0.
s(s 1) + 2s = `(` + 1)
s(s + 1) = `(` + 1)
There are actually two solutions to this equation, s = ` and s = ` 1. The first
solution, s = `, is well behaved at the origin ( regular solution). The second solution,
s = ` 1, causes normalization problems at the origin ( irregular solution).

29.4

Spherical Bessel Functions *

We will now give the full solutions in terms of

= kr.
These are written for E > V but can be are also valid for E < V where k becomes
imaginary.
= kr ir
The full regular solution of the radial equation for a constant potential for a given
` is

`
1 d
sin
`
j` () = ()
d

the spherical Bessel function. For small r, the Bessel function has the following
behavior.
`
j` ()
1 3 5 ...(2` + 1)
The full irregular solution of the radial equation for a constant potential for a given
` is

`
1 d
cos
n` () = ()`
d

the spherical Neumann function. For small r, the Neumann function has the
following behavior.
1 3 5 ...(2` + 1)
n` ()
`+1
The lowest ` Bessel functions (regular at the origin) solutions are listed below.
547

29. Scattering

TOC

j0 () =
j1 () =
j2 () =

sin

sin cos

3 sin 3 cos sin

3
2

The lowest ` Neumann functions (irregular at the origin) solutions are listed below.

n0 () =
n1 () =
n2 () =

cos

cos sin

3 cos 3 sin cos

+
3
2

The most general solution is a linear combination of the Bessel and Neumann
functions. The Neumann function should not be used in a region containing the origin.
The Bessel and Neumann functions are analogous the sine and cosine functions of
the 1D free particle solutions. The linear combinations analogous to the complex
exponentials of the 1D free particle solutions are the spherical Hankel functions.
(1)
h` ()

= j` () + in` () = ()
(2)

1 d
d

`
sin i cos
i
ei( 2 )

(1)

h` () = j` () in` () = h`

()

The functional for for large r is given. The Hankel functions of the first type are
the ones that will decay exponentially as r goes to infinity if E < V , so it is right for
bound state solutions.
The lowest ` Hankel functions of the first type are shown below.

548

29. Scattering

TOC

ei
i

ei
i
(1)
h1 () =
1+

(1)

h0 () =

(1)
h2 ()

iei
=

3i
3
1+
2

We should also give the limits for large r, ( >> `),of the Bessel and Neumann
functions.

sin `
2
j` ()

cos `
2
n` ()

Decomposing the sine in the Bessel function at large r, we see that the Bessel function
is composed of an incoming spherical wave and an outgoing spherical wave of the same
magnitude.

1 i(kr`/2)
j` ()
e
ei(kr`/2)
2ikr
This is important. If the fluxes were not equal, probability would build up at the origin.
All our solutions must have equal flux in and out.

29.5

Particle in a Sphere *

This is like the particle in a box except now the particle is confined to the inside of
a sphere of radius a. Inside the sphere, the solution is a Bessel function. Outside the
sphere, the wavefunction is zero. The boundary condition is that the wave function go
to zero on the sphere.
j` (ka) = 0
There are an infinite number of solutions for each `. We only need to find
q the zeros of
2E
the Bessel functions. The table below gives the lowest values of ka = 2ma
which
~2
satisfy the boundary condition.

549

29. Scattering

TOC

`
0
1
2
3
4
5

n=1
3.14
4.49
5.72
6.99
8.18
9.32

n=2
6.28
7.73
9.10
10.42

n=3
9.42

We can see both angular and radial excitations.

29.6

Bound States in a Spherical Potential Well *

We now wish to find the energy eigenstates for a spherical potential well of radius
a and potential V0 .

We must use the Bessel function near the origin.

Rn` (r) = Aj` (kr)
r
2(E + V0 )
k=
~2
We must use the Hankel function of the first type for large r.
= kr ir
r
2E
=
~2
(1)

Rn` = Bh` (ir)

550

29. Scattering

TOC

To solve the problem, we have to match the solutions at the boundary. First match
the wavefunction.
A [j` ()]=ka = B [h` ()]=ia
Then match the first derivative.

dh` ()
dj` ()
= B(i)
Ak
d =ka
d
=ia
We can divide the two equations to eliminate the constants to get a condition on the
energies.
" dj` () #
" dh` () #
k

j` ()

= (i)
=ka

h` ()

=ia

This is often called matching the logarithmic derivative.

Often, the ` = 0 term will be sufficient to describe scattering. For ` = 0, the boundary
condition is
" cos sin #
" iei
#
ei
2
i i2
k
= (i)
.
sin
ei

=ka

=ia

Dividing and substituting for , we get

1
1
k cot(ka)
= i i
.
ka
ia
ka cot(ka) 1 = a 1
k cot(ka) =

This is the same transcendental equation that we had for the odd solution
in one dimension.
! r
r
2(E + V0 )
E
a =
cot
~2
V0 + E
The number of solutions depends on the depth and radius of the well. There can even
be no solution.

551

29. Scattering

29.7

TOC

Partial Wave Analysis of Scattering *

We can take a quick look at scattering from a potential in 3D We assume that

V = 0 far from the origin so the incoming and outgoing waves can be written in terms
of our solutions for a constant potential.
In fact, an incoming plane wave along the z direction can be expanded in Bessel
functions.

ikz

ikr cos

p
X

4(2` + 1)i` j` (kr)Y`0

`=0

Each angular momentum (`) term is called a partial wave. The scattering for each
partial wave can be computed independently.
For large r the Bessel function becomes

1 i(kr`/2)
j` ()
e
ei(kr`/2) ,
2ikr
so our plane wave becomes
eikz

p

X
1 i(kr`/2)
4(2` + 1)i`
e
ei(kr`/2) Y`0
2ikr
`=0

The scattering potential will modify the plane wave, particularly the outgoing part. To
maintain the outgoing flux equal to the incoming flux, the most the scattering can do
552

29. Scattering

TOC

is change the relative phase of the incoming an outgoing waves.

R` (r)

1 i(kr`/2)
e
e2i` (k) ei(kr`/2)
2ikr

sin(kr `/2 + ` (k)) i` (k)

e
kr

The ` (k) is called the phase shift for the partial wave of angular momentum `. We
can compute the differential cross section for scattering
scattered flux into d
d

d
incident flux
in terms of the phase shifts.

d
1

i
`

(2`
+
1)e
sin(
)P
(cos
)
= 2

`
`

d
k

`=0

The phase shifts must be computed by actually solving the problem for the particular
potential.
In fact, for low energy scattering and short range potentials, the first term ` = 0 is
often enough to solve the problem.

Only the low ` partial waves get close enough to the origin to be affected by the
potential.
553

29. Scattering

29.8

TOC

Scattering from a Spherical Well *

For the scattering problem, the energy is greater than zero. We must choose the Bessel
function in the region containing the origin.
R` = Aj` (k 0 r)
r
2(E + V0 )
k0 =
~2
For large r, we can have a linear combination of functions.
R` = Bj` (kr) + Cn` (kr)
r
2E
k=
~2

Matching the logarithmic derivative, we get

" dj` () #
" dj` ()
#
` ()
B d + C dnd
d
0
k
=k
j` ()
Bj` () + Cn` ()
0
=k a

=ka

Recalling that for r ,

sin(

cos(
n`

`
2 )
`
2 )

554

29. Scattering

TOC

and that our formula with the phase shift is

R(r)
=

sin

`
2

+ ` (k)

1
`
`
cos ` sin( ) + sin ` cos( ) ,

2
2

we can identify the phase shift easily.

tan ` =

C
B

We need to use the boundary condition to get this phase shift.

For ` = 0, we get
k0

B cos(ka) + C sin(ka)
cos(k 0 a)
=k
sin(k 0 a)
B sin(ka) C cos(ka)

k0
cot(k 0 a) (B sin(ka) C cos(ka)) = B cos(ka) + C sin(ka)
k

0
k0
k
cot(k 0 a) sin(ka) cos(ka) B = sin(ka) + cot(k 0 a) cos(ka) C
k
k

We can now get the phase shift.

tan 0 =

C
k cos(ka) sin(k 0 a) k 0 cos(k 0 a) sin(ka)
=
B
k sin(ka) sin(k 0 a) + k 0 cos(k 0 a) cos(ka)

With just the ` = 0 term, the differential scattering cross section is.
d
sin2 (` )

d
k2
The cross section will have zeros when
k0
= cot(ka) tan(k 0 a)
k
k 0 cot(k 0 a) = k cot(ka).

There will be many solutions to this and the cross section will look like diffraction.

555

29. Scattering

29.9

TOC

The Radial Equation for u(r) = rR(r) *

It is sometimes useful to use

un` (r) = rRn` (r)
to solve a radial equation problem. We can rewrite the equation for u.
2

d
2 d u(r)
d 1 du
u
2 du 2u
+
=
2 + 2
3
dr2
r dr
r
dr r dr
r
r dr
r
2
2
1d u
1 du
1 du 2u
2 du 2u
1d u
=
2
2
+ 3 + 2
3 =
r dr2
r dr
r dr
r
r dr
r
r dr2
1 d2 u(r) 2
`(` + 1)~2 u(r)
+ 2 E V (r)
=0
r dr2
~
2r2
r

d2 u(r) 2
`(` + 1)~2
+
E

V
(r)

u(r) = 0
dr2
~2
2r2
This now looks just like the one dimensional equation except the pseudo potential due
to angular momentum has been added.
We do get the additional condition that
u(0) = 0
to keep R normalizable.
For the case of a constant potential V0 , we define k =

2(EV0 )
~2

and = kr, and the

556

29. Scattering

TOC

radial equation becomes.

`(` + 1)~2
d2 u(r) 2
+
E

u(r) = 0
0
dr2
~2
2r2
d2 u(r)
`(` + 1)
+ k 2 u(r)
u(r) = 0
dr2
r2
d2 u() `(` + 1)

u() + u() = 0
d2
2
For ` = 0, its easy to see that sin and cos are solutions. Dividing by r to get R(),
we see that these are j0 and n0 . The solutions can be checked for other `, with some
work.

29.10

Partial Wave Analysis

We have already studied one approximation method for scattering called a partial
wave analysis. It is good for scattering potentials of limited range and for low energy
scattering. It divides the incoming plane wave in to partial waves with definite angular
momentum. The high angular momentum components of the wave will not scatter
(much) because they are at large distance from the scattering potential where that
potential is very small. We may then deal with just the first few terms (or even just
the ` = 0 term) in the expansion. We showed that the incoming partial wave and the
outgoing wave can differ only by a phase shift for elastic scattering. If we calculate this
phase shift ` , we can then determine the differential scattering cross section.
Lets review some of the equations. A plane wave can be decomposed into a sum of
spherical waves with definite angular momenta which goes to a simple sum of incoming
and outgoing spherical waves at large r.
eikz = eikr cos =

p
p
X
X
1 i(kr`/2)
4(2` + 1)i` j` (kr)Y`0
4(2` + 1)i`
e
ei(k
2ikr
`=0

`=0

A potential causing elastic scattering will modify the phases of the outgoing spherical
waves.
lim =

p
X

4(2` + 1)i`

`=0

1 i(kr`/2)
e
e2i` (k) ei(kr`/2) Y`0
2ikr

We can compute the differential cross section for elastic scattering.

2

d
1 X

= 2 (2` + 1)ei` (k) sin(` (k))P` (cos )

d
k
`

557

29. Scattering

TOC

It is useful to write this in terms of the amplitudes of the scattered waves.

d
d

= |f (, )|
X
1X
f (, ) =
(2` + 1)ei` (k) sin(` (k))P` (cos ) =
f` (, )
k
`

As an example, this has been used to compute the cross section for scattering from
a spherical potential well assuming only the ` = 0 phase shift was significant. By
matching the boundary conditions at the boundary of the spherical well, we determined
the phase shift.
tan 0 =

C
k cos(ka) sin(k 0 a) k 0 cos(k 0 a) sin(ka)
=
B
k sin(ka) sin(k 0 a) + k 0 cos(k 0 a) cos(ka)

The differential cross section is

d
sin2 (0 )

d
k2
which will have zeros if
k 0 cot(k 0 a) = k cot(ka).

We can compute the total scattering cross section using the relation dP` (cos )P`0 (cos ) =
4
0
2`+1 `` .

2
tot =
d |f (, )|
#"
"

0
0
1X
1X 0
i` (k)
=
d
(2` + 1)e
sin(` (k))P` (cos )
(2` + 1)ei` (k) sin(`0 (k))P`0 (
k
k
`
`
X
4
=
(2` + 1) sin(` (k))2
k2
`

It is interesting that we can relate the total cross section to the scattering amplitude
at = 0, for which P` (1) = 1.
f ( = 0, )

Im [f ( = 0, )]

1X
(2` + 1)ei` (k) sin(` (k))
k
`
1X
(2` + 1) sin2 (` (k))
k
`

tot

4
Im [f ( = 0, )]
k
558

29. Scattering

TOC

The total cross section is related to the imaginary part of the forward elastic scattering
amplitude. This seemingly strange relation is known as the Optical Theorem. It
can be understood in terms of removal of flux from the incoming plane wave. Remember we have an incoming plane wave plus scattered spherical waves. The total cross
section corresponds to removal of flux from the plane wave. The only way to do this is
destructive interference with the scattered waves. Since the plane wave is at = 0 it is
only the scattered amplitude at = 0 that can interfere. It is therefore reasonable that
a relation like the Optical Theorem is correct, even when elastic and inelastic processes
are possible.
We have not treated inelastic scattering. Inelastic scattering can be a complex
and interesting process. It was with high energy inelastic scattering of electrons from
protons that the quark structure of the proton was seen. In fact, the electrons
appeared to be scattering from essentially free quarks inside the proton. The proton
was broken up into sometimes many particles in the process but the data could be
simply analyzed using the scatter electron. In a phase shift analysis, inelastic scattering
removes flux from the outgoing spherical waves.
lim =

p
X
`=0

4(2` + 1)i`

1 i(kr`/2)
e
` (k)e2i` (k) ei(kr`/2) Y`0
2ikr

Here 0 < ` < 1, with 0 represent complete absorption of the partial wave and 1
representing purely elastic scattering. An interesting example of the effect of absorption
(or inelastic production of another state) is the black disk. The disk has a definite
radius a and absorbs partial waves for ` < ka. If one works out this problem, one finds
that there is an inelastic scattering cross section of inel = a2 . Somewhat surprisingly
the total elastic scattering cross section is elas = a2 . The disk absorbs part of the
beam and there is also diffraction around the sharp edges. That is, the removal of the
outgoing spherical partial waves modifies the plane wave to include scattered waves.

29.10.1

Scattering from a Hard Sphere

Assume a low energy beam is incident upon a small, hard sphere of radius r0 . We will
assume that ~kr0 < ~ so that only the ` = 0 partial wave is significantly affected by
the sphere. As with the particle in a box, the boundary condition on a hard surface
is that the wavefunction is zero. Outside the sphere, the potential is zero and the

559

29. Scattering

TOC

wavefunction solution will have reached its form for large r. So we set

ei(kr0 `/2) e2i` (k) ei(kr0 `/2) = eikr0 e2i0 (k) eikr0 = 0
e2i0 (k) = e2ikr0
0 (k) = kr0

2
d
1

= 2 ei` (k) sin(` (k))P0 (cos )
d
k
2
1
d
= 2 eikr0 sin(kr0 )
d
k
d
sin2 (kr0 )
=
d
k2
For very low energy, kr0 << 1 and
(kr0 )2
d

= r02
d
k2
The total cross section is then = 4r02 which is 4 times the area of the hard sphere.

29.11

Homework

1. The deuteron, a bound state of a proton and neutron with ` = 0, has a binding
energy of -2.18 MeV. Assume that the potential is a spherical well with potential
of V0 for r < 2.8 Fermis and zero potential outside. Find the approximate value
of V0 using numerical techniques.
2. Calculate the ` = 0 phase shift for the spherical potential well for both and
attractive and repulsive potential.
3. Calculate the ` = 0 phase shift for a hard sphere V = for r < a and V = 0 for
r > a. What are the limits for ka large and small?
4. Show that at large r, the radial flux is large compared to the angular components
ikr
of the flux for wave-functions of the form C e r Y`m (, ).
5. Calculate the difference in wavelengths of the 2p to 1s transition in Hydrogen and
Deuterium. Calculate the wavelength of the 2p to 1s transition in positronium.
6. Photons from the 3p 1s transition are observed coming from the sun. Quantitatively compare the natural line width to the widths from Doppler broadening
and collision broadening expected for radiation from the suns surface.

560

29. Scattering

29.12

TOC

Sample Test Problems

1. A particle has orbital angular momentum quantum number l = 1 and is bound

in the potential well V (r) = V0 for r < a and V (r) = 0 elsewhere. Write
down the form of the solution (in terms of known functions) in the two regions.
Your solution should satisfy constraints at the origin and at infinity. Be sure to
include angular dependence. Now use the boundary condition at r = a to get
one equation, the solution of which will quantize the energies. Do not bother to
solve the equation.
2. A particle of mass m with 0 total angular momentum is in a 3 dimensional
potential well V (r) = V0 for r < a (otherwise V (r) = 0).
a) Write down the form of the (l = 0) solution, to the time independent
Schr
odinger equation, inside the well, which is well behaved at at r = 0.
Specify the relationship between the particles energy and any parameters in
your solution.
b) Write down the form of the solution to the time independent Schrodinger
equation, outside the well, which has the right behavior as r . Again
specify how the parameters depend on energy.
c) Write down the boundary conditions that must be satisfied to match the
two regions. Use u(r) = rR(r) to simplify the calculation.
d) Find the transcendental equation which will determine the energy eigenvalues.
3. A particle has orbital angular momentum quantum number l = 1 and is bound
in the potential well V (r) = V0 for r < a and V (r) = 0 elsewhere. Write
down the form of the solution (in terms of known functions) in the two regions.
Your solution should satisfy constraints at the origin and at infinity. Be sure to
include angular dependence. Now use the boundary condition at r = a to get
one equation, the solution of which will quantize the energies. Do not bother to
solve the equation.
4. A particle is confined to the inside of a sphere of radius a. Find the energies
of the two lowest energy states for ` = 0. Write down (but do not solve) the
equation for the energies for ` = 1.
d
5. Calculate the differential cross section, d
, for high energy scattering of particles
of momentum p, from a spherical shell delta function

V (r) = (r r0 )
Assume that the potential is weak so that perturbation theory can be used. Be
sure to write your answer in terms of the scattering angles.
6. Assume that a heavy nucleus attracts K0 mesons with a weak Yakawa potential
d
V (r) = Vr0 er . Calculate the differential cross section, d
, for scattering high
561

29. Scattering

TOC

energy K0 mesons (mass mK ) from that nucleus. Give your answer in terms of
the scattering angle .

562

30. Classical Scalar Fields

TOC

Classical Scalar Fields

The non-relativistic quantum mechanics that we have studied so far developed

largely between 1923 and 1925, based on the hypothesis of Planck from the late 19th
century. It assumes that a particle has a probability that integrates to one over all space
and that the particles are not created or destroyed. The theory neither deals with the
quantized electromagnetic field nor with the relativistic energy equation.
It was not long after the non-relativistic theory was completed that Dirac introduced a
relativistic theory for electrons. By about 1928, relativistic theories, in which the
electromagnetic field was quantized and the creation and absorption of particles was
possible, had been developed by Dirac.
Quantum Mechanics became a quantum theory of fields, with the fields for bosons
and fermions treated in a symmetric way, yet behaving quite differently. In
1940, Pauli proved the spin-statistics theorem which showed why spin one-half particles should behave like fermions and spin zero or spin one particles should have the
properties of bosons.
Quantum Field Theory (QFT) was quite successful in describing all detailed experiments in electromagnetic interactions and many aspects of the weak interactions.
Nevertheless, by the 1960s, when our textbook was written, most particle theorists
were doubtful that QFT was suitable for describing the strong interactions and some
aspects of the weak interactions. This all changed dramatically around 1970 when very
successful Gauge Theories of the strong and weak interactions were introduced. By now, the physics of the electromagnetic, weak, and strong interactions are
well described by Quantum Field (Gauge) Theories that together from the Standard
Model.
Diracs relativistic theory of electrons introduced many new ideas such as antiparticles
and four component spinors. As we quantize the EM field, we must treat the propagation of photons relativistically. Hence we will work toward understanding relativistic
QFT.
In this chapter, we will review classical field theory, learn to write our equations in a
covariant way in four dimensions, and recall aspects of Lagrangian and Hamiltonian
formalisms for use in field theory. The emphasis will be on learning how all these things
work and on getting practice with calculations, not on mathematical rigor. While we
already have a good deal of knowledge about classical electromagnetism, we will start
with simple field theories to get some practice.

563

30. Classical Scalar Fields

30.1

TOC

Simple Mechanical Systems and Fields

This section is a review of mechanical systems largely from the point of view of
Lagrangian dynamics. In particular, we review the equations of a string as an example
of a field theory in one dimension.
We start with the Lagrangian of a discrete system like a single particle.
L(q, q)
=T V
Lagranges equations are
d
dt

L
qi

L
=0
qi

where the qi are the coordinates of the particle. This equation is derivable from the
principle of least action.
t2

L(qi , qi )dt = 0
t1

Similarly, we can define the Hamiltonian

X
H(qi , pi ) =
pi qi L
i

where pi are the momenta conjugate to the coordinates qi .

pi =

L
qi

For a continuous system, like a string, the Lagrangian is an integral of a Lagrangian

density function.

L = Ldx
For example, for a string,
"
2 #
1

L=
2 Y
2
x
where Y is Youngs modulus for the material of the string and is the mass density.
The Euler-Lagrange Equation for a continuous system is also derivable from the
principle of least action states above. For the string, this would be.

L
L
+

=0
x (/x)
t (/t)

Recall that the Lagrangian is a function of and its space and time derivatives.
564

30. Classical Scalar Fields

TOC

The Hamiltonian density can be computed from the Lagrangian density and is a
function of the coordinate and its conjugate momentum.
H =

L
L

In this example of a string, (x, t) is a simple scalar field. The string has a displacement at each point along it which varies as a function of time.
If we apply the Euler-Lagrange equation, we get a differential equation that
the strings displacement will satisfy.
"
2 #

1
2
L =
Y
2
x

L

L
L

= 0
x (/x)
t (/t)

TOC

With this, lets work on the Euler-Lagrange equation to get it into covariant shape.
Remember that the field is a Lorentz scalar.

X
L

L
L
+

=0
xk (/xk )
t (/t)

k

X
L
L

L
+

=0
xk (/xk )
(ict) (/(ict))

L
L
=0

x (/x )

This is the Euler-Lagrange equation for a single scalar field. Each term in this
equation is a Lorentz scalar, if L is a scalar.

L
(/x )

L
=0

Now we want to find a reasonable Lagrangian for a scalar field. The Lagrangian
depends on the and its derivatives. It should not depend explicitly on the
coordinates x , since that would violate translation and/or rotation invariance. We
also want to come out with a linear wave equation so that high powers of the field
should not appear. The only Lagrangian we can choose (up to unimportant constants)
is

1
2 2
+
L=
2 x x
The one constant sets the ratio of the two terms. The overall constant is not important except to match the T V definition. Remember that is a function of the
coordinates.
With this Lagrangian, the Euler-Lagrange equation is.

+ 2 = 0
x
x

2 = 0
x x
2 2 = 0

This is the known as the Klein-Gordon equation. It is a good relativistic equation

for a massive scalar field. It was also an early candidate for the relativistic equivalent
568

30. Classical Scalar Fields

TOC

of the Schr
odinger equation for electrons because it basically has the relativistic
analog of the energy relation inherent in the Schr
odinger equation. Writing that
relation in the order terms appear in the Klein-Gordon equation above we get (letting
c = 1 briefly).
p2 + E 2 m2 = 0
Its worth noting that this equation, unlike the non-relativistic Schrodinger equation,
relates the second spatial derivative of the field to the second time derivative.
(Remember the Schr
odinger equation has i times the first time derivative as the energy
operator.)
So far we have the Lagrangian and wave equation for a free scalar field. There are
no sources of the field (the equivalent of charges and currents in electromagnetism.)
Lets assume the source density is (x ). The source term must be a scalar function so,
we add the term to the Lagrangian.

1
2 2
L=
+ +
2 x x
This adds a term to the wave equation.
2 2 =
Any source density can be built up from point sources so it is useful to understand the
field generated by a point source as we do for electromagnetism.
(~x, t) = (x ) = G 3 (~x)
This is a source of strength G at the origin. It does not change with time so we expect
a static field.

=0
t
The Euler-Lagrange equation becomes.
2 2 = G 3 (~x)
We will solve this for the field from a point source below and get the result
(~x) =

Ger
.
4r

This solution should be familiar to us from the scalar potential for an electric point
charge which satisfies the same equation with = 0, 2 = = Q 3 (~x). This is
a field that falls off much faster than 1r . A massive scalar field falls off exponentially and the larger the mass, the faster the fall off. (We also get a mathematical
result which is useful in several applications. This is worked out another way in the
section on hyperfine splitting) 2 1r = 4 3 (~x).
569

30. Classical Scalar Fields

TOC

Now we solve for the scalar field from a point source by Fourier transforming the wave
equation. Define the Fourier transforms to be.

1
~
~k) =
d3 x eik~x (~x)
(
3
(2) 2

1
~ ~
(~x) =
d3 k eik~x (
k)
3
(2) 2
We now take the transform of both sides of the equation.

1
3
2

(2)

1
3

(2) 2

1
(2)

2 2

d3 x eik~x (2 2 )

3
2

d3 x e

i~
k~
x

(2 2 )

d3 x eik~x (k 2 2 )

(k 2 2 ) =
=

G 3 (~x)

1
(2)
G

3
2

d3 x eik~x G 3 (~x)

(2) 2
G
3

(2) 2
G
(2)

3
2

(k 2

1
+ 2 )

To deal with the 2 , we have integrated by parts twice assuming that the field falls off
fast enough at infinity.
We now have the Fourier transform of the field of a point source. If we can
transform back to position space, we will have the field. This is a fairly standard

570

30. Classical Scalar Fields

TOC

type of problem in quantum mechanics.

1
1
G
~
(~x) =
d3 k eik~x
3
3
2
(2) 2
(2) 2 (k + 2 )

~
G
eik~x
3
=
d
k
(2)3
(k 2 + 2 )

1 ikr cos k
2G
e
2
=
k dk
d cos k
(2)3
(k 2 + 2 )
1

G
(2)2

k2
(k 2 + 2 )

=
=

eikr cos k d cos k dk

1
G
k
1 ikr cos k
e
dk
(2)2
(k 2 + 2 ) ikr
1

k
G
eikr eikr dk
2
2
2
(2) ir
(k + )

G
k
eikr eikr dk
(2)2 ir
(k + i)(k i)
2

This is now of a form for which we can use Cauchys theorem for contour integrals.
The theorem says that an integral around a closed contour in the complex
plane is equal to 2i times the sum of the residues at the poles enclosed in the contour.
Contour Integration is a powerful technique often used in quantum mechanics.
If the integrand is a function of the complex variable k, a pole is of the form kr where r
is called the residue at the pole. The integrand above has two poles, one at k = i
and the other at k = i. The integral we are interested in is just along the
real axis so we want the integral along the rest of the contour to give zero. Thats easy
to do since the integrand goes to zero at infinity on the real axis and the exponentials
go to zero either
at positive or negative infinity for the imaginary part of k. Examine

k
ikr
dk around a contour in the upper half plane as shown
the integral
(k+i)(ki) e
below.

571

30. Classical Scalar Fields

TOC

The pole inside the contour is at k = i. The residue at the pole is

ikr
ier
1
ke
=
= er .
k + i k=i
i + i
2
The integrand goes to zero exponentially on the semicircle at infinity so only the real
axis contributes to the integral along the contour. The integral along the real axis can

572

30. Classical Scalar Fields

TOC

be manipulated to do the whole problem for us.

keikr
2i
k + i

k
eikr dk
(k + i)(k i)

1
2i er
2

k
eikr dk
(k + i)(k i)

ier

k
eikr dk
+ 2

ier

k
eikr dk
+ 2

ier

k
eikr dk
k 2 + 2

ier

eikr eikr dk

ier

ier
i + i

k
eikr dk +
(k + i)(k i)

0
(k 0

0
k 0
eik r (dk 0 ) +
0
+ i)(k i)

0
k
eik r dk 0 +
02
2
k +

k2
0

k
eikr dk +
+ 2

k2
0

k
eikr dk
k 2 + 2

k
k 2 + 2

This is exactly the integral we wanted to do.

Plug the integral into the Fourier transform we were computing.
(~x)

G
(2)2 ir

k
+ 2

eikr eikr dk

(~x)

G
ier
(2)2 ir
Ger
4r

In this case, it is simple to compute the interaction Hamiltonian from the interaction Lagrangian and the potential between two particles. Lets assume we have
573

k=i

30. Classical Scalar Fields

TOC

two particles, each with the same interaction with the field.
(~x, t) = (x) = G 3 (~x)
Now compute the Hamiltonian.
Lint
Hint
Hint
(1,2)

Hint

=
Lint
=
Lint = Lint =

=
Hint d3 x2

G2 er12
Ger 3
3
G (~x2 )d3 x2 =
=
1 2 d x2 =
4r
4r12

We see that this is a short-range, attractive potential.

This was proposed by Yukawa as the nuclear force. He predicted a scalar particle
with a mass close to that of the pion before the pion was discovered. His prediction of
the mass was based on the range of the nuclear force, on the order of one Fermi. In
some sense, his prediction is approximately correct. Pion exchange can explain much
of the nuclear force but does not explain all the details. Pions and nucleons have since
been show to be composite particles with internal structure. Other composites with
masses larger than the pion also play a role in the force between nucleons.
Pions were also found to come in three charges: + , , and 0 . This would lead us to
develop a complex scalar field as done in the text. Its not our goal right now so we will
skip this. Its interesting to note that the Higgs Boson is also represented by a complex
scalar field.
We have developed a covariant classical theory for a scalar field. The Lagrangian
density is a Lorentz scalar function. We have included an interaction term to provide
a source for the field. Now we will attempt to do the same for classical electromagnetism.

574

31. Classical Maxwell Fields

31
31.1

TOC

Classical Maxwell Fields

Rationalized Heaviside-Lorentz Units

The SI units are based on a unit of length of the order of human size originally
related to the size of the earth, a unit of time approximately equal to the time between
heartbeats, and a unit of mass related to the length unit and the mass of water.
None of these depend on any even nearly fundamental physical quantities. Therefore
many important physical equations end up with extra (needless) constants in them like
c. Even with the three basic units defined, we could have chosen the unit of charge
correctly to make 0 and 0 unnecessary but instead a very arbitrary choice was made
0 = 4 107 and the Ampere is defined by the current in parallel wires at one
meter distance from each other that gives a force of 2 107 Newtons per meter. The
Coulomb is set so that the Ampere is one Coulomb per second. With these choices
SI units make Maxwells equations and our filed theory look very messy.
Physicists have more often used CGS units in which the unit of charge and definition
of the field units are set so that 0 = 1 and 0 = 1 so they need not show up in the
equations. The CGS units are not perfect, however, and we will want to change them
slightly to make our theory of the Maxwell Field simple. The mistake made in defining
CGS units was in removing the 4 that show up in Coulombs law. Coulombs law is
not fundamental and the 4 belonged there.
We will correct this little mistake and move to Rationalized Heaviside-Lorentz
Units by making a minor modification to the unit of charge and the units of fields.
With this modification, our field theory will have few constants to carry around. As
the name of the system of units suggests, the problem with CGS has been with . We
dont need to change the centimeter, gram or second to fix the problem.
In
Rationalized Heaviside-Lorentz units we decrease the field strength by a factor
of 4 and increase the charges by the same factor, leaving the force unchanged.
~
E

~
B

~
A
e
2

e
~c

~
E

4
~
B

4
~
A

e 4
e2
1

4~c
137
575

31. Classical Maxwell Fields

TOC

Its not a very big change but it would have been nice if Maxwell had started with
this set of units. Of course the value of cannot change, but, the formula for it does
because we have redefined the charge e.
Maxwells Equations in CGS units are
~ B
~

B
1
~ E
~+

c t
~ E
~

~ B
~ 1 E

c t
The Lorentz Force is

4
4 ~
=
j.
c

~ + 1 ~v B).
~
F~ = e(E
c

When we change to Rationalized Heaviside-Lorentz units, the equations become

~ B
~ =0

~ E
~ + 1 B = 0

c t
~ E
~ =

~ B
~ 1 E = 1~j

c t
c
~ + 1 ~v B)
~
F~ = e(E
c

That is, the equations remain the same except the factors of 4 in front of the source
terms disappear. Of course, it would still be convenient to set c = 1 since this has been
confusing us about 4D geometry and c is the last unnecessary constant in Maxwells
equations. For our calculations, we can set c = 1 any time we want unless we need
answers in centimeters.

31.2

The Electromagnetic Field Tensor

The transformation of electric and magnetic fields under a Lorentz boost we established
even before Einstein developed the theory of relativity. We know that E-fields can
576

31. Classical Maxwell Fields

TOC

transform into B-fields and vice versa. For example, a point charge at rest gives an
Electric field. If we boost to a frame in which the charge is moving, there is an Electric
and a Magnetic field. This means that the E-field cannot be a Lorentz vector. We need
to put the Electric and Magnetic fields together into one (tensor) object to properly
handle Lorentz transformations and to write our equations in a covariant way.
The simplest way and the correct way to do this is to make the Electric and Magnetic
fields components of a rank 2 (antisymmetric) tensor.

0
Bz
=
By
iEx

Bz
0
Bx
iEy

By
Bx
0
iEz

iEx
iEy

iEz
0

The fields can simply be written in terms of the vector potential, (which is a Lorentz
~ i).
vector) A = (A,

F =

A
A

x
x

Note that this is automatically antisymmetric under the interchange of the indices.
As before, the first two (sourceless) Maxwell equations are automatically
satisfied for fields derived from a vector potential. We may write the other two
Maxwell equations in terms of the 4-vector j = (~j, ic).

F
j
=
x
c

Which is why the T-shirt given to every MIT freshman when they take Electricity and
Magnetism should say
... and God said

A
x

j
c

and there was light.

Of course he or she hadnt yet quantized the theory in that statement.

577

31. Classical Maxwell Fields

TOC

For some peace of mind, lets verify a few terms in the equations. Clearly all the
diagonal terms in the field tensor are zero by antisymmetry. Lets take some example
off-diagonal terms in the field tensor, checking the (old) definition of the fields in terms
of the potential.
~
B

~ A
~
=

~
E

~
=

F12

F13

F4i

~
1 A
c t
A2
A1
~ A)
~ z = Bz

= (
x1
x2
A3
A1
~ A)
~ y = By

= (
x1
x3

(i)

A4
1 Ai
1 Ai

1 Ai
Ai

=
= i
= i
+
= iEi
x4
xi
ic t
xi
c t
xi
xi
c t

Lets also check what the Maxwell equation says for the last row in the tensor.
F4
x
F4i
xi
(iEi )
xi
Ei
xi
~ E
~

=
=

j4
c
ic
c

= i
=
=

We will not bother to check the Lorentz transformation of the fields here. Its right.

31.3

The Lagrangian for Electromagnetic Fields

There are not many ways to make a scalar Lagrangian from the field tensor. We
already know that
j
F
=
x
c
and we need to make our Lagrangian out of the fields, not just the current. Again, x
cannot appear explicitly because that violates symmetries of nature. Also we want
a linear equation and so higher powers of the field should not occur. A term of the
form mA A is a mass term and would cause fields to fall off faster than 1r . So, the
only reasonable choice is
F F = 2(B 2 E 2 ).
578

31. Classical Maxwell Fields

TOC

One might consider

~ E
~
e F F = B
but that is a pseudo-scalar, not a scalar. That is, it changes sign under a parity
transformation. The EM interaction is known to conserve parity so this is not a real
option. As with the scalar field, we need to add an interaction with a source
term. Of course, we know electromagnetism well, so finding the right Lagrangian is
not really guess work. The source of the field is the vector j , so the simple scalar we
can write is j A .
The Lagrangian for Classical Electricity and Magnetism we will try is.

1
1
LEM = F F + j A
4
c

In working with this Lagrangian, we will treat each component of A as an independent field.
The next step is to check what the Euler-Lagrange equation gives us.

L
L

= 0
x (A /x )
A

1
1 A
A
A
A
1
L = F F + j A =

4
c
4 x
x
x
x

L
1

A
A
A
A
=

(A /x )
4 A /x x
x
x
x

1

A A
A A
=
2
2
4 A /x
x x
x x

A
1
A
= 4

4
x
x
= F = F

L
F
= 0
x
A

j
F
= 0
x
c

j
F =
x
c
Note that, since we have four independent components of A as independent fields,
we have four equations; or one 4-vector equation. The Euler-Lagrange equation
579

31. Classical Maxwell Fields

TOC

gets us back Maxwells equation with this choice of the Lagrangian. This clearly
justifies the choice of L.
It is important to emphasize that we have a Lagrangian based, formal classical field
theory for electricity and magnetism which has the four components of the 4vector potential as the independent fields. We could not treat each component
of F as independent since they are clearly correlated. We could have tried using the
six independent components of the antisymmetric tensor but it would not have given
the right answer. Using the 4-vector potentials as the fields does give the right answer.
Electricity and Magnetism is a theory of a 4-vector field A .
We can also calculate the free field Hamiltonian density, that is, the Hamiltonian density in regions with no source term. We use the standard definition of the
Hamiltonian in terms of the Lagrangian.

L
A
L
A
H=
L=
L
(A /dt) dt
(A /x4 ) x4
We just calculated above that
L
= F
(A /x )
which we can use to get
L
(A /x4 )
H

= F4
A
L
x4
1
A
= F4
+ F F
x4
4

H = F4

(F4 )

A
1
+ F F
x4
4

We will use this once we have written the radiation field in a convenient form. In the
meantime, we can check what this gives us in general in a region with no sources.

580

31. Classical Maxwell Fields

=
=
=
=
=
=

TOC

A4
1
F4 F4 +
+ F F
x
4

1
A4
+ F F
F4 F4 +
x
4
1
A4
F4 F4 F4
+ F F
x
4
A
1
4
+ (B 2 E 2 )
E 2 F4i
xi
2
(i)
1 2
(E + B 2 ) iEi
2
xi

1 2
2
(E + B ) + Ei
2
xi

If we integrate the last term by parts, (and the fields fall to zero at infinity), then that term
~ E
~ which is zero with no sources in the region. We can therefore drop it and are
contains a
left with
1
H = (E 2 + B 2 ).
2
This is the result we expected, the energy density and an EM field. (Remember the fields have

been decreased by a factor of 4 compared to CGS units.)

We will study the interaction between electrons and the electromagnetic field with the
Dirac equation. Until then, the Hamiltonian used for non-relativistic quantum mechanics will be sufficient. We have derived the Lorentz force law from that Hamiltonian.
1
e ~ 2
H=
p~ + A
+ eA0
2m
c

31.4

Gauge Invariance can Simplify Equations

We have already studied many aspects of gauge invariance in electromagnetism and the
corresponding invariance under a phase transformation in Quantum Mechanics. One
point to note is that, with our choice to treat each component of A as an independent
field, we are making a theory for the vector field A with a gauge symmetry,
not really a theory for the field F .
Recall that the gauge symmetry of Electricity and Magnetism and the phase symmetry of electron wavefunctions are really one and the same. Neither the phase of the
wavefunction nor the vector potential are directly observable, but the symmetry is.
We will not go over the consequences of gauge invariance again here, but, we do want
to use gauge invariance to simplify our equations.
581

31. Classical Maxwell Fields

TOC

Maxwells equation is
F
=
x

A
A

=
x x
x

j
c
j
c

A
j
2 A
=

x x
x2
c
A
2 A
j

=
x2
x x
c
We can simplify this basic equation by setting the gauge according to the Lorentz
condition.

A
=0
x

The gauge transformation needed is

A A +
x

A
2 =
x old
The Maxwell equation with the Lorentz condition now reads
2A =

j
.
c

There is still substantial gauge freedom possible. The second derivative of Lambda
is set by the Lorentz condition but there is still freedom in the first derivative which
will modify A. Gauge transformations can be made as shown below.

x
2 = 0

A A +

This transformation will not disturb the Lorentz condition which simplifies
our equation. We will use a further gauge condition in the next chapter to work with
transverse fields.
582

32. Quantum Theory of Radiation

32
32.1

TOC

Quantum Theory of Radiation

Transverse and Longitudinal Fields

In non-relativistic Quantum Mechanics, the static Electric field is represented by a

scalar potential, magnetic fields by the vector potential, and the radiation field also
through the vector potential. It will be convenient to keep this separation between the
large static atomic Electric field and the radiation fields, however, the equations we
have contain the four-vector A with all the fields mixed. When we quantize the field,
all E and B fields as well as electromagnetic waves will be made up of photons. It is
useful to be able to separate the E fields due to fixed charges from the EM
radiation from moving charges. This separation is not Lorentz invariant, but it is
still useful.
Enrico Fermi showed, in 1930, that Ak together with A0 give rise to Coulomb interactions between particles, whereas A gives rise to the EM radiation from moving charges.
With this separation, we can maintain the form of our non-relativistic Hamiltonian,

2 X
X 1
ei ej
e~
xj ) +
+ Hrad
p~ A
(~
2mj
c
4|~xi ~xj |
j
i>j

~ ), and A
~ is
where Hrad is purely the Hamiltonian of the radiation (containing only A
~
~
~
the part of the vector potential which satisfies A = 0. Note that Ak and A4 appear
nowhere in the Hamiltonian. Instead, we have the Coulomb potential. This separation
allows us to continue with our standard Hydrogen solution and just add radiation. We
will not derive this result.
In a region in which there are no source terms,
j = 0
we can make a gauge transformation which eliminates A0 by choosing such
that
1
= A0 .
c t
Since the fourth component of A is now eliminated, the Lorentz condition now implies
that
~ A
~ = 0.

Again, making one component of a 4-vector zero is not a Lorentz invariant way of
working. We have to redo the gauge transformation if we move to another frame.
583

32. Quantum Theory of Radiation

TOC

If j 6= 0, then we cannot eliminate A0 , since 2A0 = jc0 and we are only allowed to
make gauge transformations for which 2 = 0. In this case we must separate the
vector potential into the transverse and longitudinal parts, with
~ = A
~ + A
~k
A
~ A
~ = 0

~ A
~k = 0

~ A
~ = 0.
We will now study the radiation field in a region with no sources so that
We will use the equations

~
2 A

32.2

~
B

~
E

~
1 A
2
c t2

~ A
~

~
1 A

c t
0

Fourier Decomposition of Radiation Oscillators

Our goal is to write the Hamiltonian for the radiation field in terms of a sum of harmonic
oscillator Hamiltonians. The first step is to write the radiation field in as simple a way
as possible, as a sum of harmonic components. We will work in a cubic volume V = L3
and apply periodic boundary conditions on our electromagnetic waves. We also
assume for now that there are no sources inside the region so that we can make
~ A
~ = 0. We decompose the
a gauge transformation to make A0 = 0 and hence
field into its Fourier components at t = 0
2

XX
~
~
~ x, t = 0) = 1
() ck, (t = 0)eik~x + ck, (t = 0)eik~x
A(~
V k =1

2ni
L

due to the periodic boundary conditions. The factor out front is set to normalize the
states nicely since

1
~
~0
d3 xeik~x eik ~x = ~k~k0
V
584

32. Quantum Theory of Radiation

and

TOC

() ( ) = 0 .
We know the time dependence of the waves from Maxwells equation,
ck, (t) = ck, (0)eit
where = kc. We can now write the vector potential as a function of position
and time.
2

1 X X ()
~
~
~

ck, (t)eik~x + ck, (t)eik~x
A(~x, t) =
V k =1
We may write this solution in several different ways, and use the best one for the
calculation being performed. One nice way to write this is in terms 4-vector k , the
wave number,

p
= (kx , ky , kz , ik) = (kx , ky , kz , i )
k =
~
c
so that
k x = k x = ~k ~x t.
We can then write the radiation field in a more covariant way.
2
XX

~ x, t) = 1
A(~
() ck, (0)eik x + ck, (0)eik x
V k =1

A convenient shorthand for calculations is possible by noticing that the second term is
just the complex conjugate of the first.
2

1 X X ()
~

A(~x, t) =

ck, (0)eik x + c.c.
V k =1
2
XX
~ x, t) = 1
A(~
() ck, (0)eik x + c.c.
V k =1

Note again that we have made this a transverse field by construction. The unit vectors
() are transverse to the direction of propagation. Also note that we are working in a
gauge with A4 = 0, so this can also represent the 4-vector form of the potential. The
Fourier decomposition of the radiation field can be written very simply.

2
1 X X ()
A =
ck, (0)eik x + c.c.
V k =1

585

32. Quantum Theory of Radiation

TOC

This choice of gauge makes switching between 4-vector and 3-vector expressions for the
potential trivial.
Lets verify that this decomposition of the radiation field satisfies the Maxwell
equation, just for some practice. Its most convenient to use the covariant form of the
equation and field.

2
X

2A = 0
!

1 X
ik x

()
+ c.c.
ck, (0)e
V k =1

2
1 X X ()

ck, (0)2eik x + c.c.

V k =1

2
1 X X ()

ck, (0)(k k )eik x + c.c. =

V k =1

The result is zero since k k = k 2 k 2 = 0.

~ A
~ = 0.
Lets also verify that
~

!
2
1 X X ()
i~
k~
x

ck, (t)e
+ c.c.
V k =1

2
1 XX
~ i~k~x + c.c.

ck, (t)
() e
V k =1

2
1 XX
~

ck, (t)
() ~keik~x + c.c. = 0
V k =1

The result here is zero because () ~k = 0.

32.3

The Hamiltonian for the Radiation Field

We now wish to compute the Hamiltonian in terms of the coefficients ck, (t).
This is an important calculation because we will use the Hamiltonian formalism
to do the quantization of the field. We will do the calculation using the covariant
notation (while Sakurai outlines an alternate calculation using 3-vectors). We have
already calculated the Hamiltonian density for a classical EM field.

H = F4

A
1
+ F F
x4
4

586

32. Quantum Theory of Radiation

TOC

A4
A A
1 A
A
A
A

x
x4 x4
4 x
x
x
x

A A
1 A A
A A

x4 x4
2 x x
x x

Now lets compute the basic element of the above formula for our decomposed
radiation field.
=

2

1 X X ()

ck, (0)eik x + ck, (0)eik x

V k =1

A
x

2

1 X X ()

ck, (0)(ik )eik x + ck, (0)(ik )eik x

V k =1

A
x

2

1 X X ()
k ck, (0)eik x ck, (0)eik x
= i
V k =1

A
x4

2

1 X X ()
=

ck, (0)eik x ck, (0)eik x
c
V k =1

We have all the elements to finish the calculation of the Hamiltonian. Before pulling
this all together in a brute force way, its good to realize that almost all the terms
will give zero. We see that the derivative of A is proportional to a 4-vector, say k
()
and to a polarization vector, say . The dot products of the 4-vectors, either k with
itself or k with are zero. Going back to our expression for the Hamiltonian density,
we can eliminate some terms.

1 A A
A A
A A
+

H =
x4 x4
2 x x
x x
1
A A
+ (0 0)
H =
x4 x4
2
A A
H =
x4 x4
The remaining term has a dot product between polarization vectors which will be
nonzero if the polarization vectors are the same. (Note that this simplification is
possible because we have assumed no sources in the region.)
The total Hamiltonian we are aiming at, is the integral of the Hamiltonian density.

H=
d3 x H
587

32. Quantum Theory of Radiation

TOC

When we integrate over the volume only products like eik x eik x will give a nonzero
result. So when we multiply one sum over k by another, only the terms with the same k
will contribute to the integral, basically because the waves with different wave number
are orthogonal.

0
1
d3 x eik x eik x = kk0
V

d3 xH

A
x4
H
H

A A
x4 x4
2

1 X X ()

=
ck, (0) eik x ck, (0) eik x
c
c
V k =1

A A
=
d3 x
x4 x4

2
2
1 XX

=
d3 x
ck, (0) eik x ck, (0) eik x
V
c
c
=1
k

2 2
XX

ck, (t)ck, (t) ck, (t)ck, (t)
=
c
=1
k

H
H

2 2
XX

=
ck, (t)ck, (t) + ck, (t)ck, (t)
c
k =1
X 2

=
ck, (t)ck, (t) + ck, (t)ck, (t)
c
k,

This is the result we will use to quantize the field. We have been careful not
to commute c and c here in anticipation of the fact that they do not commute.
It should not be a surprise that the terms that made up the Lagrangian gave a zero
contribution because L = 21 (E 2 B 2 ) and we know that E and B have the same
magnitude in a radiation field. (There is one wrinkle we have glossed over; terms with
~k 0 = ~k.)

32.4

Canonical Coordinates and Momenta

We now have the Hamiltonian for the radiation field.

588

32. Quantum Theory of Radiation

TOC

X 2

ck, (t)ck, (t) + ck, (t)ck, (t)
c
k,

It was with the Hamiltonian that we first quantized the non-relativistic

motion of particles. The position and momentum became operators which
did not commute. Lets define ck, to be the time dependent Fourier coefficient.
ck, = 2 ck,
We can then simplify our notation a bit.
X 2

H=
ck, ck, + ck, ck,
c
k,

This now clearly looks like the Hamiltonian for a collection of uncoupled
oscillators; one oscillator for each wave vector and polarization.
We wish to write the Hamiltonian in terms of a coordinate for each oscillator and
the conjugate momenta. The coordinate should be real so it can be represented by
a Hermitian operator and have a physical meaning. The simplest choice for a real
coordinates is c + c . With a little effort we can identify the coordinate
Qk, =

1
(ck, + ck, )
c

and its conjugate momentum for each oscillator,

Pk, =

i
(ck, ck, ).
c

The Hamiltonian can be written in terms of these.

1 X 2
Pk, + 2 Q2k,
H =
2
k,

2
X
1
2
=

(ck, ck, )2 +
(ck, + ck, )2
2
c
c
k,

1 X 2
=
(ck, ck, )2 + (ck, + ck, )2
2
c
k,

1 X 2
=
2 ck, ck, + ck, ck,
2
c
k,

X 2

=
ck, ck, + ck, ck,
c
k,

589

32. Quantum Theory of Radiation

TOC

This verifies that this choice gives the right Hamiltonian. We should also check that this
choice of coordinates and momenta satisfy Hamiltons equations to identify
them as the canonical coordinates. The first equation is
H
Qk,
2 Qk,
2
(ck, + ck, )
c
2
(ck, + ck, )
c

= Pk,
=
=
=

i
(ck, ck, )
c
i
(ick, ick, )
c
2
(ck, + ck, )
c

This one checks out OK.

The other equation of Hamilton is
H
Pk,

Pk,

i
(ck, ck, )
c
i
(ck, ck, )
c

=
=

Q k,
1
(ck, + ck, )
c
1
(ick, + ick, )
c
i
(ck, ck, )
c

This also checks out, so we have identified the canonical coordinates and
momenta of our oscillators.
We have a collection of uncoupled oscillators with identified canonical coordinate and
momentum. The next step is to quantize the oscillators.

32.5

Quantization of the Oscillators

To summarize the result of the calculations of the last section we have the Hamiltonian for the radiation field.

590

32. Quantum Theory of Radiation

TOC

X 2

ck, ck, + ck, ck,
c
k,

Qk, =
Pk, =
H=

1
(ck, + ck, )
c
i
(ck, ck, )
c

1 X 2
Pk, + 2 Q2k,
2
k,

Soon after the development of non-relativistic quantum mechanics, Dirac proposed

that the canonical variables of the radiation oscillators be treated like p and x in the
quantum mechanics we know. The place to start is with the commutators.
The
coordinate and its corresponding momentum do not commute. For example
[px , x] = ~i . Coordinates and momenta that do not correspond, do commute. For
example [py , x] = 0. Different coordinates commute with each other as do different
momenta. We will impose the same rules here.
[Qk, , Pk0 ,0 ]

= i~kk0 0

[Qk, , Qk0 ,0 ]

[Pk, , P

k0 ,0

By now we know that if the Q and P do not commute, neither do the c and c so we
should continue to avoid commuting them.
Since we are dealing with harmonic oscillators, we want to find the analog of the
raising and lowering operators. We developed the raising and lowering operators

591

32. Quantum Theory of Radiation

TOC

by trying to write the Hamiltonian as H = A A~. Following the same idea, we get
ak,

ak,

ak, ak,

=
=
=
=

1
ak, ak, +

2
1
ak, ak, +
~
2

=
=

1
(Qk, + iPk, )
2~
1

(Qk, iPk, )
2~
1
(Qk, iPk, )(Qk, + iPk, )
2~
1
2
( 2 Q2k, + Pk,
+ iQk, Pk, iPk, Qk, )
2~
1
~
2
( 2 Q2k, + Pk,
+ iQk, Pk, i(Qk, Pk, + ))
2~
i
1
2
( 2 Q2k, + Pk,
~)
2~
1
2
( 2 Q2k, + Pk,
)
2~
1 2 2
2
( Qk, + Pk,
)=H
2

ak, ak, +

1
2

This is just the same as the Hamiltonian that we had for the one dimensional
harmonic oscillator. We therefore have the raising and lowering operators, as long
as [ak, , ak, ] = 1, as we had for the 1D harmonic oscillator.
h

ak, , ak,

=
=
=
=
=

1
1
(Qk, + iPk, ),
(Qk, iPk, )]
2~
2~

1
[Qk, + iPk, , Qk, iPk, ]
2~
1
(i[Qk, , Pk, ] + i[Pk, , Qk, ])
2~
1
(~ + ~)
2~
1

So these are definitely the raising and lowering operators. Of course the commutator
would be zero if the operators were not for the same oscillator.
[ak, , ak0 ,0 ] = kk0 0

592

32. Quantum Theory of Radiation

TOC

(Note that all of our commutators are assumed to be taken at equal time.) The
Hamiltonian is written in terms a and a in the same way as for the 1D harmonic
oscillator. Therefore, everything we know about the raising and lowering operators
applies here, including the commutator with the Hamiltonian, the raising and lowering
of energy eigenstates, and even the constants.

nk, |nk, 1i
ak, |nk, i =
p

ak, |nk, i =
nk, + 1 |nk, + 1i
The nk, can only take on integer values as with the harmonic oscillator we know.
As with the 1D harmonic oscillator, we also can define the number operator.

1
1
H =
ak, ak, +
~ = Nk, +
~
2
2
= ak, ak,

Nk,

The last step is to compute the raising and lowering operators in terms of
the original coefficients.
ak,

Qk,

Pk,

ak,

=
=
=
=
=

1
(Qk, + iPk, )
2~

1
(ck, + ck, )
c
i
(ck, ck, )
c
1
i
1

( (ck, + ck, ) i (ck, ck, ))

c
c
2~
1

((ck, + ck, ) + (ck, ck, ))

2~ c
1

(ck, + ck, + ck, ck, )

2~ c
r

(2ck, )
2~c2
r
2
ck,
~c2

r
ck, =

~c2
ak,
2
593

32. Quantum Theory of Radiation

TOC

Similarly we can compute that

r
ck,

~c2
a
2 k,

Since we now have the coefficients in our decomposition of the field equal to a constant times the raising or lowering operator, it is clear that these coefficients have
themselves become operators.

32.6

Photon States

It is now obvious that the integer nk, is the number of photons in the volume
with wave number ~k and polarization () . It is called the occupation number for
the state designated by wave number ~k and polarization () . We can represent the
state of the entire volume by giving the number of photons of each type (and some
phases). The state vector for the volume is given by the direct product of the states
for each type of photon.
|nk1 ,1 , nk2 ,2 , ..., nki ,i , ...i = |nk1 ,1 i|nk2 ,2 i..., |nki ,i i...
The ground state for a particular oscillator cannot be lowered. The state in which all
the oscillators are in the ground state is called the vacuum state and can be
written simply as |0i. We can generate any state we want by applying raising operators
to the vacuum state.
|nk1 ,1 , nk2 ,2 , ..., nki ,i , ...i =
The factorial on the bottom cancels all the

Y (ak , )nki ,i
i
i
p
|0i
n
ki ,i !
i

n + 1 we get from the raising operators.

Any multi-photon state we construct is automatically symmetric under the interchange of pairs of photons. For example if we want to raise two photons out of
the vacuum, we apply two raising operators. Since [ak, , ak0 ,0 ] = 0, interchanging the
photons gives the same state.
ak, ak0 ,0 |0i = ak0 ,0 ak, |0i
So the fact that the creation operators commute dictates that photon states
are symmetric under interchange.
594

32. Quantum Theory of Radiation

32.7

TOC

Fermion Operators

At this point, we can hypothesize that the operators that create fermion states
do not commute. In fact, if we assume that the operators creating fermion
states anti-commute (as do the Pauli matrices), then we can show that fermion
states are antisymmetric under interchange. Assume br and br are the creation and
annihilation operators for fermions and that they anti-commute.

{br , br0 } = 0

The states are then antisymmetric under interchange of pairs of fermions.

br br0 |0i = br0 br |0i
Its not hard to show that the occupation number for fermion states is either
zero or one.

32.8

Quantized Radiation Field

The Fourier coefficients of the expansion of the classical radiation field should now be
replaced by operators.
r
~c2
ck,
ak,
2
r
~c2
ck,
a
2 k,
r

1 X ~c2 ()
~
~
A =
ak, (t)eik~x + ak, (t)eik~x
2
V
k

A is now an operator that acts on state vectors in occupation number space. The
operator is parameterized in terms of ~x and t. This type of operator is called a field
operator or a quantized field.
The Hamiltonian operator can also be written in terms of the creation and annihilation

595

32. Quantum Theory of Radiation

TOC

operators.
H

X 2

ck, ck, + ck, ck,
c
k,

i
X 2 ~c2 h
ak, ak, + ak, ak,
=
c
2
k,
h
i
1X
~ ak, ak, + ak, ak,
=
2
k,

X
k,

1
~ Nk, +
2

For our purposes, we may remove the (infinite) constant energy due to the ground state
energy of all the oscillators. It is simply the energy of the vacuum which we may define
as zero. Note that the field fluctuations that cause this energy density, also cause the
spontaneous decay of excited states of atoms. One thing that must be done is to cut
off the sum at some maximum value of k. We do not expect electricity and magnetism
to be completely valid up to infinite energy. Certainly by the gravitational or grand
unified energy scale there must be important corrections to our formulas. The energy
density of the vacuum is hard to define but plays an important role in cosmology.
At this time, physicists have difficulty explaining how small the energy density in the
vacuum is. Until recent experiments showed otherwise, most physicists thought it was
actually zero due to some unknown symmetry. In any case we are not ready to consider
this problem.
X
H=
~Nk,
k,

With this subtraction, the energy of the vacuum state has been defined to be zero.
H |0i = 0
The total momentum in the (transverse) radiation field can
the classical formula for the Poynting vector).

X
1
~ B
~ d3 x =
P~ =
E
~~k Nk, +
c
k,

also be computed (from

1
2

This time the 12 can really be dropped since the sum is over positive and negative ~k,
so it sums to zero.
X
P~ =
~~k Nk,
k,

596

32. Quantum Theory of Radiation

TOC

We can compute the energy and momentum of a single photon state by operating on
that state with the Hamiltonian and with the total momentum operator. The state for
a single photon with a given momentum and polarization can be written as ak, |0i.

Hak, |0i = ak, H + [H, ak, ] |0i = 0 + ~ak, |0i = ~ak, |0i
The energy of single photon state is ~.

P ak, |0i = ak, P + [P, ak, ] |0i = 0 + ~~kak, |0i = ~~kak, |0i
The momentum of the single photon state is ~~k. The mass of the photon can be
computed.
E2
mc2

= p2 c2 + (mc2 )2
p
p
(~)2 (~k)2 c2 = ~ 2 2 = 0
=

So the energy, momentum, and mass of a single photon state are as we would expect.
The vector potential has been given two transverse polarizations as expected from
classical Electricity and Magnetism. The result is two possible transverse polarization
vectors in our quantized field. The photon states are also labeled by one of two polarizations, that we have so far assumed were linear polarizations. The polarization
vector, and therefore the vector potential, transform like a Lorentz vector. We know
that the matrix element of vector operators is associated with an angular momentum
of one. When a photon is emitted, selection rules indicate it is carrying away an angular momentum of one, so we deduce that the photon has spin one. We need not add
anything to our theory though; the vector properties of the field are already included
in our assumptions about polarization.
Of course we could equally well use circular polarizations which are related to the linear
set we have been using by
1 (1)
() = (
i
(2) ).
2
The polarization () is associated with the m = 1 component of the photons spin.
These are the transverse mode of the photon, ~k () = 0. We have separated the field
into transverse and longitudinal parts. The longitudinal part is partially responsible
for static E and B fields, while the transverse part makes up radiation. The m = 0
component of the photon is not present in radiation but is important in understanding
static fields.
By assuming the canonical coordinates and momenta in the Hamiltonian have commutators like those of the position and momentum of a particle, led to an understanding
that radiation is made up of spin-1 particles with mass zero. All fields correspond to
a particle of definite mass and spin. We now have a pretty good idea how to quantize
the field for any particle.
597

32. Quantum Theory of Radiation

32.9

TOC

The Time Development of Field Operators

The creation and annihilation operators are related to the time dependent coefficients in our Fourier expansion of the radiation field.
r
~c2
ak,
ck, (t) =
2
r
~c2
ck, (t) =
a
2 k,
This means that the creation, annihilation, and other operators are time dependent
operators as we have studied the Heisenberg representation. In particular, we derived
the canonical equation for the time dependence of an operator.
d
B(t)
dt

a k,

i
[H, B(t)]
~
i
i
[H, ak, (t)] = (~)ak, (t) = iak, (t)
~
~
i
[H, ak, (t)] = iak, (t)
~

So the operators have the same time dependence as did the coefficients in the
Fourier expansion.
ak, (t)

= ak, (0)eit

ak, (t)

= ak, (0)eit

We can now write the quantized radiation field in terms of the operators at t = 0.

1 X
A =
V k

~c2 ()
ak, (0)eik x + ak, (0)eik x
2

Again, the 4-vector x is a parameter of this field, not the location of a photon.
The field operator is Hermitian and the field itself is real.

598

32. Quantum Theory of Radiation

32.10

TOC

Uncertainty Relations and RMS Field Fluctuations

Since the fields are a sum of creation and annihilation operators, they do not commute with the occupation number operators
Nk, = ak, ak, .
Observables corresponding to operators which do not commute have an uncertainty
principle between them. So we cant fix the number of photons and know the
fields exactly. Fluctuations in the field take place even in the vacuum state, where we
know there are no photons.
Of course the average value of the Electric or Magnetic field vector is zero by symmetry.
To get an idea about the size of field fluctuations, we should look at the mean square
~ E|0i.
~
value of the field, for example in the vacuum state. We compute h0|E
~
E

~
1 A
c t r

1 X ~c2 ()

ak, (0)eik x + ak, (0)eik x

2
V k
r

1 X ~c2 ()

ak, (0)eik x + ak, (0)eik x
2
V k
r

1 1 X ~c2 ()
i

ak, (0)eik x + ak, (0)eik x
c V
2
k
r

i X ~ ()

ak, (0)eik x ak, (0)eik x
2
V k
r
i X ~ () ik x

ak, e
|0i
2
V k
1 X ~
1
V
2
k
1 X
~
V

~
A

~
E

~
E|0i
=
~ E|0i
~
h0|E
=
~ E|0i
~
h0|E
=

~
(Notice that we are basically taking the absolute square of E|0i
and that the
orthogonality of the states collapses the result down to a single sum.)
The calculation is illustrative even though the answer is infinite. Basically, a term
proportional to aa first creates one photon then absorbs it giving a nonzero
contribution for every oscillator mode. The terms sum to infinity but really its the
infinitesimally short wavelengths that cause this. Again, some cut off in the maximum
energy would make sense.
599

32. Quantum Theory of Radiation

In the previous section, we derived the same formulas for matrix elements that we had
earlier used to study decays of Hydrogen atom states with no applied EM field, that is
zero photons in the initial state.
in

(2)2 e2
~
|hn |eik~r p~|i i|2 (En Ei + ~)
2
m V

With the inclusion of the phase space integral over final states this became

e2 (Ei En ) X
~
tot =
dp |hn |eik~r () p~e |i i|2
2~2 m2 c3

The quantity ~k ~r is typically small for atomic transitions

1 2 2
mc
2
~
r a0 =
mc
1 2 mc ~

kr
=
2 ~ mc
2

E = pc = ~kc

Note that we have take the full binding energy as the energy difference between states
so almost all transitions will have kr smaller than this estimate. This makes ~k ~r an
excellent parameter in which to expand decay rate formulas.
~

The approximation that eik~r 1 is a very good one and is called the electric dipole
or E1 approximation. We previously derived the E1 selection rules.
` = 1.
m

s =

0, 1.
0.

The general E1 decay result depends on photon direction and polarization.

If information about angular distributions or polarization is needed, it can be pried out

602

32. Quantum Theory of Radiation

of this formula.
tot

e2 (Ei En ) X
2~2 m2 c3

TOC

d |hn |eik~r () p~e |i i|2

3 X
4
in
x
x + iy
3

d
r drRnn `n Rni `i dY`n mn z Y10 +
Y11 +
2
2c
2
3

Summing over polarization and integrating over photon direction, we get a

simpler formula that is quite useful to compute the decay rate from one initial atomic
state to one final atomic state.
3
4in
|~rni |2
3c2
is the matrix element of the coordinate vector between final and initial states.

tot =

Here ~rni

For single electron atoms, we can sum over the final states with different m and
get a formula only requires us to do a radial integral.

2

`+1
3

4in 2`+1
`+1

3
0

R
R
r
dr
for
`
=
tot =
0
0
n`
`
n`

`1
3c2

2`+1
0

The decay rate does not depend on the m of the initial state.

32.12.1

Beyond the Electric Dipole Approximation

Some atomic states have no lower energy state that satisfies the E1 selection rules to
decay to. Then, higher order processes must be considered. The next order term in
~
the expansion of eik~r = 1 i~k ~r + ... will allow other transitions to take place but
at lower rates. We will attempt to understand the selection rules when we include
the i~k ~r term.
The matrix element is proportional to ihn |(~k ~r)(
() p~e )|i i which we will split
up into two terms. You might ask why split it. The reason is that we will essentially
be computing matrix elements of at tensor and dotting it into two vectors that
do not depend on the atomic state.
~k hn |~rp~e )|i i ()
Putting these two vectors together is like adding to ` = 1 states. We can get total
angular momentum quantum numbers 2, 1, and 0. Each vector has three components.
The direct product tensor has 9. Its another case of
3 3 = 5S 3A 1S .
603

32. Quantum Theory of Radiation

TOC

The tensor we make when we just multiply two vectors together can be
reduced into three irreducible (spherical) tensors. These are the ones for which
we can use the Wigner-Eckart theorem to derive selection rules. Under rotations of
the coordinate axes, the rotation matrix for the 9 component Cartesian tensor will be
block diagonal. It can be reduced into three spherical tensors. Under rotations the 5
component (traceless) symmetric tensor will always rotate into another 5 component
symmetric tensor. The 3 component anti symmetric tensor will rotate into another
antisymmetric tensor and the part proportional to the identity will rotate into the
identity.
1
1
() p~) + (~k p~)(
() ~r)] + [(~k ~r)(
() p~) (~k p~)(
() ~r)]
(~k ~r)(
() p~e ) = [(~k ~r)(
2
2
The first term is symmetric and the second anti-symmetric by construction.
The first term can be rewritten.
1
hn |[(~k ~r)(
() p~) + (~k p~)(
() ~r)]|i i =
2
=
=

1~
k hn |[~rp~ + p~~r]|i i ()
2
1 ~ im
k
hn |[H0 , ~r~r]|i i ()
2
~
im ~

k hn |~r~r|i i ()
2

This makes the symmetry clear. Its normal to remove the trace of the tensor:

~r~r ~r~r 3ij r2 . The term proportional to ij gives zero because ~k = 0. The traceless
symmetric tensor has 5 components like an ` = 2 operator; The anti-symmetric tensor
has 3 components; and the trace term has one. This is the separation of the Cartesian
tensor into irreducible spherical tensors. The five components of the traceless
symmetric tensor can be written as a linear combination of the Y2m .
Similarly, the second (anti-symmetric) term can be rewritten slightly.
1 ~
[(k ~r)(
() p~) (~k p~)(
() ~r)] = (~k () ) (~r p~)
2
The atomic state dependent part of this, ~r p~, is an axial vector and therefore
has three components. (Remember and axial vector is the same thing as an antisymmetric tensor.) So this is clearly an ` = 1 operator and can be expanded in
terms of the Y1m . Note that it is actually a constant times the orbital angular
~
momentum operator L.
So the first term is reasonably named the Electric Quadrupole term because
it depends on the quadrupole moment of the state. It does not change parity and
gives us the selection rule.
|`n `i | 2 `n + `i
604

32. Quantum Theory of Radiation

TOC

The second term dots the radiation magnetic field into the angular momentum of the
atomic state, so it is reasonably called the magnetic dipole interaction. The
interaction of the electron spin with the magnetic field is of the same order and
should be included together with the E2 and M1 terms.
e~ ~
(k () ) ~
2mc
Higher order terms can be computed but its not recommended.
Some atomic states, such as the 2s state of Hydrogen, cannot decay by any of
these terms basically because the 2s to 1s is a 0 to 0 transition and there is no way to
conserve angular momentum and parity. This state can only decay by the emission of
two photons.
While E1 transitions in hydrogen have lifetimes as small as 109 seconds, the E2 and
M1 transitions have lifetimes of the order of 103 seconds, and the 2s state
has a lifetime of about 17 of a second.

32.13

Black Body Radiation Spectrum

We are in a position to fairly easily calculate the spectrum of Black Body radiation.
Assume there is a cavity with a radiation field on the inside and that the field
interacts with the atoms of the cavity. Assume thermal equilibrium is reached.
Lets take two atomic states that can make transitions to each other: A
B + and B + A. From statistical mechanics, we have
eEb /kT
NB
= E /kT = e~/kT
NA
e A
and for equilibrium we must have
NB absorb
NB
NA

= NA emit
emit
=
absorb

We have previously calculated the emission and absorption rates. We can

calculate the ratio between the emission and absorption rates per atom:

2
P

~
(n~k, + 1) hB|eikr~i () p~i |Ai
NB
emit
i
=
=

2
NA
absorb
P

~
i
k
r
~
()

i
n~k, hA|e
p~i |Bi
i
605

32. Quantum Theory of Radiation

TOC

where the sum is over atomic electrons. The matrix elements are closely related.
~

pi () eikr~i |Bi = hA|eikr~i () p~i |Bi

hB|eikr~i () p~i |Ai = hA|~
We have used the fact that ~k = 0.
The two matrix elements are simple
complex conjugates of each other so that when we take the absolute square, they
are the same. Therefore, we may cancel them.
(n~k, + 1)
NB
= e~/kT
=
NA
n~k,
1 = n~k, (e~/kT 1)
n~k, =

1
e~/kT

Now suppose the walls of the cavity are black so that they emit and absorb
photons at any energy. Then the result for the number of photons above is true for all
the radiation modes of the cavity. The energy in the frequency interval (, +d)
per unit volume can be calculated by multiplying the number of photons by the
energy per photon times the number of modes in that frequency interval and dividing
by the volume of the cavity.
U ()d =

~
e~/kT 1

U () = 8

L
2

4k 2 dk

1
L3

dk
d

1
2

1
8~ 3
1
U () = 3
~/kT
c
2 e
1
d
8
h 3
U () = U ()
= 3 ~/kT
d
c e
1
e~/kT

This was the formula Planck used to start the revolution.

606

33. Scattering of Photons

TOC

Scattering of Photons

In the scattering of photons, for example from an atom, an initial state photon with
wave-number ~k and polarization is absorbed by the atom and a final state photon
with wave-number ~k 0 and polarization 0 is emitted. The atom may remain in the same
state (elastic scattering) or it may change to another state (inelastic). Any calculation
we will do will use the matrix element of the interaction Hamiltonian between
initial and final states.
Hni

Hint

0
hn; ~k 0 ( ) |Hint |i; ~k
() i
e ~
e2 ~ ~

A(x) p~ +
AA
mc
2mc2

The scattering process clearly requires terms in Hint that annihilate one photon
e2 ~
~
and create another. The order does not matter. The 2mc
2 A A is the square of
the Fourier decomposition of the radiation field so it contains terms like ak0 ,0 ak,
e ~
A p~ term has both creation
and ak, ak0 ,0 which are just what we want. The mc
and annihilation operators in it but not products of them. It changes the number
of photons by plus or minus one, not by zero as required for the scattering process.
Nevertheless this part of the interaction could contribute in second order perturbation
theory, by absorbing one photon in a transition from the initial atomic state to an
intermediate state, then emitting another photon and making a transition to the final
atomic state. While this is higher order in perturbation theory, it is the same order in
the electromagnetic coupling constant e, which is what really counts when expanding
e2 ~
~
in powers of . Therefore, we will need to consider the 2mc
2 A A term in first
e ~
order and the mc A p~ term in second order perturbation theory to get an order
calculation of the matrix element.
Start with the first order perturbation theory term. All the terms in the sum that do
not annihilate the initial state photon and create the final state photon give zero. We
will assume that the wavelength of the photons is long compared to the size of the

607

33. Scattering of Photons

TOC

atom so that eik~r 1.

A (x)

0
e2
~ A|i;
~ ~k
hn; ~k 0 ( ) |A
() i =
2
2mc

=
=
=

1 X

V k
e2 1
2mc2 V
e2 1
2mc2 V
e2 1
2mc2 V
e2 1
2mc2 V

~c2 ()
ak, (0)eik x + ak, (0)eik x
2
~c2 () (0 ) ~ 0 (0 )

hn; k | ak, ak0 ,0 + ak0 ,0 ak,

2 0
0
~c2 () (0 ) i(0 )t ~ 0 (0 )

e
hn; k |2|i; k~0 ( ) i
2 0
~c2 () (0 ) i(0 )t

e
2hn|ii
2 0
~c2 () (0 ) i(0 )t

e
ni
0

This is the matrix element Hni (t). The amplitude to be in the final state
0
|n; ~k 0 ( ) i is given by first order time dependent perturbation theory.
c(1)
n (t)

1
i~

eini t Hni (t0 )dt0

(1)
c ~ 0 (0 ) (t)
n;k

1 e2 1 ~c2 () (0 )

ni
i~ 2mc2 V 0
0
e2

() ( ) ni
0
2imV

eini t ei( )t dt0

0
0

ei(ni + )t dt0
0

Recall that the absolute square of the time integral will turn into 2t(ni + 0 ).
We will carry along the integral for now, since we are not yet ready to square it.
Now we very carefully put the interaction term into the formula for second order
~
time dependent perturbation theory, again using eik~x 1. Our notation is that
the intermediate state of atom and field is called |Ii = |j, n~k, , n~k0 ,0 i where j
represents the state of the atom and we may have zero or two photons, as indicated in

608

33. Scattering of Photons

TOC

the diagram.
V

e ~
e 1 X

=
A p~ =
mc
mc V

~
k

c(2)
n (t)

1 X
~2

t2
dt2 VnI (t2 )e

inj t2

j,~
k, 0

(2)
(t)
n;~
k0 (0 )

e2 X 1 ~c2

m2 c2 ~2
V 2 0
I
t2

~c2 ()
p~ ak, eit + ak, eit
2
dt1 eiji t1 VIi (t1 )

dt2 hn; ~k 0 ( ) |(
() ak, eit2 + ( ) ak0 ,0 ei t2 ) p~|
0
0

dt1 eiji t1 hI|(

() ak, eit1 + ( ) ak0 ,0 ei t1 ) p~|i; ~k
() i

We can understand this formula as a second order transition from state |ii to state |ni
through all possible intermediate states. The transition from the initial state to the
intermediate state takes place at time t1 . The transition from the intermediate state
to the final state takes place at time t2 .
The space-time diagram below shows the three terms in cn (t) Time is assumed to
run upward in the diagrams.

Diagram (c) represents the A2 term in which one photon is absorbed and one emitted
at the same point. Diagrams (a) and (b) represent two second order terms. In diagram
(a) the initial state photon is absorbed at time t1 , leaving the atom in an intermediate
state which may or may not be the same as the initial (or final) atomic state. This
intermediate state has no photons in the field. In diagram (b), the atom emits the final
state photon at time t1 , leaving the atom in some intermediate state. The intermediate
state |Ii includes two photons in the field for this diagram. At time t2 the atom absorbs
the initial state photon.
609

33. Scattering of Photons

TOC

Looking again at the formula for the second order scattering amplitude, note that we
integrate over the times t1 and t2 and that t1 < t2 . For diagram (a), the annihilation
operator ak, is active at time t1 and the creation operator is active at time t2 . For
diagram (b) its just the opposite. The second order formula above contains four terms
as written. The a a and aa terms are the ones described by the diagram. The aa
and a a terms will clearly give zero. Note that we are just picking the terms that will
survive the calculation, not changing any formulas.
Now, reduce to the two nonzero terms. The operators just give a factor of 1 and make
the photon states work out. If |ji is the intermediate atomic state, the second order
term reduces to.

h
X
0
0
e2

dt2
dt1 ei( +nj )t2 hn|
( ) p~|jihj|
() p~|iiei(ji )
2V m2 ~ 0 j
0
0
i
0
0
i(nj )t2
()
e
hn|
p~|jihj|
() p~|iiei( +ji )t1
"
t
i(ji )t1 t2
X
e
e2
i( 0 +nj )t2
0

dt2 e
hn|
p~|jihj|
p~|ii
2
0
i(ji ) 0
2V m ~ j
0

"
#t2
0
ei( +ji )t1
i(nj )t2
0

e
hn|
p~|jihj|
p~|ii
i( 0 + ji )
t

(2)
c ~ 0 (0 ) (t)
n;k

=
+

(2)
(t)
n;~
k0 0

i(ji )t2

X
e
1
e2
i( 0 +nj )t2
0

dt2 e
hn|
p~|jihj|
p~|ii
i(ji )
2V m2 ~ 0 j
0
"
##
0
ei( +ji )t2 1
ei(nj )t2 hn|
p~|jihj|
0 p~|ii
i( 0 + ji )

The 1 terms coming from the integration over t1 can be dropped. We can anticipate
that the integral over t2 will eventually give us a delta function of energy conservation,
going to infinity when energy is conserved and going to zero when it is not. Those
1 terms can never go to infinity and can therefore be neglected. When the energy
conservation is satisfied, those terms are negligible and when it is not, the whole thing

610

33. Scattering of Photons

TOC

goes to zero.

X
0
e2
1

0 p~|jihj|
dt2 ei(ni + )t2 hn|
p~|ii
i(ji )
2V m2 ~ 0 j
0

0
1
ei(ni + )t2 hn|
p~|jihj|
0 p~|ii
0
i( + ji )

X hn|
p~|jihj|
0 p~|ii
0 p~|jihj|
p~|ii hn|
e2

+
ji
0 + ji
2iV m2 ~ 0 j
t

(2)
c ~ 0 0 (t)
n;k

+
(2)
(t)
n;~
k0 0

dt2 ei(ni + )t2

We have calculated all the amplitudes. The first order and second order amplitudes should be combined, then squared.
cn (t)
(1)
c ~ 0 0 (t)
n;k

(2)
= c(1)
n (t) + cn (t)

0 ni
0
2iV m

ei(ni + )t dt0
0

t
0
e2
p~|jihj|
0 p~|ii
hn|
0 p~|jihj|
p~|ii hn|

=
+
dt2 ei(ni +
0+
2
0

2iV m ~ j
ji
ji
0

0
0
X hn|
1

p
~
|jihj|

p
~
|ii
hn|

p
~
|jihj|

p
~
|ii

cn;~k0 0 (t) = ni 0
+
m~ j
ji
0 + ji
X

(2)
c ~ 0 0 (t)
n;k

|c(t)|2

t
0
e2

dt2 ei(ni + )t2

0
2iV m
0

2
0
0
X

hn|
p~|jihj|
p~|ii hn|
p~|jihj|
p~|ii
ni 0 1
+

0
m~ j
ji
+ ji

2
t

e4
dt2 ei(ni +0 )t2

2
2
0
4V m

0

2
0
0
X

hn|
p~|jihj|
p~|ii hn|
p~|jihj|
p~|ii
ni 0 1
+

0
m~ j
ji
+ ji

e4
2t(ni + 0 )
4V 2 m2 0
611

33. Scattering of Photons

d
d

V d3 k 0
(2)3

TOC

2
0
0
X hn|

1

p
~
|jihj|

p
~
|ii
hn|

p
~
|jihj|

p
~
|ii
ni 0

+

0
m~ j
ji
+ ji

e4
2(ni + 0 )
4V 2 m2 0

2

0
0
X hn|

1

p
~
|jihj|

p
~
|ii
hn|

p
~
|jihj|

p
~
|ii
V 02 d 0 d

ni 0
+

3
0
(2c)
m~ j
ji
+ ji

e4
2(ni + 0 )
4V 2 m2 0

2

0
0
X hn|

p
~
|jihj|

p
~
|ii
hn|

p
~
|jihj|

p
~
|ii
1

d ni 0
+

0
m~ j
ji
+ ji

V 02
e4
2
3
2
(2c) 4V m2 0

2
0
4 0
0
X

hn|
p~|jihj|
p~|ii hn|
e
p~|jihj|
p~|ii
ni 0 1
+

2
2
3
0
(4) V m c
m~ j
ji
+ ji

Note that the delta function has enforced energy conservation requiring that 0 =
ni , but we have left 0 in the formula for convenience.
The final step to a differential cross section is to divide the transition rate by the
incident flux of particles. This is a surprisingly easy step because we are using
plane waves of photons. The initial state is one particle in the volume V moving
with a velocity of c, so the flux is simply Vc .
d
e4 0
=
d
(4)2 m2 c4

2
0
0
X hn|

1

p
~
|jihj|

[H, ~x]

~
p~
im

~
(
p~)ij
im

(
[H, ~x])ij

~ij (
~x)ij
i
=
(
p~)ij
mij

X i
=
(
p~)ij (
0 p)ji
m
ij
j
X i
=
(
p~)ij (
0 p)ji +
m
ij
j
=

(
~x)ij
0

(Reminder: ij =
between states.)

i 0
(
p~)ij (
p~)ji
mji

i 0
(
p~)ij (
p~)ji
mij

X i
[(
p~)ij (
0 p)ji + (
0 p~)ij (
p~)ji ]
m
ij
j

1 X 1
[(
0 p~)ij (
p~)ji + (
p~)ij (
0 p~)ji ]
m~ j ij

Ei Ej
~

is just a number. (
p~)ij = hi|
p~|ji is a matrix element

We may now combine the terms for elastic scattering.

2

2
X hi|

delas
e
0 p~|jihj|
p~|ii hi|
p~|jihj|
0 p~|ii
ii 0 1
=
+

d
4mc2
m~ j
ji
ji +

0 p~|jihj|
p~)|ii hi|
p~|jihj|
0 p~|ii
1 X hi|
+
ii 0 =
m~ j
ij
ij
1
1
+
ij
ji
delas
d

ji + ij

=
ij (ji )
ji (ji )

2
2 X
2
2
0
0

e
1
hi|
p~|jihj|
p~|ii hi|
p~|jihj|
p~|ii

=

4mc2
m~ j
ji (ji )
ji (ji + )

=

This is a nice symmetric form for elastic scattering. If computation of the

matrix elements is planned, it useful to again use the commutator to change p~ into ~x.
615

33. Scattering of Photons

delas
=
d

33.3

e
4mc2

TOC

X

2

0
0
2

m
hi|
~x|ji hj|
~x|ii hi|
~x|ji hj|
~x|ii
ji

~
ji
ji +
j

Rayleigh Scattering

Lord Rayleigh calculated low energy elastic scattering of light from atoms using
classical electromagnetism. If the energy of the scattered photon is much less than the
energy needed to excite an atom, << ji , then the cross section may be approximated.
ji
ji
delas
d

= (1
) = 1 +
ji (1 ji )
ji
ji

2
2

e2
m 2 X ji hi|
0 ~x|ji hj|
~x|ii ji hi|
~x|ji hj|
0 ~x|ii
=

4mc2
~
ji
ji +

j

2
e2
m 2 X
=
[(hi|
0 ~x|ji hj|
~x|ii hi|
~x|ji hj|
0 ~x|ii)

4mc2
~
j
2

0
0
+
(hi|
~x|ji hj|
~x|ii + hi|
~x|ji hj|
~x|ii)
ji

2
2

e2
m 2 4 X 1
=

(hi|
0 ~x|ji hj|
~x|ii + hi|
~x|ji hj|
0 ~x|ii)
2
4mc
~
j ji

=

For the colorless gasses (like the ones in our atmosphere), the first excited state in
the UV, so the scattering of visible light with be proportional to 4 , which explains
why the sky is blue and sunsets are red. Atoms with intermediate states in the visible
will appear to be colored due to the strong resonances in the scattering. Rayleigh
got the same dependence from classical physics.

33.4

Thomson Scattering

If the energy of the scattered photon is much bigger than the binding energy of the
atom, >> 1 eV. then cross section approaches that for scattering from a free
electron, Thomson Scattering. We still neglect the effect of electron recoil so we
616

33. Scattering of Photons

TOC

should also require that ~ << me c2 . Start from the Kramers-Heisenberg formula.

2 0

2
2
0
0
X

1
d
e

hn|
p~|jihj|
p~|ii hn|
p~|jihj|
p~|ii
0

=
+
ni

d
4mc2

m~ j
ji
ji + 0

0

~
p|ii
The ~ = ~ 0 denominators are much larger than hn| ~p|jihj|
which is of the order
m
of the electrons kinetic energy, so we can ignore the second two terms. (Even if the
intermediate and final states have unbound electrons, the initial state wave function
will keep these terms small.)

d
=
d

e2
4mc2

|
0 |

This scattering cross section is of the order of the classical radius of the
electron squared, and is independent of the frequency of the light.
The only dependence is on polarization. This is a good time to take a look at the
meaning of the polarization vectors weve been carrying around in the calculation
and at the lack of any wave-vectors for the initial and final state. A look back at the
calculation shows that we calculated the transition rate from a state with one pho0
ton with wave-vector ~k and polarization () to a final state with polarization ( ) .
We have integrated over the final state wave vector magnitude, subject to
the delta function giving energy conservation, but, we have not integrated over
d
final state photon direction yet, as indicated by the d
. There is no explicit angular dependence but there is some hidden in the dot product between initial
and final polarization vectors, both of which must be transverse to the direction of propagation. We are ready to compute four different differential cross sections
corresponding to two initial polarizations times two final state photon polarizations.
Alternatively, we average and/or sum, if we so choose.
In the high energy approximation we have made, there is no dependence on the state
of the atoms, so we are free to choose our coordinate system any way we want.
Set the z-axis to be along the direction of the initial photon and set the x-axis
so that the scattered photon is in the x-z plane ( = 0). The scattered photon is
at an angle to the initial photon direction and at = 0. A reasonable set of initial
state polarization vectors is
(1)

= x

(2)

= y

Pick (1) to be in the scattering plane (x-z) defined as the plane containing both ~k
0
0
and ~k 0 and (2) to be perpendicular to the scattering plane. (1) is then at an angle
617

33. Scattering of Photons

TOC

to the x-axis. (2) is along the y-axis. We can compute all the dot products.
(1) (1)

(1) (2)

(2)

(1)0

(2)0

cos

From these, we can compute any cross section we want. For example, averaging over
initial state polarization and summing over final is just half the sum of the
squares of the above.
d
=
d

e2
4mc2

1
(1 + cos2 )
2

Even if the initial state is unpolarized, the final state can be polarized. For
0
example, for = 2 , all of the above dot products are zero except (2) (2) = 1. That
means only the initial photons polarized along the y direction will scatter and that the
scattered photon is 100% polarized transverse to the scattering plane (really
just the same polarization as the initial state). The angular distribution could also be
used to deduce the polarization of the initial state if a large ensemble of initial state
photons were available.
For a definite initial state polarization (at an angle to the scattering plane, the
component along (1) is cos and along (2) is sin . If we dont observe final state
polarization we sum (cos cos )2 + (sin )2 and have
d
=
d

e2
4mc2

1
(cos2 cos2 + sin2 )
2

For atoms with more than one electron, this cross section will grow as Z 4 .

33.5

Raman Effect

The Kramers-Heisenberg formula clearly allows for the initial and final state to be
different. The atom changes state and the scattered photon energy is not equal to
the initial photon energy. This is called the Raman effect. Of course, total energy is
still conserved. A given initial photon frequency will produce scattered photons with
definite frequencies, or lines in the spectrum.

618

34. Electron Self Energy Corrections

TOC

Electron Self Energy Corrections

If one calculates the energy of a point charge using classical electromagnetism,

the result is infinite, yet as far as we know, the electron is point charge. One can
calculate the energy needed to assemble an electron due, essentially, to the interaction
of the electron with its own field. A uniform charge distribution with the classical
radius of an electron, would have an energy of the order of me c2 . Experiments
have probed the electrons charge distribution and found that it is consistent with a
point charge down to distances much smaller than the classical radius. Beyond classical
calculations, the self energy of the electron calculated in the quantum theory
of Dirac is still infinite but the divergences are less severe.
At this point we must take the unpleasant position that this (constant) infinite
energy should just be subtracted when we consider the overall zero of energy (as
we did for the field energy in the vacuum). Electrons exist and dont carry infinite
amount of energy baggage so we just subtract off the infinite constant. Nevertheless,
we will find that the electrons self energy may change when it is a bound
state and that we should account for this change in our energy level calculations. This
calculation will also give us the opportunity to understand resonant behavior
in scattering.
We can calculate the lowest order self energy corrections represented by the
two Feynman diagrams below.

In these, a photon is emitted then reabsorbed. As we now know, both of these amplitudes are of order e2 . The first one comes from the A2 term in which the number of
~ p~ term in second
photons changes by zero or two and the second comes from the A
order time dependent perturbation theory. A calculation of the first diagram will give
the same result for a free electron and a bound electron, while the second diagram
will give different results because the intermediate states are different if an electron is
619

34. Electron Self Energy Corrections

TOC

bound than they are if it is free. We will therefore compute the amplitude from
the second diagram.

Hint
~
A

e ~
A p~
mc r

1 X ~c2 ()
~
~

a~k, ei(k~xt) + a~ ei(k~xt)
k,
2
V ~

=
=

This contains a term causing absorption of a photon and another term causing emission.
We separate the terms for absorption and emission and pull out the time dependence.

X
abs it
emit it
Hint =
H~k,
e
+ H~k,
e
~
k,

abs

H emit

~e2
~
a~ eik~x p~ ()
2m2 V k,
r
~e2
~
=
a eik~x p~ ()
2m2 V ~k,
=

The initial and final state is the same |ni, and second order perturbation theory will
involve a sum over intermediate atomic states, |ji and photon states. We will
use the matrix elements of the interaction Hamiltonian between those states.
Hjn

emit
= hj|H~k,
|ni

Hnj

abs
= hn|H~k,
|ji

Hnj

= Hjn

We have dropped the subscript on Hjn specifying the photon emitted or absorbed
leaving a reminder in the sum. Recall from earlier calculations that the creation and
annihilation operators just give a factor of 1 when a photon is emitted or absorbed.
From time dependent perturbation theory, the rate of change of the amplitude to be
in a state is given by
i~

cj (t)
t

Hjk (t)ck (t)eijk t

In this case, we want to use the equations for the the state we are studying, n , and all
intermediate states, j plus a photon. Transitions can be made by emitting a photon
from n to an intermdiate state and transitions can be made back to the state n from
620

34. Electron Self Energy Corrections

TOC

any intermediate state. We neglect transitions from one intermediate state to another
as they are higher order. (The diagram is emit a photon from n then reabsorb it.)
The differential equations for the amplitudes are then.
i~

dcj
dt

dcn
dt

Hjn eit cn einj t

~
k,

Hnj eit cj einj t

In the equations for cn , we explicitly account for the fact that an intermediate
state can make a transition back to the initial state. Transitions through
another intermediate state would be higher order and thus should be neglected. Note
that the matrix elements for the transitions to and from the initial state are closely
related. We also include the effect that the initial state can become depleted
as intermediate states are populated by using cn (instead of 1) in the equation for cj .
Note also that all the photon states will make nonzero contributions to the sum.
Our task is to solve these coupled equations. Previously, we did this by integration,
but needed the assumption that the amplitude to be in the initial state was 1.
Since we are attempting to calculate an energy shift, let us make that assumption and plug it into the equations to verify the solution.
cn = e

iEn t
~

En will be a complex number, the real part of which represents an energy

shift, and the imaginary part of which represents the lifetime (and energy

621

34. Electron Self Energy Corrections

TOC

width) of the state.

dcj
dt

1 X
i~

Hjn eit cn einj t

~
k,

cj (t)

iEn t
~

dt0 Hjn eit e

iEn t0
~

einj t

~
k, 0

cj (t)

1 X
i~

dt0 Hjn ei(nj n +)t

~
k, 0

"
cj (t)

Hjn

~
k,

cj (t)

ei(nj n +)t
~(nj + n )

#t
0

i(nj n +)t

Hjn

~
k,

e
1
~(nj + n )

Substitute this back into the differential equation for cn to verify the solution and to
find out what En is. Note that the double sum over photons reduces to a single
sum because we must absorb the same type of photon that was emitted. (We have not
explicitly carried along the photon state for economy.)
dcn
dt

cj (t)

XX
~
k,

Hjn

~
k,

dcn
dt

ei(nj n +)t 1
~(nj + n )

En eiEn t/~ =

XX
~
k,

Hnj eit cj einj t

~
k,

|Hnj |2 ei(nj +n )t

Hnj Hjn eit einj t

|Hnj |2
~(nj )

abs
= hn|H~k,
|ji
r
~e2
~
eik~x p~ ()
=
2m2 V
~
~e2 X X |hn|eik~x p~ () |ji|2
=
2m2 V
~(nj )
j
~
k,

~
V d3 k X X |hn|eik~x p~ () |ji|2
(2)3 j
(nj )
~
XX
e2
k 2 dk |hn|eik~x p~ () |ji|2
d
(2)3 2m2 j

(nj )

~
XX
e2
|hn|eik~x p~|ji () |2
d
d
(2)3 2m2 c3 j
(nj )

e2
2m2 V

In our calculation of the total decay rate summed over polarization and integrated over
photon direction we computed the cosine of the angle between each polarization vector
and the (vector) matrix element. Summing these two and integrating over photon
624

34. Electron Self Energy Corrections

TOC

direction we got a factor of 8

3 and the polarization is eliminated from the matrix
element. The same calculation applies here.
En

X 8 |hn|ei~k~x p~|ji|2
e2
d
(2)3 2m2 c3 j 3
(nj )
X |hn|ei~k~x p~|ji|2
e2
d
6 2 m2 c3 j
(nj )

~
2~ X |hn|eik~x p~|ji|2
d
3m2 c2 j
(nj )

Note that we wish to use the electric dipole approximation which is not valid for
large k = c . It is valid up to about 2000 eV so we wish to cut off the calculation around
there. While this calculation clearly diverges, things are less clear here because of the
~
eventually rapid oscillation of the eik~x term in the integrand as the E1 approximation
fails. Nevertheless, the largest differences in corrections between free electrons and
bound electrons occur in the region in which the E1 approximation is valid. For now
we will just use it and assume the cut-off is low enough.
It is the difference between the bound electrons self energy and that for a
free electron in which we are interested. Therefore, we will start with the free
electron with a definite momentum p~. The normalized wave function for the free
electron is 1V ei~p~x/~ .
Ef ree

=
=

2
X

2~

ei~p~x/~ p~ei~p0 ~x/~ d3 x d

2
2
2
3m c V
(nj )
p
~0

2
X

2~

2
ei(~p0 ~x/~~p~x/~) d3 x d
|~
p
|

2
2
2
3m c V
(nj )
p
~0
X

2~
2
|~
p
|
|V p~0 ,~p |2 d
3m2 c2 V 2
(

)
nj
0
p
~

2~
|~
p|2
3m2 c2

d
(nj )

2~
|~
p|2
3m2 c2

d
(nj )

It easy to see that this will go to negative infinity if the limit on the integral is infinite. It
is quite reasonable to cut off the integral at some energy beyond which the theory
we are using is invalid. Since we are still using non-relativistic quantum mechanics,
625

34. Electron Self Energy Corrections

TOC

the cut-off should have ~ << mc2 . For the E1 approximation, it should be ~ <<
2~c/1 = 10keV . We will approximate (nj) 1 since the integral is just giving us
a number and we are not interested in high accuracy here. We will be more interested
in accuracy in the next section when we compute the difference between free electron
and bound electron self energy corrections.

Ef ree

2~
|~
p |2
3m2 c2

Ecutof
f /~

d
(nj )

2~
|~
p|2
3m2 c2

Ecutof
f /~

d
0

2~
=
|~
p|2 Ecutof f /~
3m2 c2
2
=
|~
p|2 Ecutof f
3m2 c2
= C|~
p|2

If we were hoping for little dependence on the cut-off we should be disappointed. This
self energy calculated is linear in the cut-off.
2

p
For a non-relativistic free electron the energy 2m
decreases as the mass of the
electron increases, so the negative sign corresponds to a positive shift in the
electrons mass, and hence an increase in the real energy of the electron. Later, we
will think of this as a renormalization of the electrons mass. The electron starts
off with some bare mass. The self-energy due to the interaction of the electrons
charge with its own radiation field increases the mass to what is observed.

Note that the correction to the energy is a constant times p2 , like the nonrelativistic formula for the kinetic energy.
C
p2
2mobs
1
mobs
mobs

=
=
=
=

2
Ecutof f
3m2 c2
p2
Cp2
2mbare
1
2C
mbare
mbare
(1 + 2Cmbare )mbare (1 + 2Cm)mbare
1 2Cmbare
4Ecutof f
(1 +
)mbare
3mc2

If we cut off the integral at me c2 , the correction to the mass is only about
626

34. Electron Self Energy Corrections

TOC

0.3%, but if we dont cut off, its infinite. It makes no sense to trust our non-relativistic
calculation up to infinite energy, so we must proceed with the cut-off integral.
If we use the Dirac theory, then we will be justified to move the cut-off up to very high
energy. It turns out that the relativistic correction diverges logarithmically (instead of
linearly) and the difference between bound and free electrons is finite relativistically
(while it diverges logarithmically for our non-relativistic calculation).
Note that the self-energy of the free electron depends on the momentum of the electron,
so we cannot simply subtract it from our bound state calculation. (What p2 would we
choose?) Rather me must account for the mass renormalization. We used the
observed electron mass in the calculation of the Hydrogen bound state energies.
In so doing, we have already included some of the self energy correction and we must
not double correct. This is the subtraction we must make.
Its hard to keep all the minus signs straight in this calculation, particularly if we consider the bound and continuum electron states separately. The free particle correction
to the electron mass is positive. Because we ignore the rest energy of the electron
in our non-relativistic calculations, This makes a negative energy correction to both
p2
). Bound states and continuum
the bound (E = 2n1 2 2 mc2 ) and continuum (E 2m
states have the same fractional change in the energy. We need to add back in a positive
term in En to avoid double counting of the self-energy correction. Since the bound
state and continuum state terms have the same fractional change, it is convenient to
p2
just use 2m
for all the corrections.

p2
2mobs

p2
Cp2
2mbare

En(obs)

En + Chn|p2 |ni = En +

2
Ecutof f hn|p2 |ni
3m2 c2

Because we are correcting for the mass used to calculate the base energy of the state
|ni, our correction is written in terms of the electrons momentum in that state.

34.1

The Lamb Shift

In 1947, Willis E. Lamb and R. C. Retherford used microwave techniques to determine

the splitting between the 2S 12 and 2P 12 states in Hydrogen to have a frequency
of 1.06 GHz, (a wavelength of about 30 cm). (The shift is now accurately measured to
be 1057.864 MHz.) This is about the same size as the hyperfine splitting of the ground
state.
627

34. Electron Self Energy Corrections

TOC

The technique used was quite interesting. They made a beam of Hydrogen atoms in
the 2S 12 state, which has a very long lifetime because of selection rules. Microwave
radiation with a (fixed) frequency of 2395 MHz was used to cause transitions to the
2P 32 state and a magnetic field was adjusted to shift the energy of the states
until the rate was largest. The decay of the 2P 23 state to the ground state was
observed to determine the transition rate. From this, they were able to deduce the
shift between the 2S 12 and 2P 12 states.
Hans Bethe used non-relativistic quantum mechanics to calculate the selfenergy correction to account for this observation.

628

34. Electron Self Energy Corrections

TOC

We now can compute the correction the same way he did.

En(obs)
En

En(obs)

2
Ecutof f hn|p2 |ni
En + Chn|p2 |ni = En +
3m2 c2

~
2~ X |hn|eik~x p~|ji|2
=
d
3m2 c2 j
(nj )

cutof
f X

2~
~

|hn|eik~x p~|ji|2 + hn|p2 |ni d

=
3m2 c2
(

)
nj
j

2~
3m2 c2
2~
3m2 c2

0
cutof
f

j
0
cutof
f

i~
k~
x
2
|hn|e p~|ji| + hn|~
p|jihj|~
p|ni d
(nj )

~
|hn|eik~x p~|ji|2 + |hn|~
p|ji|2 d
(nj )

It is now necessary to discuss approximations needed to complete this calculation.

In particular, the electric dipole approximation will be of great help, however, it is
certainly not warranted for large photon energies. For a good E1 approximation we
need E << 1973 eV. On the other hand, we want the cut-off for the calculation to be
of order wcutof f mc2 /~. We will use the E1 approximation and the high cut-off, as
Bethe did, to get the right answer. At the end, the result from a relativistic calculation
629

34. Electron Self Energy Corrections

TOC

can be tacked on to show why it turns out to be the right answer. (We arent aiming
for the worlds best calculation anyway.)
En(obs)

2~
3m2 c2
2~
3m2 c2

cutof
f

j
0
cutof
f

|hn|~
p|ji|2 + |hn|~
p|ji|2 d
(nj )

X + (nj )
|hn|~
p|ji|2 d
(

)
nj
j

cutof
f

2~ X

3m2 c2 j

2~ X

nj [log( nj )]0 cutof f |hn|~

p|ji|2
3m2 c2 j

2~ X
nj [log(|nj |) log(cutof f nj )] |hn|~
p|ji|2
3m2 c2 j

nj
|hn|~
p|ji|2 d
nj

2~ X
nj [log(|nj |) log(cutof f )] |hn|~
p|ji|2
3m2 c2 j

2~ X
|nj |
nj log
|hn|~
p|ji|2
3m2 c2 j
cutof f

The log term varies more slowly than does the rest of the terms in the sum. We can
approximate it by an average. Bethe used numerical calculations to determine
that the effective average of ~nj is 8.92 mc2 . We will do the same and pull the log
term out as a constant.
X

|
nj |
2~
nj |hn|~
p|ji|2
log
En(obs) =
3m2 c2
cutof f
j
This sum can now be reduced further to a simple expression proportional to the |n (0)|2
using a typical clever quantum mechanics calculation. The basic Hamiltonian
p2
for the Hydrogen atom is H0 = 2m
+ V (r).
[~
p, H0 ]

hj|[~
p, H0 ]|ni =
X
hn|~
p|jihj|[~
p, H0 ]|ni =
j

X
j

(Ei En )hn|~
p|jihj|~
p|ni =

[~
p, V ] =

~~
V
i

~ ~
hj|V |ni
i
~X
~ |ni
hn|~
p|ji hj|V
i j
~X
~ |ni
hn|~
p|jihj|V
i j
630

34. Electron Self Energy Corrections

TOC

This must be a real number so we may use its complex conjugate.

X
X
(Ei En )hn|~
(Ei En )hn|~
p|jihj|~
p|ni
p|jihj|~
p|ni =
j

~X ~
hn|V |jihj|~
p|ni
i j

~ X
~ |ni hn|V
~ |jihj|~
hn|~
p|jihj|V
p|ni
2i j

(Ei En )hn|~
p|jihj|~
p|ni =

~
~ ]|ni
hn|[~
p, V
2i
~2
= hn|2 V |ni
2
~2
= hn|e2 3 (~x)|ni
2
e2 ~2
=
|n (0)|2
2
=

Only the s states will have a non-vanishing probability to be at the origin

~
with |n00 (0)|2 = n13 a3 and a0 = mc
. Therefore, only the s states will shift in energy
0
appreciably. The shift will be.

2
2~
|
nj |
e ~ 1 mc 3
(obs)
En
=
log
3m2 c2
cutof f
2 n3
~

4 2
e mc
cutof f
=
log
3 2 ~n3
|
nj |

5
2
4 mc
cutof f
=
log
3n3
|
nj |

5
2
mc
mc2
(obs)
E2s
=
log
6
8.92 mc2

(obs)
5 mc2 c
1
E2s
=
log
= 1.041 GHz
=
2~
12 2 ~c
8.92
This agrees far too well with the measurement, considering the approximations made
and the dependence on the cut-off. There is, however, justification in the relativistic
calculation. Typically, the full calculation was made by using this non-relativistic
approach up to some energy of the order of mc2 , and using the relativistic calculation
above that. The relativistic free electron self-energy correction diverges only
logarithmically and a very high cutoff can be used without a problem. The
mass of the electron is renormalized as above. The Lamb shift does not depend
631

34. Electron Self Energy Corrections

TOC

on the cutoff and hence it is well calculated. We only need the non-relativistic part
of the calculation up to photon energies for which the E1 approximation is OK. The
relativistic part of the calculation down to min yields.

45
mc2
11 1
log
mc2

En =
+
3n3
2~min
24 5
The non-relativistic calculation gave.
En =

45
log
3n3

min
|
nj |

mc2

So the sum of the two gives.

En(obs)

45
=
3n3

mc2
11 1

log
+
mc2
2~
nj
24 5

The dependence on min cancels. In this calculation, the mc2 in the log is the
outcome of the relativistic calculation, not the cutoff. The electric dipole approximation
is even pretty good since we did not need to go up to large photon energies nonrelativistically and no E1 approximation is needed for the relativistic part. Thats how
we (and Bethe) got about the right answer.
The Lamb shift splits the 2S 21 and 2P 12 states which are otherwise degenerate.
Its origin is purely from field theory. The experimental measurement of the
Lamb shift stimulated theorists to develop Quantum ElectroDynamics. The
correction increases the energy of s states. One may think of the physical origin as
the electron becoming less pointlike as virtual photons are emitted and reabsorbed.
Spreading the electron out a bit decreases the effect of being in the deepest part of the
potential, right at the origin. Based on the energy shift, I estimate that the electron
in the 2s state is spread out by about 0.005 Angstroms, much more than the
size of the nucleus.
The anomalous magnetic moment of the electron, g 2, which can also be
calculated in field theory, makes a small contribution to the Lamb shift.

632

35. Dirac Equation

35
35.1

TOC

Dirac Equation
Diracs Motivation

The Schr
odinger equation is simply the non-relativistic energy equation
operating on a wavefunction.
p2
+ V (~r)
E=
2m
The natural extension of this is the relativistic energy equation.
E 2 = p2 c2 + (mc2 )2
This is just the Klein-Gordon equation that we derived for a scalar field. It did
not take physicists long to come up with this equation.
Because the Schr
odinger equation is first order in the time derivative, the initial
conditions needed to determine a solution to the equation are just (t = 0). In an
equation that is second order in the time derivative, we also need to specify some
information about the time derivatives at t = 0 to determine the solution at a later
time. It seemed strange to give up the concept that all information is contained in the
wave function to go to the relativistically correct equation.
If we have a complex scalar field that satisfies the (Euler-Lagrange = Klein-Gordon)
equations
2 m2
2 m

it can be shown that the bilinear quantity

s =

2mi
x
x
satisfies the flux conservation equation

s
~

~
=
+ 2 (2 )
=
m2 ( ) = 0
x
2mi x x
x x
2mi
and reduces to the probability flux we used with the Schrodinger equation, in the nonrelativistic limit. The fourth component of the vector is just c times the probability
2
density, so thats fine too (using eimc t/~ as the time dependence.).
The perceived problem with this probability is that it is not always positive.
Because the energy operator appears squared in the equation, both positive energies
633

35. Dirac Equation

TOC

and negative energies are solutions. Both solutions are needed to form a complete
set. With negative energies, the probability density is negative. Dirac thought this
was a problem. Later, the vector s was reinterpreted as the electric current and
charge density, rather than probability. The Klein-Gordon equation was indicating
that particles of both positive and negative charge are present in the complex
scalar field. The negative energy solutions are needed to form a complete set, so they
cannot be discarded.
Dirac sought to solve the perceived problem by finding an equation that was somehow
linear in the time derivative as is the Schr
odinger equation. He managed to do this
but still found negative energy solutions which he eventually interpreted to predict
antimatter. We may also be motivated to naturally describe particles with spin onehalf.

35.2

The Schr
odinger-Pauli Hamiltonian

In the homework on electrons in an electromagnetic field, we showed that the Schr

odingerPauli Hamiltonian gives the same result as the non-relativistic Hamiltonian
we have been using and automatically includes the interaction of the electrons spin with the magnetic field.

2
e~
1
~ [~
p + A(~
r, t)] e(~r, t)
2m
c

634

35. Dirac Equation

TOC

The derivation is repeated here. Recall that {i , j } = 2ij , [i , j ] = 2ijk k , and

~ and the wavefunction.
that the momentum operator differentiates both A

2

e2
e~
e
= i j pi pj + 2 Ai Aj + (pi Aj + Ai pj )
~ [~
p + A(~r, t)]
c
c
c

1
e2
e
~ Aj
=
(i j + j i ) pi pj + 2 Ai Aj + i j Aj pi + Ai pj +
2
c
c
i xi

e~
Aj
e2
e
= ij pi pj + 2 Ai Aj + (i j Aj pi + j i Aj pi ) + i j
c
c
ic
xi
2
e
e
e~
1
A
j
= p2 + 2 A2 + {i , j }Aj pi +
(i j + i j )
c
c
ic 2
xi
e
e~ 1
Aj
e2
(i j j i + 2ij )
= p2 + 2 A2 + 2ij Aj pi +
c
c
ic 2
xi
2
2e
e~
A
e
j
~ p~ +
(iijk k + ij )
= p2 + 2 A2 + A
c
c
ic
xi
2
e
2e ~
e~
Aj
= p2 + 2 A2 + A
p~ + iijk k
c
c
ic
xi
2
e
2e ~
e~
~
= p2 + 2 A2 + A
p~ + ~ B
c
c
c

e ~
e2
e~
p2
~
+
~ B
A p~ +
A2 e +
2
2m mc
2mc
2mc
1
e~
e~
~ r, t)
[~
p + A(~
r, t)]2 e(~r, t) +
~ B(~
2m
c
2mc

We assume the Lorentz condition applies. This is a step in the right direction. The
wavefunction now has two components (a spinor) and the effect of spin is included.
Note that this form of the NR Hamiltonian yields the coupling of the electron spin to
a magnetic field with the correct g factor of 2. The spin-orbit interaction can be
correctly derived from this.

635

35. Dirac Equation

35.3

TOC

The Dirac Equation

We can extend this concept to use the relativistic energy equation (for now
with no EM field). The idea is to replace p~ with ~ p~.
2
E
p2 = (mc)2
c

E
E
~ p~
+ ~ p~ = (mc)2
c
c

i~
i~
~
~
+ i~~
i~~ = (mc)2
c t
c t

~
~
i~
+ i~~
i~
i~~ = (mc)2
x0
x0

This is again written in terms of a 2 component spinor .

This equation is clearly headed toward being second order in the time derivative. As
with Maxwells equation, which is first order when written in terms of the field tensor,
we can try to write a first order equation in terms of a quantity derived from . Define
(L)
(R)

=

1

~
=
i~
i~~ (L)
mc
x0

Including the two components of (L) and the two components of (R) , we now have
four components which satisfy the equations.

~
~
+ i~~
i~
i~~ (L) = (mc)2 (L)
i~
x0
x0

~
i~
+ i~~ mc(R) = (mc)2 (L)
x0

~ (R) = mc(L)
i~
+ i~~
x0

~ (L) = mc(R)
i~
i~~
x0
These (last) two equations couple the 4 components together unless m = 0. Both of
the above equations are first order in the time derivative. We could continue with this
set of coupled equations but it is more reasonable to write a single equation in
terms of a 4 component wave function. This will also be a first order equation.
636

35. Dirac Equation

TOC

First rewrite the two equations together, putting all the terms on one side.

~
i~~ i~
(L) + mc(R) = 0
x0

~
(R) + mc(L) = 0
i~~ i~
x0
Now take the sum and the difference of the two equations.

((R) + (L) ) + mc((R) + (L) ) = 0

x0
~ (R) + (L) ) i~ ((L) (R) ) + mc((R) (L) ) = 0
i~~ (
x0
~ (L) (R) ) i~
i~~ (

Now rewriting in terms of A = (R) + (L) and B = (R) (L) and ordering it as
a matrix equation, we get.
~ (R) (L) ) i~ ((R) + (L) ) + mc((R) + (L) ) = 0
i~~ (
x0

(R)
(L)
~
i~~ (
+ ) + i~
((R) (L) ) + mc((R) (L) ) = 0
x0

~ (R) (L) ) + mc((R) + (L) ) = 0

i~
((R) + (L) ) i~~ (
x0
~ (R) + (L) ) + i~ ((R) (L) ) + mc((R) (L) ) = 0
i~~ (
x0

~ B + mcA = 0
i~
A i~~
x0
~ A + i~ B + mcB = 0
i~~
x
! 0

~
i~ x0 i~~
A
A
+ mc
=0

i~~
i~
B
B
x0

Remember that A and B are two component spinors so this is an equation in 4

components.
We can rewrite the matrix above as a dot product between 4-vectors. The matrix has
a dot product in 3 dimensions and a time component
!

~
~
i~ x
i~~

0
0
i~
x4
0
=
~
+

~
~
0
x
i~~
i~ x
i~
0
4
0

0 ii
1 0
= ~
+
= ~
ii
0
0 1 x4
xi
x
637

35. Dirac Equation

TOC

The 4 by 4 matrices are given by.

0
ii

1
0

ii
0

0
1

With this definition, the relativistic equation can be simplified a great deal.
!

~ A
i~~
i~ x
A
0
+ mc
=0

i~~
i~ x0
B
B

1

2
A

3
B
4

+ mc = 0
~
x
The Dirac equation in the absence of EM fields is

+
= 0.

x
~

is a 4-component Dirac spinor and, like the spin states we are used to, represents
a coordinate different from the spatial ones.
The gamma matrices are 4 by 4 matrices operating in this spinor space. Note that
there are 4 matrices, one for each coordinate but that the row or column of the matrix
does not correlate with the coordinate.

0
0
1 =
0
i

0
0
i
0

0
i
0
0

i
0

0
0

0
0
2 =
0
1

0 0 1
0 1 0

1 0 0
0 0 0

0
0
3 =
i
0

0
0
0
i

i 0
0 i

0 0
0 0

1
0
4 =
0
0

Like the Pauli matrices, the gamma matrices form a vector, (this time a 4vector).
638

0
1
0
0

35. Dirac Equation

TOC

It is easy to see by inspection that the matrices are Hermitian and traceless. A
little computation will verify that they anticommute as the Pauli matrices did.

{ , } = 2

Sakurai shows that the anticommutation is all that is needed to determine the physics.
That is, for any set of 4 by 4 matrices that satisfy { , } = 2 ,

+
=0
x
~
will give the same physical result, although the representation of may be different. This is truly an amazing result.
There are a few other representations of the Dirac matrices that are used. We will try
hard to stick with this one, the one originally proposed by Dirac.
It is interesting to note that the primary physics input was the choice of the
Schr
odinger-Pauli Hamiltonian
e~
e~
r, t)] ~ [~
p + A(~
r, t)]
[~
p + A(~
c
c
that gave us the correct interaction with the electrons spin. We have applied this
same momentum operator relativistically, not much of a stretch. We have also written
the equation in terms of four components, but there was no new physics in that since
everything could be computed from two components, say (L) since

1
(R)
~
i~
i~~ (L) .

=
mc
x0
Diracs paper did not follow the same line of reasoning. Historically, the SchrodingerPauli Hamiltonian was derived from the Dirac equation. It was Dirac who produced
the correct equation for electrons and went on to interpret it to gain new insight into
physics.

639

35. Dirac Equation

TOC

Dirac Biography

35.4

The Conserved Probability Current

We now return to the nagging problem of the probability density and current which
prompted Dirac to find an equation that is first order in the time derivative. We derived
the equation showing conservation of probability for 1D Schrodinger theory by using
the Schr
odinger equation and its complex conjugate to get an equation of the
form
P (x, t) j(x, t)
+
= 0.
t
x
We also extended it to three dimensions in the same way.
Our problem to find a similar probability and flux for Dirac theory is similar
but a little more difficult. Start with the Dirac equation.

+
=0
x
~
640

35. Dirac Equation

TOC

Since the wave function is a 4 component spinor, we will use the Hermitian conjugate of the Dirac equation instead of the complex conjugate. The matrices are
Hermitian.

mc
=0

+
x
~

mc
=0
+
(x )
~
The complex conjugate does nothing to the spatial component of x but does change
the sign of the fourth component. To turn this back into a 4-vector expression, we can
change the sign back by multiplying the equation by 4 (from the right).

mc

4 +
=0
k +

xk
(x4 )
~

mc
4 = 0
k 4
4 4 +
xk
x4
~
4
4
mc

4 = 0
k
4 +
xk
x4
~
Defining = 4 , the adjoint spinor, we can rewrite the Hermitian conjugate
equation.

mc

k
4 +
=0
xk
x4
~
mc

+
=0

x
~
This is the adjoint equation. We now multiply the Dirac equation by from the left
and multiply the adjoint equation by from the right, and subtract.

+ mc
+ mc
=0

x
~
x
~

+ = 0

x
x

= 0

x

j =

j = 0
x
641

35. Dirac Equation

TOC

We have found a conserved current. Some interpretation will be required as we learn

more about the solutions to the Dirac equation and ultimately quantize it. We may
choose an overall constant to set the normalization. The fourth component of the
current should be ic times the probability density so that the derivative with respect
to x4 turns into P
t . Therefore let us set the properly normalized conserved
4-vector to be

.
j = ic

35.5

The Non-relativistic Limit of the Dirac Equation

One important requirement for the Dirac equation is that it reproduces

what we know from non-relativistic quantum mechanics. Note that we have
derived this equation from something that did give the right answers so we expect the
Dirac equation to pass this test. Perhaps we will learn something new though.
We know that our non-relativistic Quantum Mechanics only needed a two component
spinor. We can show that, in the non-relativistic limit, two components of the Dirac
spinor are large and two are quite small. To do this, we go back to the equations written
in terms of A and B , just prior to the introduction of the matrices. We make the
substitution to put the couplings to the electromagnetic field into the Hamiltonian.

35.5.1

The Two Component Dirac Equation

First, we can write the two component equation that is equivalent to the Dirac
equation. Assume that the solution has the usual time dependence eiEt/~ . We
start from the equation in A and B .
!

~ A
i~ x
i~~

A
0
+
mc
=0

~
B
B
i~~
i~ x
0
E

c
~ p~
A
A
+ mc
=0
E

~ p~
B
B
c

642

35. Dirac Equation

TOC

~ and adding the

Turn on the EM field by making the usual substitution p~ p~ + ec A
scalar potential term.

~
~ p~ + ec A
1c (E + eA0 mc2 )

A = 0

1
2
B
~
~ p~ + ec A
c (E + eA0 + mc )

e ~
1
(E + eA0 mc2 )A = ~ p~ + A
B
c
c

e~
1
(E + eA0 + mc2 )B = ~ p~ + A
A
c
c
These two equations can be turned into one by eliminating B .

e ~
e ~
1
c
(E + eA0 mc2 )A = ~ p~ + A
~

p
~
+
A A
c
c
(E + eA0 + mc2 )
c
This is the two component equation which is equivalent to the Dirac equation for
energy eigenstates. The one difference from our understanding of the Dirac equation is
in the normalization. We shall see below that the normalization difference is small
for non-relativistic electron states but needs to be considered for atomic fine structure.

35.5.2

The Large and Small Components of the Dirac Wavefunction

Returning to the pair of equations in A and B . Note that for E mc2 , that is
non-relativistic electrons, A is much bigger than B .

e ~
1
(E + eA0 + mc2 )B = ~ p~ + A
A
c
c

c
pc
e~
B
A
~ p~ + A
A
2
2mc
c
2mc2
In the Hydrogen atom, B would be of order 2 times smaller, so we call A
the large component and B the small component. When we include relativistic corrections for the fine structure of Hydrogen, we must consider the effect B has on the
normalization. Remember that the conserved current indicates that the normalization
condition for the four component Dirac spinor is.
4 = 4 4 =
j0 =

35.5.3

The Non-Relativistic Equation

Now we will calculate the prediction of the Dirac equation for the non-relativistic
coulomb problem, aiming to directly compare to what we have done with the
643

35. Dirac Equation

TOC

Schr
odinger equation for Hydrogen. As for previous Hydrogen solutions, we will set
~ = 0 but have a scalar potential due to the nucleus = A0 . The energy we have
A
been using in our non-relativistic formulation is E (N R) = E mc2 . We will work
with the equation for the large component A . Note that A0 is a function of
the coordinates and the momentum operator will differentiate it.

c2
e ~
e ~
~ p~ + A
A
~ p~ + A
2
c
(E + eA0 + mc )
c
c2
~ p~
~ p~A
(E + eA0 + mc2 )

(E + eA0 mc2 )A

(E (N R) + eA0 )A

Expand the energy term on the left of the equation for the non-relativistic case.

1
2mc2
c2
=
E + eA0 + mc2
2m mc2 + E (N R) + eA0 + mc2

1
2mc2
=
2m 2mc2 + E (N R) + eA0
!
1
1
=
0
2m 1 + E (N R) +eA
2mc2
!
(N R)
2
1
E (N R) + eA0
E
+ eA0

1
+
+ ...
2m
2mc2
2mc2
We will be attempting to get the correct Schr
odinger equation to order 4 , like the
one we used to calculate the fine structure in Hydrogen. Since this energy term we are
expanding is multiplied in the equation by p2 , we only need the first two terms in the
expansion (order 1 and order 2 ).
1
~ p~
2m

E (N R) + eA0
1
2mc2

~ p~A

(E (N R) + eA0 )A

644

35. Dirac Equation

TOC

The normalization condition we derive from the Dirac equation is

4 = 4 4 = = A + B = 1.
j0 =
A
B
pc
B

2 A
2mc

pc 2

A
A = 1
B 1 +
A + B
A
2mc2

p2

A = 1
1+
A
4m2 c2

p2
1+
A
8m2 c2

p2

A 1
8m2 c2
Weve defined , the 2 component wavefunction we will use, in terms of A so that it is
properly normalized, at least to order 4 . We can now replace A in the equation.

1
E (N R) + eA0
p2
p2
(N R)
~ p~
1
~

p
~
1

=
(E
+
eA
)
1

0
2m
2mc2
8m2 c2
8m2 c2
This equation is correct, but not exactly what we want for the Schrodinger equation.
In particular, we want to isolate the non-relativistic energy on the right of
the equation without other
operators.
We can solve the problem by multiplying both
p2
sides of the equation by 1 8m2 c2 .

1

p2
8m2 c2

~ p~

1
2m

E (N R) + eA0
2mc2

~ p~ 1

8m2 c2

= 1

p2
8m2 c2

(E (N R) + eA0 ) 1

~ p~~ p~
p2 ~ p~~ p~ ~ p~ E (N R) + eA0
~ p~~ p~ p2

~ p~

2m
8m2 c2 2m
2m
2mc2
2m 8m2 c2

p2
p2
(N R)
E

eA0 e
= (E (N R) + eA0 )
4m2 c2
8m2 c2
2

p
p2 p2
p2 E (N R)
e~ p~A0~ p~
p2 p2

2m 8m2 c2 2m 2m 2mc2
4m2 c2
2m 8m2 c2

p2
p2
(N R)
= (E (N R) + eA0 )
E

eA0 e
4m2 c2
8m2 c2
2

p
p4
e~ p~A0~ p~
p2
p2
(N R)

=
E

0
0
0
2m 8m3 c2
4m2 c2
8m2 c2
8m2 c2

645

35. Dirac Equation

TOC

We have only kept terms to order 4 . Now we must simplify two of the terms in
the equation which contain the momentum operator acting on the field.
p2 A0

~ ((A
~ 0 ) + A0 )
~
~ 0 ) ()
~
= ~2 2 A0 = ~2
= ~2 ((2 A0 ) + 2(A
~ p~ + A0 p2
= ~2 (2 A0 ) + 2i~E

~
~ 0 )~ p~ + A0~ p~~ p~ = ~ ~ E~
~ p~ + A0 p2 = i~(i Ei j pj ) + A0 p
~ (A
i
i
~ p~ + A0 p2 = ~~ E
~ p~ + i~E
~
= ~(i j Ei pj ) + A0 p2 = i~iijk k Ei pj + i~E

~ p~A0~ p~ =

Plugging this back into the equation, we can cancel several terms.
!
~ p~ ie~E
~ p~ eA0 p2
p2
p4
e~~ E
eA0
+

2m
8m3 c2
4m2 c2

!
~ p~ + A0 p2
~2 (2 A0 ) + 2i~E
p2
eA0
= E
e
8m2 c2
8m2 c2
!
!
~ p~ ie~E
~ p~
~ p~
p2
p4
e~~ E
~2 (2 A0 ) + 2i~E
(N R)
eA0
+
= E
e
2m
8m2 c2
4m2 c2
8m2 c2
!
~ E
~
~ p~ e~2
p4
e~~ E
p2
eA0
+
+
= E (N R)
3
2
2
2
2
2
2m
8m c
4m c
8m c
(N R)

Ze
Now we can explicitly put in the potential due to the nucleus 4r
in our new units.
~
~
We identify ~r p~ as the orbital angular momentum. Note that E = = Ze 3 (~r).
The equation can now be cast in a more familiar form.

eA0

~ p~
~ ~r p~
e~~ E
Ze2 S
=
=
2
2
4m c
8m2 c2 r3
~ E
~
e~2
e~2 Ze 3
=
(~r) =
8m2 c2
8m2 c2

e = e

Ze
4r

~ S
~
Ze2 L
2
2
8m c r3
Ze2 ~2 3
(~r)
8m2 c2

!
~ S
~
p2
Ze2
p4
Ze2 L
Ze2 ~2 3

+
+
(~r) = E (N R)
2m
4r
8m3 c2
8m2 c2 r3
8m2 c2

This Schr
odinger equation, derived from the Dirac equation, agrees well
with the one we used to understand the fine structure of Hydrogen. The first two terms
646

35. Dirac Equation

TOC

are the kinetic and potential energy terms for the unperturbed Hydrogen Hamiltonian.
Note that our units now put a 4 in the denominator here. (The 4 will be absorbed
into the new formula for .) The third term is the relativistic correction to the
kinetic energy. The fourth term is the correct spin-orbit interaction, including
the Thomas Precession effect that we did not take the time to understand when
we did the NR fine structure. The fifth term is the so called Darwin term which we
said would come from the Dirac equation; and now it has.
This was an important test of the Dirac equation and it passed with flying colors. The
Dirac equation naturally incorporates relativistic corrections, the interaction with
electron spin, and gives an additional correction for s states that is found to be correct
experimentally. When the Dirac equation is used to make a quantum field theory of
electrons and photons, Quantum ElectroDynamics, we can calculate effects to very
high order and compare the calculations with experiment, finding good agreement.

35.6

Solution of Dirac Equation for a Free Particle

As with the Schr

odinger equation, the simplest solutions of the Dirac equation are
those for a free particle. They are also quite important to understand. We will find
that each component of the Dirac spinor represents a state of a free particle
at rest that we can interpret fairly easily.
We can show that a free particle solution can be written as a constant spinor times
the usual free particle exponential. Start from the Dirac equation and attempt
to develop an equation to show that each component has the free particle exponential.
We will do this by making a second order differential equation, which turns out to be

647

35. Dirac Equation

TOC

the Klein-Gordon equation.

mc
+

x
~

mc

x
x
~

+
x x
x ~

x x
~
x
mc 2

x x
~
mc 2

x x
~
mc 2

( + )

2
x x
~
mc 2

2

2
x x
~
mc 2

2
2

x x
~

=0
=0
=0
=0
=0
=0
=0
=0
=0

The free electron solutions all satisfy the wave equation.

mc 2
2
=0
~

Because we have eliminated the matrices from the equation, this is an equation
for each component of the Dirac spinor . Each component satisfies the wave
(Klein-Gordon) equation and a solution can be written as a constant spinor times
the usual exponential representing a wave.

p~ = up~ ei(~p~xEt)/~

Plugging this into the equation, we get a relation between the momentum and
648

35. Dirac Equation

TOC

the energy.
mc 2
E2
p2
+ 2 2
=0
2
~
~ c
~
p2 c2 + E 2 m2 c4 = 0
E 2 = p2 c2 + m2 c4
p
E = p2 c2 + m2 c4

Note that the momentum operator is clearly still

i~ t
.

~~
i

and the energy operator is still

There is no coupling between the different components in this equation, but, we will
see that (unlike the equation differentiated again) the Dirac equation will give us
relations between the components of the constant spinor. Again, the solution
can be written as a constant spinor, which may depend on momentum up~ , times the
exponential.
p~ (x) = up~ ei(~p~xEt)/~
We should normalize the state if we want to describe one particle per unit volume:
= V1 . We havent learned much about what each component represents yet. We
p
also have the plus or minus in the relation E = p2 c2 + m2 c4 to deal with. The
solutions for a free particle at rest will tell us more about what the different components
mean.

35.6.1

Dirac Particle at Rest

To study this further, lets take the simple case of the free particle at rest. This is just
the p~ = 0 case of the the solution above so the energy equation gives E = mc2 . The
Dirac equation can now be used.

+
=0
x
~

mc

+
0 eiEt/~ = 0
4
(ict)
~
E
mc
4
0 =
0
~c
~
mc2
mc
4 0 =
0
~c
~
4 0 = 0

This is a very simple equation, putting conditions on the spinor 0 .

649

35. Dirac Equation

TOC

Lets take the case of positive energy first.

1
0

0
0

0
1
0
0

0
0
1
0

= +
4 0 0
0
A1
A1
A2
A2
0
= +
B1
0 B1
1
B2
B2

A1
A1
A2
A2

B1 = + B1
B2
B2
B1 = B2 = 0

A1
A2

0 =
0
0

We see that the positive energy solutions, for a free particle at rest, are described
by the upper two component spinor. what we have called A . We are free to
choose each component of that spinor independently. For now, lets assume that the
two components can be used to designate the spin up and spin down states
according to some quantization axis.
For the negative energy

1
0

0
0

solutions we have.

0 0
0
A1
A1
A2
A2
1 0
0
=
B1
0 1 0 B1
0 0 1
B2
B2
A1 = A2 = 0

0
0

0 =
B1
B2

We can describe two spin states for the negative energy solutions.
Recall that we have demonstrated that the first two components of are large compared to the other two for a non-relativistic electron solution and that the first two
components, A , can be used as the two component spinor in the Schrodinger equation
(with a normalization factor). Lets identify the first component as spin up along the
z axis and the second as spin down. (We do still have a choice of quantization axis.)
650

35. Dirac Equation

TOC

Define a 4 by 4 matrix which gives the z component of the spin.

=
=
=
=
=

1
(1 2 2 1 )
2i

1
0 ix
0 iy
0

ix
0
iy
0
iy
2i

1
x y
0
y x
0

0
x y
0
y x
2i

1 [x , y ]
0
0
[x , y ]
2i

z 0
0 z

iy
0

0
ix

ix
0

With this matrix defining the spin, the third component is the one with spin up
along the z direction for the negative energy solutions. We could also define 4 by 4
matrices for the x and y components of spin by using cyclic permutations of the above.
So the four normalized solutions for a Dirac particle at rest are.

(1) = E=+mc2 ,+~/2

1
1
0
eimc2 t/~
=

0
V
0

(3) = E=mc2 ,+~/2

0
1
0
e+imc2 t/~
=

1
V
0

1
(2) = E=+mc2 ,~/2 =
V

1
(4) = E=mc2 ,~/2 =
V

The first and third have spin up while the second and fourth have spin down.
The first and second are positive energy solutions while the third and fourth are
negative energy solutions, which we still need to understand.

35.6.2

Dirac Plane Wave Solution

We now have simple solutions for spin up and spin down for both positive energy and
negative energy particles at rest. The solutions for nonzero momentum are
651

35. Dirac Equation

TOC

not as simple.

mc
+
x
~

p~ (x) = up~ ei(p x )/~

mc
ip
+
=0

~
~

0 ii
i =
ii
0

1 0
4 =
0 1

E
p~ ~
c 0
+ mc up~ ei(p x )/~ = 0
+
E
0
0
c

E + mc2
c~
p ~
up~ = 0
c~
p ~
E + mc2

0 1
0 i
1 0
x =
y =
z =
1 0
i 0
0 1

E + mc2
0
pz c
(px ipy )c
2

0
E
+
mc
(p
+
ip
)c
pz c
x
y
up~ = 0

pz c
(px ipy )c E + mc2
0
(px + ipy )c
pz c
0
E + mc2

0
~
p ~

We should find four solutions. Lets start with one that gives a spin up electron in
the first two components and plug it into the Dirac equation to see what the third and
fourth components can be for a good solution.

E + mc2
0
pz c
(px ipy )c
1
2
0

0
E
+
mc
(p
+
ip
)c
p
c
x
y
z

= 0

B1
0
pz c
(px ipy )c E + mc2
(px + ipy )c
pz c
0
E + mc2
B2

2
E + mc + B1 pz c + B2 (px ipy )c

B1 (px + ipy )c B2 pz c

=0
2

pz c + B1 (E + mc )
(px + ipy )c + B2 (E + mc2 )

652

35. Dirac Equation

TOC

Use the third and fourth components to solve for the coefficients and plug them in for
a check of the result.
pz c
E + mc2
(px + ipy )c
B2 =
E + mc2

2 2
p c
E + mc2 + E+mc
2
pz (px +ipy )c2 pz (px +ipy )c2

E+mc2
=0

0
0
E 2 +(mc2 )2 +p2 c2
B1 =

E+mc2

0
0
0

This will be a solution as long as E 2 = p2 c2 + (mc2 )2 , not a surprising condition.

Adding a normalization factor, the solution is.

pz c
up~ = N
E+mc
2
(px +ipy )c
E+mc2

up~=0

1
0

=N
0
0

This reduces to the spin up positive energy solution for a particle at rest as the momentum goes to zero. We can therefore identify this as that same solution boosted to
have momentum p~. The full solution is.

i(~p~xEt)/~
0
(1)

pz c
p~ = N
E+mc2 e
(px +ipy )c
E+mc2

We again see that for a non-relativistic electron, the last two components are small
compared to the first. This solution is that for a positive energy electron. The
fact that the last two components are non-zero does not mean it contains
negative energy solutions.
If we make the upper two components those of a spin down electron, we get the
653

35. Dirac Equation

TOC

next solution following the same procedure.

E + mc2
0
pz c
(px ipy )c
0
2
1

0
E
+
mc
(p
+
ip
)c
p
c
x
y
z
= 0

B1
pz c
(px ipy )c E + mc2
0
2
(px + ipy )c
pz c
0
E + mc
B2

B1 pz c + B2 (px ipy )c
E + mc2 + B1 (px + ipy )c B2 pz c
=0

(px ipy )c + B1 (E + mc2 )

2
pz c + B2 (E + mc )

(px ipy )c
E + mc2
pz c
B2 =
E + mc2

B1 =

pz (px ipy )c2 pz (px ipy )c

E+mc2

2 2
E + mc2 + p c 2
E+mc

0
0

2 2

+(mc ) +p c
E+mc2

0
0

E 2 = p2 c2 + (mc2 )2

up~ = N
(px ipy2)c
E+mc
pz c
E+mc2

up~=0

0
1

=N
0
0

This reduces to the spin down positive energy solution for a particle at rest as the
momentum goes to zero. The full solution is.

0
1

i(~p~xEt)/~
(2)

p~ = N
(px ipy )c e
E+mc2
pz c
E+mc2

654

35. Dirac Equation

TOC

Now we take a look at the negative energy spin up solution in the same way.

A1
E + mc2
0
pz c
(px ipy )c
2

0
E
+
mc
(p
+
ip
)c
p
c
x
y
z

A2 = 0

1
pz c
(px ipy )c E + mc2
0
0
(px + ipy )c
pz c
0
E + mc2

A1 (E + mc2 ) + pz c

A2 (E + mc2 ) + (px + ipy )c

A1 pz c A2 (px ipy )c + (E + mc2 ) = 0

A1 (px + ipy )c + A2 pz c
pz c
A1 =
E + mc2
(px + ipy )c
A2 =
E + mc2

=0
p2 c 2
2

E+mc2 + (E + mc )
(px +ipy )c
pz c
E+mc
p
c
2 (px + ipy )c +
2
z
E+mc

0
E 2 +(mc2 )2 +p2 c2 = 0

2
E+mc

0
2

E = p2 c2 + (mc2 )2
pz c
E+mc2

(px +ipy )c
E+mc2
up~ = N

1
0

up~=0

0
0

=N
1
0

This reduces to the spin up negative energy solution for a particle at rest as the
momentum goes to zero. The full solution is
pz c
E+mc2

(3)
p~

(px +ipy )c i(~p~xEt)/~

E+mc2 e
=N

1
0

with E being a negative number. We will eventually understand the negative energy
solutions in terms of anti-electrons otherwise known as positrons.
655

35. Dirac Equation

TOC

Finally, the negative energy, spin down solution follows the same pattern.

A1
E + mc2
0
pz c
(px ipy )c
2

A2
0
E
+
mc
(p
+
ip
)c
p
c
x
y
z

= 0

0
pz c
(px ipy )c E + mc2
0
2
(px + ipy )c
pz c
0
E + mc
1
(px ipy )c
(4)
p~

E+mc2
pz c
E+mc2

0
1

i(~p~xEt)/~
e

with E being a negative number.

We will normalize the states so that there is one particle per unit volume.
1
=

V
p 2 c2
1
2
N 1+
=
(|E| + mc2 )2
V
2

E + m2 c4 + 2|E|mc2 + p2 c2
1
N2
=
(|E| + mc2 )2
V
2

2E + 2|E|mc2
1
2
N
=
(|E| + mc2 )2
V

2|E|
1
N2
=
(|E| + mc2 )
V
s
|E| + mc2
N=
2|E|V

We have the four solutions with for a free particle with momentum p~. For
solutions 1 and 2, E is a positive number. For solutions 3 and 4, E is negative.

0
1
s
s

1
0
|E| + mc2
|E| + mc2
(1)
(2)
pz c ei(~p~xEt)/~
(px ipy )c ei(
p~ =
p~ =

2|E|V
2|E|V
E+mc2
E+mc2

s
(3)
p~

(px +ipy )c
E+mc2
pz c
E+mc2
(px +ipy )c
mc2
E+mc2 ei(~p~xEt)/~

|E| +
2|E|V

1
0

s
(4)
p~

pz c
E+mc2
(px ipy )c
E+mc2
pz c

mc2
E+mc2 ei(

|E| +
2|E|V

0
1

The spinors are orthogonal for states with the same momentum and the free particle
waves are orthogonal for different momenta. Note that the orthogonality condition is
656

35. Dirac Equation

TOC

the same as for non-relativistic spinors

(r)

(r 0 )

p~0

= rr0 (~
p p~0 )

(r)

It is useful to write the plane wave states as a spinor up~ times an exponential. Sakurai
picks a normalization of the spinor so that u u transforms like the fourth component
of a vector. We will follow the same convention.
s
mc2 (r) i(~p~xEt)/~
(r)
u e
p~
|E|V p~

r
(1)

up~ =

(3)

up~ =

E + mc2

2mc2

1
0

pz c

E+mc2
(px +ipy )c
E+mc2
pz c
r
E+mc2
(px +ipy )c
E + mc2
E+mc2

2mc2
1
0

0
1

r
(2)

up~ =

(4)

up~ =

E + mc2
(px ipy )c

2
2mc
E+mc2

pz c
E+mc2
(px ipy )c
r
E+mc2
pz c

E + mc2
E+mc2

2
2mc

0
1

(r)

up~

(r 0 )

up~

|E|
rr0
mc2

z 0
Are the free particle states still eigenstates of z =
as were the states of a
0 z
particle at rest? In general, they are not. To have an eigenvalue of +1, a spinor must
have zero second and fourth components and to have an eigenvalue of -1, the first and
third components must be zero. So boosting our Dirac particle to a frame in
which it is moving, mixes up the spin states.
There is one case for which these are still spin eigenstates. If the particles momentum
is in the z direction, then we have just the spinors we need to be eigenstates of z . That
is, if we boost along the quantization axis, the spin eigenstates are preserved. These
are called helicity eigenstates. Helicity is the spin component along the direction
of the particle. While it is possible to make definite momentum solutions which are
eigenstates of helicity, it is not possible to make definite momentum states which are
eigenstates of spin along some other direction (except in the trivial case of p~ = 0 as we
have shown).
To further understand these solutions, we can compute the conserved probability cur657

35. Dirac Equation

TOC

rent. First, compute it in general for a Dirac spinor

A1
A2

=
B1 .
B2

j = ic
j
4

=
=

ic 4

1 0 0
0 1 0

0 0 1
0 0 0

0
0

0
1

0
0
=
0
i

0
0
3 =
i
0

0 0 0 1
0 0 1 0

2 =
0 1 0 0
1 0 0 0

A1 B2 + A1 B2 + A2 B1 + A2 B1
i[A1 B2 A1 B2 A2 B1 + A2 B1 ]

c
A1 B1 + A1 B1 A2 B2 A2 B2
i[A1 A1 + A2 A2 + B1 B1 + B2 B2 ]

= ic A1

A1
A2

B2
B1
B2

0
0
i
0

i
0

0
0

0
i
0
0

0
0
0
i

i 0
0 i

0 0
0 0

(1)

Now compute it specifically for a positive energy plane wave, p~ , and a negative
(3)

energy plane wave, p~ .

(1)

= N

j(1)

ei(~p~xEt)/~
pz c

E+mc2
(px +ipy )c
E+mc2
2
N c ([B2 + B2 , i[B2 B2 ], [B1

j(1)

1
0

+ B1 ], i[1 + B1 B1 + B2 B2 ])

c
p2 c2
2
[2p
c],
[2p
c],
[2p
c],
i[E
+
mc
+
]
= N2
x
y
z
E + mc2
E + mc2
2c
= N2
(px c, py c, pz c, iE)
E + mc2

658

35. Dirac Equation

TOC

pz c
E+mc2
(px +ipy )c i(~p~xEt)/~
E+mc2 e

(3)

= N

j(3)

= N 2 c ([A2 + A2 ], i[A2 + A2 ], [A1 + A1 ], i[A1 A1 + A2 A2 + 1])

p2 c2
c
2
2
2px c, 2py c, 2pz c, i[E + mc +
]
= N
E + mc2
E + mc2
2c
= N 2
(px c, py c, pz c, iE)
E + mc2

j(3)
j(3)

1
0

Since E is negative for the negative energy solution, the probability density is
positive but the probability flux is in the opposite direction of the momentum.

35.6.3

Alternate Labeling of the Plane Wave Solutions

Start from the four plane wave solutions: 1 and 2 with positive energy and 3 and 4
with negative. There are four solutions for each choice of momentum p~.

0
1
s
s

1
0
|E| + mc2
|E| + mc2
(2)
(1)
pz c ei(~p~xEt)/~
(px ipy )c ei(
p~ =
p~ =

2|E|V
2|E|V
E+mc2
E+mc2

s
(3)
p~

(px +ipy )c
E+mc2
pz c
E+mc2
(px +ipy )c
mc2
E+mc2 ei(~p~xEt)/~

|E| +
2|E|V

1
0

s
(4)
p~

pz c
E+mc2
(px ipy )c
E+mc2
pz c

mc2
E+mc2 ei(

|E| +
2|E|V

0
1

Concentrate on the exponential which determines the wave property. For solutions
3 and 4, both the phase and group velocity are in the opposite direction to the momentum, indicating we have a problem that was not seen in non-relativistic quantum
mechanics.
p
E
p2 c2 + m2 c4
~vphase = k = p =
p
k
p
p
d dE
pc2
~vgroup =
k=
p = p
p
dk
dp
p2 c2 + m2 c4
Clearly, we want waves that propagate in the right direction. Perhaps the momentum
and energy operators we developed in NR quantum mechanics are not the whole story.
For solutions 3 and 4, pick the solution for ~
p to classify with solutions 1 and 2
with
momentum
p
~
write
everything
in
terms
of the positive square root E =
p
p2 c2 + m2 c4 .
659

35. Dirac Equation

TOC

1
0

r
(1)

p~ =

E + mc2

2EV

e
pz c

E+mc2
(px +ipy )c
E+mc2

i(~
p~
xEt)/~

(2)

p~ =

(3)
p~

pz c
E+mc2
(px +ipy )c
mc2
E+mc2 ei(~p~xEt)/~

E+
2EV

1
0

E + mc2
(px ipy )c e

2EV
E+mc2
(px ipy )c

s
(4)
p~

|E| + mc2

2|E|V

E+mc2
pz c
E+mc2

0
1

We have plane waves of the form

ei(p x )/~
which is not very surprising. In fact we picked the + sign somewhat randomly in the
development of NR quantum mechanics. For relativistic quantum mechanics, both
solutions are needed. We have no good reason to associate the ei(p x ) solution with
negative energy. Lets assume it also has positive energy but happens to have the - sign
on the whole exponent.
Consider the Dirac equation with the EM field term included. (While we are dealing
with free particle solutions, we can consider that nearly free particles will have a very
similar exponential term.)

+
x

ie
+ A +
x
~c

mc
=0
~
mc
=0
~

The x operating on the exponential produces ip /~. If we change the charge on

the electron from e to +e and change the sign of the exponent, the equation remains
the same. Thus, we can turn the negative exponent solution (going backward in time)
into the conventional positive exponent solution if we change the charge to +e. Recall
that the momentum has already been inverted (and the spin also will be inverted).
The negative exponent electron solutions can be recast as conventional
exponent positron solutions.

35.7

pz c
E+mc2

0
1

Negative Energy Solutions: Hole Theory

(x) =
(x) +
x
~

mc 0 0
0 0 (x0 ) +
(x ) =
x
~

0 (x0 )

mc
S
a
S +
x
~

mc

1
S +
S
S
a
x
~

mc 1
S 1 Sa
+
S S
x
~

mc
S 1 Sa

+
x
~

=
=

0
0
S(x)

a
x

The transformed equation will be the same as the Dirac equation if S 1 Sa = .

Multiply by the inverse Lorentz transformation.

a (a)1

a a

Sa a

= a

S 1 S = a

This is the requirement on S for covariance of the Dirac equation.

Rotations and boosts are symmetry transformations of the coordinates in 4 dimensions.
Consider the cases of rotations about the z axis and boosts along the x direction, as

663

35. Dirac Equation

TOC

examples.

a(rot)

a(boost)

cos
sin

0
0

0
i

sin
cos
0
0
0 0
1 0
0 1
0 0

0 0
0 0

1 0
0 1

cosh
i

0
0
=
0
0
i sinh

0 0
1 0
0 1
0 0

i sinh
0

0
cosh

The boost is just another rotation in Minkowski space through and angle
i = i tanh1 . For example a boost with velocity in the x direction is like a
rotation in the 1-4 plane by an angle i. Let us review the Lorentz transformation
for boosts in terms of hyperbolic functions. We define tanh = .
tanh =
cosh =
sinh =
cos(i)

sin(i)

e e
=
e + e
e + e
2
e e
2
i(i)
e + e
e
+ ei(i)
=
= cosh
2
2
ei(i) ei(i)
e e
e e
=
=i
= i sinh
2i
2i
2
1
1
1
p
=p
=q
=p
2
(1
+
tanh
)(1
tanh )
1 2
1 tanh

1
2e
2e
e +e e +e

tanh cosh = sinh

a(boost)

x01
x

x04

cosh

0
=

0
i sinh

0 0
1 0
0 1
0 0

i sinh
0

0
cosh

= x1 cosh + ix4 sinh

= x + i(ict)
= x4 cosh ix1 sinh

We verify that a boost along the i direction is like a rotation in the i4 plane through
an angle i.
664

35. Dirac Equation

TOC

We need to find the transformation matrices S that satisfy the equation S 1 S =

a for the Dirac equation to be covariant. Recalling that the 4 component equivalent
of z is z = [12i,2 ] = 1i2 , we will show that these matrices are (for a rotation in the
xy plane and a boost in the x direction).
Srot

Sboost

+ 1 2 sin
2
2

cosh + i1 4 sinh
2
2
cos

Note that this is essentially the transformation that we derived for rotations of spin
one-half states extended to 4 components. For the case of the boost the angle is now
i.
Lets verify that this choice works for a boost.

+ i1 4 sinh
cosh + i1 4 sinh
2
2
2
2

+ i1 4 sinh
cosh + i1 4 sinh
cosh
2
2
2
2

cosh i1 4 sinh
cosh + i1 4 sinh
2
2
2
2

cosh cosh + cosh i1 4 sinh i1 4 sinh cosh i1 4 sinh i1 4 sinh

2
2
2
2
2
2
2
2

2
2
cosh
+ i 1 4 cosh sinh i1 4 cosh sinh + 1 4 1 4 sinh
2
2
2
2
2
2

cosh

The equation we must satisfy can be checked for each matrix. First check 1 . The
operations with the matrices all come from the anticommutator, { , } = 2 ,
which tells us that the square of any gamma matrix is one and that commuting a pair
of (unequal) matrices changes the sign.
1 cosh2

+ i1 1 4 cosh sinh i1 4 1 cosh sinh + 1 4 1 1 4 sinh2

2
2
2
2
2

1 cosh2 + i4 cosh sinh + i4 cosh sinh + 1 sinh2

2
2
2
2
2

2
1 cosh
+ 2i4 cosh sinh + 1 sinh2
2
2
2

= a1
= a1
= a1

665

35. Dirac Equation

cosh2

TOC

1
1
+ sinh2 = ((e 2 + e 2 )2 + (e 2 e 2 )2 ) = (e + 2 + e + e 2 + e )
2
2
4
4
1
= (e + e ) =
2

1
1
1

cosh sinh = ((e 2 + e 2 )(e 2 e 2 )) = (e e ) = sinh

2
2
4
4
2

1
2
2
2
cosh
sinh
= ((e 2 + e 2 ) (e 2 e 2 )2 )
2
2
4
1
= (e + 2 + e e + 2 e ) =
4
1 cosh + i4 sinh =

cosh
0 0 i sinh

0
1 0
0

a =

0
0 1
0
i sinh 0 0 cosh
1 cosh + i4 sinh = 1 cosh + i4 sinh

That checks for 1 . Now, try 4 .

4 cosh2

+ i4 1 4 cosh sinh i1 4 4 cosh sinh + 1 4 4 1 4 sinh2

2
2
2
2
2
2

2
2
4 cosh
i1 cosh sinh i1 cosh sinh + 4 sinh
2
2
2
2
2
2

2
2
4 cosh
2i1 cosh sinh + 4 sinh
2
2
2
2
4 cosh i1 sinh

= a4

= i sin

That one also checks. As a last test, try 2 .

2 cosh2

+ i2 1 4 cosh sinh i1 4 2 cosh sinh + 1 4 2 1 4 sinh2

2
2
2
2
2

2
2 cosh
+ i2 1 4 cosh sinh i2 1 4 cosh sinh 2 sinh2
2
2
2
2
2

2
2

= 2
= a2
= 2

The Dirac equation is therefore shown to be invariant under boosts along the xi
direction if we transform the Dirac spinor according to 0 = Sboost with the
matrix

Sboost = cosh

+ ii 4 sinh
2
2
666

35. Dirac Equation

TOC

and tanh = .
The pure rotation about the z axis should also be verified.

1

cos + 1 2 sin
cos + 1 2 sin
= a
2
2
2
2

cos 1 2 sin
cos + 1 2 sin
= a
2
2
2
2

= a
cos2 + 1 2 cos sin 1 2 cos sin 1 2 1 2 sin2
2
2
2
2
2
2
For = 3 or 4, a = and the requirement is fairly obviously satisfied. Checking
the requirement for = 1, we get.
1 cos2

+ 1 1 2 cos sin 1 2 1 cos sin 1 2 1 1 2 sin2

2
2
2
2
2
2
a

1 cos2

+ 22 cos sin 1 sin2

2
2
2
2
1 cos + 2 sin

cos
sin

0
0

cos 1 + sin 2

This also proves to be the right transformation of so that the Dirac equation
is invariant under rotations about the k axis if we transform the Dirac spinor
according to 0 = Srot with the matrix

Srot = cos

sin
cos
0
0

+ i j sin
2
2

and ijk is a cyclic permutation.

Despite the fact that we are using a vector of constant matrices, , the Dirac equation
is covariant if we choose the right transformation of the spinors. This allows us to move
from one coordinate system to another.
As an example, we might try our solution for a free electron with spin up along the z
667

35. Dirac Equation

TOC

axis at rest.

(1) = E=+mc2 ,+~/2

1
1
imc2 t/~
0 ip x /~
1
1
0
e
e
=
=
V 0
V 0
0
0

The solution we found for a free particle with momentum p~ was.

1
r

0
E + mc2
(1)
pz c eip x /~
p~ =

2
2EV
E+mc
(px +ipy )c
E+mc2

Imagine we boost the coordinate system along the x direction with vc = . We can
transform the momentum of the electron to the new frame.

mc
0

0 0 i

0
1 0
0
0 = 0
p0 = a(boost)
p =

0
0
0
0 1
0
imc
imc
i 0 0

The momentum along the x direction is px = mc = mc sinh . We now have two

ways to get the free particle state with momentum in the x direction. We can use our
free particle state

1
r

E + mc2
0 ei(~p~xEt)/~
(1) =

0
0
2EV
px c
E+mc2

1
0
0

r
=

E + mc2

2EV 0

mc2
mc2 +mc2

i(~p~xEt)/~
e

E + mc2
0 ei(~p~xEt)/~

0
0
2EV

r
=

E + mc2

2EV 0

1
0
0
sinh
cosh +1

i(~p~xEt)/~
e

668

35. Dirac Equation

TOC

r
=

E + mc2
2EV 0

cosh2

r
=

E + mc2
2EV 0

1
0
0

2 sinh
2 cosh 2

2 +cosh2 sinh2
+sinh
2
2
2
2

1
0
0
2 sinh
2 cosh
2 cosh2
2

i(~p~xEt)/~
e

i(~p~xEt)/~
e

2

r
=

E + mc2
0 ei(~p~xEt)/~

0
0
2EV
tanh 2

where the normalization factor is now set to be

1 ,
V0

defining this as the primed system.

We can also find the same state by boosting the at rest solution. Recall that we are
boosting in the x direction with , implying .
Sboost

i1 4 sinh
2
2

0 0 0 i
1 0 0
0 0 i 0 0 1 0

cosh i
0 i 0
0 0 0 1
2
i 0 0
0
0 0 0

0 0 0 i
0 0 i 0

cosh i
0 i 0 0 sinh 2
2
i 0 0 0

0 0 0 1
0 0 1 0
sinh
cosh +
2 0 1 0 0
2
1 0 0 0

cosh 2
0
0
sinh 2
0
cosh 2 sinh 2
0

0
sinh 2 cosh 2
0
sinh 2
0
0
cosh 2
cosh

0
0
sinh
0
2
1

669

35. Dirac Equation

TOC

cosh 2
0
0
sinh 2
0
cosh 2 sinh 2
0
1
=

0
0 V
sinh 2 cosh 2
0
0
cosh 2
sinh 2

cosh 2

1
0
eip x /~
=

0
V
sinh 2

1
0
1
eip x /~
= cosh
2 0
V
tanh 2

(1)

1 + cosh
2

cosh

(1)

e +e
2

1
0 ip x /~
e
0
0

e + 2 + e
e2 + e
=
=
2
4
2
r
r
r
1 + cosh
1+
E + mc2
=
=
2
2
2mc2

1
r

1
E + mc2
0 eip x /~

2mc2 0
V
tanh
2
1
r

1
E + mc2
0 eip x /~
0

0
2mc2
V
tanh 2

1
r

E + mc2
0 eip x /~

0
0
2EV
tanh 2
1+

!2
= cosh2

In the last step the simple Lorentz contraction was used to set V 0 = V . This boosted
state matches the plane wave solution including the normalization.

35.10

Parity

It is useful to understand the effect of a parity inversion on a Dirac spinor.

Again work with the Dirac equation and its parity inverted form in which xj xj

670

35. Dirac Equation

TOC

and x4 remains unchanged (the same for the vector potential).

mc
(x) +
(x)
x
~
mc 0 0

(x )
0 0 (x0 ) +
x
~

0 (x0 ) = SP (x)

=
0
xj
xj

x04

mc
j
+ 4
SP
SP +
xj
x4
~

SP1 j

+ 4
SP +
xj
x4
~

Since 4 commutes with itself but anticommutes with the i , it works fine.
SP = 4
(We could multiply it by a phase factor if we want, but there is no point to it.)
Therefore, under a parity inversion operation

0 = SP = 4

1 0 0
0
0 1 0
0

35.11

Bilinear Covariants

We have seen that the constant matrices can be used to make a conserved vector
current

j = ic
671

35. Dirac Equation

TOC

that transforms correctly under Lorentz transformations. With 4 by 4 matrices, we

should be able to make up to 16 components. The vector above represents 4 of those.
The Dirac spinor is transformed by the matrix S.
0 = S
This implies that = 4 transforms according to the equation.
0 = (S) 4 = S 4
Looking at the two transformations, we can write the inverse transformation.

+ i j sin
2
2

Sboost = cosh + ii 4 sinh

2
2

1
Srot = cos i j sin
2
2

1
Sboost = cosh ii 4 sinh
2
2

Srot
= cos + j i sin = cos i j sin
2
2
2
2

Sboost
= cosh i4 i sinh = cosh + ii 4 sinh
2
2
2
2

1
4 Srot 4 = cos i j sin = Srot
2
2

1
4 Sboost 4 = cosh ii 4 sinh = Sboost
2
2
4 S 4 = S 1
0

1
= (S) 4 = 4 4 S 4 = 4 S 1 = S
Srot = cos

This also holds for SP .

SP = 4
SP = 4
SP1 = 4
4 SP 4 = 4 4 4 = 4 = SP1

is invariant under Lorentz transformations and

From this we can quickly get that
hence is a scalar.
1 S =

0 0 = S
672

35. Dirac Equation

TOC

we have
Repeating the argument for
1 S = a

0 0 = S
according to our derivation of the transformations S. Under the parity transformation
1 S =
4 4
0 0 = S
the spacial components of the vector change sign and the fourth component doesnt.
It transforms like a Lorentz vector under parity.
Similarly, for 6= ,
i

forms a rank 2 (antisymmetric) tensor.

We now have 1+4+6 components for the scalar, vector and rank 2 antisymmetric
tensor. To get an axial vector and a pseudoscalar, we define the product of all
gamma matrices.
5 = 1 2 3 4
which obviously anticommutes with all the gamma matrices.
{ , 5 } = 0
For rotations and boosts, 5 commutes with S since it commutes with the pair of
gamma matrices. For a parity inversion, it anticommutes with SP = 4 . Therefore
5 transforms like a pseudoscalar and i
5 transforms
its easy to show that
like an axial vector. This now brings our total to 16 components of bilinear (in the
spinor) covariants. Note that things like 5 12 = i1 2 3 4 1 2 = i3 4 is just a
constant times another antisymmetric tensor element, so its nothing new.
Classification
Scalar
Pseudoscalar
Vector
Axial Vector
Rank 2 antisymmetric tensor
Total

Covariant Form

no. of Components

1
1
4
4
6
16

The matrices can be used along with Dirac spinors to make a Lorentz scalar, pseudoscalar, vector, axial vector and rank 2 tensor. This is the complete set of covariants, which of course could be used together to make up Lagrangians for physical
quantities. All sixteen quantities defined satisfy 2 = 1.
673

35. Dirac Equation

35.12

TOC

Constants of the Motion for a Free Particle

We know that operators representing constants of the motion commute with

the Hamiltonian. The form of the Dirac equation we have been using does not have
a clear Hamiltonian. This is true essentially because of the covariant form we have
been using. For a Hamiltonian formulation, we need to separate the space and time
derivatives. Lets find the Hamiltonian in the Dirac equation.

+
=0
x
~

mc

+
=0
+ 4
j
xj
ict
~

~
j pj 4
imc = 0
c t
~
(j pj imc) = 4
c t
~
(4 j pj imc4 ) =
c t

ic4 j pj + mc2 4 = E

H = ic4 j pj + mc2 4

Its easy to see the pk commutes with the Hamiltonian for a free particle so that momentum will be conserved.
The components of orbital angular momentum do not commute with H.
[H, Lz ] = ic4 [j pj , xpy ypx ] = ~c4 (1 py 2 px )
The components of spin also do not commute with H.
[1 , 2 ]
1 2
=
2i
i
~
~
~
[H, Sz ] = [H, z ] = c [4 j pj , 1 2 ] = c pj [4 j 1 2 1 2 4 j ]
2
2
2
~
~
= c pj [4 j 1 2 4 1 2 j ] = c pj 4 [j 1 2 1 2 j ] = ~c4 [2 px 1 py ]
2
2
z

674

35. Dirac Equation

TOC

However, the helicity, or spin along the direction of motion does commute.
~ p~] = [H, S]
~ p~ = ~c4 p~ ~ p~ = 0
[H, S
From the above commutators [H, Lz ] and [H, Sz ], the components of total angular
momentum do commute with H.
[H, Jz ] = [H, Lz ] + [H, Sz ] = ~c4 (1 py 2 px ) + ~c4 [2 px 1 py ] = 0

The Dirac equation naturally conserves total angular momentum but not conserve
the orbital or spin parts of it.
We will need another conserved quantity for the solution to the Hydrogen atom; something akin to the in j = ` 12 we used in the NR solution. We can show that
[H, K] = 0 for
~ J~ ~ 4 .
K = 4
2
It is related to the spin component along the total angular momentum direction. Lets
compute the commutator recalling that H commutes with the total angular momentum.
ic4~ p~ + mc2 4
~ J~ ~ ) + 4 [H, ]
~ J~
[H, K] = [H, 4 ](
2
[H, 4 ] = ic[4~ p~, 4 ] = 2ic~ p~
H

[H, z ] =
h
i
~
H,
=

2c4 [2 px 1 py ]

[H, K]

~ J~ ~ ) 2c~ p~ J~
2ic(~ p~)(
2
0
~ + ~
~
L
2

~ =
p~ L
J~ =

2c4~ p~

675

35. Dirac Equation

TOC

[H, K]

~ J)
~
(~ p~)(

i m n mnj =
~ J)
~ =
(~ p~)(
[H, K]

[H, K]

~ J)
~ ~ p~ J~ i ~ (~ p~)
2c i(~ p~)(
2
i
i pi j Jj =
pi Jj i m n mnj
2
2(ij 5 4 + ijk k )
ipi Jj (ij 5 4 + ijk k ) = i5 4 p~ J~ i~ p~ J~

~
2c 5 4 p~ J~ + ~ p~ J~ ~ p~ J~ i (~ p~)
2

~
2c 1 2 3 4 4 p~ J~ i (~ p~)
2

~
~
2c 1 2 3 p~ J i (~ p~)
2

~~
~
~
2c 1 2 3 p~ (L + ) i (~ p~)
2
2

~
~ i(~ p~)
2c
1 2 3 p~
2
~
p ~ i(~ p~)) = 0
2c (i~
2

~ = 0 so that we have a mutually commuting set of

It is also useful to show that [K, J]
operators to define our eigenstates.
~ = [4
~ J~ ~ 4 , J]
~ = [4 , J]
~
~ J~ + 4 [
~ J,
~ J]
~ ~ [4 , J]
~
[K, J]
2
2
~ = 0 and [
~ J,
~ J]
~ = 0.
This will be zero if [4 , J]
h

i
4 , J~
h
i
~ J,
~ J~

h
i
~ L,
~ J~

=
=

~~
~
~ =0
] = [4 , ]
2
2
~ L,
~ J]
~ + [
~ ,
~ J]
~ = [
~ L,
~ J]
~ + [3, J]
~ = [
~ L,
~ J]
~
[

~+
[4 , L

~
~
~
i , j Lj ] = [Li , j Lj ] + [i , j Lj ] = j [Li , Lj ] + [i , j ]Lj
2
2
2
~
= i~ijk j Lk + 2i ijk k Lj = i~(ijk j Lk + ijk k Lj ) = i~(ijk j Lk ikj
2

[Li +

So for the Hydrogen atom, H, J 2 , Jz , and K form a complete set of mutually commuting operators for a system with four coordinates x, y, z and electron spin.

676

35. Dirac Equation

35.13

TOC

The Relativistic Interaction Hamiltonian

The interaction Hamiltonian for the Dirac equation can be deduced in several
ways. The simplest for now is to just use the same interaction term that we had for
electromagnetism
1
Hint = j A
c
and identify the probability current multiplied by the charge (-e) as the current that
couples to the EM field.

j(EM ) = eic
Removing the from the left and from the right and dotting into A, we have the
interaction Hamiltonian.
Hint = ie4 A
Note the difference between this interaction and the one we used in the non-relativistic
case. The relativistic interaction has just one term, is linear in A, and is naturally
proportional to the coupling e. There is no longer an A2 term with a different
power of e. This will make our perturbation series also a series in powers of .
We may still assume that A is transverse and that A0 = 0 by choice of gauge.

Hint = ie4 k Ak

35.14

Phenomena of Dirac States

35.14.1

Velocity Operator and Zitterbewegung

We will work for a while in the Heisenberg representation in which the operators depend
on time and we can see some of the general behavior of electrons. If we work in a state
of definite energy, the time dependence of the operators is very simple, just the usual
exponentials.
The operator for velocity in the x direction can be computed from the commutator
with the Hamiltonian.
x =

i
i
[H, x] = ic[4 j pj , x] = ic4 1
~
~
vj = ic4 j
677

35. Dirac Equation

TOC

The velocity operator then is vj = ic4 j .

Its not hard to compute that the velocity eigenvalues (any component) are c.

1 0 0
0
0 0 0 i
0 0 0 1
0 1 0

0
0 0 i 0 = c 0 0 1 0
ic4 1 = ic
0 0 1 0 0 i 0

0
0 1 0 0
0 0 0 1
i 0 0
0
1 0 0 0

a
a
0 0 0 1
b
0 0 1 0 b

c
0 1 0 0 c = c c
d
d
1 0 0 0

0
0
1

0 1
0

=0
0
1 0

1
0
0
[(2 ) 1()] 1[() + 1(1)] = 0
4 22 + 1 = 0
(2 1)2 = 0
(2 1) = 0
= 1
vx = c

Thus, if we measure the velocity component in any direction, we should either

get plus or minus c. This seems quite surprising, but we should note that a component
of the velocity operator does not commute with momentum, the Hamiltonian, or even
the other components of the velocity operator. If the electron were massless, velocity
operators would commute with momentum. (In more speculative theories of particles,
electrons are actually thought to be massless, getting an effective mass from interactions
with particles present in the vacuum state.)
The states of definite momentum are not eigenstates of velocity for a massive electron.

678

35. Dirac Equation

TOC

The velocity eigenstates mix positive and

0 0
0 0

0 1
1 0

1
0

u=
0
1

0
1

u=
1
0

negative energy states equally.

a
a
0 1
b
b
1 0
=
c
0 0 c
d
d
0 0

a
d
b
c
=
c
b
d
a

0
1
1
0

u=
u=
1
0
0
1

Thus, while momentum is a constant of the motion for a free electron and behaves as
it did in NR Quantum Mechanics, velocity behaves very strangely in the Dirac theory,
even for a free electron. Some further study of this effect is in order to see if there are
physical consequences and what is different about the Dirac theory in this regard.
We may get the differential equation for the velocity of a free electron by computing
the derivative of velocity. We attempt to write the derivative in terms of the constants
of the motion E and p~.
v j

=
=

i
i
i
i
[H, vj ] = (2Hvj + {H, vj }) = (2Hvj + {vj pj + mc2 4 , vj }) = (2Hvj + {
~
~
~
~
i
i
i
i
(2Hvj + {vj pj , vj }) = (2Hvj + 2vj pj ) = (2Hvj + 2vj2 pj ) = (2Hvj + 2c
~
~
~
~

This is a differential equation for the Heisenberg operator vj which we may solve.
vj (t) = c2 pj /E + (vj (0) c2 pj /E)e2iEt/~
To check, differentiate the above
v j (t) =

2iE
i
(vj (0) c2 pj /E)e2iEt/~ = (2Evj (0) + 2c2 pj )e2iEt/~
~
~

and compare it to the original derivative.

v j (t)

=
=

i
i
(2Evj + 2c2 pj ) = (2E(c2 pj /E + (vj (0) c2 pj /E)e2iEt/~ ) + 2c2 pj )
~
~
i
i
2
((2c pj + (2Evj (0) + 2c2 pj )e2iEt/~ ) + 2c2 pj ) = (2Evj (0) + 2c2 pj )e2iEt
~
~
679

35. Dirac Equation

TOC

This checks so the solution for the velocity as a function of time is correct.
vj (t) = c2 pj /E + (vj (0) c2 pj /E)e2iEt/~
There is a steady motion in the direction of the momentum with the correct magnitude c. There are also very rapid oscillations with some amplitude. Since the
2~
~c
energy includes mc2 , the period of these oscillations is at most 2mc
= mc
2
2c =
13
(3.14)(197.3M eV F )
120010
21
= 1200F/c = 31010 = 4 10
seconds. This very rapid oscilla0.5M eV (c)
tion is known as Zitterbewegung. Obviously, we would see the same kind of oscillation
in the position if we integrate the above solution for the velocity. This very rapid
motion of the electron means we cannot localize the electron extremely well and gives
rise to the Darwin term. This operator analysis is not sufficient to fully understand
the effect of Zitterbewegung but it illustrates the behavior.

35.14.2

Expansion of a State in Plane Waves

To show how the negative energy states play a role in Zitterbewegung, it is convenient
to go back to the Schr
odinger representation and expand an arbitrary state in terms of
plane waves. As with non-relativistic quantum mechanics, the (free particle) definite
momentum states form a complete set and we can expand any state in terms of
them.
s
4
XX
mc2
(r)
(~x, t) =
cp~,r up~ ei(~p~xEt)/~
|E|V
r=1
p
~

The r = 1, 2 terms are positive energy plane waves and the r = 3, 4 states are negative
energy. The differing signs of the energy in the time behavior will give rise to rapid
oscillations.
The plane waves can be purely either positive or negative energy, however, localized states have uncertainty in the momentum and tend to have both
positive and negative energy components. As the momentum components
become relativistic, the negative energy amplitude becomes appreciable.
pc
c3,4

c1,2
E + mc2
Even the Hydrogen bound states have small negative energy components.
The cross terms between positive and negative energy will give rise to very rapid
oscillation of the expected values of both velocity and position. The amplitude of the
oscillations is small for non-relativistic electrons but grows with momentum (or with
localization).
680

35. Dirac Equation

35.14.3

TOC

The Expected Velocity and Zitterbewegung

The expected value of the velocity in a

hvk i = (ic4 k ) d3 x

0
(1)
(ic4 1 )up~ = c
0
1

plane wave state can be simply calculated.

0
0
1
0

0
1
0
0

1 r
2
0
E + mc
0
2EV
0
r

(1)
(ic4 1 )up~

1
0

pz c

E+mc2
(px +ipy )c
E+mc2
(px +ipy )c
E+mc2
pz c

mc2
E+mc2

E+
2EV

0
1

(px +ipy )c
(1)
(1)
up~ (ic4 1 )up~

E + mc2
=
c 1
2EV
(1)

up~

pz c
E+mc2

(1)

(ic4 1 )up~ =

(px ipy )c
E+mc2

E+mc2
pz c
E+mc2

0
1

E + mc2
2px c
px c
c
c
=
2
2EV
E + mc
EV
p k c2
hvk i =
E

The expected value of a component of the velocity exhibits strange behavior when
negative and positive energy states are mixed. Sakurai (equation 3.253) computes this.
Note that we use the fact that u(3,4) have negative energy.

hvk i = (ic4 k ) d3 x
hvk i =

4
XX
p
~ r=1

|cp~,r |2

2
4
pk c2 X X X mc3 h
(r 0 )
(r)
i4 k up~ e2i|E|t/~
+
cp~,r0 cp~,r up~
|E|
|E|
0
p
~ r=1 r =3
i
(r)
(r 0 )
+cp~,r0 cp~,r up~ i4 k up~ e2i|E|t/~

The last sum which contains the cross terms between negative and positive energy
represents extremely high frequency oscillations in the expected value of the
velocity, known as Zitterbewegung. The expected value of the position has similar
rapid oscillations.
The Zitterbewegung again keeps electrons from being well localized in a deep potential
681

35. Dirac Equation

TOC

raising the energy of s states. Its effect is already included in our calculation as it is
the source of the Darwin term.

35.15

Solution of the Dirac Equation for Hydrogen

The standard Hydrogen atom problem can be solved exactly using relativistic quantum
mechanics. The full solution is a bit long but short compared to the complete effort
we made in non-relativistic QM. We have already seen that (even with no applied
fields), while the total angular momentum operator commutes with the Dirac
Hamiltonian, neither the orbital angular momentum operator nor the spin operators
do commute with H. The addition of a spherically symmetric potential does not change
these facts.
We have shown in the section on conserved quantities that the operator
~ J~ 4 ~
K = 4
2
~ K is a measure of the
also commutes with the Hamiltonian and with J.
component of spin along the total angular momentum direction. We will use K to help
solve problems with spherical symmetry and ultimately the problem of hydrogen. We
therefore have four mutually commuting operators the eigenvalues of which can
completely label the eigenstates:
H, J 2 , Jz , K nr , j, mj , .
The operator K may be written in several ways.
K

~ J~ 4 ~ = 4
~ L
~ + ~ 4
~
~ 4 ~ = 4
~ L
~ + 4 3~ 4 ~
= 4
2
2
2
2
2

~ +~
~

L
0
~ L
~ + ~4 =
= 4
~ ~
0
~ L

Assume that the eigenvalues of K are given by

K = ~.

682

35. Dirac Equation

TOC

We now compare the K 2 and J 2 operators.

=
=

~ L
~ + ~)4 (
~ L
~ + ~4 ) = (
~ L
~ + ~4 )2
4 (
~ L
~
~ L
~ + 2~
~ L
~ + ~2 = i Li j Lj + 2~
~ L
~ + ~2

1 1

2 3 2 3 = 2 2 3 3 = 1

1 2

2 3 3 1 = 2 1 = 1 2 = i3

i j

ij + iijk k

~ L
~ + ~2
Li Lj (ij + iijk k ) + 2~
~ (L
~ L)
~ + 2~
~ L
~ + ~2
L2 + i

~ L)
~ o = xi pj xm pn ijk mnl klo
(L

=
=
=

2 ~2

~
~
xi jm pn ijk mnl klo = xi pn ijk jnl klo
i
i
~
~
xi pn jik jnl klo = xi pn (in kl il kn )klo
i
i
~
~
~
xi pi llo + xi pk kio = xi pk iko = i~Lo
i
i
i
~ L
~ + 2~
~ L
~ + ~2 = L2 + ~
~ L
~ + ~2
L2 ~
2
~ L
~ + 3 ~2 = K 2 ~
L2 + ~
4
4
0+

~2
4

= j(j + 1)~2

= j2 + j +

1
4

1
= (j + )
2
1
K = (j + )
2

The eigenvalues of K are

1
= j+
2

We may explicitly write out the eigenvalue equation for K for = j + 21 ~.

~ +~
1
~ L
0
A
A
~
K = ~ =
=

j
+
~ ~
B
B
2
0
~ L

683

35. Dirac Equation

TOC

~
The difference between J 2 and L2 is related to ~ L.
~ 3~
L2 = J 2 ~~ L
4
~ on the spinor , then, solve for the effect of L2 .
We may solve for the effect of ~ L
~ they are eigenstates of L2
Note that since A and B are eigenstates of J 2 and ~ L,
but have different eigenvalues.

~ +~
~ L
0
A
((j + 12 )A
=
~
~ ~
B
((j + 12 )B
0
~ L

~
~ L
0
A
((j + 12 1)A
= ~
~
B
((j + 12 1)B
0
~ L

A
((j +
~
~ L
=~
B
((j +

A
B

=
=
=
=
=

1
2
1
2

1)A
1)B

3 2 A
((j + 21 1)A
j(j + 1)~
~
~
B
((j + 1 1)B
4

2

1
3
A
((j + 2 1)A
~2
j(j + 1)

((j
+ 12 1)B
4
B

(j(j + 1) 34 (j + 12 1))A
~2
(j(j + 1) 34 (j + 12 1))B
2

1
3
2 (j + j j 2 + 1 4 )A
~
(j 2 + j j 12 + 1 34 )B
2

1
1
2 (j + j j 2 + 4 )A
~
(j 2 + j j 12 + 14 )B
2

A
B

Note that the eigenvalues for the upper and lower components have the same possible
values, but are opposite for energy eigenstates. We already know the relation ` = j 21
from NR QM. We simply check that it is the same here.
1
1
1
1
1 1
1 1
`(` + 1) = (j )(j + 1 ) = j 2 + j j j + = j 2 + j j +
2
2
2
2
2 4
2 4
It is correct. So A and B are eigenstates of L2 but with different eigenvalues.

684

35. Dirac Equation

TOC

A
B

= ~2

` (` + 1)A
` (` + 1)B

` = j

1
2

Now we apply the Dirac equation and try to use our operators to help solve the
problem.

mc

+ A +
=0
x
~c
~

mc

ie
=0
i
+ 4
+ 4 iA0 +
xi
x4
~c
~

e e mc2

ci
+
=0
i4 4
xi
t
~r
~

~c4 i
i~ + V (r) + mc2 4 = 0
xi
t

2
~c4 i
= i~ V (r) mc 4 = 0
xi
t

1 0
0 ii
2
~c
= i~ V (r) mc 4 = 0
0 1
ii
0
xi
t

0
i~i
2
c
= i~ V (r) mc 4 = 0
i~i
0
xi
t

2
0
0
i pi
A
i~ t V (r) mc
A
c
=

i p i
0
B
B
0
i~ t
V (r) + mc2

0
i pi
A
E V (r) mc2
0
A
c
=
i p i
0
B
0
E V (r) + mc2
B
The Dirac Equation then is.

c~ p~

B
E V (r) mc2
=
A
0

0
E V (r) + mc2

A
B

We can use commutation and anticommutation relations to write ~ p~ in terms of

685

35. Dirac Equation

TOC

separate angular and radial operators.

i x
i j x
j n pn
n pn

1
1
(i j + j i )
xi x
j n pn = 2ij x
i x
j n pn = n pn
2
2
1 i xi 1
1 i xi
(j n xj pn ) =
(j n xj pn + n j xn pj )
i x
i j x
j n pn =
r r
r r 2
1 i xi 1
1 i xi 1
(j n xj pn + (j n + 2injk k )xn pj ) =
( (j n xj pn + j
r r 2
r r 2
1 i xi 1
1 i xi 1
( (j n xj pn + n j xj pn ) + ik njk xn pj ) =
( (j n + n
r r 2
r r 2
1 i xi 1
1 i xi
( 2jn xj pn + ik Lk ) =
(xj pj + ik Lk )
r r 2
r r

1 ~ ~x

~
+ i~ L
i~r
r r
r

= i j x
i x
j n pn =
=
=
=
=

~ p~ =

1 ~ ~x
~ p~ =
r r

~
+ i~ L
i~r
r

~ act only on the angular momentum parts of

Note that the operators ~r~x and i~ L

the state. There are no radial derivatives so they commute with i~r r
. Lets pick a
shorthand notation for the angular momentum eigenstates we must use. These have
quantum numbers j, mj , and `. A will have ` = À and B must have the other
possible value of ` which we label `B . Following the notation of Sakurai, we will call
m
the state |jmj À i YjÀj = YÀ ,mj 21 + + YÀ ,mj + 12 . (Note that our previous
functions made use of m = m` particularly in the calculation of and .)

1 ~ ~x

0
A
B
E V (r) mc2
~
c
i~r
+ i~ L
=
0
E V (r) + mc2
B
A
r r
r

m
if (r)Yj`Bj
g(r)Y
1 ~ ~x

E V (r) mc2
0
~
c
i~r
+ i~ L
=
mj
0
E V (r) + mc2
g(r)Yj`A
if (r)
r r
r

The effect of the two operators related to angular momentum can be deduced.
~ is related to K. For positive , A has ` = j + 1 . For negative , A has
First, ~ L
2
1
` = j 2 . For either, B has the opposite relation for `, indicating why the full spinor

686

35. Dirac Equation

TOC

is not an eigenstate of L2 .

~ +~
~ L
0
K =
~ + ~
0
~ L

~ +~
1
~ L
0
A
A
K = ~ =
= j+
~
~ ~
B
B
2
0
~ L
~ + ~)A = ~A
(~ L
~ A
~ L
~ ~)B
(~ L
~ B
~ L

( 1)~A

= ~B
=

( 1)~B

Second, ~r~x is a pseudoscalar operator. It therefore changes parity and the parity of
the state is given by (1)` ; so it must change `.
~ ~x mj
m
Yj`A = CYj`Bj
r
2
The square of the operator ~r~x is one, as is clear from the derivation above, so we
know the effect of this operator up to a phase factor.
~ ~x mj
m
Yj`A = ei Yj`Bj
r
m

The phase factor depends on the conventions we choose for the states Yj` j . For our
conventions, the factor is 1.
~ ~x mj
m
Yj`A = Yj`Bj
r
We now have everything we need to get to

m
if (r)Yj`Bj
1 ~ ~x

~
c
i~r
+ i~ L
m
g(r)Yj`Aj
r r
r

~ if (r)Y mj
1 ~ ~x i~r r + i~ L
j`B

~ g(r)Y mj
r r
i~r r
+ i~ L
j`A

mj

~r r ( 1)~ f (r)Yj`B
1 ~ ~x

c
m

i~r r
+ i( 1)~ g(r)Yj`Aj
r r
~c

the radial equations.

g(
E V (r) mc2
0
=
0
E V (r) + mc2
if (

g(
E V (r) mc2
0
=
0
E V (r) + mc2
if (

=

E V (r) mc2
0

0
E V (r) + mc2

1 ~ ~x
r r
687

g(
if (

35. Dirac Equation

TOC

( 1) f (r)Yj`Bj
r r

m

ir r
i(1 + ) g(r)Yj`Aj

m

+ ( 1) f (r)Yj`Aj
1 r r

~c
m

ir r
+ i(1 + ) g(r)Yj`Bj
r

1 r r
+ ( 1) f (r)
~c

+ (1 + ) g(r)
r r
r

(1)
+
f
f
r
r
~c g (1+)
+
r
r g

E V (r) mc2
0

E V (r) mc2
0

E V (r) mc2
0

(E V mc2 )g
(E V + mc2 )f

=
=

m
g(r)Yj`Aj
m
if (r)Yj`Bj

m
g(r)Yj`Aj
0
m
E V (r) + mc2
if (r)Yj`Bj

0
g(r)
E V (r) + mc2
f (r)

0
E V (r) + mc2

This is now a set of two coupled radial equations. We can simplify them a bit by
making the substitutions F = rf and G = rg. The extra term from the derivative
cancels the 1s that are with s.

1 F
F
(E V mc2 ) G
r r + rF2 + F
2 r2
r
r

=
~c
1 G
G
G
G
(E V + mc2 ) Fr
r r r 2 + r 2 + r 2

F

2
r + F
(E

mc
)G
r

~c
=
G
G
(E V + mc2 )F
r + r

F
r
G
r

mc2 E+V
~c
mc2 +EV
~c

G
F

These equations are true for any spherically symmetric potential. Now it is time to
V
mc2 +E
specialize to the hydrogen atom for which ~c
= Z
and
r . We define k1 =
~c

2
mc E
k2 = ~c and the dimensionless = k1 k2 r. The equations then become.

F
F
k2 Z
r r =
r G
G
G
k1 + Z
F
r
r + r

q

k2
F
Z
F
G

k1

q

=
G
G
k1
Z
F
+
k2 +

q

k2

F
Z
G

q k1
=0
k1

Z
F
+ G
k2 +

With the guidance of the non-relativistic solutions, we will postulate a solution of

688

35. Dirac Equation

TOC

the form
F = e s
G = e s

X
m=0

am m = e

am s+m

m=0

bm m = e

m=0

bm s+m .

m=0

The exponential will make everything go to zero for large if the power series terminates. We need to verify that this is a solution near = 0 if we pick the right a0 ,
b0 , and s. We now substitute these postulated solutions into the equations to obtain
recursion relations.
!
r

k2
Z

G=0
F

!
r

k1
Z
+
G
F =0
+

k2

!
r

X
k2 s+m
s+m
s+m1
s+m1
s+m1
am
+ am (s + m)
am
bm

+ bm Z
=0
k1
m=0
!
r

X
k1 s+m
s+m
s+m1
s+m1
s+m1
bm
+ bm (s + m)
+ bm
am

am Z
=0
k2
m=0
!
r
k2
am + am+1 (s + m + 1) am+1 bm
+ bm+1 Z = 0
k1
!
r
k1
am+1 Z = 0
bm + bm+1 (s + m + 1) + bm+1 am
k2
r
k2
am + (s + m + 1 )am+1
bm + Zbm+1 = 0
k
r 1
k1
bm + (s + m + 1 + )bm+1
am Zam+1 = 0
k2

r
am + (s + m + 1 )am+1
r

k2
bm + Zbm+1 = 0
k1

k1
am Zam+1 bm + (s + m + 1 + )bm+1 = 0
k2

For the lowest order term s , we need to have a solution without lower powers. This
means that we look at the m = 1 recursion relations with am = bm = 0 and
689

35. Dirac Equation

TOC

solve the equations.

(s )a0 + Zb0 = 0

Za0 + (s + )b0 = 0

(s )
Z
a0
=0
Z (s + )
b0
s2 2 + Z 2 2 = 0
s2 = 2 Z 2 2
p
s = 2 Z 2 2

Note that while is a non-zero integer, Z 2 2 is a small non-integer number. We need

to take the positive root in order to keep the state normalized.
p
s = + 2 Z 2 2

As usual, the series must terminate at some m = nr for the state to normalizable.
This can be seen approximately by assuming either the as or the bs are small and
noting that the series is that of a positive exponential.
Assume the series for F and G terminate at the same nr . We can then take the
equations in the coefficients and set anr +1 = bnr +1 = 0 to get relationships between
anr and bnr .
r
k2
anr =
bn
k r
r 1
k1
an
bn r =
k2 r
These are the same equation, which is consistent with our assumption.
The final step is to use this result in the recursion relations for m = nr 1 to find
a condition on E which must be satisfied for the series to terminate. Note that this
choice of m connects anr and bnr to the rest of the series giving nontrivial conditions
on E. We already have the information from the next step in the recursion which gives

690

35. Dirac Equation

TOC

anr +1 = bnr +1 = 0.
r

k2
am + (s + m + 1 )am+1
bm + Zbm+1
k1
r
k1

am Zam+1 bm + (s + m + 1 + )bm+1
k2
r
k2
anr 1 + (s + nr )anr
bn 1 + Zbnr
k1 r
r
k1

an 1 Zanr bnr 1 + (s + nr + )bnr

k2 r
r
r
r
k1
k1
k1
anr 1 + (s + nr )
anr bnr 1 + Z
bn

k2
k2
k2 r
r
k1

an 1 Zanr bnr 1 + (s + nr + )bnr

k2 r

=0
=0
=0
=0
=0
=0

691

35. Dirac Equation

TOC

At this point we take the difference between the two equations to get one condition.
!
!
r
r
k1
k1
(s + nr )
+ Z anr + Z
(s + nr + ) bnr = 0
k2
k2
!r
!
r
r
k1
k2
k1
Z
bn + Z
(s + nr + ) bnr = 0
(s + nr )
k2
k1 r
k2
r
r
k2
k1
(s + nr ) Z
+ Z
(s + nr + ) = 0
k1
k2
p
p
(s + nr ) k1 k2 Zk2 + Zk1 (s + nr + ) k1 k2 = 0
p
2(s + nr ) k1 k2 + Z(k1 k2 ) = 0
p
2(s + nr ) k1 k2 = Z(k1 k2 )
p
2(s + nr ) m2 c4 E 2 = 2ZE
p
(s + nr ) m2 c4 E 2 = ZE
(s + nr )2 (m2 c4 E 2 ) = Z 2 2 E 2
(s + nr )2 (m2 c4 ) = (Z 2 2 + (s + nr )2 )E 2
(s + nr )2
(m2 c4 ) = E 2
((s + nr )2 + Z 2 2 )
m2 c4
E2 =
Z 2 2
(1 + (s+n
2)
r)
E=q
E=q
1+
E=s

mc2
1+

Z 2 2
(nr +s)2

mc2
2 2
Z
(nr + 2 Z 2 2 )2

mc2
1+

Z 2 2
2
q
2
nr + (j+ 12 ) Z 2 2

Using the quantum numbers from four mutually commuting operators, we have solved
the radial equation in a similar way as for the non-relativistic case yielding the exact
energy relation for relativistic Quantum Mechanics.

692

35. Dirac Equation

TOC

E=s

mc2
1+

Z 2 2
2
q
2
nr + (j+ 21 ) Z 2 2

We can identify the standard principle quantum number in this case as n = nr + j + 21 .

35.16

Thomson Scattering

The cross section for Thomson scattering illustrates the need for negative
energy states in our calculations. Recall that we got the correct cross section
from the non-relativistic calculation and that Thomson also got the correct result from
classical E&M.
In the Dirac theory, we have only one term in the interaction Hamiltonian,
Hint = ie4 k Ak
Because it is linear in A it can create a photon or annihilate a photon. Photon scattering
is therefore second order (and proportional to e2 ). The quantized photon field is
r

1 X ~c2 ()
A (x) =
ak, (0)eik x + ak, (0)eik x .
2
V
k

The initial and final states are definite momentum states, as are the intermediate
electron states. We shall first do the calculation assuming no electrons from
the negative energy sea participate, other than to exclude transitions to those
negative energy states. The initial and final states are therefore the positive energy
(r)
plane wave states p~ for r = 1, 2. The intermediate states must also be positive
energy states since the negative energy states are all filled.
The computation of the scattering cross section follows the same steps made in the
development of the Krammers-Heisenberg formula for photon scattering. There is no
693

35. Dirac Equation

TOC

A2 term so we are just computing the two second order terms.

(2)
c ~0 0 ~0 0 (t)
p ,r ;k ,

t
0
0
e2 X 1 ~c2
i(~
k~
xt2 )
)

dt2 hp~0 r0 ; k~0 ( ) |i4 n (()

+ (
n ak, e
n ak0 ,0 e
2
0
~
V 2
I
0

t2
dt1 ei(E

E)t1 /~

i(k~
xt1 )
)
i(k~
x+ t1 )
hI|i4 n (()
+ (
)|~
pr;
n ak, e
n ak0 ,0 e

0
(2)
c ~0 0 ~0 0 (t)
p ,r ;k

e2 ~c2

2V 0

"
X
p~00 r 00 =1,2

~0
~
hp~0 r0 |i4 n 0n eik ~x |p~00 r00 ihp~00 r00 |i4 n n eik~x |~
pri
E 00 E ~

# t
~0
~
0
0
pri i
hp~0 r0 |i4 n n eik~x |p~00 r00 ihp~00 r00 |i4 n 0n eik ~x |~
dt2 ei(E E+~ ~)t
00
0
E E + ~
~
0

As in the earlier calculation, the photon states have been eliminated from the equation
since they give a factor of 1 with the initial state photon being annihilated and the
final state photon being created in each term.
Now lets take a look at one of the matrix elements. Assume the initial state
electron is at rest and that the photon momentum is small.
~

hp~00 r00 |i4 n n eik~x |~

pri
For p~ = 0 and ~k = 0, a delta function requires that p~00 = 0. It turns out that
(r)
(r 00 )
u0
4 n u0 = 0, so that the cross section is zero in this limit.

0 ii
i =
ii
0

1 0
4 =
0 1

0
ii
4 i =
ii
0
This matrix only connects r = 1, 2 spinors to r = 3, 4 spinors because of its off diagonal
nature. So, the calculation yields zero for a cross section in contradiction to the other
two calculations. In fact, since the photon momentum is not quite zero, there is a small
contribution, but far too small.
The above calculation misses some important terms due to the negative energy sea. There are additional terms if we consider the possibility that the photon
can elevate a negative energy electron to have positive energy.

694

35. Dirac Equation

TOC

In one term, the initial state photon is absorbed by a negative energy electron, then
the initial state electron fills the hole in the negative energy sea emitting the final
state photon. In the other term, even further from the mass shell, a negative energy
electron emits the final state photon and moves to a positive energy state, then the
initial state electron absorbs the initial photon and fills the hole left behind in the sea.
These terms are larger because the 4 i matrix connects positive energy and negative
energy states.
"
~0
~
X
ie2 c2
hp~00 r00 |i4 n 0n eik ~x |~
prihp~0 r0 |i4 n n eik~x |p~00 r00 i
(2)

c ~0 0 ~0 0 (t) =
p ,r ;k
E 0 E 00 ~
2V 0 ~00 00
p r =3,4
~

hp~00 r00 |i4 n n eik~x |~

prihp~0 r0 |i4 n 0n eik ~x |p~00 r00 i
E 00 E 0 + ~ 0

(~ )(~ ) + (~ )(~ )

0
4mV
0

(2)
(t)
p ,r 0 ;k~0 0

c ~0

(2)
(t)
p ,r 0 ;k~0 0

c ~0

0
ie

(r ) [i j + j i ] i 0 j (r)
0
4mV

0
ie2

(r ) 0 (r)
2mV 0

t
dt2 ei(E

E+~ 0 ~)t2 /~

(2)
(t)
p ,r 0 ;k~0 0

(2)
(t)|2
p ,r 0 ;k~0 0

d
d

c ~0
|c ~0

0
0
ie2

0 rr0
dt2 ei(E E+~ ~)t2 /~
2mV 0
0
2 2
e
1
|
0 |2 rr0 2t(E 0 /~ E/~ + 0 )
2mV
0
2 2

2
V k 02 dk 0 d
e
|
0 |2 rr0 ( 0 )
2mV
0
(2)3
2 2

e
1
V d
2
|
0 |2 rr0
2mcV
c
(2)3
2 2
e
V 1 V
2
|
0 |2 rr0
2mcV
c c (2)3

2
e2
|
0 |2 rr0
4mc2

This agrees with the other calculations and with experiment. The negative energy
sea is required to get the right answer in Dirac theory. There are alternatives to the
negative energy sea. Effectively we are allowing for the creation of electron positron
pairs through the use of the filled negative energy states. The same result could be
obtained with the possibility of pair creation, again with the hypothesis that a positron
is a negative energy electron going backward in time.

35.17

Hole Theory and Charge Conjugation

Dirac postulated that the negative energy sea was entirely filled with electrons and
that an anti-electron would be formed if one of the negative energy electrons were
elevated to a positive energy state. This would yield a positive energy electron plus a
hole in the negative energy sea. The hole also has positive energy compared to the
vacuum since it is lacking the negative energy present in the vacuum state. Therefore,
both the electron and the positron would have positive energy. This describes the
process of pair creation.
697

35. Dirac Equation

TOC

Similarly, any positive energy electron could make a transition to the now empty negative energy state. Energy would be given off, for example by the emission of two
photons, and the system would return to the vacuum state. This is the process of pair
annihilation.
The tables below compare an electron and a positron with the same momentum
and spin. For simplicity, the tables assume the momentum is along the z direction so
that we can have spin up and spin down eigenstates. The electron and positron have
opposite charge so the standard Electromagnetic currents are in opposite directions.
The last row of the table shows the negative energy electron state that must be
unoccupied to produce the positron state shown in the row above it. A hole in the
vacuum naturally produces a positron of opposite charge, momentum, and
spin. Because the probability flux for the negative energy electron states is in the
opposite direction of the momentum, (and the charge and momentum are opposite the
positron) the EM current of the positron and of the negative energy state are in
opposite directions, due the product of three signs. This means the velocities are in
the same direction.

spin up, positive energy electron

spin up, positive energy positron
(spin down negative energy hole)
spin down, negative energy electron

charge
e
+e

mom.
p
z
p
z

p
z

p Energy
+pp2 c2 + m2 c4
+ p2 c2 + m2 c4

p2 c2 + m2 c4

Sz
+ ~2
+ ~2

~j (EM )

z
+
z

We have defined the positron spinor v (1) to be the one with positive momentum
and spin up. Note that the minus sign on u(4) is conventional and will come from our
future definition of the charge conjugation operator.
Similarly we can make a table starting from a spin down electron.

spin down, positive energy electron

spin down, positive energy positron
(spin up negative energy hole)
spin up, negative energy electron

charge
e
+e

mom.
p
z
p
z

p
z

p Energy
+pp2 c2 + m2 c4
+ p2 c2 + m2 c4

p2 c2 + m2 c4

Sz
~2
~2

~j (EM )

z
+
z

~v
+
+

+ ~2

We have now also defined the spinor, v (2) , for the spin down positron.

698

35. Dirac Equation

35.18

TOC

Charge Conjugate Waves

Assume that, in addition to rotation, boost and parity symmetry, the Dirac equation
also has a symmetry under charge conjugation. We wish to write the Dirac equation
in a way that makes the symmetry between electron and positron clear. Start from
the Dirac
equation
and include the coupling to the EM field with the substitution that

e ~
p~ p~ + c A .

+
x

ie
+ A +
x
~c

mc
=0
~
mc
=0
~

The strategy is to try to write the charge conjugate of this equation then show that it is
equivalent to the Dirac equation with the right choice of charge conjugation operator for
. First of all, the sign of eA is expected to change in the charge conjugate equation.
(Assume the equation, including the constant e is the same but the sign of the EM
field A changes.) Second assume, for now, that the Dirac spinor is transformed
to its charge conjugate by the operation
C = SC
where we are motivated by complex scalar field experience. SC is a 4 by 4 matrix. The
charge conjugate equation then is

ie
mc
A SC +
SC = 0.
x
~c
~
Take the complex conjugate carefully remembering that x4 and A4 will change
signs.

ie

ie
mc

+ Ai i SC +
A4 4 SC
+
S =0
xi
~c
x4
~c
~ C
1
Multiply from the left by SC
.

ie
mc
1
1
+ Ai SC i SC +
A4 SC
4 SC +
=0
xi
~c
x4
~c
~

Compare this to the original Dirac equation,

ie
+ A +
x
~c

ie
+ Ai i +
+ A4 4 +
xi
~c
x4
~c

mc
=0
~
mc
=0
~
699

35. Dirac Equation

TOC

The two equations will be the same if the matrix SC satisfies the conditions.
1
SC
i SC = i
1
SC
4 SC = 4 .

Recalling the matrices in our representation,

1
0 0 i 0
0 0 0 1
0 0 0 i

0
0 0
0 0 1 0
0 0 i 0
0
i
4 =

2 =
1 =
0
0 1 0 0 3 = i 0
0 i 0
0 0
0
0
0 i 0 0
1 0 0 0
i 0 0
0
note that 1 and 3 are completely imaginary and will change sign upon complex
conjugation, while 2 and 4 are completely real and will not. The solution in our
representation (only) is
1
1

SC
= SC
= SC = SC
= 2 .

0
1
0
0

It anti-commutes with 1 and 3 producing a minus sign to cancel the one from complex
conjugation. It commutes with 2 giving the right + sign. It anti-commutes with 4
giving the right - sign.
The charge conjugate of the Dirac spinor is given by.

0 = 2

Of course a second charge conjugation operation takes the state back to the original .

Applying this to the plane wave solutions gives.

s
s
s
mc2 (1) i(~p~xEt)/~
mc2 (4) i(~p~x+Et)/~
mc2 (1) i(~p~x+Et
(1)

up~ e
u~p e
v e
p~ =
|E|V
|E|V
|E|V p~
s
s
s
2
2
mc
mc
mc2 (2) i(~p~x+Et)/
(2)
(2)
(3)
p~ =
up~ ei(~p~xEt)/~
u~p ei(~p~x+Et)/~
v e
|E|V
|E|V
|E|V p~
s
s
mc2 (3) i(~p~x+|E|t)/~
mc2 (2) i(~p~x|E|t)/~
(3)
p~ =
up~ e

u e
|E|V
|E|V ~p
s
s
mc2 (4) i(~p~x+|E|t)/~
mc2 (1) i(~p~x|E|t)/~
(4)
p~ =
up~ e

u e
|E|V
|E|V ~p
The charge conjugate of an electron state is the negative energy electron
state, the absence of which would produce a positron of the same energy,
700

35. Dirac Equation

TOC

momentum, spin, and velocity as the electron. That is, the conjugate is the
hole needed to make a positron with the same properties as the electron except that it
has opposite charge.

Let us take one more look at a plane wave solution to the Dirac equation, for example
(1)
p~ and its charge conjugate, from the point of view that a positron is an electron
moving backward in time. Discard the idea of the negative energy sea. Assume
that we have found a new solution to the field equations that moves backward in time
rather than forward.
s
s
s
mc2 (1) i(~p~xEt)/~
mc2 (4) i(~p~x+Et)/~
mc2 (1) i(~p~x+Et)
(1)
up~ e

u~p e

v e
p~ =
|E|V
|E|V
|E|V p~
The charge conjugate of the electron solution is an electron with the same charge e,
opposite momentum ~
p, and spin opposite to the original state. It satisfies the equation
with the signs of the EM fields reversed and, because the sign of the Et term in the
exponential is reversed, it behaves as a positive energy solution moving backward in
time, with the right momentum and spin.
Our opinion of the negative energy solutions has been biased by living in a world
of matter. We know about matter waves oscillating as ei(~p~xEt)/~ . There is a
symmetric set of solutions for the same particles moving backward in time
oscillating as ei(~p~x+Et)/~ . These solutions behave like antiparticles moving
forward in time. Consider the following diagram (which contributes to Thomson
scattering) from two points of view. From one point of view, an electron starts out
(1)
at t1 , lets say in the state p~ . At time t3 , the electron interacts with the field and
(4)

makes a transition to the state p~00 which travels backward in time to t2 where it again
(1)

interacts and makes a transition to p~0 . From the other point of view, the electron
starts out at t1 , then, at time t2 , the field causes the creation of an electron positron
pair both of which propagate forward in time. At time t3 , the positron and initial
electron annihilate interacting with the field. The electron produced at t2 propagates
on into the future.

701

35. Dirac Equation

TOC

No reference to the negative energy sea is needed. No change in the

negative energy solutions is needed although it will be helpful to relabel them
with the properties of the positron rather than the properties of the electron moving
backward in time.
The charge conjugation operation is similar to parity. A parity operation changes the
system to a symmetric one that also satisfies the equations of motion but is different
from the original system. Both parity and charge conjugation are good symmetries of the Dirac equation and of the electromagnetic interaction. The
charge conjugate solution is that of an electron going backward in time that can also
be treated as a positron going forward in time.

35.19

Quantization of the Dirac Field

The classical free field Lagrangian density for the Dirac electron field is.

mc2
L = c~
x
The independent fields are considered to be the 4 components of and the four
This Lagrange density is a Lorentz scalar that depends only on
components of .
the fields. The Euler-Lagrange equation using the independent fields is simple

702

35. Dirac Equation

TOC

since there is no derivative of in the Lagrangian.

L
L
=0

x ( /x
)

L
=0

( /x
)
L
=0

c~
mc2 = 0
x

mc

+
x
~
This gives us the Dirac equation indicating that this Lagrangian is the right
one. The Euler-Lagrange equation derived using the fields is the Dirac adjoint
equation,

L
L

= 0
x (/x )

+ mc2 = 0
c~
x

mc

+
= 0
x
~
again indicating that this is the correct Lagrangian if the Dirac equation is
assumed to be correct.
To compute the Hamiltonian density, we start by finding the momenta conjugate
to the fields .
L
4 1 = i~ 4 4 = i~
= = c~

ic
t
There is no time derivative of so those momenta are zero. The Hamiltonian can then

703

35. Dirac Equation

TOC

be computed.
H

=
=
=
=
=

L
t

+ mc2

+ c~
i~
t
x

k + mc2

c~
+ c~ 4 4
+ c~
x4
x4
xk

~c 4 k
+ mc2 4
xk

~c4 k
+ mc2 4
xk

+ mc2 4 d3 x
~c4 k
xk

We may expand the field in the complete set of plane waves either using the
(r)
(r)
four spinors up~ for r = 1, 2, 3, 4 or using the electron and positron spinors up~ and
(r)

vp~ for r = 1, 2. For economy of notation, we choose the former with a plan to change
to the later once the quantization is completed.
s
4
XX
mc2
(r)
cp~,r up~ ei(~p~xEt)/~
(~x, t) =
|E|V
r=1
p
~

The conjugate can also be written out.

s
4
XX
mc2
(r)
cp~,r up~ ei(~p~xEt)/~
(~x, t) =
|E|V
r=1
p
~

Writing the Hamiltonian in terms of these fields, the formula can be simplified

704

35. Dirac Equation

TOC

as follows

2
H =
~c4 k
+ mc 4 d3 x
xk
s
XX
4 X X
4
mc2
(r 0 )
cp~0 ,r0 up~0
H =
0
|E |V
p
~ r=1 p~0 r 0 =1
s
XX
4
4 X X
mc2
(r 0 )
H =
cp~0 ,r0 up~0
0 |V
|E
p
~ r=1 p~0 r 0 =1
s
XX
4
4 X X
mc2
(r 0 )
cp~0 ,r0 up~0
H =
0
|E |V
p
~ r=1 p~0 r 0 =1

ic4 j pj + mc2 4 = E
s
XX
4 X X
4
mc2
(r 0 )
H =
cp~0 ,r0 up~0
0
|E |V
0
r=1

4 X X
4
XX

p
~ r=1 p~0 r 0 =1

(r)
up~

(E)

mc2
(r 0 )
cp~0 ,r0 up~0
(E)
0
|E |

mc2
(r)
cp~,r up~ ei(~p~xEt)/~
|E|V

mc2
(r)
cp~,r up~ p~p~0
|E|

r =1

(r 0 )
up~

|E|
rr0
mc2

4 X
4
XX
|E|
mc2 E
cp~,r0 cp~,r
rr0
|E|
mc2
r=1 0
p
~

4 X
4
XX
mc2
(r 0 )
(r)
cp~,r0 cp~,r up~ (E) up~
|E|
r=1 0
p
~

s
i(~
p~
xEt)/~

p~0 r =1

p
~

mc2
ei(~p~xEt)/~ ~c4 k
+ mc2 4
xk
|E|V
s

ipk
mc2
i(~
p~
xEt)/~
2
e
~c4 k
+ mc 4
c
~
|E|V
s
mc2
i(~
p~
xEt)/~
2
e
ic4 k pk + mc 4
cp~,r
|E|V

4
XX

The creation an annihilation operators bp~

relations.
(r 0 )

(r)

rr0 p~p~0

{bp~ , bp~ }

(r)
Np~

bp~

{bp~ , bp~0
(r)

(r)

{bp~

(r)

, bp~

(r) (r)
bp~

(r)

Np~ is the occupation number operator. The anti-commutation relations constrain the
occupation number to be 1 or 0.
A state of the electrons in a system can be described by the occupation numbers
(0 or 1 for each plane wave). The state can be generated by operation on the vacuum
state with the appropriate set of creation operators.
706

35. Dirac Equation

35.20

TOC

The Quantized Dirac Field with Positron Spinors

The basis states in our quantized Dirac field can be changed eliminate the negative
energy states and replace them with positron states. Recall that we can replace
(4)
(1)
(3)
(2)
u~p with the positron spinor vp~ and u~p with vp~ such that the new spinors are
charge conjugates of the electron spinors.
(s)

SC up~

(s)

= vp~

s = 1, 2

The positron spinor is actually just the same as the negative energy spinor when the
momentum is reversed.
We name the creation and annihilation operators for the positron states to
(s)
(s)
be dp~ and dp~ and identify them to be.
(4)

(1)

bp~

(2)

bp~

dp~
dp~

(3)

These anti-commute with everything else with the exception that

(s)

(s0 )

{dp~ , dp~0

The LaGrangian for electrons, photons, and the interaction between the two is the
LaGrangian of Quantum ElectroDynamics.

mc
1
A
L = ~c
+
F F ie
x
~
4
QED is our first complete example of an interacting Quantum Field Theory. It taught
us a great deal about the laws of physics.
The primary difference between Quantum Mechanics and Quantum Field Theory is
that particles can be created and destroyed. The probability to find an electron or
a photon integrated over space does not have to be one. It can change with time.
We have written the fields of the photon and the electron in terms of creation and
annihilation operators.
r

1 X ~c2 ()
~
~
ak, (t)eik~x + ak, (t)eik~x
A =
2
V k
r
2

XX
mc2 (s) (s) i(~p~xEt)/~
(s) (s)
(~x, t) =
bp~ up~ e
+ dp~ vp~ ei(~p~xEt)/~
EV
p
~ s=1
r
2

XX
mc2 (s) (s) i(~p~xEt)/~
(s) (s)
bp~ up~ e
+ dp~ vp~ ei(~p~xEt)/~
(~x, t) =
EV
s=1
p
~

A photons can be created or destroyed

Note that in the interaction term ie
singly but that electrons must be created and destroyed along with a positron.
Phase (or Gauge) symmetry can be studied very simply from this LaGrangian. We
have shown that the phase transformation
ei(x)
~c (x)
A A
e x
leaves the Schr
odinger equation invariant. This can be most directly studied using the
LaGrangian. We can deduce from the above transformation that

F =

A
A
~c

F
x
x
e

ei(x)

(x)
(x)

= F
x x
x x
710

35. Dirac Equation

TOC

The transformed LaGrangian then can be computed easily.

mc
1
A
L = ~c
+
F F ie
x
~
4
The exponentials from and cancel except for the term in which is differentiated.
ie
~c = L
L L i~c
x
e x
This all may seem fairly simple but imagine that we add a mass term for the EM field,
m2 A A . The LaGrangian is no longer gauge invariant. Gauge invariance implies
zero mass photons and even maintains the massless photon after radiative corrections.
Gauge invariance also implies the existence of a conserved current. Remember that
electric current in 4D also includes the charge density. Gauge invariance implies conservation of charge, another important result.
This simple transformation ei(x) is called a local U(1) symmetry where the U
stands for unitary.
The Weak interactions are based on an SU(2) symmetry. This is just a local phase
symmetry times an arbitrary local rotation in SU(2) space. The SU(2) group is familiar
to us since angular momentum is based on SU(2). In the weak interactions, there are
two particles that are the symmetric (much like a spin up and a spin down electron
but NOT a spin up and spin down electron). We can rotate our states into different
linear combinations of the symmetric particles and the LaGrangian remains invariant.
Given this local SU(2) symmetry of the fermion wave functions, we can easily deduce
what boson fields are required to make the LaGrangian gauge invariant. It turns out
we need a triplet of bosons. (The weak interactions then get messy because of the
Higgs mechanism but the underlying gauge theory is still correct.)
The Strong interactions are based on the SU(3) group. Instead of having 3 sigma
matrices to do rotations in the lowest dimension representation of the group, SU(3)
has eight lambda matrices. The SU(3) symmetry for the quark wavefunctions requires
an octet of massless vector boson called gluons to make the LaGrangian gauge invariant.
So the Standard Model is as simple as 1 2 3 in Quantum Field Theories.

35.23

Interaction with a Scalar Field

while couplings to a
Yukawa couplings to a scalar field would be of the form G
5 .
pseudoscalar field would be of the form iG
711

36. Formulas with Hyperlinks to Course

TOC

Formulas with Hyperlinks to Course

~ = 1.055 1034 J s
1eV = 1.602 1019 J R
1
A= 0.1 nm

1
137

e2
40 ~c

mn = 939.6 MeV/c

kB = 1.38 1023 J/ K

ge = 2 +

= 5.79 10

dx eax =

p
a

eA =

P
n=0

P (x) =
V (r) =

An
n!

Z~c
r

use

Ze
40 r

gp = 5.6

f (x)dx(g(x)) =

for other forms R

n=1,3...

1
2~

1
dg
| dx
|

f (x)

R
g(x)=0

dx f (x) (x a) = f (a)

dr rn ear =

= 0.529 108 cm

~
me c

me = 9.11 1031 kg = 0.511 MeV/c2

eV/Tesla

sin =

2
2
1
ex /2 R
2 2

~c = 197.3 eV nm = 197.3 MeV fm

a0 =

mp = 938.3 MeV/c

Bohr =

e2
cgs

1 Fermi = 1 fm

e~
2me c

e = 1.602 1019 Coulomb

c = 3.00 108 m/sec

n1
n
(1) 2
n!

n!
an+1

cos =

n=0,2,4...

n
n
(1) 2
n!

p
m2 c4 + p2 c2

ei(p1 p2 )x/~ dx (p1 p2 )

GENERAL WAVE MECHANICS

E = h = ~

= h/p

p = ~k

p x

(A)2 (B)2 h 2i [A, B]i2 R

(x) =
p =

~
2

1
2~

dp (p) eipx/~

~
i x

= i~
E
t

Huj (x) = Ej uj (x) R

d
dx

(x) continuous
d
dx
P

2m
~2 (a)

|ui ihui | = 1

p
hA2 i hAi2

1
(p) = 2~
dx (x) eipx/~

continous if V finite

for V (x) = (x a)
R

hui |uj i = ij

R
R
R

x
= i~ p

h|i

dx (x)(x)

R
R
R
R

x2 + V (x) = i~ t
2
j
x2 + V (x)j = Ej j
j (x, t) = uj (x)eiEj t/~

~2
2m
~2
2m

ai ui

ai = hui |i

(x) = hx|i

(p) = hp|i

h|A|i = h|Ai = hA |i = h|A |i

~ 2 + V (~r)](~r) = E(~r)
[ 1 (~
p + e A)
2m

[px , x] =

c
~
i

i = hui |i

H = E

R
R

Aij = hui |A|uj i

dhAi
dt

i
= h A
t i + ~ h[H, A]i

712

36. Formulas with Hyperlinks to Course

TOC

HARMONIC OSCILLATOR
H=

p2
2m

1
m 2 x2
2

= ~A A + 12 ~

k y 2 /2

un (x) =
ak y e
R
k=0
p m
p
A = ( 2~ x + i 2m~
) R
p

A |ni = (n + 1) |n + 1iR

En = (n + 12 )~
p
y = m
x
~

R
2(kn)
a
(k+1)(k+2) k

ak+2 =
R
p
p
A = ( m
x i 2m~
)
2~
p
A |ni = (n) |n 1i R

n = 0, 1, 2...R
R

[A, A ] = 1
u0 (x) =

1
( m
)4
~

mx2 /2~

ANGULAR MOMENTUM
[Li , Lj ] = i~ijk Lk

L2 Y`m = `(` + 1)~2 Y`m R

L = Lx iLy
Y00 =
Y22 =

1
q4

R
R

e2i sin2

15
32

Y`0 m0 d = ``0 mm0

Y`m

Lz Y`m = m~Y`m
R ` m `
p
L Y`m = ~ `(` + 1) m(m 1) Y`,m1
q
q
3
3
Y11 = 8
ei sin
R Y10 = 4
cos
q
q
15 i
5
Y21 = 8
e sin cos Y20 = 16
(3 cos2 1)

[L2 , Li ] = 0

R
R
R

Y`` = Cei` sin`

R Y`(m) = (1)m Y`m
R P Y`m (, ) = (1)` Y`m (, ) R
h 2
i

`(`+1)~2
~2

2
+ r r Rn` (r) + V (r) + 2r2
Rn` (r) = ERn` (r)
R
2
r 2

~
H = H0
~ B
Si =

2 i

x =

[i , j ] = 2iijk k
!
0 i
y =
i
0

0
2

i
Ly = ~
0
2
i
0
2

!
R

1 0

1
Lx = ~
2
0

1
2

1 R
2

0
1
2

e ~
L
2mc

~=
R
R
0

R
2

ge ~
S
2mc

= B ~

{i , j } = 0
z =

R
!

0 1

1 0

Lz = ~
0 0
0 0

R
0

HYDROGEN ATOM
p2
2

Z~c
r

n`m = Rn` (r)Y`m (, )

n = nr + ` + 1
q
= 8E
r=
~2

a0 =

ak+1 =

3
2

R10 = 2( aZ0 ) e
Rn,n1 r

2r
na0
Zr
a0

Zr/na0

H1 = 8mp3 c2
H3 =

e2 gp
~
S
3mMp c2

HB =

eB
(Lz
2mc

3
~
I4
(~r)

+ 2Sz )

~
c
k+`+1n
a
(k+1)(k+2`+2) k
3
2

Z
R20 = 2( 2a
) (1
0
m2
= mm11+m
2
2
~
H2 = 2m2ec2 r3 S

E3 =
EB =

` = 0, 1, ..., n 1

P
Rn` () = `
ak k e/2

Zr
2a0

R
~
L

2gp m4 mc2
(f (f
3Mp n3
e~B
(1
2mc

2 c2
2n2

En = Z

Zr
)
2a0

= 13.6
eV
n2

k=0
Zr
3
Z
R21 = 13 ( 2a
) 2 ( Zr
) e 2a0
R
a0
0
Z2 c2
e2
Ze2
hn`m | r |n`m i = n2 a0 = n2
3
E12 = 2n1 3 4 mc2 ( j+1 1 4n
)
2

+ 1) I(I + 1) 43 )

1
)mj
2`+1

for j = `

1
2

713

36. Formulas with Hyperlinks to Course

TOC

ADDITION OF ANGULAR MOMENTUM

~ +S
~
J~ = L

~ S
~ = 1 (J 2 L2 S 2 )
|` s| j ` + s
L
2
P
P
jmj `s =
C(jmj ; `m` sms )Y`m` sms =
hjmj `s|`m` sms iY`m` sms
m` ms
q
q m` ms
`+m+1
`m
1
1
j,mj = `+ 2 ,m+ 2 =
Y`m + + 2`+1
Y`,m+1 for s = 21 and any `
q 2`+1
q
`m
j,mj = ` 12 ,m+ 12 = 2`+1
for s = 21 and any `
Y`m + `+m+1
2`+1 Y`,m+1

PERTURBATION THEORY AND RADIATIVE DECAYS

|hk |H1 |n i|2

P
(1)
En

(2)
En

= hn |H1 |n i

cn (t) =

1
i~

dt0 ei(En Ei )t

(1)

k6=n
(0)

cnk =

(0)

En Ek

hk |H1 |n i
(0)

(0)

En Ek

hn |V (t0 )|i i

Fermis Golden Rule:

if = 2
|hf |V |i i|2 f (E)
~

Q
3
d pk
2 3
if = 2
( V(2~)
3 ) |Mf i | (momentum conservation) (Energy
~
k

dp km |hm |eik~r p
~|k i|2
rad
mk = 2m2 c2

dp km
|hm |
~r|k i|2
E1
mk = 2c2
q
+i
+i
r = 4
(z Y10 + x2 y Y11 + x2 y Y11 )
3
q
3
I() (0 /2

=
P

collision
2
2
mkT
) +(/2)

~ r
p
d
~ )|2 V (
~ ) = d3~r ei~
( d
)BORN = 412 ~4 pfi mf mi |V (
V (~r )

conservation)

l = 1, s = 0
~k = 0
(
)
=
Dopler
~ =

kT
mc2

p~f p~i
~

ATOMS AND MOLECULES

Hund: 1) max s
Erot =

`(`+1)~
2I

2)max ` (allowed)
1
2000

Evib = (n +

1
2 )~

3) min j (

1
50

1
2

shell) else max j

714

Statistical Thermophysics, Robertson
100% (1)
Statistical Thermophysics, Robertson
594 pages
Workstation Installation Checklist
No ratings yet
Workstation Installation Checklist
13 pages
Basic Space Plasma Physics - Revised Edition
No ratings yet
Basic Space Plasma Physics - Revised Edition
17 pages
20th Century Physics
100% (1)
20th Century Physics
313 pages
PHYS122 University of Waterloo
No ratings yet
PHYS122 University of Waterloo
174 pages
Felix Boehm, Petr Vogel Physics of Massive Neutrinos, Second Edition
No ratings yet
Felix Boehm, Petr Vogel Physics of Massive Neutrinos, Second Edition
259 pages
Intro Modern Physics
No ratings yet
Intro Modern Physics
157 pages
2016 Book MeltingHadronsBoilingQuarks-Fr
No ratings yet
2016 Book MeltingHadronsBoilingQuarks-Fr
457 pages
Exploring Black Holes PDF
No ratings yet
Exploring Black Holes PDF
2 pages
Teaching Electromagnetic Field Theory Using Differential Forms PDF
100% (1)
Teaching Electromagnetic Field Theory Using Differential Forms PDF
18 pages
2009 Book ECGStueckelbergAnUnconventiona
No ratings yet
2009 Book ECGStueckelbergAnUnconventiona
414 pages
Nuclear Physics and Radioactivity
No ratings yet
Nuclear Physics and Radioactivity
58 pages
Physics Outreach Program Physics Olympiad Preparation 2012-2013
No ratings yet
Physics Outreach Program Physics Olympiad Preparation 2012-2013
2 pages
Jim Baggott - The Quantum Cookbook - Mathematical Recipes For The Foundations of Quantum Mechanics (2020, Oxford University Press)
100% (1)
Jim Baggott - The Quantum Cookbook - Mathematical Recipes For The Foundations of Quantum Mechanics (2020, Oxford University Press)
314 pages
Concepts in Particle Physics A Concise Introduction To The Standard Model by Nair, V. P.
No ratings yet
Concepts in Particle Physics A Concise Introduction To The Standard Model by Nair, V. P.
326 pages
Interpretations of Quantum Mechanics
100% (1)
Interpretations of Quantum Mechanics
15 pages
Quantum Entanglement
100% (1)
Quantum Entanglement
2 pages
General Theory of Relativity
No ratings yet
General Theory of Relativity
45 pages
Errors of Wald On General Relativity
100% (1)
Errors of Wald On General Relativity
16 pages
Instructions and Demonstrations
No ratings yet
Instructions and Demonstrations
32 pages
Physics 2008
100% (1)
Physics 2008
85 pages
Kluwer - 2001 - Problems in Laser Physics
100% (2)
Kluwer - 2001 - Problems in Laser Physics
323 pages
Chen Tang PrimeNumberTheorem
No ratings yet
Chen Tang PrimeNumberTheorem
81 pages
9 Quantum Field Theory For Children
No ratings yet
9 Quantum Field Theory For Children
8 pages
Iit Jee Physics Book Features PDF
No ratings yet
Iit Jee Physics Book Features PDF
8 pages
978-3-031-59501-1
No ratings yet
978-3-031-59501-1
174 pages
QM PDF
100% (1)
QM PDF
341 pages
CH 29 Knight 4th
No ratings yet
CH 29 Knight 4th
118 pages
String Theory: String Theory Is A Developing Theory in Particle Physics That Attempts To
No ratings yet
String Theory: String Theory Is A Developing Theory in Particle Physics That Attempts To
16 pages
Intro Physics 2.a4
No ratings yet
Intro Physics 2.a4
527 pages
Ising Model 2d
100% (1)
Ising Model 2d
11 pages
Entropy in Statistical Mechanics
No ratings yet
Entropy in Statistical Mechanics
8 pages
Modern Physics For High School
No ratings yet
Modern Physics For High School
3 pages
Supathorn Phongikaroon
No ratings yet
Supathorn Phongikaroon
306 pages
Quantum Mechanics
50% (4)
Quantum Mechanics
34 pages
Cosmology Part I: The Homogeneous Universe: Hiranya V. Peiris
No ratings yet
Cosmology Part I: The Homogeneous Universe: Hiranya V. Peiris
54 pages
Lamb Shift Presentation For Leyman
100% (1)
Lamb Shift Presentation For Leyman
18 pages
Physics Experiments en
100% (3)
Physics Experiments en
123 pages
Ten Myths of Science
No ratings yet
Ten Myths of Science
8 pages
Townes, Charles Logic and Uncertainties in Science and Religion 2001 14pp
No ratings yet
Townes, Charles Logic and Uncertainties in Science and Religion 2001 14pp
14 pages
Master Thesis Optical Properties of Pentacene and Picene: University of The Basque Country WWW - Mscnano.eu
No ratings yet
Master Thesis Optical Properties of Pentacene and Picene: University of The Basque Country WWW - Mscnano.eu
51 pages
Chapter 26 Quantum Physics
No ratings yet
Chapter 26 Quantum Physics
92 pages
Wolfgang Demtröder Atoms, Molecules and Photons - An Introduction To Atomic-, Molecular - and Quantum-Physics
No ratings yet
Wolfgang Demtröder Atoms, Molecules and Photons - An Introduction To Atomic-, Molecular - and Quantum-Physics
10 pages
University Technology
No ratings yet
University Technology
16 pages
Where Do Students "Go Wrong" When Solving Newton's Second Law Problems
No ratings yet
Where Do Students "Go Wrong" When Solving Newton's Second Law Problems
19 pages
Summary: PC1144, Introduction To Modern Physics
No ratings yet
Summary: PC1144, Introduction To Modern Physics
41 pages
Lecture Notes On General Relativity - Mathias Blau
No ratings yet
Lecture Notes On General Relativity - Mathias Blau
229 pages
Julian Schwinger
No ratings yet
Julian Schwinger
23 pages
Relativistic Quantum Mechanics: 6.1 The Klein-Gordon Equation
No ratings yet
Relativistic Quantum Mechanics: 6.1 The Klein-Gordon Equation
17 pages
1970 Zwiebach Part1
100% (1)
1970 Zwiebach Part1
143 pages
Natario Broeck Warp Drive
No ratings yet
Natario Broeck Warp Drive
88 pages
Mathematical Methods Quantum Mechanics
100% (6)
Mathematical Methods Quantum Mechanics
314 pages
Solid State Physics
50% (2)
Solid State Physics
18 pages
Quasi Particles Kaganov Lifshits
100% (1)
Quasi Particles Kaganov Lifshits
100 pages
Monte Carlo Simulation in Statistical Physics An I
No ratings yet
Monte Carlo Simulation in Statistical Physics An I
36 pages
Instead of the ITER project and the TOKAMAK principle: – a new type of fusion machine
From Everand
Instead of the ITER project and the TOKAMAK principle: – a new type of fusion machine
Åke Hedberg
5/5 (1)
Applications of Model Theory to Functional Analysis
From Everand
Applications of Model Theory to Functional Analysis
José Iovino
No ratings yet
Lectures on the Coupling Method
From Everand
Lectures on the Coupling Method
Torgny Lindvall
No ratings yet
130A Classnotes With Solutions
No ratings yet
130A Classnotes With Solutions
675 pages
Quantum Physics 130 - Notes - James Branson
0% (1)
Quantum Physics 130 - Notes - James Branson
548 pages
Quantum Physics
No ratings yet
Quantum Physics
553 pages
Commutative K-Theory: C. Norris
No ratings yet
Commutative K-Theory: C. Norris
12 pages
Infinite Moduli and An Example of D Escartes: C. Norris
No ratings yet
Infinite Moduli and An Example of D Escartes: C. Norris
12 pages
Multiply Anti-Algebraic, Semi-Almost Surely χ-p-Adic, Freely Additive Paths over Naturally Left-Meager, Ultra-Noetherian, Pairwise Riemann Subalegebras
No ratings yet
Multiply Anti-Algebraic, Semi-Almost Surely χ-p-Adic, Freely Additive Paths over Naturally Left-Meager, Ultra-Noetherian, Pairwise Riemann Subalegebras
8 pages
An Example of Cartan: C. Norris
No ratings yet
An Example of Cartan: C. Norris
7 pages
Locality Methods in Introductory Combinatorics: C. Norris
No ratings yet
Locality Methods in Introductory Combinatorics: C. Norris
8 pages
Mathgen 1742088573
No ratings yet
Mathgen 1742088573
8 pages
Mathgen 374984038
No ratings yet
Mathgen 374984038
11 pages
Geometry and Its Applications, 2: Math 150A - Prof. Rabin - Fall 2015 - Homework Assignments
No ratings yet
Geometry and Its Applications, 2: Math 150A - Prof. Rabin - Fall 2015 - Homework Assignments
1 page
Stock Level
No ratings yet
Stock Level
14 pages
Motion Class 9
No ratings yet
Motion Class 9
16 pages
Ielts Speaking A Collection of Common Topics: Unit 1 People Lesson 7 An Old Person You Respect
No ratings yet
Ielts Speaking A Collection of Common Topics: Unit 1 People Lesson 7 An Old Person You Respect
8 pages
Pccoe College Information
No ratings yet
Pccoe College Information
11 pages
TDD With Django
No ratings yet
TDD With Django
27 pages
How Mixed Must A Mixed System Be?
No ratings yet
How Mixed Must A Mixed System Be?
17 pages
Role of Library in Research
100% (1)
Role of Library in Research
6 pages
Fabm 1 Lesson 4
67% (3)
Fabm 1 Lesson 4
3 pages
TCS34725 Color Sensor User Manual
No ratings yet
TCS34725 Color Sensor User Manual
16 pages
Auto Car Lock Using GSM
No ratings yet
Auto Car Lock Using GSM
15 pages
Design For Subjective Well-Being in Interior Architecture: Petermans, Ann
No ratings yet
Design For Subjective Well-Being in Interior Architecture: Petermans, Ann
13 pages
Slang Word Formation in Pitch Perfect Movie
No ratings yet
Slang Word Formation in Pitch Perfect Movie
11 pages
A-1 Bca 101
No ratings yet
A-1 Bca 101
2 pages
The Role of Informal Mechanism in The Co
No ratings yet
The Role of Informal Mechanism in The Co
28 pages
How To Set Up Reports in The Page Layout
No ratings yet
How To Set Up Reports in The Page Layout
2 pages
Evidence For Slavic Lead Mining and Trade Early Rus - 2024 - Journal of Archaeo
No ratings yet
Evidence For Slavic Lead Mining and Trade Early Rus - 2024 - Journal of Archaeo
10 pages
Writing Task 1 Week 1
No ratings yet
Writing Task 1 Week 1
7 pages
Sys Eqns Applications 3x3
No ratings yet
Sys Eqns Applications 3x3
11 pages
Transformer Company CV
No ratings yet
Transformer Company CV
1 page
Catalog 16 Februarie 20111
No ratings yet
Catalog 16 Februarie 20111
27 pages
Khảo Sát 126 1
No ratings yet
Khảo Sát 126 1
5 pages
ADM 4615 & MBA 6623 - Fall 2018 H.W. #1
No ratings yet
ADM 4615 & MBA 6623 - Fall 2018 H.W. #1
11 pages
Magtungtong Tayo
No ratings yet
Magtungtong Tayo
66 pages
Influence The Psychology of Persuasion - Graphic
0% (1)
Influence The Psychology of Persuasion - Graphic
1 page
Jurisprudence-I Class Notes
100% (2)
Jurisprudence-I Class Notes
23 pages
Essay Muet
No ratings yet
Essay Muet
2 pages
Hypothesis Testing
No ratings yet
Hypothesis Testing
65 pages
STEP7 ErrorCode English
No ratings yet
STEP7 ErrorCode English
31 pages
SAP Overview NTL
No ratings yet
SAP Overview NTL
15 pages