43 tayangan

Diunggah oleh jhoan gaviria

- Dhaka University EMBA Hand Book
- Abstract Algebra Cheat Sheet
- Basic Theory of Affine Group Schemes
- Mnl Dec2013 Crc 18thjan14
- Linear Algebra (Math1014) Course Outline
- F5 Maths Annual Scheme of Work_2010
- QuTiP-2.0.0-DOC
- Contextual
- fisika komputasi matlab
- Group
- BSc Computer Science Www.alagappauniversity.ac.in Alagappa University Model Question Paper Dist (1)
- Magic Squares
- Operations on Matrices - Linear Algebra
- Aimsun Macroscopic Modelling Manual v6_1
- Trident 360-Degree Matirx System, Jan 2012 Dentistry
- JEE Mathematics Syllabus
- 3P2Lb.pdf
- Science (1)
- 0702842.pdf
- MIT2_019S11_SK4

Anda di halaman 1dari 364

(under construction)

Peter Woit

Department of Mathematics, Columbia University

woit@math.columbia.edu

ii

Contents

Preface xi

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Basic axioms of quantum mechanics . . . . . . . . . . . . . . . . 2

1.3 Unitary group representations . . . . . . . . . . . . . . . . . . . . 5

1.4 Representations and quantum mechanics . . . . . . . . . . . . . . 6

1.5 Symmetry groups and representations on function spaces . . . . 7

1.6 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1 Some representation theory . . . . . . . . . . . . . . . . . . . . . 11

2.2 The group U (1) and its representations . . . . . . . . . . . . . . 14

2.3 The charge operator . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.4 Conservation of charge . . . . . . . . . . . . . . . . . . . . . . . . 18

2.5 Suggestions for further reading . . . . . . . . . . . . . . . . . . . 19

3.1 The two-state quantum system . . . . . . . . . . . . . . . . . . . 22

3.1.1 The Pauli matrices: observables of the two-state quantum

system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.1.2 Exponentials of Pauli matrices: unitary transformations

of the two-state system . . . . . . . . . . . . . . . . . . . 24

3.2 Commutation relations for Pauli matrices . . . . . . . . . . . . . 26

3.3 Dynamics of a two-state system . . . . . . . . . . . . . . . . . . . 28

4.1 Dual spaces and inner products . . . . . . . . . . . . . . . . . . . 31

4.2 Bases, linear operators and matrix elements . . . . . . . . . . . . 33

4.3 Adjoint operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.4 Orthogonal and unitary transformations . . . . . . . . . . . . . . 36

4.4.1 Orthogonal groups . . . . . . . . . . . . . . . . . . . . . . 37

4.4.2 Unitary groups . . . . . . . . . . . . . . . . . . . . . . . . 38

4.5 Eigenvalues and eigenvectors . . . . . . . . . . . . . . . . . . . . 38

iii

4.6 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 40

5.1 Lie algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5.2 Lie algebras of the orthogonal and unitary groups . . . . . . . . . 44

5.2.1 Lie algebra of the orthogonal group . . . . . . . . . . . . . 45

5.2.2 Lie algebra of the unitary group . . . . . . . . . . . . . . 46

5.3 Lie algebra representations . . . . . . . . . . . . . . . . . . . . . 47

5.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 51

6.1 The rotation group in three dimensions . . . . . . . . . . . . . . 53

6.2 Spin groups in three and four dimensions . . . . . . . . . . . . . 56

6.2.1 Quaternions . . . . . . . . . . . . . . . . . . . . . . . . . . 57

6.2.2 Rotations and spin groups in four dimensions . . . . . . . 58

6.2.3 Rotations and spin groups in three dimensions . . . . . . 58

6.2.4 The spin group and SU (2) . . . . . . . . . . . . . . . . . 62

6.3 Spin groups in higher dimension . . . . . . . . . . . . . . . . . . 65

6.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 65

7.1 The spinor representation . . . . . . . . . . . . . . . . . . . . . . 67

7.2 The spin 1/2 particle in a magnetic field . . . . . . . . . . . . . . 68

7.3 The Heisenberg picture . . . . . . . . . . . . . . . . . . . . . . . 72

7.4 The Bloch sphere and complex projective space . . . . . . . . . . 73

7.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 77

8.0.1 The spin 0 representation . . . . . . . . . . . . . . . . . . 79

8.0.2 The spin 21 representation . . . . . . . . . . . . . . . . . . 79

8.0.3 The spin 1 representation . . . . . . . . . . . . . . . . . . 80

8.1 Representations of SU (2): classification . . . . . . . . . . . . . . 80

8.1.1 Weight decomposition . . . . . . . . . . . . . . . . . . . . 80

8.1.2 Lie algebra representations: raising and lowering operators 82

8.2 Representations of SU (2): construction . . . . . . . . . . . . . . 86

8.3 Representations of SO(3) and spherical harmonics . . . . . . . . 89

8.4 The Casimir operator . . . . . . . . . . . . . . . . . . . . . . . . 94

8.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 96

9.1 Tensor products . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

9.2 Composite quantum systems and tensor products . . . . . . . . . 99

9.3 Indecomposable vectors and entanglement . . . . . . . . . . . . . 100

9.4 Tensor products of representations . . . . . . . . . . . . . . . . . 101

9.4.1 Tensor products of SU (2) representations . . . . . . . . . 102

9.4.2 Characters of representations . . . . . . . . . . . . . . . . 102

iv

9.4.3 Some examples . . . . . . . . . . . . . . . . . . . . . . . . 104

9.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 105

10.1 Energy, momentum and space-time translations . . . . . . . . . . 107

10.2 Periodic boundary conditions and the group U (1) . . . . . . . . . 111

10.3 The group R and the Fourier transform . . . . . . . . . . . . . . 115

10.3.1 Delta functions . . . . . . . . . . . . . . . . . . . . . . . . 117

10.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 119

11.1 The position operator and the Heisenberg Lie algebra . . . . . . 122

11.1.1 Position space representation . . . . . . . . . . . . . . . . 122

11.1.2 Momentum space representation . . . . . . . . . . . . . . 123

11.1.3 Physical interpretation . . . . . . . . . . . . . . . . . . . . 123

11.2 The Heisenberg Lie algebra . . . . . . . . . . . . . . . . . . . . . 125

11.3 The Heisenberg group . . . . . . . . . . . . . . . . . . . . . . . . 126

11.4 The Schrodinger representation . . . . . . . . . . . . . . . . . . . 127

11.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 129

12.1 Poisson brackets . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

12.2 The Lie subalgebra of quadratic polynomials . . . . . . . . . . . 135

12.3 The symplectic group and its action on the Heisenberg Lie algebra137

12.4 The Poisson bracket on the dual of a Lie algebra . . . . . . . . . 139

12.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 140

13 Quantization 141

13.1 Canonical quantization . . . . . . . . . . . . . . . . . . . . . . . . 141

13.2 The Groenewold-van Hove no-go theorem . . . . . . . . . . . . . 144

13.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Atom 147

14.1 Angular momentum . . . . . . . . . . . . . . . . . . . . . . . . . 148

14.2 Quantum particle in a central potential . . . . . . . . . . . . . . 150

14.3 so(4) symmetry and the Coulomb potential . . . . . . . . . . . . 154

14.4 The hydrogen atom . . . . . . . . . . . . . . . . . . . . . . . . . . 157

14.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 158

15.1 The harmonic oscillator with one degree of freedom . . . . . . . . 160

15.2 Creation and annihilation operators . . . . . . . . . . . . . . . . 162

15.3 The Bargmann-Fock representation . . . . . . . . . . . . . . . . . 165

15.4 Multiple degrees of freedom . . . . . . . . . . . . . . . . . . . . . 167

15.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 168

v

16 The Harmonic Oscillator as a Representation of the Heisenberg

Group 169

16.1 Complex structures and quantization . . . . . . . . . . . . . . . . 170

16.2 The Bargmann transform . . . . . . . . . . . . . . . . . . . . . . 173

16.3 Coherent states and the Heisenberg group action . . . . . . . . . 174

16.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 176

Group 177

17.1 The metaplectic representation for d = 1 . . . . . . . . . . . . . . 178

17.2 The metaplectic representation and the choice of complex structure181

17.3 The metaplectic representation for d degrees of freedom . . . . . 183

17.3.1 Two degrees of freedom and SU (2) . . . . . . . . . . . . . 186

17.3.2 Three degrees of freedom and SO(3) . . . . . . . . . . . . 189

17.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 190

18.1 Canonical commutation relations and the bosonic oscillator . . . 191

18.2 Canonical anti-commutation relations and the fermionic oscillator 192

18.3 Multiple degrees of freedom . . . . . . . . . . . . . . . . . . . . . 194

18.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 196

19.1 The Complex Weyl and Clifford algebras . . . . . . . . . . . . . . 197

19.1.1 One degree of freedom, bosonic case . . . . . . . . . . . . 197

19.1.2 One degree of freedom, fermionic case . . . . . . . . . . . 198

19.1.3 Multiple degrees of freedom . . . . . . . . . . . . . . . . . 200

19.2 Real Clifford algebras . . . . . . . . . . . . . . . . . . . . . . . . 201

19.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 203

20.1 Non-degenerate bilinear forms . . . . . . . . . . . . . . . . . . . . 205

20.2 Clifford algebras and geometry . . . . . . . . . . . . . . . . . . . 207

20.2.1 Rotations as iterated orthogonal reflections . . . . . . . . 209

20.2.2 The Lie algebra of the rotation group and quadratic ele-

ments of the Clifford algebra . . . . . . . . . . . . . . . . 210

20.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 212

21.1 The Grassmann algebra of polynomials on anti-commuting gen-

erators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

21.2 Pseudo-classical mechanics . . . . . . . . . . . . . . . . . . . . . . 216

21.3 Lie superalgebras . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

21.4 Examples of pseudo-classical mechanics . . . . . . . . . . . . . . 218

21.4.1 The classical spin degree of freedom . . . . . . . . . . . . 218

21.4.2 The classical fermionic oscillator . . . . . . . . . . . . . . 219

vi

21.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 220

22.1 Quantization of pseudo-classical systems . . . . . . . . . . . . . . 221

22.2 The Schr odinger representation for fermions: ghosts . . . . . . . 223

22.3 Spinors and the Bargmann-Fock construction . . . . . . . . . . . 224

22.4 Parallels between bosonic and fermionic . . . . . . . . . . . . . . 227

22.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 228

23.1 The supersymmetric oscillator . . . . . . . . . . . . . . . . . . . . 229

23.2 Supersymmetric quantum mechanics with a superpotential . . . . 232

23.3 Supersymmetric quantum mechanics and differential forms . . . . 234

23.4 Supersymmetric quantum mechanics and the Dirac operator . . . 235

23.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 236

24.1 Lagrangian mechanics . . . . . . . . . . . . . . . . . . . . . . . . 237

24.2 Path integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

24.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 245

25.1 Multi-particle quantum systems as quanta of a harmonic oscillator248

25.1.1 Bosons and the quantum harmonic oscillator . . . . . . . 248

25.1.2 Fermions and the fermionic oscillator . . . . . . . . . . . . 250

25.2 Solutions to the free particle Schrodinger equation . . . . . . . . 250

25.2.1 Box normalization . . . . . . . . . . . . . . . . . . . . . . 251

25.2.2 Continuum normalization . . . . . . . . . . . . . . . . . . 254

25.3 Quantum field operators . . . . . . . . . . . . . . . . . . . . . . . 255

25.4 Quantization of classical fields . . . . . . . . . . . . . . . . . . . . 260

25.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 261

26.1 Dynamics of the free quantum field . . . . . . . . . . . . . . . . . 263

26.2 Spatial symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . 268

26.3 Internal symmetries . . . . . . . . . . . . . . . . . . . . . . . . . 269

26.3.1 U (1) symmetry . . . . . . . . . . . . . . . . . . . . . . . . 269

26.3.2 U (m) symmetry . . . . . . . . . . . . . . . . . . . . . . . 272

26.4 Fermions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

26.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 274

27.1 Minkowski space . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

27.2 The Lorentz group and its Lie algebra . . . . . . . . . . . . . . . 278

27.3 Spin and the Lorentz group . . . . . . . . . . . . . . . . . . . . . 281

27.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 283

vii

28 Representations of the Lorentz Group 285

28.1 Representations of the Lorentz group . . . . . . . . . . . . . . . . 285

28.2 Dirac matrices and Cliff(1,3) . . . . . . . . . . . . . . . . . . . 290

28.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 294

29.1 Semi-direct products . . . . . . . . . . . . . . . . . . . . . . . . . 295

29.2 Representations of semi-direct products . . . . . . . . . . . . . . 298

29.3 Representations of the Euclidean group E(n) . . . . . . . . . . . 300

29.4 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 302

30.1 The Poincare group and its Lie algebra . . . . . . . . . . . . . . . 303

30.2 Representations of the Poincare group . . . . . . . . . . . . . . . 305

30.2.1 Positive energy time-like orbits . . . . . . . . . . . . . . . 307

30.2.2 Negative energy time-like orbits . . . . . . . . . . . . . . . 308

30.2.3 Space-like orbits . . . . . . . . . . . . . . . . . . . . . . . 308

30.2.4 The zero orbit . . . . . . . . . . . . . . . . . . . . . . . . 308

30.2.5 Positive energy null orbits . . . . . . . . . . . . . . . . . . 309

30.2.6 Negative energy null orbits . . . . . . . . . . . . . . . . . 309

30.3 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 310

31.1 The Klein-Gordon equation and its solutions . . . . . . . . . . . 312

31.2 Classical relativistic scalar field theory . . . . . . . . . . . . . . . 315

31.3 Quantization of the real scalar field . . . . . . . . . . . . . . . . . 318

31.4 The propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

31.5 For further reading . . . . . . . . . . . . . . . . . . . . . . . . . . 322

Field 325

A Conventions 335

viii

B Problems 339

B.1 Problem Set 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339

B.2 Problem Set 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340

B.3 Problem Set 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341

B.4 Problem Set 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

B.5 Problem Set 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

B.6 Problem Set 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

B.7 Problem Set 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

B.8 Problem Set 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

B.9 Problem Set 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347

ix

x

Preface

These are the course notes prepared for a class taught at Columbia during the

2012-13 academic year. The intent was to cover the basics of quantum me-

chanics, from a point of view emphasizing the role of unitary representations of

Lie groups in the foundations of the subject. The approach to this material is

simultaneously rather advanced, using crucially some fundamental mathemati-

cal structures normally only discussed in graduate mathematics courses, while

at the same time trying to do this in as elementary terms as possible. The

Lie groups needed are relatively simple ones that can be described purely in

terms of small matrices. Much of the representation theory will also just use

standard manipulations of such matrices. The only prerequisite for the course

was linear algebra and multi-variable calculus. My hope is that this level of

presentation will simultaneously be useful to mathematics students trying to

learn something about both quantum mechanics and representation theory, as

well as to physics students who already have seen some quantum mechanics,

but would like to know more about the mathematics underlying the subject,

especially that relevant to exploiting symmetry principles.

This document will continue to be updated as time permits with errors fixed,

notation improved, minor additions of material, and whatever other improve-

ments I can think of, with the latest version available at

http://www.math.columbia.edu/~woit/QM/qmbook.pdf

Corrections, comments and suggestions about this are warmly encouraged,

with the best way to contact me email to woit@math.columbia.edu

xi

xii

Chapter 1

1.1 Introduction

A famous quote from Richard Feynman goes I think it is safe to say that no

one understands quantum mechanics.[16]. In this course well try to do the

best we can to come to grips with the subject, emphasizing its deep connections

to fundamental ideas and powerful techniques of modern mathematics. The

strangeness inherent in quantum theory that Feynman was referring to has two

rather different sources. One of them is the inherent disjunction and incom-

mensurability between the conceptual framework of the classical physics which

governs our everyday experience of the physical world, and the very different

framework which governs physical reality at the atomic scale. Familiarity with

the powerful formalisms of classical mechanics and electromagnetism provides

deep understanding of the world at the distance scales familiar to us. Supple-

menting these with the more modern (but still classical in the sense of not

quantum) subjects of special and general relativity extend our understanding

into other less accessible regimes.

Read in context though, Feynman was pointing to a second source of diffi-

culty, contrasting the mathematical formalism of quantum mechanics with that

of the theory of general relativity, a supposedly equally hard to understand

subject. General relativity can be a difficult subject to master, but its math-

ematical and conceptual structure involves a fairly straight-forward extension

of structures that characterize 19th century physics. The fundamental physical

laws (Einsteins equations for general relativity) are expressed as partial differ-

ential equations, a familiar if difficult mathematical subject. The state of the

system is determined by the set of fields satisfying these equations, and observ-

able quantities are functionals of these fields. The mathematics is just that of

the usual calculus: differential equations and their real-valued solutions.

In quantum mechanics, the state of a system is best thought of as a different

sort of mathematical object: a vector in a complex vector space, the so-called

state space. One can sometimes interpret this vector as a function, the wave-

1

function, although this comes with the non-classical feature that wave-functions

are complex-valued. Whats truly completely different is the treatment of ob-

servable quantities, which correspond to self-adjoint linear operators on the state

space. This has no parallel in classical physics, and violates our intuitions about

how physics should work.

During the earliest days of quantum mechanics, the mathematician Hermann

Weyl quickly recognized that the mathematical structures being used were ones

he was quite familiar with from his work in the field of representation theory.

From the point of view that takes representation theory as a fundamental struc-

ture, the framework of quantum mechanics looks perfectly natural. Weyl soon

wrote a book expounding such ideas[52], one which got a mixed reaction from

physicists unhappy with the penetration of unfamiliar mathematical structures

into their subject (with some of them referring to it as the Gruppenpest, the

group theory plague). One goal of this course will be to try and make some of

this mathematics as accessible as possible, boiling down Weyls exposition to its

essentials while updating it in the light of many decades of progress and better

understanding of the subject.

Weyls insight that quantum mechanics crucially involves understanding the

symmetry groups of a physical system and the unitary representations of these

groups has been vindicated by later developments which dramatically expanded

the scope of these ideas. The use of representation theory to exploit the sym-

metries of a problem has become a powerful tool that has found uses in many

areas of science, not just quantum mechanics. From this course I hope that

students whose main interest is physics will learn to appreciate the mathemat-

ical structures that lie behind the calculations of standard textbooks, helping

them understand how to effectively exploit them in other contexts. Students

whose main interest is mathematics will hopefully gain some understanding of

fundamental physics, at the same time as seeing some crucial examples of groups

and representations. These should provide a good grounding for appreciating

more abstract presentations of the subject that are part of the standard mathe-

matical curriculum. Anyone curious about the relation of fundamental physics

to mathematics, and what Eugene Wigner described as The Unreasonable Ef-

fectiveness of Mathematics in the Natural Sciences[53] should benefit from an

exposure to this remarkable story at the intersection of the two subjects.

The following sections give an overview of the fundamental ideas behind

much of the material well be covering in the class. They may initially be rather

mystifying, but should become much clearer as we work through basic examples

this semester.

In classical physics, the state of a system is given by a point in a phase space

(the space of coordinates and momenta), and observable quantities are just

functions on this space (i.e. functions of the coordinates and momenta). The

basic structure of quantum mechanics is completely different, with the formalism

2

built on the following simple axioms:

Axiom (States). The state of a quantum mechanical system is given by a vector

in a complex vector space H with Hermitian inner product < , >.

Well review in an early class some linear algebra, including the properties

of inner products on complex vector spaces. H may be finite or infinite dimen-

sional, with further restrictions required in the infinite-dimensional case (e.g.

we may want to require H to be a Hilbert space). Note two very important

differences with classical mechanical states:

The state space is always linear: a linear combination of states is also a

state.

The state space is a complex vector space: these linear combinations can

and do crucially involve complex numbers, in an inescapable way. In the

classical case only real numbers appear, with complex numbers used only

as an inessential calculational tool.

In this course we will often use the notation introduced by Dirac for vectors

in the state space H: such a vector with a label is denoted

|i

given by self-adjoint linear operators on H.

Well also review the notion of self-adjointness in our review of linear algebra.

When H is infinite-dimensional, further restrictions will be needed on the class

of linear operators to be used.

Axiom (Dynamics). There is a distinguished observable, the Hamiltonian H.

Time evolution of states |(t)i H is given by the Schr

odinger equation

d i

|(t)i = H|(t)i

dt ~

The Hamiltonian observable H will have a physical interpretation in terms

of energy, and one may also want to specify some sort of positivity property on

H, in order to assure the existence of a stable lowest energy state.

~ is a dimensional constant, the value of which depends on what units you

use for time and for energy. It has the dimensions [energy] [time] and its

experimental values are

ing through a one-Volt electric potential). The most natural units to use for

quantum mechanical problems would be energy and time units chosen so that

~ = 1. For instance you could use seconds and measure energies in the very

3

small units of 6.6 1016 eV or energies in eV and time in the very small units

of 6.6 1016 seconds. Schrodingers equation implies that if you are looking

at a system where the typical energy scale is an eV, your state-vector will be

changing on the very short time scale of 6.6 1016 seconds. When we do

computations, often we will just set ~ = 1, implicitly going to a unit system

natural for quantum mechanics. When we get our final result, we can insert

appropriate factors of ~ to allow one to get answers in more conventional unit

systems.

It is sometimes convenient however to carry along factors of ~, since this

can help make clear which terms correspond to classical physics behavior, and

which ones are purely quantum mechanical in nature. We will see that typically

classical physics comes about in the limit where

~

is large. This is true for the energy and time scales encountered in everyday

life, but it can also always be achieved by taking ~ 0, and this is what will

often be referred to as the classical limit.

The above axioms characterize the mathematical structure of a quantum

theory, but they dont address the measurement problem. This is the ques-

tion of how to apply this structure to a physical system interacting with some

sort of macroscopic, human-scale experimental apparatus that measures what

is going on. Since a macroscopic apparatus will involve an exponentially large

number of degrees of freedom, this question is extremely hard to analyze purely

within the quantum mechanical framework. Instead one generally assumes

something like the following, which works very well in practice, and allows one

to make precise statistical predictions using quantum theory

Principle (Measurements).

defined number are the states that are eigenvectors for the corresponding

self-adjoint operator. The value of the observable in such a state will be a

real number, the eigenvalue of the operator.

O with eigenvalues 1 and 2 (i.e. O|1 i = 1 |1 i and O|2 i = 2 |2 i)

, the complex linear combination state

c1 |1 i + c2 |2 i

may not have a well-defined value for the observable O. If one attempts

to measure this observable, one will get either 1 or 2 , with probabilities

|c21 |

|c21 | + |c22 |

4

and

|c22 |

|c21 | + |c22 |

respectively.

This principle is sometimes raised to the level of an axiom of the theory,

but it is better to consider it as a phenomenological over-simplified description

of what happens in typical experimental set-ups. A full understanding of how

classical behavior emerges in experiments is a very challenging topic, with the

notion of decoherence playing an important role. See the end of this chapter

for some references useful in pursuing this question.

Note that the state c|i will have the same eigenvalues and probabilities as

the state |i, for any complex number c. So states are only really determined

up to multiplication by a complex number. It is conventional to work with

states of norm fixed to the value 1, which fixes the amplitude of c, but leaving

a remaining ambiguity which is a phase ei .

The mathematical framework of quantum mechanics is closely related to what

mathematicians describe as a Unitary group representation. We will be ex-

amining this notion in great detail and working through many examples, but

heres a quick summary of the relevant definitions, and an indication of the

relationship to the formalism of quantum theory.

Definition (Group). A group G is a set with an associative multiplication, such

that the set contains an identity element, as well as the multiplicative inverse

of each element.

Many different kinds of groups are of interest in mathematics, an example of

the sort that we will be interested in would be the group of all rotations about

a point in 3-dimensional space. Most of the groups we will consider are matrix

groups, i.e. subgroups of a group of n by n invertible matrices (with real or

complex coefficients).

Definition (Representation). A (complex) representation (, V ) of a group G

is a homomorphism

: g G (g) GL(V )

where GL(V ) is the group of invertible linear maps V V , with V a complex

vector space.

Saying the map is a homomorphism means

(g1 )(g2 ) = (g1 g2 )

for all g1 , g2 G. When V is finite dimensional and we have chosen a basis of

V , then we have an identification of linear maps and matrices

GL(V ) ' GL(n, C)

5

where GL(n, C) is the group of invertible n by n complex matrices. For the

first part of this course we will just be studying representations that are finite

dimensional and will try to make rigorous statements. Later on we will get to

representations on function spaces, which are infinite dimensional, but from then

on will be rather cavalier about the serious analytical difficulties that arise when

one tries to make mathematically precise statements in the infinite-dimensional

case.

To make matters confusing, representations (, V ) are sometimes referred to

by the map , leaving implicit the vector space V that the matrices (g) act on,

but at other times referred to by specifying the vector space V , leaving implicit

the map . One reason for this is that the map may be the identity map:

often G is a matrix group, so a subgroup of GL(n, C), acting on V ' Cn by the

standard action of matrices on vectors. One should keep in mind though that

just specifying V is generally not enough to specify the representation, since it

may not be the standard one. For example, it could very well carry the trivial

representation, where

(g) = 1n

i.e. each element of G acts on V as the identity.

In mathematics, it turns out that the most interesting classes of complex

representations are unitary, i.e. preserving the standard notion of length in

a complex vector space. Similarly in physics, the group representations we will

be interested in are those corresponding to physical symmetries, so preserving

lengths in H, since these correspond to probabilities of various observations. We

have the definition

Definition (Unitary representation). A representation (, V ) on a complex vec-

tor space V with Hermitian inner product < , > is a unitary representation if

it preserves the inner product, i.e.

For a unitary representation, the matrices (g) take values in a subgroup

U (n) GL(n, C). In our review of linear algebra we will see that U (n) can be

characterized as the group of n by n complex matrices U such that

U 1 = U

to mean the adjoint or conjugate-transpose matrix. This notation is pretty

universal in physics, whereas mathematicians prefer to use instead of .

The fundamental relationship between quantum mechanics and representation

theory is that whenever we have a physical quantum system with a group G of

6

symmetries acting on it, the space of states H will carry a unitary representation

of G. For physicists working with quantum mechanics, this implies that repre-

sentation theory provides information about quantum mechanical state spaces .

For mathematicians studying representation theory, this means that physics is

a very fruitful source of unitary representations to study: any physical system

with a symmetry group G will provide one.

For a representation and group elements g that are close to the identity,

one can use exponentiation to write (g) GL(n, C) as

(g) = eA

We will study this situation in much more detail and work extensively with

examples, showing in particular that if (g) is unitary (i.e. in the subgroup

U (n) GL(n, C)) , then A will be skew-adjoint:

A = A

defining A = iB, we find that B is self-adjoint

B = B

representation of G on H coming from a symmetry G of our physical system

gives us not just unitary matrices, but also corresponding self-adjoint operators

B on H. Symmetries give us quantum mechanical observables, with the fact that

these are self-adjoint linear operators corresponding to the fact that symmetries

are realized as unitary representations on the state space.

In this course well see many examples of this phenomenon. A fundamental

example that we will study in detail is that of time-translation symmetry. Here

the group G = R and we get a unitary representation of R on the space of states

H. The corresponding self-adjoint operator is the Hamiltonian operator H. This

unitary representation gives the dynamics of the theory, with the Schrodinger

equation just the statement that iHt is the skew-adjoint operator that gets

exponentiated to give the unitary transformation that moves states (t) ahead

in time by an amount t.

tion spaces

It is conventional to refer to the groups that appear in this subject as symmetry

groups, which emphasizes the phenomenon of invariance of properties of objects

under sets of transformations that form a group. This is a bit misleading though,

since we are interested in not just invariance, but the more general phenomenon

of groups acting on sets, according to the following definition:

7

Definition (Group action on a set). An action of a group G on a set M is

given by a map

(g, x) G M g x M

such that

g1 (g2 x) = (g1 g2 ) x

and

ex=x

where e is the identity element of G

R3 with the standard inner product. This comes with an action of the group

G1 = R3 on X = M by translations, and of the group G2 = O(3) of 3-

dimensional orthogonal transformations (rotations about the origin). Note that

order matters: we will often be interested in non-commutative groups where

g1 g2 6= g2 g1 .

A fundamental principle of modern mathematics is that the way to under-

stand a space X, given as some set of points, is to look at F un(M ), the set of

functions on this space. This linearizes the problem, since the function space

is a vector space, no matter what the geometrical structure of the original set

is. If our original set is finite dimensional, the function space will be a finite

dimensional vector space. In general though it will be infinite dimensional and

we will need to further specify the space of functions (i.e. continuous functions,

differentiable functions, functions with finite integral, etc.).

Given a group action of G on M , taking complex functions on M provides

a representation (, F un(M )) of G, with defined on functions f by

((g)f )(x) = f (g 1 x)

Note the inverse that is needed to get the group homomorphism property to

work since one has

= f (g21 (g11 x))

= f ((g21 g11 ) x)

= f ((g1 g2 )1 x)

= (g1 g2 )f (x)

a calculation that would not work out properly for non-commutative G if one

defined ((g)f )(x) = f (g x).

One way to construct quantum mechanical state space H is as wave-functions,

meaning complex valued functions on space-time. The above shows that given

any group action on space-time, we get a representation on the state space H

of such wave-functions.

8

Note that only in the case of M a finite set of points will we get a finite-

dimensional representation this way, since only then will F un(M ) be a finite-

dimensional vector space (C# of points in M ). A good example to consider to

understand this construction is the following:

f F un(M ), f is a vector in C3 , with components (f (x1 ), f (x2 ), f (x3 )).

3! = 6 elements.

(g, xi ) g xi

F un(M ) as above

((g)f )(xi ) = f (g 1 xi )

This construction gives 6 three by three complex matrices, which under multipli-

cation of matrices satisfy the same relations as the elements of the group under

group multiplication. In this very special case, all the entries of the matrix will

be 0 or 1, but that is special to the permutation representation.

The discussion here has been just a quick sketch of some of the ideas behind

the material we will cover later in this course. These ideas will be covered

in much greater detail as the course unfolds, beginning with the next couple

lectures where they will appear very concretely when we discuss the simplest

possible quantum systems, those with one and two-complex dimensional state

spaces.

We are approaching the subject from a different direction than the conventional

one, starting with the role of symmetry and with the simplest possible finite-

dimensional quantum systems, systems which are purely quantum mechanical,

with no classical analog. This means that the early discussion youll find in

most textbooks is rather different than the one here. You may want to look

though at any one of the textbooks listed in the course web-page, and see how

they handle the description of the fundamental principles of quantum mechan-

ics. They will include the same structure described here, but typically begin

with the theory of motion of a quantized particle, trying to motivate it from

classical mechanics. The state space is then a space of wave-functions, which is

infinite-dimensional and necessarily brings some analytical difficulties. Quan-

tum mechanics is inherently a quite different conceptual structure than classical

mechanics. The relationship of the two subjects is rather complicated, but it is

9

clear that quantum mechanics cannot be derived from classical mechanics, so

attempts to motivate it that way are of necessity rather unclear, although they

correspond to the very interesting historical story of how the subject evolved.

As we get further along in the course, we will come to the topic of the quantized

motion of a particle, at which point it will become much easier to follow the

standard books.

A good standard quantum mechanics textbook is [40]. Three useful text-

books on the subject aimed at mathematicians are [45], [12], and [27], as well as

the forthcoming [26]. The first few chapters of [19] provide an excellent while

very concise summary of both basic physics and quantum mechanics. One im-

portant topic we wont discuss is that of the application of the representation

theory of finite groups in quantum mechanics. For this as well as a discussion

that overlaps quite a bit with the point of view of this course while emphasizing

different areas, see [42].

The field of Quantum Information Theory gives a perspective on quantum

theory that puts the elementary two-state quantum system (the qubit) front

and center. You might want to try reading the first part of Chapter 2 of John

Preskills notes on quantum computation[34] which give similar axioms to ours,

and starts off with the same example we will soon be considering.

On the whole, in this class well be avoiding the difficult issue of how mea-

surements work and how classical physics emerges from quantum theory. Some

suggested references for these subjects are Wojciech Zureks updated version

of his 1991 Physics Today article[58] and his more recent work on quantum

darwinism[59], as well as books by Schlosshauer[38] and Haroche/Raimond[28].

10

Chapter 2

electromagnetic field is given by the electric charge q of the object, and q can

take on any real value. Quantum states instead carry an electric charge that is

quantized: it can be labeled by an integer rather than a real number. We will

see that these integers come about because the complex vector space of states

H is a representation of the symmetry group U (1) of complex phase transfor-

mations. This is a purely quantum phenomenon, with no classical analog: the

representation does not come from a group action on anything known to clas-

sical physics. When integers like this labeling U (1) representations occur in

mathematics, they are generally called weights. In physics, they are often

but not always called charges.

Besides the U (1) symmetry group that corresponds to electric charge, there

are other physically relevant U (1) symmetry groups. In particle physics for

instance, such a symmetry group is responsible for the existence of an integer

baryon number. We will see later in this course that the fact that the energy

levels of a quantum harmonic oscillator are quantized comes about because in

such a situation there is a U (1) symmetry, with the Hamiltonian itself the analog

of the charge observable. In a very real sense, quantum mechanics is quan-

tum because of the importance of U (1) symmetries and their corresponding

integer-quantized observables.

A caveat to keep in mind: at this point in the course we are still far from

being able to discuss actual electric charge, which involves much more structure,

including specifically the electromagnetic field. Much later in this course we will

get to this, and only then see the quantization of electric charge as a specific

example of an implication of a U (1) symmetry.

Recall the definition given last time of a group representation

Definition (Representation). A (complex) representation (, V ) of a group G

11

on a complex vector space V (with a chosen basis identifying V ' Cn ) is a

homomorphism

: G GL(n, C)

This is basically a set of n by n matrices, one for each group element, satis-

fying the multiplication rules of the group elements. n is called the dimension of

the representation. We will be mainly interested in continuous groups with an

infinite number of elements, but the subject of representations of finite groups

is a very rich one, with a great deal of interesting structure.

The groups G we are interested in will be examples of what mathematicians

call Lie groups. If you are familiar with differential geometry, such groups are

examples of smooth manifolds, and one can define derivatives of functions on G

and more generally of maps from G into spaces like GL(n, C). We will assume

that our representations are given by differentiable maps . Some difficult gen-

eral theory shows that considering the more general case of continuous maps

gives nothing new since the homomorphism property of these maps is highly

constraining. In any case, our goal in this course will be to study quite explic-

itly certain specific groups and representations which are central in quantum

mechanics, and these representations will always be differentiable.

Given two representations one can form their direct sum

Definition (Direct sum representation). Given representations 1 and 2 of

dimensions n1 and n2 , one can define another representation, of dimension

n1 + n2 called the direct sum of the two representations, denoted by 1 2 .

This representation is given by the homomorphism

1 (g) 0

(1 2 ) : g G

0 2 (g)

In other words, one just takes as representation matrices block-diagonal

matrices with 1 and 2 giving the blocks.

To understand the representations of a group G, one proceeds by first iden-

tifying the irreducible ones, those that cannot be decomposed into two repre-

sentations of lower dimension.

Definition (Irreducible representation). A representation is called irreducible

if it cannot be put in the form 1 2 , for 1 and 2 of dimension greater than

zero.

This criterion is not so easy to check, and the decomposition of an arbitrary

reducible representation into irreducible components can be a very non-trivial

problem. Recall that one gets explicit matrices for the (g) of a representation

(, V ) only when a basis for V is chosen. To see if the representation is reducible,

one needs to see not whether the (g) are all in block-diagonal form, but whether

there is some basis for V with respect to which they are all in such form,

something very non-obvious from just looking at the matrices themselves.

Digression. Another approach to this would be to check to see if the repre-

sentation has no proper non-trivial sub-representations (such a representation

12

is often called indecomposable). This is not necessarily equivalent to irre-

ducibility, since a sub-representation may have no complement that is also a

sub-representation. A simple example of this occurs for the action of upper

triangular matrices on column vectors. Such representations are however non-

unitary. In the unitary case indecomposability and irreducibility are equivalent.

In these notes unless otherwise specified, one should assume that all represen-

tations are unitary, so, in particular, this issue will not arise.

is irreducible or not:

the only linear maps M : V V commuting with all the (g) are the scalar

matrices 1.

Proof. Since we are working over the algebraically closed field C, we can solve

the eigenvalue equation

det(M 1) = 0

to find the eigenvalues of M . The eigenspaces

V = {v V : M v = v}

are vector subspaces of V and can also be described as ker(M 1), the kernel

of the operator M 1. Since and this operator and all the (g) commute, we

have

v ker(M 1) = (g)v ker(M 1)

so ker(M 1) V is a representation of G. If V is irreducible, we must

have either ker(M 1) = V or ker(M 1) = 0. Since is an eigenvalue,

ker(M 1) 6= 0, so M = 1 as a linear operator on V .

matrix M commutes with all the representation matrices (g), Schurs lemma

says that M must be a scalar multiple of the unit matrix.

Note that the proof crucially uses the fact that one can solve the eigenvalue

equation. This will only be true in general if one works with C and thus with

complex representations. For the theory of representations on real vector spaces,

Schurs lemma is no longer true.

An important corrollary of Schurs lemma is the following characterization

of irreducible representations of G when G is commutative.

dimensional.

(g)(h) = (h)(g)

13

for all h G. If is irreducible, Schurs lemma implies that, since they commute

with all the (g), the matrices (h) are all scalar matrices, i.e. (h) = h 1 for

some h C. is then irreducible exactly when it is the one-dimensional

representation given by (h) = h .

One can think of the group U (1) as the unit circle, with the multiplication rule

on its points given by addition of angles. More explicitly:

Definition (The group U (1)). The elements of the group U (1) are points on

the unit circle, which can be labeled by the unit complex number ei , for

R. Note that and + N 2 label the same group element. Multiplication of

group elements is just complex multiplication, which by the properties of the

exponential satisfies

ei1 ei2 = ei(1 +2 )

By our theorem from the last section, since U (1) is a commutative group,

all irreducible representations will be one-dimensional. Such an irreducible rep-

resentation will be given by a map

: U (1) GL(1, C)

will denote the group of these as C . Such a map must satisfy homomorphism

and periodicity properties which can be used to show:

Theorem. All irreducible representations of the group U (1) are unitary, and

given by

for k Z.

k (1 + 2 ) = k (1 )k (2 )

k (2) = k (0) = 1

f : U (1) C

14

satisfying the homomorphism and periodicity properties is of this form. Com-

df

puting the derivative f 0 () = d we find

f ( + ) f ()

f 0 () = lim

0

(f () 1)

= f () lim (using the homomorphism property)

0

= f ()f 0 (0)

Denoting the constant f 0 (0) by C, the only solutions to this differential equation

satisfying f (0) = 1 are

f () = eC

Requiring periodicity we find

The representations we have found are all unitary, with k taking values not

just in C , but in U (1) C . You can check that the complex numbers eik

satisfy the condition to be a unitary 1 by 1 matrix, since

These representations are restrictions to the unit circle U (1) of the irre-

ducible representations of the group C , which are given by

k : z C k (z) = z k C

Such representations are not unitary, but they have an extremely simple form,

so it sometimes is convenient to work with them, later restricting to the unit

circle, where the representation is unitary.

Digression (Fourier analysis of periodic functions). Well later on discuss

Fourier analysis more seriously when we come to the case of the translation

groups and of state-spaces that are spaces of wave-functions on space-time.

For now though, it might be worth pointing out an important example of a rep-

resentation of U (1), the space F un(S 1 ) of complex-valued functions on the circle

S 1 . We will evade completely for now the very non-trivial analysis involved here,

by not specifying what class of functions we are talking about (e.g. continuous,

integrable, differentiable, etc.). Periodic functions can be studied by rescaling

the period to 2, thus looking at complex-valued functions of a real variable

satisfying

f ( + N 2) = f ()

for integer N , which we can think of as functions on a circle, parametrized by

angle . We have an action of the group U (1) on the circle by rotation:

(, ) U (1) S 1 + S 1

15

Last time, we saw that given an action of a group on a space X, we can lin-

earize and get a representation (, F un(X)) of the group on the functions on

the space, by taking

((g)f )(x) = f (g 1 x)

for f F un(X), x X. Here X = S 1 , the action is the rotation action and we

find

(()f )() = f ( )

since the inverse of a rotation by is a rotation by .

This representation (, F un(S 1 )) is infinite-dimensional, but one can still

ask how it decomposes into the one-dimensional irreducible representations (k , C)

of U (1). What we learn from the subject of Fourier analysis is that each (k , C)

occurs exactly once in the decomposition of F un(S 1 ) into irreducibles, i.e.

M

(, F un(S 1 )) = (k , C)

kZ

where we have matched the sin of not specifying the class of functions in F un(S 1 )

on the left-hand side with the sin of not explaining how to handle the infinite

direct sum on the right-hand side. What can be specified precisely is how

L

the irreducible sub-representation (k , C) sits inside F un(S 1 ). It is the set of

functions f satisfying

tional to eik .

One part of the relevance of representation theory to Fourier analysis is

that the representation theory of U (1) picks out a distinguished basis of the

infinite-dimensional space of periodic functions by using the decomposition of

the function space into irreducible representations. One can then effectively

study functions by expanding them in terms of their components in this special

basis, writing an f S 1 as

X

f () = ck eik

kZ

questions about the convergence of this series.

Recall from our overview last time the general principle that since the state space

H is a unitary representation of a Lie group, we get an associated self-adjoint

operator on H. For the case of G = U (1), this operator is just the operator

that acts by multiplication by the integer q on the representation space C of

the irreducible representation (q , C). Since the irreducible representations of

16

G = U (1) are all one-dimensional, this means that as a U (1) representation, we

have

H = Hq1 Hq2 Hqn

be distinct). We will call this operator the charge operator:

acts by multiplication by qi on the irreducible representation Hqi . It acts on H

as the matrix

q1 0 0

0 q2 0

0 0 qn

subspaces Hqi will have a well-defined numerical value for this observable, the

integer qi . A general state will be a linear superposition of state vectors from

different Hqi and will not have such a well-defined numerical value for the ob-

servable Q on such a state.

From the action of Q on H, one can recover the representation, i.e. the action

of the symmetry group U (1) on H, by multiplying by i and exponentiating, to

get

iq

e 1 0 0

0 eiq2 0

() = eiQ =

U (n) GL(n, C)

0 0 eiqn

transformations.

The general abstract high-powered mathematical point of view is that the

representation is a map between manifolds, from the Lie group U (1) to the

Lie group GL(n, C) that takes the identity of U (1) to the identity of GL(n, C).

As such it has a differential, 0 which is a map from the tangent space at

the identity of U (1) (which here is iR) to the tangent space at the identity

of GL(n, C), which is the space M (n, C), the n by n complex matrices. The

tangent space at the identity of a Lie group is called a Lie algebra, and we

will later study not the general theory of these, but the examples that occur for

the specific Lie groups that will appear in this course.

Here the relation between the differential of and the operator Q is

0 : i iR 0 (i) = iQ

17

The right-hand side of the picture is supposed to somehow represent GL(n, C),

which is the 2n2 dimensional space of n by n complex matrices, minus the locus

of matrices with zero determinant, that cant be inverted. It has a distinguished

point, the identity. The charge Q map is the differential d of at the identity

In this very simple example, this abstract picture is over-kill and likely con-

fusing. We will see the same picture though occurring in many other examples

as the course goes on, ones where the abstract technology is increasingly useful.

Mostly though, keep in mind that these maps will be just explicit exponentia-

tion operators in the examples we care about, so have very concrete incarnations

as matrices we can do explicit calculations with.

The way we have defined observable operators in terms of the action of symmetry

groups on H, the action of these operators has nothing to do with the dynamics,

and if we start at time t = 0 in a state in Hqi , with definite numerical value qi

for the observable, there is no reason that time evolution should preserve this.

Recall from one of our basic axioms that time evolution of states is given by the

Schrodinger equation

d i

|(t)i = H|(t)i

dt ~

We will later more carefully study the relation of this equation to the symmetry

of time translation (basically the Hamiltonian operator H generates an action of

the group R of time translations, just as the operator Q generates an action of

18

the group U (1)). For now though, note that for time-independent Hamiltonian

operators H, the solution to this equation is given by exponentiating H, with

|(s)i = U (s)|(0)i

for

i

U (s) = e ~ sH

If the Hamiltonian operator H and the charge operator Q commute

[H, Q] = HQ QH = 0

[H k , Q] = 0

[U (s), Q] = 0

This condition

U (s)Q = QU (s)

implies that if a state has a well-defined value qi for the observable Q at time

t = 0, it will continue to have the same value at time t = s, since

tonian observable, we get a conservation law. This conservation law says that

if we start in a state with a well-defined numerical value for the observable, we

will remain in such a state, with the value not changing, i.e. conserved.

Ive had trouble finding another source that covers the material here. Most

quantum mechanics books consider it somehow too trivial to mention, starting

their discussion of symmetries with more complicated examples. Id be curious

to hear about it if you find a good example of a discussion of these topics at

this level elsewhere.

19

20

Chapter 3

Spin 1/2

The simplest truly non-trivial quantum systems have state spaces that are in-

herently two-complex dimensional, and have significantly more structure than

the charge structure of the previous section, which could be analyzed by break-

ing up the space of states into one-dimensional subspaces of given charge. Well

study these in this section, encountering for the first time the implications of

working with representations of non-commutative groups. Since they give the

simplest non-trivial realization of quantum phenomena, such systems are the

fundamental objects of quantum information theory (the qubit) and the fo-

cus of attempts to build a quantum computer (which would be built out of

copies of this sort of fundamental object). Many different possible two-state

quantum systems could potentially be used as the physical implementation of a

qubit.

One of the simplest possibilities to take would be the idealized situation

of a single electron, somehow fixed so that its spatial motion could be ignored,

leaving its quantum state described just by its so-called spin degree of freedom,

which takes values in H = C2 . The term spin is supposed to call to mind

the angular momentum of an object spinning about about some axis, but such

classical physics has nothing to do with the qubit, it is a purely quantum system.

We will see later that quantum systems can be defined with spin quantum

number N/2, for

N = 0, 1, 2, 3, . . .

with dimension of the state space given by N + 1. The N = 0 case is the trivial

case, and the case at hand is N = 1, the spin 1/2 case. In the limit N

one can make contact with classical notions of spinning objects and angular

momentum, but the spin 1/2 case is at the other limit, where the behavior is

purely quantum.

21

3.1 The two-state quantum system

3.1.1 The Pauli matrices: observables of the two-state

quantum system

For a quantum system with two-dimensional state space, observables are self-

adjoint linear operators on C2 . With respect to a chosen basis of C2 , these are 2

by 2 complex matrices M satisfying the condition M = M (M is the conjugate

transpose matrix). Any such matrix will be a (real) linear combination of four

matrices:

M = c0 1 + c1 1 + c2 2 + c3 3

with ca R and the standard choice of the basis elements given by

1 0 0 1 0 i 1 0

1= , 1 = , 2 = , 3 =

0 1 1 0 i 0 0 1

The a are called the Pauli matrices and are a pretty universal choice of basis

in this subject. They reflect an arbitary convention, in particular the convention

of diagonalizing the basis element in the 3 or z-direction of R3 . One could just

as easily have decided to choose a basis diagonal in some other direction.

Recall that the basic principle of how measurements are supposed to work

in quantum theory says that the only states that have well-defined values for

these four observables are the eigenvectors for these matrices. The first matrix

gives a trivial observable, whereas the last one, 3 , has the two eigenvectors

1 1

3 =

0 0

and

0 0

3 =

1 1

with eigenvalues +1 and 1. In quantum information theory, where this is

the qubit system, these two eigenstates are labeled |0i and |1i because of the

analogy with a classical bit of information. Later on when we get to the theory of

spin, we will see that 21 3 is the observable corresponding to the SO(2) = U (1)

symmetry group of rotations about the third spatial axis, and the eigenvalues

12 , + 12 of this operator will be used to label the two eigenstates

1 1 1 0

|+ i= and | i =

2 0 2 1

A crucial and easy to check fact is that |+ 21 i and | 12 i are NOT eigenvectors

for the operators 1 and 2 . An exercise in the first problem set is to compute

the eigenvectors and eigenvalues for these operators. It can be checked that no

pair of the three a commute, so one cannot find vectors that are simultaneous

eigenvectors for more than one a . This non-commutativity of the operators

is responsible for the characteristic classically paradoxical property of quantum

22

observables: one can find states with a well defined number for the measured

value of one observable a , but such states will not have a well-defined num-

ber for the measured value of the other two non-commuting observables. The

physics language for this phenomenon is that if we prepare states with a well-

defined spin component in the a-direction, the two other components of the

spin cant be assigned a numerical value in such a state. Any attempt to

prepare states that simultaneously have specific chosen numerical values for the

3 observables corresponding to the a is doomed. So is any attempt to si-

multaneously measure such values: if you thought you knew the value for one

observable, measuring another would ensure that this first measurement was no

longer valid.

This non-classical behavior of observables takes some getting used to, and

there are many subtleties in the theory of measurement, but the basic phe-

nomenon is captured accurately by the very simple mathematics of the eigenvec-

tors and eigenvalues of the three 2 by 2 matrices a . Many quantum mechanics

textbooks (a good example is Volume III of the Feynman lectures[15]) contain

an extensive discussion of the physical implications of this specific example.

The choice we have made for the a corresponds to a choice of basis such

that the basis vectors are eigenvectors of 3 . 1 and 2 take these basis vectors

to non-trivial linear combinations of basis vectors. It turns out that there are

two specific linear combinations of 1 and 2 that do something very simple to

the basis vectors, since

0 2 0 0

(1 + i2 ) = and (1 i2 ) =

0 0 2 0

we have

0 1 1 0

(1 + i2 ) =2 (1 + i2 ) =

1 0 0 0

and

1 0 0 0

(1 i2 ) =2 (1 i2 ) =

0 1 1 0

(1 + i2 ) is called a raising operator: on eigenvectors of 3 it either

increases the eigenvalue by 2, or annihilates the vector. (1 i2 ) is called

a lowering operator: on eigenvectors of 3 it either decreases the eigenvalue

by 2, or annihilates the vector. Note that these linear combinations are not

self-adjoint, (1 + i2 ) is the adjoint of (1 i2 ) and vice-versa.

Well recall soon in our review of linear algebra that an arbitrary self-adjoint

2 by 2 complex matrix can be diagonalized by a unitary change of basis, i.e.

there is a unitary matrix U such that

1 0

U M U 1 =

0 2

where 1 , 2 R are the eigenvalues of the matrix. If c0 = 0, so we just have a

linear combination of the three Pauli matrices, we must have 2 = 1 , since

the Pauli matrices all have trace zero.

23

3.1.2 Exponentials of Pauli matrices: unitary transforma-

tions of the two-state system

Exponentiating the identity matrix gives the diagonal unitary matrix

i

e 0

ei1 =

0 ei

This matrix commutes with any other 2 by 2 matrix and just acts on our state

space C2 by multiplication by a phase, as studied in the previous section.

If we exponentiate Pauli matrices, it turns out that since all the j satisfy

j2 = 1, their exponentials also take a simple form:

1 1

eij = 1 + ij + (i)2 j2 + (i)3 j3 +

2 3!

1 1

= 1 + ij 2 1 i 3 j +

2 3!

1 1

= (1 2 + )1 + i( 3 + )j

2 3!

= (cos )1 + ij (sin )

the space of unitary 2 by 2 matrices, starting and ending at the unit matrix.

This circle is a group, isomorphic to U (1). So, we have found three different,

non-commuting, U (1) subgroups inside the unitary 2 by 2 matrices.

To exponentiate linear combinations of Pauli matrices, one can check that

these matrices satisfy the following relations, useful in general for doing calcu-

lations with them instead of multiplying out explicitly the 2 by 2 matrices:

{j , k } = j k + k j = 2jk 1

and distinct j anti-commute (e.g. j k = k j for j 6= k).

Notice that the anti-commutation relations imply that, if we take a vector

v = (v1 , v2 , v3 ) R3 and define a 2 by 2 matrix by

v3 v1 iv2

v = v1 1 + v2 2 + v3 3 =

v1 + iv2 v3

(

1 n even

(v )n =

(v ) n odd

24

We can use this to compute the exponential of the matrix iv ( R, v

a unit vector) as follows, using the Taylor series formula for the exponential:

(iv )2 (iv )3

eiv = 1 + iv + + +

2! 3!

2 3

= (1 + )1 + i( + )v

2 3!

= (cos )1 + i(sin )v

Notice that you can easily compute the inverse of this matrix:

since

Well review linear algebra and the notion of a unitary matrix in one of the next

classes, but one form of the condition for a matrix M to be unitary is

M = M 1

= ((cos )1 i(sin )v )

= ((cos )1 i(sin )v )

= (eiv )1

cos + i sin v3 i sin (v1 iv2 )

= det

i sin (v1 + iv2 ) cos i sin v3

= cos2 + sin2 (v12 + v22 + v32 )

=1

adjoint Pauli matrices (which all have trace zero), we get unitary matrices of

determinant one. These are invertible, and form the group named SU (2), the

group of unitary 2 by 2 matrices of determinant one. If we exponentiated not

just iv , but i(1 + v ) for some real constant (which does not have

trace zero unless = 0), we would get a unitary matrix with determinant ei2 .

The group of unitary 2 by 2 matrices with arbitrary determinant is called U (2).

In our review of linear algebra to come we will encounter the groups SU (n) and

U (n), groups of unitary n by n complex matrices.

25

To get some more insight into the structure of the group SU (2), consider an

arbitrary 2 by 2 complex matrix

Unitarity implies that the rows are orthonormal. One can see this explicitly

from the condition that the matrix times its conjugate-transpose is the identity

1 0

=

0 1

+ = 0 = =

The condition that the first row has length one gives

+ = ||2 + ||2 = 1

Using these two relations and computing the determinant (which has to be 1)

gives

= = ( + ) = = 1

so one must have

= , =

and an SU (2) matrix will have the form

where (, ) C2 and

||2 + ||2 = 1

So, the elements of SU (2) are parametrized by two complex numbers, with

the sum of their length-squareds equal to one. Identifying C2 = R4 , these are

just vectors of length one in R4 . Just as U (1) could be identified as a space with

the unit circle S 1 in C = R2 , SU (2) can be identified with the three-sphere S 3

in R4 .

The anti-commutation relations for Pauli matrices show that when we multiply

a linear combination of them by i and exponentiate, we get elements of SU (2),

26

in particular a U (1) subgroup of SU (2). Another important set of relations

satisfied by Pauli matrices are their commutation relations:

3

X

[j , k ] = j k k j = 2i jkl l

l=1

subscripts, and vanishes if two of the subscripts take the same value. More

explicitly, this says:

One can easily check these relations by explicitly computing with the matrices.

Putting together the anticommutation and commutation relations, one gets a

formula for the product of two Pauli matrices:

3

X

j k = jk 1 + i jkl l

l=1

While physicists prefer to work with self-adjoint Pauli matrices and their

real eigenvalues, one can work instead with the following skew-adjoint matrices

j

Xj = i

2

which satisfy the slightly simpler commutation relations

3

X

[Xj , Xk ] = jkl Xl

l=1

or more explicitly

If these commutators were zero, the SU (2) elements one gets by exponentiat-

ing linear combinations of the Xj would be commuting group elements. The

non-triviality of the commutators reflects the non-commutativity of the group.

Group elements U SU (2) near the identity satisfy

U ' 1 + 1 X1 + 2 X2 + 3 X2

for a small and real, just as group elements z U (1) near the identity satisfy

z ' 1 + i

imal version of the full group and its group multiplication law, valid near

the identity. In terms of the geometry of manifolds, recall that SU (2) is the

27

space S 3 . The Xj give a basis of the tangent space R3 to the identity of

SU (2), just as i gives a basis of the tangent space to the identity of U (1).

Recall that the time dependence of states in quantum mechanics is given by the

Schr

odinger equation

d i

|(t)i = H|(t)i

dt ~

where H is a particular self-adjoint linear operator on H, the Hamiltonian op-

erator. The most general such operator on C2 will be given by

H = h0 1 + h1 1 + h2 2 + h3 3

is just given by exponentiation:

|(t)i = U (t)|(0)i

where

itH

U (t) = e ~

ih0 t

The h0 1 term in H just contributes an overall phase factor e ~ , with the

remaining factor of U (t) an element of the group SU (2) rather than the larger

group U (2) of all 2 by 2 unitaries.

Using our earlier equation

28

h

valid for a unit vector v, our U (t) is given by taking h = (h1 , h2 , h3 ), v = |h|

and = t|h|

~ , so we find

U (t) =e ~ ((cos())1 + i(sin( )) )

~ ~ |h|

ih0 t t|h| t|h| h1 1 + h2 2 + h3 3

=e ~ ((cos( ))1 i(sin( )) )

~ ~ |h|

ih0 t !

e ~ (cos( t|h| h3 t|h|

~ ) i |h| sin( ~ )) i sin( t|h| h1 ih2

~ ) |h|

= ih0 t

i sin( t|h| h1 +ih2

~ ) |h| e ~ (cos( t|h| h3 t|h|

~ ) + i |h| sin( ~ ))

t(h0 +h3 )

!

ei ~ 0

U (t) = t(h0 h3 )

0 ei ~

1 1

|(0)i = | + i + | i

2 2

for , C, at later times the state will be

t(h0 +h3 ) 1 t(h0 h3 ) 1

|(t)i = ei ~ | + i + ei ~ | i

2 2

In this special case, one can see that the eigenvalues of the Hamiltonian are

h0 h3 . A bit later on in this class we will study the behavior of the observables

of this system as a function of time, for this Hamiltonian.

In the physical realization of this system by a spin 1/2 particle (ignoring its

spatial motion), the Hamiltonian is given by

ge~

H= (B1 1 + B2 2 + B3 3 )

4mc

where the Ba are the components of the magnetic field, and the physical con-

stants are the gyromagnetic ratio (g), the electric charge (e), the mass (m) and

the speed of light (c), so we have solved the problem of the time evolution of

ge~

such a system, setting ha = 4mc Ba . For magnetic fields of size |B| in the 3-

direction, we see that the two different states with well-defined energy (| + 12 i

and | 12 i) will have an energy difference between them of

ge~

|B|

mc

This is known as the Zeeman effect and is readily visible in the spectra of atoms

subjected to a magnetic field.

29

30

Chapter 4

Unitary and Orthogonal

Groups

A course in linear algebra is a prerequisite for this course, and well need a

range of specific facts from that subject. We will also need some aspects of

linear algebra that might not have made an appearance (at least not explicitly)

in whatever course you may have taken. These include the use of complex

vector spaces, as opposed to the real vector spaces that a typical linear algebra

course concentrates on. We also will consider matrices in their role as elements

of matrix groups, a topic often not emphasized in typical courses. For now our

vector spaces will be finite-dimensional. Later in the course we will come to

state spaces that are infinite dimensional, and will address the various issues

that this raises at that time.

A vector space V over a field k is just a set such that one can consistently take

linear combinations of elements with coefficients in k. We will only be using

the cases k = R and k = C, mostly the latter. To any vector space V one can

associate a new vector space, its dual:

Definition. Given a vector space V over a field k, the dual vector space V is

the set of all linear maps V k, i.e.

for , k, v, w V .

Physicists have a useful notation for elements of vector space and their duals.

31

An element of a vector space V is written as a ket vector

|vi

where v is a label for a vector. An element of the dual vector space V is written

as a bra vector

hl|

Evaluating l V on v V gives an element of k, written

hl|vi

If : V V is a linear map

a notion of length for vectors, of angle between vectors, and identifies V ' V .

One has, in the real case:

Definition (Inner Product, real case). An inner product on a real vector space

V is a map

h, i : V V R

that is linear in both variables and symmetric (hv, wi = hw, vi).

In the complex case, one has

Definition (Inner Product, complex case). An Hermitian inner product on a

complex vector space V is a map

h, i : V V C

that is linear in the second variable, antilinear in the first variable, i.e. for

, C and u, v, w V

hu + v, wi = hu, wi + hv, wi

hv, wi = hw, vi

Our inner products will always be positive-definite (hv, vi 0 and hv, vi =

0 = v = 0), with indefinite inner products only appearing in physics when

one considers an inner product on space-time in the context of special or general

relativity.

An inner product gives a notion of length || || for vectors, with

||v||2 = hv, vi

matter of convention. The choice we are making is universal among physicists,

with the opposite choice common among mathematicians.

32

An inner product provides an isomorphism V ' V by the map

v V lv V

where lv is defined by

lv (w) = hv, wi

In the bra-ket notation, one denotes the dual vector lv by hv|. Note that in

the inner product the angle bracket notation means something different than in

the bra-ket notation. The similarity is intentional though, since in the bra-ket

notation one has

hv|wi = hv, wi

Note that our convention of linearity in the second variable of the inner product,

anti-linearity in the first, implies

for C.

The above discussion of vectors has been in terms of abstract elements of V .

More concretely, one can identify V with Rn or Cn , by choosing a basis, in

particular an orthonormal basis {ei }, i = 1, . . . , n satisfying

hei , ej i = ij

|ii = ei

v = v1 e1 + v2 e2 + + vn en

Digression. In this course, all our indices will be lower indices. One way to

keep straight the difference between vectors and dual vectors is to use upper

indices for components of vectors, lower indices for components of dual vectors.

This is quite useful in Riemannian geometry and general relativity, where the

inner product is given by a metric that can vary from point to point, causing

the isomorphism between vectors and dual vectors to also vary. For quantum

mechanical state spaces, we will be using a single, standard, fixed inner product,

so there will be a single isomorphism between vectors and dual vectors. The

bra-ket notation will take care of the notational distinction between vectors and

dual vectors as necessary.

33

Note that one can also think of the vi as coordinate functions on V . In this

interpretation of the symbol vi , it is a linear function on V , thus an element of

the dual V . Specifically, it is the linear function on V which takes value the

coefficient vi of the basis vector ei on a vector v V .

Because of orthonormality, these coefficients of vectors can be calculated as

vi = hei , vi

vi = hi|vi

and

n

X

|vi = |iihi|vi

i=1

n

X n

X

hv| = vi hi| = hv|iihi|

i=1 i=1

With respect to the chosen orthonormal basis ei , one can explicitly represent

vectors v as column vectors

v1

v2

..

.

vn

with

0

..

.

ei = |ii =

1

.

..

0

(i.e. a 1 in the ith position, 0s elsewhere).

The operation of taking a vector |vi to a dual vector hv| corresponds to

taking a column vector to the row vector that is its conjugate-transpose.

hv| = v1 v2 vn

Then one has

w1

w2

hv|wi = v1 v2 vn . = v1 w1 + v2 w2 + + vn wn

..

wn

34

If is a linear operator : V V , then with respect to the chosen basis it

becomes a matrix, with matrix elements

ji = hj|ii

v1 11 12 1n v1

v2 21 22 2n v2

.. ..

.. .. .. ..

. . . . . .

vn n1 n2 nn vn

n

X

|vi = |iihi|vi

i=1

n

X

1= |iihi|

i=1

tion, since it requires the set of |ii to be a basis. The operator

Pi = |iihi|

is called the projection operator onto the ith basis vector, it corresponds to the

matrix that has 0s everywhere except in the ii component.

Given a linear operator , we have seen that v can be written as a column

vector, with v given by multiplication of the column vector by the matrix ji

corresponding to . Considering the dual vectors, one can define the adjoint of

by

the operator satisfying

hv, wi = hv, wi

hv|wi = hv| wi

for all v, w V .

35

Generalizing the fact that

hv| = hv|

for C, one can write

hv| = hv|

Note that mathematicians tend to favor the notation over the physicists

notation . On this one Im siding with the physicists.

In terms of explicit matrices, < v| is the conjugate-transpose of |v >, so

the matrix with respect to a basis for will be given by the conjugate transpose

T of the matrix for :

( )ij = ji

In the real case, the matrix for the adjoint is just the transpose matrix. We

will say that a linear transformation is self-adjoint if = , skew-adjoint if

= .

A special class of linear transformations will be invertible transformations that

preserve the inner product, i.e. satisfying

hv, wi = hv|wi = hv, wi = hv|wi

for all v, w V . Such transformations take orthonormal bases to orthonormal

bases, so they will appear in one role as change of basis transformations.

In terms of adjoints, this condition becomes

hv, wi = hv, wi = hv, wi

so

= 1

or equivalently

= 1

In matrix notation this first condition becomes

n

X n

X

( )ij jk = ji jk = ik

j=1 j=1

which says that the column vectors of the matrix for are orthonormal vectors.

Using instead the equivalent condition

= 1

one finds that the row vectors of the matrix for are also orthornormal.

Since such linear transformations can be composed and are invertible, they

form a group, and some of the basic examples of Lie groups are given by these

groups for the cases of real and complex vector spaces.

36

4.4.1 Orthogonal groups

Well begin with the real case, where these groups are called orthogonal groups:

is the group of invertible transformations preserving an inner product on a real

n-dimensional vector space V . This is also the group of n by n real invertible

matrices satisfying

(1 )ij = ji

preserving orientation of orthonormal bases) is called SO(n).

Note that since the determinant of the transpose of a matrix is the same as

the determinant of the matrix, we have

so

det() = 1

O(n) is a continuous Lie group, with two components: SO(n), the subgroup of

orientation-preserving transformations, which include the identity, and a com-

ponent of orientation-changing transformations.

The simplest non-trivial example is for n = 2, where all elements of SO(2)

are given by matrices of the form

cos sin

sin cos

component of O(2) will be given by matrices of the form

cos sin

sin cos

cos sin

ei

sin cos

so the representation theory of SO(2) is just as for U (1), with irreducible com-

plex representations one-dimensional and classified by an integer.

We will go on in the next section of the class to examine in detail the case

of SO(3), which is crucial for physical applications because it is the group of

rotations in the physical number of three space dimensions.

37

4.4.2 Unitary groups

In the complex case, groups of invertible transformations preserving the Hermi-

tian inner product are called unitary groups:

Definition (Unitary group). The unitary group U (n) in n-dimensions is the

group of invertible transformations preserving an Hermitian inner product on a

complex n-dimensional vector space V . This is also the group of n by n complex

invertible matrices satisfying

(1 )ij = ji = ( )ij

The same calculation as in the real case here gives

is a group homomorphism.

We have already seen the examples U (1), U (2) and SU (2). For general

values of n, the case of U (n) can be split into the study of its determinant,

which lies in U (1) so is easy to deal with, and the subgroup SU (n), which is

a much more complicated story. As a manifold, we saw that SU (2) can be

identified with the three-sphere S 3 , since an arbitrary group element can be

constructed by specifying one row (or one column), which must be a vector of

length one in C2 . For the case n = 3, the same sort of construction starts by

picking a row of length one in C3 , which will be a point in S 5 . The second

row must be orthornormal, and one can show that the possibilities lie in a

three-sphere S 3 . Once the first two rows are specified, the third row is uniquely

determined. So as a manifold, SU (3) is eight-dimensional, and one might think

it could be identified with S 5 S 3 . It turns out that this is not the case, since

the S 3 varies in a topologically non-trivial way as one varies the point in S 5 .

As spaces, the SU (n) are topologically twisted products of odd-dimensional

spheres, providing some of the basic examples of quite non-trivial topological

manifolds.

In the study of linear transformations of a vector space V , the matrix M cor-

responding to a linear transformation depends on the choice of a basis for V .

Under a change of basis, the matrix will change by conjugation. To better un-

derstand the underlying linear transformation, one can use conjugation to put

the matrix in a canonical form, determining its eigenvalues and eigenvectors.

38

Complex matrices behave in a much simpler fashion than real matrices, since in

the complex case the eigenvalue equation

det(1 M ) = 0

can always be factored into linear factors, and solved for the eigenvalues . For

an arbitrary n by n complex matrix there will be n solutions, counting repeated

eigenvalues with multiplicity. Changing basis and conjugating such a matrix M ,

one can always find a basis for which the matrix is in upper triangular form.

For the case of self-adjoint matrices, things are much more constrained, and

one has the following theorem, proved in most linear algebra textbooks:

Theorem (Spectral theorem for self-adjoint matrices). Given a self-adjoint

complex n by n matrix M , one can always find a unitary matrix U such that

U M U 1 = D

where D is a diagonal matrix with entries Dii = i , i R.

Given M , one finds the eigenvalues i by solving the eigenvalue equation.

One can then go on to solve for the eigenvectors and use these to find U . For

distinct eigenvalues one finds that the corresponding eigenvectors are orthogo-

nal.

This theorem is of crucial importance in quantum mechanics, where for M

an observable, the eigenvectors are the states in the state space with well-defined

numerical values that can be observed experimentally for the observable, and

these numerical values are the eigenvalues. The theorem also tells us that given

an observable, we can use it to choose distinguished orthonormal bases for the

state space: a basis of eigenvectors, normalized to length one. This is a theorem

about finite-dimensional vector spaces, but later on in the course we will see

that something similar will be true even in the case of infinite-dimensional state

spaces.

One can also diagonalize unitary matrices themselves by conjugation by

another unitary. The diagonal entries will all be complex numbers of unit length,

so of the form eij , j R.

For the simplest examples, consider the cases of the groups SU (2) and U (2).

Any matrix in U (2) can be conjugated by a unitary matrix to the diagonal

matrix i

e 1 0

0 ei2

which is the exponential of a corresponding diagonalized skew-adjoint matrix

i1 0

0 i2

For matrices in the subgroup SU (2), one has 2 = 1 = so in diagonal form

an SU (2) matrix will be i

e 0

0 ei

39

which is the exponential of a corresponding diagonalized skew-adjoint matrix

that has trace zero

i 0

0 i

Almost any of the more advanced linear algebra textbooks should cover the

material of this section.

40

Chapter 5

Algebra Representations

vector spaces V of complex dimension n as homomorphisms

: G U (n)

Recall that in the case of G = U (1), we could use the homomorphism property

of to determine in terms of its derivative at the identity. This turns out to

be a general phenomenon for Lie groups G: we can study their representations

by considering the derivative of at the identity, which we will call 0 . Because

of the homomorphism property, knowing 0 is often sufficient to characterize

the representation it comes from. 0 is a linear map from the tangent space

to G at the identity to the tangent space of U (n) at the identity. The tangent

space to G at the identity will carry some extra structure coming from the group

multiplication, and this vector space with this structure will be called the Lie

algebra of G.

The subject of differential geometry gives many equivalent ways of defining

the tangent space at a point of manifolds like G, but we do not want to enter

here into the large subject of differential geometry in general. One of these ways

though is to define the tangent space as the space of tangent vectors, and define

tangent vectors as the possible velocity vectors of parametrized curves g(t) in

the group G.

This can be done, and more advanced treatments of Lie group theory de-

velop this point of view (see for example [49]). In our case though, since we

are interested in specific groups that are explicitly given as groups of matrices,

we can give a more concrete definition, just using the exponential map on ma-

trices. For a more detailed exposition of this subject, using the same concrete

definition of the Lie algebra in terms of matrices, see Brian Halls book[24] or

the abbreviated on-line version[25].

Note that the material of this chapter is quite general, and may be hard

41

to until one has some experience with basic examples. The next chapter will

discuss in detail the groups SU (2) and SO(3) and their Lie algebras, as well

as giving some examples of their representations, and this may be helpful in

making sense of the general theory of this chapter.

Well work with the following definition of a Lie algebra:

Definition (Lie algebra). For G a Lie group of n by n invertible matrices, the

Lie algebra of G (written Lie(G) or g) is the space of n by n matrices X such

that etX G for t R.

Notice that while the group G determines the Lie algebra g, the Lie algebra

does not determine the group. For example, O(n) and SO(n) have the same

tangent space at the identity, and thus the same Lie algebra, but elements in

O(n) not in the component of the identity cant be written in the form etX

(since then you could make a path of matrices connecting such an element to

the identity by shrinking t to zero). Note also that, for a given X, different

values of t may give the same group element, and this may happen in different

ways for different groups sharing the same Lie algebra. For example, consider

G = U (1) and G = (R, +), which both have the same Lie algebra (R), but

in the first case an infinity of values of t give the same group element, in the

second, only one does. In the next chapter well see a more subtle example of

this: SU (2) and SO(3) are different groups with the same Lie algebra.

We have G GL(n, C), and X M (n, C), the space of n by n complex

matrices. For all t R, the exponential etX is an invertible matrix (with inverse

etX ), so in GL(n, C). For each X, we thus have a path of elements of GL(n, C)

going through the identity matrix at t = 0, with velocity vector

d tX

e = XetX

dt

which takes the value X at t = 0:

d tX

(e )|t=0 = X

dt

To calculate this derivative, just use the power series expansion for the expo-

nential, and differentiate term-by-term.

For the case G = GL(n, C), we just have gl(n, C) = M (n, C), which is a

linear space of the right dimension to be the tangent space to G at the identity,

so this definition is consistent with our general motivation. For subgroups G

GL(n, C) given by some condition (e.g. preserving an inner product), we will

need to identify the corresponding condition on X M (n, C) and check that

this defines a linear space.

The existence of a linear space V = g provides us with a distinguished

representation, called the adjoint representation

42

Definition (Adjoint representation). The adjoint representation (Ad, g) is given

by the homomorphism

Ad : g G {X gXg 1 } GL(g)

meaning

(Ad(g))(X) = gXg 1

To show that this is well-defined, one needs to check that gXg 1 g when

X g, but this can be shown using the identity

1

etgXg = getX g 1

1

which implies that etgXg G if etX G. To check this, just expand the

exponential and use

One should notice that there is something tricky going on here: in general

a Lie algebra g is a real vector space, not a complex vector space (for instance,

u(1) = R and su(2) = R3 as vector spaces). Even if G is a group of complex

matrices, when it is not GL(n, C) itself but some subgroup, its tangent space

at the identity will not necessarily be a complex vector space. Consider for

examples the cases G = U (1) and G = SU (2), where u(1) = R and su(2) = R3 .

While the tangent space to the group of all invertible complex matrices is a

complex vector space, imposing some condition such as unitarity picks out a

subspace which generally is just a real vector space, not a complex one. So the

adjoint representation (Ad, g) is in general not a complex representation, but a

real representation, with

by complexifying, i.e. taking complex linear combinations of elements of g

and using the same definition of the adjoint representation in terms of matrix

conjugation. We will denote the complex vector space of such linear combina-

tions as gC , and point out later more explicitly what is going on in the special

cases we will consider.

A Lie algebra g is not just a real vector space, but comes with an extra

structure on the vector space

Definition (Lie bracket). The Lie bracket operation on g is the bilinear anti-

symmetric map given by the commutator of matrices

[, ] : (X, Y ) g g [X, Y ] = XY Y X g

43

We need to check that this is well-defined, i.e. that it takes values in g

Theorem. If X, Y g, [X, Y ] = XY Y X g

representation

Ad(etX )Y = etX Y etX g

as t varies this gives us a parametrized curve in g. Its velocity vector will also

be in g, so

d tX tX

(e Y e )g

dt

One has (by the product rule, which can easily be shown to apply in this case)

d tX tX d d

(e Y e ) = ( (etX Y ))etX + etX Y ( etX )

dt dt dt

= XetX Y etX etX Y XetX

XY Y X

which is thus shown to be in g.

for the vector space g, and use the fact that the Lie bracket can be written in

terms of this basis as

Xn

[Xj , Xk ] = cjkl Xl

l=1

where cjkl is a set of constants known as the structure constants of the Lie

algebra. For example, in the case of su(2), the Lie algebra of SU (2) one has a

basis X1 , X2 , X3 satisfying

3

X

[Xj , Xk ] = jkl Xl

l=1

so the structure constants of su(2) are just the totally anti-symmetric jkl .

groups

The groups we are most interested in are the groups of linear transformations

preserving an inner product: the orthogonal and unitary groups. We have seen

that these are subgroups of GL(n, R) or GL(n, C) of elements satisfying the

condition

= 1

44

In order to see what this condition becomes on the Lie algebra, write = etX ,

for some parameter t, and X a matrix in the Lie algebra. Since the transpose of

a product of matrices is the product (order-reversed) of the transposed matrices,

i.e.

(XY )T = Y T X T

and the complex conjugate of a product of matrices is the product of the complex

conjugates of the matrices, one has

(etX ) = etX

The condition

= 1

thus becomes

etX (etX ) = etX etX = 1

Since X and X commute, this becomes

et(X+X )

=1

or

X + X = 0

so the matrices we want to exponentiate are skew-adjoint, satisfying

X = X

Note that physicists often choose to define the Lie algebra in these cases as

self-adjoint matrices, then multiplying by i before exponentiating to get a group

element. We will not use this definition, one reason being that the situation of

real vs. complex vector spaces here is already tricky enough without introducing

another source of complex numbers into the story.

Recall that the orthogonal group O(n) is the subgroup of GL(n, R) of matrices

satisfying T = 1 . We will restrict attention to the subgroup SO(n) of

matrices with determinant 1 which is the component of the group containing

the identity, and thus elements that can be written as

= etX

These give a path connecting to the identity (taking esX , s [0, t]). We

saw above that the condition T = 1 corresponds to skew-symmetry of the

matrix X

X T = X

45

So in the case of G = SO(n), we see that the Lie algebra so(n) is the space of

skew-symmetric (X t = X) n by n real matrices, together with the bilinear,

antisymmetric product given by the commutator:

n2 n

1 + 2 + + (n 1) =

2

and a basis will be given by the matrices Eij , with i, j = 1, . . . n, i < j defined

as

+1 if i = k, j = l

(Eij )kl = 1 if j = k, i = l

0 otherwise

In the next section of the class we will examine detail the n = 3 case, where

the Lie algebra so(3) is R3 , realized as the space of anti-symmetric real 3 by 3

matrices.

Note that this is a real vector space, and its definition just involves real

matrices. Thought of as the adjoint representation, it is a real representation of

SO(n). One aspect of this is that the matrices Ad(g) of the adjoint representa-

tion are not diagonalizable as real matrices. One may want to complexify to get

a complex representation in which the Ad(g) can be diagonalized. Doing that

takes so(n), the antisymmetric n by n real matrices to so(n)C , the Lie algebra

of antisymmetric n by n complex matrices.

For the case of the group U (n), the group is connected and one can write all

group elements as etX , where now X is a complex n by n matrix. The unitarity

condition implies that X is skew-adjoint (also called skew-Hermitian), satisfying

X = X

together with the bilinear, antisymmetric product given by the commutator:

2

Note that these matrices form a subspace of C n of half the dimension, so of

real dimension n2 . u(n) is a real vector space of dimension n2 , but it is NOT the

space of real n by n matrices, but a space of real linear combinations of skew-

Hermitian matrices, which in general are complex (the real skew Hermitian

2

matrices so(n) form a subspace of lower dimension, n 2n , corresponding to the

46

subgroup SO(n) U (n)). If we take ALL complex linear combinations of skew-

Hermitian matrices, than we get the full algebra of n by n complex matrices,

i.e.

u(n)C = M (n, C) = gl(n, C)

There is an identity relating the determinant and the trace of a matrix

det(eX ) = etrace(X)

using the fact that the trace and the determinant of a matrix are conjugation-

invariant. Since the determinant of an SU (n) matrix is 1, this shows that the

Lie algebra su(n) of SU (n) will consist of matrices that are not only skew-

Hermitian, but also of trace zero. So in this case su(n) is again a real vector

space, of dimension n2 1.

In the next chapter we will consider in detail the special case of SU (2) and

its Lie algebra.

In the problem set due next week, you will use the fact that U (n) and u(n)

matrices can be diagonalized by unitaries to show that any U (n) matrix can be

written as an exponential of something in the Lie algebra. The corresponding

theorem is also true for SO(n) but requires looking at diagonalization into 2 by

2 blocks. It is not true for O(n) (you cant reach the disconnected component

of the identity by exponentiation), or for GL(n, C) in general (an example is in

one of the problems).

We have defined a group representation as a homomorphism (a map of groups

preserving group multiplication)

: G GL(n, C)

preserving the Lie bracket:

(, V ) of a Lie algebra g on an n-dimensional complex vector space V is given

by a linear map

satisfying

([X, Y ]) = [(X), (Y )]

Such a representation is called unitary if its image is in u(n), i.e it satisfies

(X) = (X)

47

More concretely, given a basis X1 , X2 , . . . , Xd of a Lie algebra g of dimension

d with structure constants cjk , a representation is given by a choice of d n by n

complex matrices (Xj ) satisfying the commutation relations

d

X

[(Xj ), (Xk )] = cjkl (Xl )

l=1

The notion of a Lie algebra is motivated by the fact that the homomorphism

property causes the map to be largely determined by its behavior infinitesi-

mally near the identity, and thus by the derivative 0 . One way to define the

derivative of such a map is in terms of velocity vectors of paths, and this sort of

definition in this case associates to a representation : G GL(n, C) a linear

map

0 : g M (n, C)

where

d

0 (X) = ((etX ))|t=0

dt

phisms U (1) GL(1, C) = C ) by looking at the derivative of the map at

the identity. For general Lie groups G, one can do something similar, show-

ing that a representation of G gives a representation of the Lie algebra (by

taking the derivative at the identity), and then trying to classify Lie algebra

representations.

Theorem. If : G GL(n, C) is a group homomorphism, then

d

0 : X g 0 (X) = ((etX ))|t=0 gl(n, C) = M (n, C)

dt

48

satisfies

1. 0

(etX ) = et (X)

2. For g G

0 (gXg 1 ) = (g) 0 (X)((g))1

3. 0 is a Lie algebra homomorphism:

0 ([X, Y ]) = [ 0 (X), 0 (Y )]

Proof. 1. We have

d d

(etX ) = (e(t+s)X )|s=0

dt ds

d

= (etX esX )|s=0

ds

d

= (etX ) (esX )|s=0

ds

= (etX ) 0 (X)

d

So f (t) = (etX ) satisfies the differential equation dt f = f 0 (X) with

0

initial condition f (0) = 1. This has the unique solution f (t) = et (X)

2. We have

0 1 1

et (gXg )

= (etgXg )

tX 1

= (ge g )

= (g)(e tX

)(g)1

0

= (g)et (X) (g)1

Differentiating with respect to t at t = 0 gives

0 (gXg 1 ) = (g) 0 (X)((g))1

3. Recall that

d tX tX

[X, Y ] = (e Y e )|t=0

dt

so

d tX tX

0 ([X, Y ]) = 0 ( (e Y e )|t=0 )

dt

d 0 tX tX

= (e Y e )|t=0 (by linearity)

dt

d

= ((etX ) 0 (Y )(etX ))|t=0 (by 2.)

dt

d 0 0

= (et (X) 0 (Y )et (X) )|t=0 (by 1.)

dt

= [ 0 (X), 0 (Y )]

49

This theorem shows that we can study Lie group representations (, V )

by studying the corresponding Lie algebra representation ( 0 , V ). This will

generally be much easier since the 0 (X) are just linear maps. We will need this

a bit later in the course when we construct and classify all SU (2) and SO(3)

representations, finding that the corresponding Lie algebra representations are

much simpler to analyze.

For any Lie group G, we have seen that there is a distingished representation,

the adjoint representation (Ad, g). The corresponding Lie algebra representation

is also called the adjoint representation, but written as (Ad0 , g) = (ad, g). From

the fact that

Ad(etX )(Y ) = etX Y etX

we can differentiate with respect to t to get the Lie algebra representation

d tX tX

ad(X)(Y ) = (e Y e )|t=0 = [X, Y ]

dt

From this we see that ad(X) is the linear map from g to itself given by

Y [X, Y ]

where these are linear maps on g, with composition of linear maps, so operating

on Z g we have

This is called the Jacobi identity. It could have been more simply derived as

an identity about matrix multiplication, but here we see that it is true for a

more abstract reason, reflecting the existence of the adjoint representation. It

can be written in other forms, rearranging terms using antisymmetry of the

commutator, with one example the sum of cyclic permutations

Definition (Abstract Lie algebra). An abstract Lie algebra over a field k is a

vector space A over k, with a bilinear operation

[, ] : (X, Y ) A A [X, Y ] A

satisfying

50

1. Antisymmetry:

[X, Y ] = [Y, X]

2. Jacobi identity:

Such Lie algebras do not need to be defined as matrices, and their Lie bracket

operation does not need to be defined in terms of a matrix commutator (al-

though the same notation continues to be used). Later on in this course we

will encounter important examples of Lie algebras that are defined in this more

abstract way.

The material of this section is quite conventional mathematics, with many good

expositions, although most aimed at a higher level than this course. An example

at the level of this course is the book Naive Lie Theory [43]. It covers basics

of Lie groups and Lie algebras, but without representations. The notes[25] and

book[24] of Brian Hall are a good source to study from. Some parts of the proofs

given here are drawn from those notes.

51

52

Chapter 6

Groups in 3 and 4

Dimensions

Among the basic symmetry groups of the physical world is the orthogonal group

SO(3) of rotations about a point in three-dimensional space. The observables

one gets from this group are the components of angular momentum, and under-

standing how the state space of a quantum system behaves as a representation

of this group is a crucial part of the analysis of atomic physics examples and

many others. Unlike some of the material covered in this course, this is a topic

you will find in some version or other in every quantum mechanics textbook.

Remarkably, it turns out that the quantum systems in nature are often

representations not of SO(3), but of a larger group called Spin(3), one that

has two elements corresponding to every element of SO(3). This is called the

Spin group, and in any dimension n it exists, always as a doubled version

of the orthogonal group SO(n), one that is needed to understand some of the

more subtle aspects of geometry in n dimensions. In the n = 3 case it turns

out that Spin(3) ' SU (2) and we will study in detail the relationship of SO(3)

and SU (2). This appearance of the unitary group SU (2) is special to 3 (and to

some extent 4) dimensions.

In R2 rotations about the origin are given by elements of SO(2), with a counter-

clockwise rotation by an angle given by the matrix

cos sin

R() =

sin cos

53

This can be written as an exponential, R() = eL = cos + L sin for

0 1

L=

1 0

Here SO(2) is a commutative Lie group with Lie algebra R (the Lie bracket is

trivial, all elements of the Lie algebra commute). Note that we have a represen-

tation on V = R2 here, but it is a real representation, not one of the complex

ones we have when we have a representation on a quantum mechanical state

space.

In three dimensions the group SO(3) is 3-dimensional and non-commutative.

One now has three independent directions one can rotate about, which one can

take to be the x, y and z-axes, with rotations about these axes given by

1 0 0 cos 0 sin

Rx () = 0 cos sin Ry () = 0 1 0

0 sin cos sin 0 cos

cos sin 0

Rz () = sin cos 0

0 0 1

A standard parametrization for elements of SO(3) is in terms of 3 Euler angles

, , with a general rotation given by

angle about the new x-axis, followed by a rotation by about the new z-

axis. Multiplying out the matrices gives a rather complicated expression for a

rotation in terms of the three angles, and one needs to figure out what range to

choose for the angles to avoid multiple counting.

The infinitesimal picture near the identity of the group, given by the Lie

algebra structure on so(3) is much easier to understand. Recall that for orthog-

onal groups the Lie algebra can be identified with the space of anti-symmetric

matrices, so one in this case has a basis

0 0 0 0 0 1 0 1 0

l1 = 0 0 1 l2 = 0 0 0 l3 = 1 0 0

0 1 0 1 0 0 0 0 0

Note that these are exactly the same commutation relations satisfied by

the basis vectors X1 , X2 , X3 of the Lie algebra su(2), so so(3) and su(2) are

isomorphic Lie algebras. They both are the vector space R3 with the same Lie

54

bracket operation on pairs of vectors. This operation is familiar in yet another

context, that of the cross-product of standard basis vectors ej in R3 :

e1 e2 = e3 , e2 e3 = e1 , e3 e1 = e2

(X, Y ) R3 R3 [X, Y ] R3

that makes R3 a Lie algebra so(3) = su(2) is just the cross-product on vectors

in R3 .

So far we have three different isomorphic ways of putting a Lie bracket on

R3 , making it into a Lie algebra:

commutator as Lie bracket.

the matrix commutator as Lie bracket.

[v, w] = v w

Something very special that happens for orthogonal groups only in 3-dimensions

is that the vector representation is isomorphic to the adjoint representation. Re-

call that any Lie group G has a representation (Ad, g) on its Lie algebra g. so(n)

can be identified with the anti-symmetric n by n matrices, so is of (real) dimen-

2

sion n 2n . Only for n = 3 is this equal to n, the dimension of the representation

on vectors in Rn . This corresponds to the geometrical fact that only in 3 di-

mensions is a plane (in all dimensions rotations are built out of rotations in

various planes) determined uniquely by a vector (the vector perpendicular to

the plane). Equivalently, only in 3 dimensions is there a cross-product v w

which takes two vectors determining a plane to a unique vector perpendicular

to the plane.

To see explicitly what is going on, lets look at the vector and adjoint Lie

algebra representations. The commutation relations for the li determine the Lie

algebra representation (ad, so(n)) using the definition of the adjoint representa-

tion, (ad(X))(Y ) = [X, Y ]. For instance, for infinitesimal rotations about the

x-axis, one has for the adjoint representation

matrix multiplication, giving

l1 e1 = 0, l1 e2 = e3 , l1 e3 = e2

55

from which one can see that one has the same Lie algebra representation, with

the isomorphism identifying li = ei . At the level of the group, rotations about

the x-axis by an angle correspond to matrices

1 0 0

el1 = 0 cos sin

0 sin cos

which act by conjugation on anti-symmetric real matrices and in the usual way

on column vectors.

Our two isomorphic representations are on column vectors (vector repre-

sentation on R3 ) and on antisymmetric real matrices (adjoint representation

on so(3)), with the isomorphism given by

v1 0 v3 v2

v2 v3 0 v1

v3 v2 v1 0

On the column vectors both the Lie algebra and the Lie group representation

are given just by matrix multiplication. On the antisymmetric matrices the Lie

group representation is given by

0 v3 v2 0 v3 v2

Ad(g) v3 0 v1 = g v3 0 v1 g 1

v2 v1 0 v2 v1 0

algebra representation is given by

0 v3 v2 0 v3 v2

ad(X) v3 0 v1 = [X, v3 0 v1 ]

v2 v1 0 v2 v1 0

A remarkable property of the orthogonal groups SO(n) is that they come with an

associated group, called Spin(n), with every element of SO(n) corresponding

to two distinct elements of Spin(n). If you have seen some topology, what

is at work here is that the fundamental group of SO(n) is non-trivial, with

1 (SO(n)) = Z2 . Spin(n) is topologically the simply-connected double-cover

of SO(n), and one can choose the covering map : Spin(n) SO(n) to be

a group homomorphism. Spin(n) is a Lie group of the same dimension, with

an isomorphic tangent space at the identity, so the Lie algebras of the two

groups are isomorphic: so(n) ' spin(n). We will construct Spin(n) and the

covering map only for the cases n = 3 and n = 4, with higher dimensional

56

cases requiring more sophisticated techniques. For n = 3 it turns out that

Spin(n) = SU (2), and for n = 4, Spin(4) = SU (2) SU (2). For perhaps the

simplest way to see how this works, it is best to not use just real and complex

numbers, but also bring in a third number system, the quaternions.

6.2.1 Quaternions

The quaternions are a number system (denoted by H) generalizing the complex

number system, with elements q H that can be written as

q = q0 + q1 i + q2 j + q3 k, qi R

with i, j, k H satisfying

i2 = j2 = k2 = 1, ij = k, ki = j, jk = i

q q = q0 q1 i q2 j q3 k

function on this R4 can be written in terms of quaternions as

Since

q q

=1

|q|2

one has a formula for the inverse of a quaternion

q

q 1 =

|q|2

From this one can also see that the conjugation operation satisfies (for u, v H)

uv = vu

|uv|2 = uvuv = uv

vu = |u|2 |v|2

The length one quaternions thus form a group under multiplication, called Sp(1).

There are also Lie groups called Sp(n), consisting of invertible matrices with

quaternionic entries that act on quaternionic vectors preserving the quaternionic

length-squared, but these play no role in quantum mechanics so we wont study

them further. Sp(1) can be identified with the three-dimensional sphere since

the length one condition on q is

57

6.2.2 Rotations and spin groups in four dimensions

Pairs (u, v) of unit quaternions give the product group Sp(1) Sp(1). An

element of this group acts on H = R4 by

q uqv

and this action preserves lengths of vectors and is linear in q, so it must corre-

spond to an element of the group SO(4). One can easily see that pairs (u, v) and

(u, v) give the same orthogonal transformation of R4 , so the same element

of SO(4). One can show that SO(4) is the group Sp(1) Sp(1), with elements

(u, v) and (u, v) identified. The name Spin(4) is given to the Lie group that

double covers SO(4), so here Spin(4) = Sp(1) Sp(1).

Later on in the course well encounter Spin(4) and SO(4) again, but for now

were interested in its subgroup Spin(3) that only acts on 3 of the dimensions,

and double-covers not SO(4) but SO(3). To find this, consider the subgroup of

Spin(4) consisting of pairs (u, v) of the form (u, u1 ) (a subgroup isomorphic

to Sp(1), since elements correspond to a single unit length quaternion u). This

subgroup acts on quaternions by conjugation

q uqu1

an action which is trivial on the real quaternions, but nontrivial on the pure

imaginary quaternions of the form

q = ~v = v1 i + v2 j + v3 k

~v u~v u1

of SO(3). We thus have a map (which can be checked to be a homomorphism)

58

Both u and u act in the same way on ~v , so we have two elements in

Sp(1) corresponding to the same element in SO(3). One can show that is a

surjective map (one can get any element of SO(3) this way), so it is what is called

a covering map, specifically a two-fold cover. It makes Sp(1) a double-cover

of SO(3), and we give this the name Spin(3). This also tells us exactly which

3-dimensional space SO(3) is. It is S 3 = Sp(1) = Spin(3) with opposite points

on the three-sphere identified. This space is known as RP(3), real projective

3-space, which can also be thought of as the space of lines through the origin in

R4 (each such line intersects S 3 in two opposite points).

59

For those who have seen some topology, note that the covering map is an

example of a topological non-trivial cover. It is just not true that topologically

S 3 ' RP3 (+1, 1). S 3 is a connected space, not two disconnected pieces.

This topological non-triviality implies that globally there is no possible homo-

morphism going in the opposite direction from (i.e. SO(3) Spin(3)). One

can do this locally, picking a local patch in SO(3) and taking the inverse of

to a local patch in Spin(3), but this wont work if we try and extend it globally

to all of SO(3).

The relationship between rotations of R3 and unit quaternions is quite sim-

ple: for

w

~ = w1 i + w2 j + w3 k

a unit vector in R3 H, conjugation by the unit quaternion

u(, w)

~ = cos + w

~ sin

gives a rotation about the w~ axis by an angle 2. The factor of 2 here reflects

the fact that unit quaternions double-cover the rotation group SO(3): taking

+ gives a different unit quaternion that implements the same rotation

in 3-space. To see how this works, one can for example take as axis of rotation

the z-axis by choosing w~ = k. The unit quaternion

u = cos + k sin

60

has inverse

u1

= cos k sin

and acts on ~v = v1 i + v2 j + v3 k as

~v u ~v u1

=(cos + k sin )(v1 i + v2 j + v3 k)(cos k sin )

=(v1 (cos2 sin2 ) v2 (2 sin cos ))i

+ (2v1 sin cos + v2 (cos2 sin2 ))j + v3 k

=(v1 cos 2 v2 sin 2)i + (v1 sin 2 + v2 cos 2)j + v3 k

cos 2 sin 2 0 v1

= sin 2 cos 2 0 v2

0 0 1 v3

Here one sees explicitly for rotations about the z-axis the double-covering map

cos 2 sin 2 0

: u = cos + k sin Sp(1) = Spin(3) sin 2 cos 2 0 SO(3)

0 0 1

homomorphism takes this to a circle in SO(3), one that gets traced out twice

as goes from 0 to 2.

In the case of U (1), the unit vectors in R2 , one can take the identity to

be in the real direction, and then the tangent space at the identity (the Lie

algebra u(1)) is iR, with basis i. For the case of Sp(1), one can again take

the identity to be in the real direction, and the tangent space (the Lie algebra

sp(1)) is isomorphic to R3 , with basis vectors i, j, k. The Lie bracket is just the

commutator, e.g.

[i, j] = ij ji = 2k

Linear combinations of these basis vectors are precisely the velocity vectors one

gets for the paths u(, w)

~ = cos + w~ sin , which are parametrized by and go

through the identity at = 0, since

d

u(, w)

~ |=0 = (sin + w

~ cos )|=0 = w

~ = w1 i + w2 j + w3 k

d

The derivative of the map will be a linear map

0 : sp(1) so(3)

can compute it easily on basis vectors, using for instance the formula above

cos 2 sin 2 0

(cos + k sin ) = sin 2 cos 2 0

0 0 1

61

Using the definition as a derivative with respect to the parameter of a path,

evaluated at the identity, one finds

d

0 (k) = (cos + k sin )|=0

d

2 sin 2 2 cos 2 0

= 2 cos 2 2 sin 2 0

0 0 0 |=0

0 2 0

= 2 0 0 = 2l3

0 0 0

i j k

, , and l1 , l2 , l3

2 2 2

i j k j k i k i j

[ , ]= , [ , ]= , [ , ]=

2 2 2 2 2 2 2 2 2

Instead of working with quaternions and their non-commutativity and special

multiplication laws, it is more conventional to choose an isomorphism between

quaternions H and a space of 2 by 2 complex matrices, and work just with

matrix multiplication and complex numbers. The Pauli matrices can be used

to gives such an isomorphism, taking

1 0 0 i 0 1

11= , i i1 = , j i2 =

0 1 i 0 1 0

i 0

k i3 =

0 i

The correspondence between H and 2 by 2 complex matrices is then given

by

q0 iq3 q2 iq1

q = q0 + q1 i + q2 j + q3 k

q2 iq1 q0 + iq3

Since

q0 iq3 q2 iq1

det = q02 + q12 + q22 + q32

q2 iq1 q0 + iq3

62

we see that the length-squared function on quaternions corresponds to the de-

terminant function on 2 by 2 complex matrices. Taking q Sp(1), so of length

one, corresponds both to the matrix having determinant one, and its rows and

columns having length one, so the complex matrix is in SU (2).

Recall that any SU (2) matrix can be written in the form

isomorphism with unit vectors in H is given by

= q0 iq3 , = q2 iq1

We see that Sp(1), Spin(3) and SU (2) are all names for the same group, geo-

metrically S 3 , the unit sphere in R4 .

Under our identification of H with 2 by 2 complex matrices, we have an

identification of Lie algebras sp(1) = su(2) between pure imaginary quaternions

and skew-Hermitian trace-zero 2 by 2 complex matrices

iw3 w2 iw1

~ = w1 i + w2 j + w3 k

w =w

w2 iw1 iw3

The basis 2i , 2j , k2 gets identified with a basis for the Lie algebra su(2) which

written in terms of the Pauli matrices is

j

Xj = i

2

with the Xj satisfying the commutation relations

We now have no less than three isomorphic Lie algebras sp(1) = su(2) =

so(3) on which we have the adjoint representation, with bases that get identified

under the following identifications

0 w3 w2

i w 3 w1 iw 2

w1 2i + w2 2j + w3 k2

w3 0 w1

2 w1 + iw2 w3

w2 w1 0

i 1

i l1

2 2

63

etc. The first of these identifications comes from the way we identify H with

2 by 2 complex matrices, these are Lie algebras of isomorphic groups. The

second identification is 0 , the derivative at the identity of the covering map :

Sp(1) = SU (2) SO(3) that is a homomorphism between two non-isomorphic

groups.

On each of these we have adjoint Lie group (Ad) and Lie algebra (ad) repre-

sentations, with Ad given by conjugation with the corresponding group elements

in Sp(1), SU (2) and SO(3), and ad given by taking commutators in the respec-

tive Lie algebras of pure imaginary quaternions, skew-Hermitian trace-zero 2 by

2 complex matrices and 3 by 3 real antisymmetric matrices.

Note that these three Lie algebras are all three-dimensional real vector

spaces, so these are real representations. If one wants a complex representa-

tion, one can complexify and take complex linear combinations of elements.

This is less confusing in the case of su(2) than for sp(1) since taking complex

linear combinations of skew-Hermitian trace-zero 2 by 2 complex matrices just

gives all trace-zero 2 by 2 matrices (the Lie algebra sl(2, C)).

In addition, recall from earlier that there is a fourth isomorphic version of

this representation, the representation of SO(3) on column vectors. This is also

a real representation, but can straightforwardly be complexified.

At the level of Lie groups we have seen that our identification of H and 2

by 2 matrices identifies Sp(1) with SU (2), taking

u(, w)

cos iw3 sin (iw1 w2 ) sin

=

(iw1 + w2 ) sin cos + iw3 sin

The relation to SO(3) rotations is that this is an SU (2) element such that,

if one identifies vectors (v1 , v2 , v3 ) R3 with complex matrices

v3 v1 iv2

v1 + iv2 v3

then

v3 v1 iv2

(cos 1 i(w ) sin ) (cos 1 i(w ) sin )1

v1 + iv2 v3

We will define

R(, w) = e(w1 X1 +w2 X2 +w3 X3 ) = ei 2 w

= cos( )1 i(w ) sin( )

2 2

and then it is conjugation by R(, w) that rotates vectors by an angle about

w.

64

In particular, rotation about the z-axis by an angle is given by conjugation

by

0

!

ei 2 0

R(, 0) =

1 0 ei 2

In terms of the group SU (2), the double covering map thus acts on diag-

onalized matrices as

i cos 2 sin 2 0

e 0

: SU (2) sin 2 cos 2 0

ei

0

0 0 1

One can write down a somewhat unenlightening formula for the map :

SU (2) SO(3) in general, getting

( ) = Re( 2 2 ) Im(2 + 2 ) 2Re()

2Re() 2Im() ||2 ||2

In this course we wont encounter again orthogonal groups above 3 or 4 dimen-

sions, but the phenomenon of spin groups occurs in every dimension. For each

n > 2, the orthogonal group SO(n) is double-covered by a group Spin(n) with

an isomorphic Lie algebra. Special phenomena relating these Spin groups oc-

cur for n < 7 (it turns out that Spin(5) = Sp(2), the 2 by 2 norm-preserving

quaternionic matrices), and Spin(6) = SU (4), but in higher dimensions these

groups have no relation to quaternions or unitary groups. The construction of

the double-covering map Spin(n) SO(n) for n > 4 requires use of a new class

of algebras that generalize the quaternion algebra, known as Clifford algebras.

For another discussion of the relationship of SO(3) and SU (2) as well as a

construction of the map , see [41], sections 4.2 and 4.3, as well as [3], chapter

8, and [43] Chapters 2 and 4.

65

66

Chapter 7

Magnetic Field

rotation group may seem to be a somewhat obscure mathematical fact but,

remarkably, it is Spin(3) rather than SO(3) that is the symmetry group of

quantum systems describing elementary particles. Studies of atomic spectra in

the early days of quantum mechanics revealed twice as many states as expected,

a phenomenon that is now understood to be due to the fact that the state space

of an electron at a point is C2 rather than C . This state space is not a repre-

sentation of the rotational symmetry group SO(3), but it is a representation of

the double-cover Spin(3) = SU (2) (the standard representation of SU (2) ma-

trices on C2 ). Particles with this degree of freedom are said to have spin 1/2,

with the origin of this 1/2 precisely the same as the 1/2 we saw in the previous

class in the discussion of the Lie algebras of the two equivalent forms Sp(1) and

SU (2) of the group Spin(3). This same C2 occurs for other matter particles

(quarks, neutrinos, etc.) and appears to be a fundamental property of nature.

Besides the doubling of the number of states, more complicated physical effects

occur when such a particle is subjected to a magnetic field, and we will examine

some of this physics in this section.

In the last section we examined in great detail various ways of looking at the

three-dimensional irreducible real representation of the groups SO(3), SU (2)

and Sp(1). For SU (2) and Sp(1) however, there is a even simpler non-trivial

irreducible representation: the representation of 2 by 2 complex matrices in

SU (2) on column vectors C2 by matrix multiplication or the representation of

unit quaternions in Sp(1) on H by scalar multiplication. Choosing an identifica-

tion C2 = H these are isomorphic representations of isomorphic groups, and for

convenience we will generally stick to using SU (2) and complex numbers rather

67

than quaternions. This irreducible representation is known as the spinor or

spin representation of Spin(3) The homomorphism spinor defining the rep-

resentation is just the identity map from SU (2) to itself.

The spin representation of Spin(3) is not a representation of SO(3). The

double cover map : Spin(3) SO(3) is a homomorphism, so given a rep-

resentation (, V ) of SO(3) one gets a representation ( , V ) of Spin(3) by

composition. But there is no homomorphism SO(3) SU (2) that would al-

low us to make the standard representation of SU (2) on C2 into an SO(3)

representation. We could try and define a representation of SO(3) by

g ) SU (2)

g ) = g.

The problem with this is that we wont quite get a homomorphism. Changing

our choice of g will introduce a minus sign, so will only be a homomorphism

up to sign

(g1 )(g2 ) = (g1 g2 )

The nontrivial nature of the double-covering ensures that there is no way to

completely eliminate all minus signs, no matter how we choose g. Something like

this which is not quite a representation, only one up to a sign ambiguity, is known

as a projective representation. So, the spinor representation of SU (2) =

Spin(3) is only a projective representation of SO(3), not a true representation

of SO(3). Quantum mechanics texts often deal with this phenomenon by noting

that physically there is an ambiguity in how one specifies the space of states H,

with multiplication by an overall scalar not changing the eigenvalues of operators

or the relative probabilities of observing these eigenvalues. As a result, the sign

ambiguity has no physical effect. It seems more straightforward though to just

work from the beginning with the larger symmetry group Spin(3), accepting

that this is the correct symmetry group reflecting the action of rotations on

three-dimensional quantum systems.

Recall from our earlier discussion of the two-state quantum system, where the

state space is H = C2 that there is a four (real)-dimensional space of observables

with a general self-adjoint linear operator on H written as

M = c0 1 + c1 1 + c2 2 + c3 3

Lie algebra u(2). Exponentiating gives elements of the unitary group U (2), with

and

ei(c1 1 +c2 2 +c3 3 ) SU (2) U (2)

68

For an arbitrary two-state quantum system, neither the operators M nor the

group SU (2) have any particular geometric significance. In some cases though,

the group SU (2) does have a geometric interpretation, reflecting its role as the

double-cover Spin(3) and the fact that the group SO(3) acts on the physical

system by rotations of three-dimensional space. In these cases, the quantum

system is said to carry spin, in particular spin one-half (we will later on

encounter state spaces of higher spin values). We will from now on assume that

we are dealing with a spin one-half state space.

A standard basis for the observables (besides the unit operator that generates

overall U (1) transformations) is taken to be the operators which give a basis for

the Lie algebra su(2) in its defining representation. We have seen these already

several times, they are the Xj , j = 1, 2, 3, where

j

Xj = i

2

and they satisfy the commutation relations

To make contact with the physics formalism, well define self-adjoint opera-

tors

j

Sj = iXj =

2

1

which will have real eigenvalues 2 .

Note that the conventional definition of these operators in physics texts

includes a factor of ~

~j

Sj = i~Xj =

2

This is because rotations of vectors are defined in physics texts using conjugation

by the matrix

R(, w) = ei ~ wS (compare to our R(, w) = eiwS = ewX )

with the convention of dividing by a factor of ~ appearing here for reasons that

have to do with the action of rotations on functions on R3 that we will encounter

later on. For now, one can either keep factors of ~ out of the definitions of Sj

and the action of rotations, or just assume that we are working in units where

~ = 1.

States in H = C2 that have a well-defined value of the observable Sj will be

the eigenvectors of Sj , with value for the observable the corresponding eigen-

value, which will be 12 . Measurement theory postulates that if we perform the

measurement corresponding to Sj on an arbitrary state |i, then

we will with probability c+ get a value of + 12 and leave the state in an

eigenvector |j, + 21 i of Sj with eigenvalue + 12

we will with probability c get a value of 12 and leave the state in an

eigenvector |j, 21 i of Sj with eigenvalue 21

69

where if

1 1

|i = |j, + i + |j, i

2 2

we have

||2 ||2

c+ = , c =

||2 + ||2 || + ||2

2

nent of S, say Sk , k 6= j will give 21 with equal probability and put the system

in a corresponding eigenvector of Sk .

If a quantum system is in an arbitrary state |i it may not have a well-

defined value for some observable A, but one can calculate the expected value

of A. This is the sum over a basis of H consisting of eigenvectors (which will

all be orthogonal) of the corresponding eigenvalues, weighted by the probability

of their occurrence. The calculation of this sum in this case A = Sj ) using

expansion in eigenvectors of Sj gives

=

h|i (hj, + 21 | + hj, 21 |)(|j, + 12 i + |j, 12 i)

||2 (+ 12 ) + ||2 ( 21 )

=

||2 + ||2

1 1

=c+ (+ ) + c ( )

2 2

One often chooses to simplify such calculations by normalizing states so that

the denominator h|i is 1. Note that the same calculation works in general

for the probability of measuring the various eigenvalues of an observable A, as

long as one has orthogonality and completeness of eigenvectors.

Recall that

R(, w) = ewX

= cos( )1 i(w ) sin( )

2 2

In the case of a spin one-half particle, the group Spin(3) = SU (2) acts on

states by the spinor representation with the element R(, w) SU (2) acting as

|i R(, w)|i

Taking adjoints and using unitarity, one has (thinking of vectors as column

vectors, elements of the dual space as row vectors) the following action on the

dual state space

h| h|R(, w)1

The operators Sj transform under this same group according to the vector

representation of SU (2). Recall that an SO(3) rotation on vectors is given by

conjugating by an SU (2) = Spin(3) group element according to

70

We see that if our observables transform as vectors, and states as spinors,

the expectation values remain invariant:

h|Sj |i h|R(, w)1 R(, w)Sj R(, w)1 R(, w)|i = h|Sj |i

and all eigenvalues of observables remain invariant. One can also interpret

these joint transformations on states and observables as simply a change of

coordinates, a rotation from the standard basis to a different one.

Next semester in this course we may get to the physics of electromagnetic

fields and how particles interact with them in quantum mechanics, but for now

all we need to know is that for a spin one-half particle, the spin degree of freedom

that we are describing by H = C2 has a dynamics described by the Hamiltonian

H = B

Here B is the vector describing the magnetic field, and

e

=g S

2mc

is an operator called the magnetic moment operator. The constants that appear

are: e the electric charge, c the speed of light, m the mass of the particle, and g,

a dimensionless number called the gyromagnetic ratio, which is approximately

2 for an electron, about 5.6 for a proton.

The Schrodinger equation is

d i

|(t)i = ( )( B)|(t)i

dt ~

with solution

|(t)i = U (t)|(0)i

where ge|B|

it ge ge B

U (t) = e ~ B = eit 2mc SB = et 2mc XB = et 2mc X |B|

The time evolution of a state is thus given at time t by a rotation about the

B

axis w = |B| by an angle

ge|B|t

2mc

a rotation taking place with angular velocity ge|B|

2mc .

The amount of non-trivial physics that is described by this simple system is

impressive, including:

The Zeeman effect: this is the splitting of atomic energy levels that occurs

when an atom is put in a constant magnetic field. With respect to the

energy levels for no magnetic field, where both states in H = C2 have the

same energy, the term in the Hamiltonian given above adds

ge|B|

4mc

to the two energy levels, giving a splitting between them proportional to

the size of the magnetic field.

71

The Stern-Gerlach experiment: here one passes a beam of spin one-half

quantum systems through an inhomogeneous magnetic field. One can

arrange this in such a way as to pick out a specific direction w, and split

the beam into two components, of eigenvalue + 12 and 21 for the operator

w S.

Nuclear magnetic resonance spectroscopy: one can subject a spin one-

half system to a time-varying magnetic field B(t), which will be described

by the same Schrodinger equation, although now the solution cannot be

found just by exponentiating a matrix. Nuclei of atoms provide spin one-

half systems that can be probed with time and space-varying magnetic

fields, allowing imaging of the material that they make up.

Quantum computing: attempts to build a quantum computer involve try-

ing to put together multiple systems of this kind (qubits), keeping them

isolated from perturbations by the environment, but still allowing inter-

action with the system in a way that preserves its quantum behavior.

The 2012 Physics Nobel prize was awarded for experimental work making

progress in this direction.

So far in this course weve been describing what is known as the Schrodinger

picture of quantum mechanics. States in H are functions of time, obeying

the Schrodinger equation determined by a Hamiltonian observable H, while

observable self-adjoint operators A are time-independent. Time evolution is

given by a unitary transformation

t

U (t) = ei ~ H , |(t)i = U (t)|(0)i

One can instead use U (t) to make a unitary transformation that puts the

time-dependence in the observables, removing it from the states, as follows:

|(t)i |(t)iH = U 1 (t)|(t)i = |(0)i, A AH (t) = U 1 (t)AU (t)

where the H subscripts for Heisenberg indicate that we are dealing with

Heisenberg picture observables and states. One can easily see that the physi-

cally observable quantities given by eigenvalues and expectations values remain

the same:

In the Heisenberg picture the dynamics is given by a differential equation

not for the states but for the operators. Recall from our discussion of the adjoint

representation the formula

d tX tX d d

(e Y e ) = ( (etX Y ))etX + etX Y ( etX )

dt dt dt

= XetX Y etX etX Y etX X

72

Using this with

H

Y = A, X = i

~

we find

d H i

AH (t) = [i , AH (t)] = [H, AH (t)]

dt ~ ~

and this equation determines the time evolution of the observables in the Heisen-

berg picture.

Applying this to the case of the spin one-half system in a magnetic field, and

taking for our observable S we find

d i eg~

SH (t) = [H, SH (t)] = i [SH (t) B, SH (t)]

dt ~ ~2mc

We know from earlier that the solution will be

for ge|B| B

U (t) = eit 2mc S |B|

and thus the spin vector observable evolves in the Heisenberg picture by rotating

about the magnetic field vector with angular velocity ge|B|

2mc .

space

There is a different approach one can take to characterizing states of a quantum

system with H = C2 . Multiplication of vectors in H by a non-zero complex

number does not change eigenvectors, eigenvalues or expectation values, so ar-

guably has no physical effect. Multiplication by a real scalar just corresponds

to a change in normalization of the state, and we will often use this freedom

to work with normalized states, those satisfying h|i = 1. With normalized

states, one still has the freedom to multiply states by a phase ei without chang-

ing eigenvectors, eigenvalues or expectation values. In terms of group theory, the

overall U (1) in the unitary group U (2) acts on H acts on H by a representation

of U (1), which can be characterized by an integer, the corresponding charge,

but this decouples from the rest of the observables and is not of much interest.

One is mainly interested in the SU (2) part of the U (2), and the observables

that correspond to its Lie algebra.

Working with normalized states in this case corresponds to working with

unit-length vectors in C2 , which are given by points on the unit sphere S 3 . If

we dont care about the overall U (1) action, we can imagine identifying all states

that are related by a phase transformation. Using this equivalence relation we

can define a new set, whose elements are the cosets, elements of S 3 C2 ,

with elements that differ just by multiplication by ei identified. The set of these

73

elements forms a new geometrical space, called the coset space, often written

S 3 /U (1). This structure is called a fibering of S 3 by circles, and is known

as the Hopf fibration. Try an internet search for various visualizations of the

geometrical structure involved, a surprising decomposition of three-dimensional

space into non-intersecting curves.

The same space can be represented in a different way, as C2 /C , by taking

all elements of C2 and identifying those related by muliplication by a non-zero

complex number. If we were just using real numbers, R2 /R can be thought of

as the space of all lines in the plane going through the origin.

One sees that each such line hits the unit circle in two opposite points, so

this set could be parametrized by a semi-circle, identifying the points at the

two ends. This space is given the name RP 1 , the real projective line, and

the analog space of lines through the origin in Rn is called RP n1 . What we

are interested in is the complex analog CP 1 , which is often called the complex

projective line.

To better understand CP 1 , one would like to put coordinates on it. A

standard way to choose such a coordinate is to associate to the vector

z1

C2

z2

drop out in this ratio, so one gets different values for the coordinate z1 /z2 for

each different coset element, and it appears that elements of CP 1 correspond to

74

points on the complex plane. There is however one problem with this coordinate:

the coset of

1

0

does not have a well-defined value: as one approaches this point one moves off

to infinity in the complex plane. In some sense the space CP 1 is the complex

plane, but with a point at infinity added.

point, but as a sphere, with the relation to the plane and the point at infinity

given by stereographic projection. Here one creates a one-to-one mapping by

considering the lines that go from a point on the sphere to the north pole of

the sphere. Such lines will intersect the plane in a point, and give a one-to-one

mapping between points on the plane and points on the sphere, except for the

north pole. Now, one can identify the north pole with the point at infinity,

and thus the space CP 1 can be identified with the space S 2 . The picture looks

like this

and the equations relating coordinates (X1 , X2 , X3 ) on the sphere and the

complex coordinate z1 /z2 = z = x + iy on the plane are given by

X1 X2

x= , y=

1 X3 1 X3

75

and

2x 2y x2 + y 2 1

X1 = , X2 = 2 , X3 = 2

x2 2

+y +1 2

x +y +1 x + y2 + 1

z1 z1

z2 z2

takes

z1 z +

z=

z2 z +

transformations known as Mobius transformations. One can check that the

corresponding transformation on the sphere is the rotation of the sphere in R3

corresponding to this SU (2) = Spin(3) transformation.

In physics language, this sphere CP 1 is known as the Bloch sphere. It

provides a useful parametrization of the states of the qubit system, up to scalar

multiplication, which is supposed to be physically irrelevant. The North pole

is the spin-up state, the South pole is the spin-down state, and along the

equator one finds the two states that have definite values for S1 , as well as the

two that have definite values for S2 .

76

Notice that the inner product on vectors in H does not correspond at all

to the inner product of unit vectors in R3 . The North and South poles of the

Bloch sphere correspond to orthogonal vectors in H, but they are not at all

orthogonal thinking of the corresponding points on the Bloch sphere as vectors

in R3 . Similarly, eigenvectors for S1 and S2 are orthogonal on the Bloch sphere,

but not at all orthogonal in H.

Many of the properties of the Bloch sphere parametrization of states in H are

special to the fact that H = C2 . In the next class we will study systems of spin

n n

2 , where H = C . In these cases there is still a two-dimensional Bloch sphere,

but only certain states in H are parametrized by it. We will see other examples

of systems with coherent states analogous to the states parametrized by the

Bloch sphere, but the case H has the special property that all states (up to

scalar multiplication) are such coherent states.

Just about every quantum mechanics textbook works out this example of a spin

1/2 particle in a magnetic field. For one example, see Chapter 14 of [40].

77

78

Chapter 8

Representations of SU (2)

and SO(3)

state space of any physical three-dimensional quantum system to provide a uni-

tary representation of this group. Remarkably, such physical systems often come

with representations not of SO(3), but of its double-cover SU (2) = Spin(3). For

the case of G = U (1) we were able to classify all irreducible representations by

an element of Z and explicitly construct each irreducible representation. We

would like to do the same thing here for representations of SU (2) and SO(3).

The end result will be that irreducible representations of SU (2) are classified

by a non-negative integer n = 0, 1, 2, 3, , and have dimension n + 1, so well

(hoping for no confusion with the irreducible representations (n , C) of U (1))

denote them (n , Cn+1 ). For even n we will get an irreducible representation

of SO(3), but not for odd n. It is common in physics to label these representa-

tions by s = n2 = 0, 12 , 1, and call the representation labeled by s the spin

s representation. We already know the first three examples:

One irreducible representation is the trivial representation on C. This is (0 , C),

with

0 (g) = 1 g SU (2)

This is also a representation of SO(3). In physics, this is sometimes called the

scalar representation. Saying that something transforms under rotations as

the scalar representation just means that it is invariant under rotations.

1

8.0.2 The spin 2

representation

This is the representation (1 , C2 ) that comes automatically from the definition

of SU (2) as 2 by 2 complex matrices, since these act on C2 . In other words, 1

79

is just the identity map on SU (2) matrices.

the spinor representation.

Since SO(3) is a group of 3 by 3 matrices, it acts on vectors in R3 . One can

act with SO(3) matrices in exactly the same manner on C3 , just by replacing

the real coordinates of vectors by complex coordinates. In other words, the

representation is (, C3 ), with a homomorphism

Recall that we defined a double-covering map : SU (2) SO(3) that is a

homomorphism, taking two points in SU (2) to one point in SO(3). We can use

the representation (, C3 ) of SO(3) to get a representation (2 , C3 ) of SU (2)

by just composing the homomorphisms and :

The same construction gives an SU (2) representation for any SO(3) representa-

tion. As we have seen though with the case of the spinor representation, there

are SU (2) representations that are NOT SO(3) representations.

8.1.1 Weight decomposition

If we pick a subgroup U (1) SU (2), then given any representation (, V ) of

SU (2) of dimension m, we get a representation (|U (1) , V ) of U (1) by restriction

to the U (1) subgroup. Since we know the classification of irreducibles of U (1),

we know that

(|U (1) , V ) = Cq1 Cq2 Cqm

for q1 , q2 , , qm Z, where Cq denotes the one-dimensional representation

of U (1) corresponding to the integer q. . These are called the weights of

the representation V . They are exactly the same thing we discussed earlier as

charges, but here since the U (1) has nothing to do with electromagnetism,

the charge terminology is best avoided.

Since our standard choice of coordinates (the Pauli matrices) picks out the

z-direction and diagonalizes the action of the U (1) subgroup corresponding to

rotation about this axis, this is the U (1) subgroup we will choose to define the

80

weights of the SU (2) representation V . This is the subgroup of elements of

SU (2) of the form i

e 0

0 ei

Our decomposition of an SU (2) representation (, V ) into irreducible repre-

sentations of this U (1) subgroup equivalently means that we can choose a basis

of V so that

iq

i e 1 0 0

eiq2

e 0 0 0

i =

0 e

0 0 eiqm

Theorem. If q is in the set of qi , so is q.

Proof. Recall that if we diagonalize a unitary matrix, the diagonal entries are

the eigenvalues, but their order is undetermined: acting by permutations on

these eigenvalues we get different diagonalizations of the same matrix. In the

case of SU (2) the matrix

0 1

P =

1 0

has the property that conjugation by it permutes the diagonal elements, in

particular i i

e 0 1 e 0

P P =

0 ei 0 ei

So i i

e 0 e 0

(P )( )(P )1 = ( )

0 ei 0 ei

and we see that (P ) gives a change of basis of V such that the representation

matrices on the U (1) subgroup are as before, with . Changing

in the representation matrices is equivalent to changing the sign of the weights

qi . The elements of the set {qi }, i = 1, , m are independent of the basis, the

additional symmetry under sign change implies that for each element in the set

there is another one with the opposite sign.

Looking at our three examples so far, we see that the scalar or spin 0 repre-

sentation of course is one-dimensional of weight 0

(0 , C) = C0

1

and the spinor or spin 2 representation decomposes into U (1) irreducibles of

weights 1, +1:

(1 , C2 ) = C1 C+1

.

81

For the vector or spin 1 representation, recall that our double-cover homo-

morphism takes

i cos 2 sin 2 0

e 0

SU (2) sin 2 cos 2 0 SO(3) U (3)

0 ei

0 0 1

From problem 3 on our first problem set, you can see that the upper left diag-

onal 2 by 2 block acts on C2 with weights 2, +2, whereas the bottom right

element acts trivially on the remaining part of C3 , which is a one-dimensional

representation of weight 0. So, the spin 1 representation decomposes as

(2 , C3 ) = C2 C0 C+2

erators

To proceed further in characterizing a representation (, V ) of SU (2) we need

to use not just the action of the chosen U (1) subgroup, but the action of

group elements in the other two directions away from the identity. The non-

commutativity of the group keeps us from simultaneously diagonalizing those

actions and assigning weights to them. We can however exploit the structure

of the Lie algebra su(2). Recall that the Lie algebra can be thought of as the

tangent space R3 to SU (2) at the identity element, with a basis given by the

three skew-adjoint 2 by 2 matrices

1

Xj = i j

2

which satisfy the commutation relations

morphism

: SU (2) U (m)

and we can take the derivative of this to get a map between the tangent spaces

of SU (2) and of U (m), at the identity of both groups, and thus a Lie algebra

representation

0 : su(2) u(m)

which takes skew-adjoint 2 by 2 matrices to skew-adjoint m by m matrices,

preserving the commutation relations.

82

We have seen that restricting the representation (, V ) to the diagonal U (1)

subgroup of SU (2) and decomposing into irreducibles tells us that we can choose

a basis of V so that

i

i2S3 e 0

e =

0 ei

with going around U (1) once as goes from 0 to 2, this means we can choose

a basis of V so that

iq

e 1 0 0

0 eiq2 0

(ei2S3 ) =

0 0 eiqm

Taking the derivative of this representation to get a Lie algebra representation,

using

d

0 (X) = (eX )|=0

d

we find for X = i2S3

iq

e 1 0 0 iq1 0 0

d iq2

0 e 0 0 iq2 0

0 (i2S3 ) =

=

d

0 0 eiqm |=0 0 0 iqm

another real vector space (u(n), the skew-Hermitian m by m complex matri-

ces). We can use complex linearity to extend any such map to a map from the

complexification of su(2) to the complexification of u(m). The complexification

of su(2) is all complex linear combinations of the skew-adjoint, trace-free 2 by

2 matrices: the Lie algebra sl(2, C) of all complex, trace-free 2 by 2 matrices.

The complexification of u(m) is M (m, C), the Lie algebra of all complex m by

m matrices.

As an example, mutiplying X = i2S3 su(2) by i 2 , we have S3 sl(2, C)

and the diagonal elements in the matrix 0 (i2S3 ) get also multiplied by i

2 (since

0 is a linear map), giving

q1

2 0 0

q2

0 0

0 (S3 ) =

2

0 0 q2m

We see that 0 (S3 ) will have half-integral values, and make the following

definitions

83

Definition (Weights and Weight Spaces). If 0 (S3 ) has an eigenvalue k

2, we

say that k is a weight of the representation (, V ).

The subspace Vk V of the representation V satisfying

k

v Vk = 0 (S3 )v = v

2

is called the kth weight space of the representation. All vectors in it are eigen-

vectors of S3 with eigenvalue k2 .

The dimension dim Vk is called the multiplicity of the weight k in the rep-

resentation (, V ).

One cant diagonalize any combination of S1 and S2 on Vk , but two specific

linear combinations of them do something interesting:

Definition (Raising and lowering operators). Let

0 1 0 0

S+ = S1 + iS2 = , S = S1 iS2 =

0 0 1 0

We have S+ , S sl(2, C). These are neither self-adjoint nor skew-adjoint, but

satisfy

(S ) = S

and similarly we have

0 (S ) = 0 (S )

We call 0 (S+ ) a raising operator for the representation (, V ), and 0 (S )

a lowering operator.

The reason for this terminology is the following calculation:

[S3 , S+ ] = [S3 , S1 + iS2 ] = iS2 + i(iS1 ) = S1 + iS2 = S+

which implies (since 0 is a Lie algebra homomorphism)

0 (S3 ) 0 (S+ ) 0 (S+ ) 0 (S3 ) = 0 ([S3 , S+ ]) = 0 (S+ )

For any v Vk , we have

k

0 (S3 ) 0 (S+ )v = 0 (S+ ) 0 (S3 )v + 0 (S+ )v = ( + 1) 0 (S+ )v

2

so

v Vk = 0 (S+ )v Vk+2

The linear operator 0 (S+ ) takes vectors with a well-defined weight to vectors

with the same weight, plus 2 (thus the terminology raising operator).

A similar calculation shows that 0 (S ) takes Vk to Vk2 , lowering the weight

by 2.

Were now ready to classify all finite dimensional irreducible unitary rep-

resentations (, V ) of SU (2). Finite dimensionality of V implies that if one

repeatedly applies 0 (S+ ) to a vector v V , sooner or later one must get 0

(otherwise one would have an infinite dimensional representation). We define

84

Definition (Highest weights and highest weight vectors). A non-zero vector

v Vn V such that

0 (S+ )v = 0

is called a highest weight vector, with highest weight n.

0 (S ) to v will give new vectors in V , with weights n 2, n 4, . Finite

dimensionality of V implies that this pattern must terminate, at a point where

one reaches a lowest weight vector, one annihilated by 0 (S ). Taking into

account the fact that the pattern of weights is invariant under change of sign

and irreducibility of the representation, one finds that the only possible pattern

of weights is

n, n + 2, , n 2, n

Since any SU (2) element can be conjugated into a diagonal one of the form

i

e 0

0 ei

on this U (1) subgroup, so, up to a change of basis for V , representations are

determined by their pattern of weights. Up to this equivalence, any irreducible

representation of SU (2) will thus be classified by a non-negative integer n, and

will be of dimension n + 1, with weights

n, n + 2, , n 2, n

(, V ) = Cn Cn+2 Cn2 Cn

basis for the representation by repeatedly applying raising or lowering operators.

The picture to keep in mind is this

where all the vector spaces are copies of C, and all the maps are isomor-

phisms (multiplications by various numbers, calculable in principle using the

commutation relations).

85

In summary, we see that all irreducible finite dimensional unitary SU (2)

representations are classified up to equivalence by a non-negative integer, the

highest weight n. These representations have dimension n + 1 and we will

denote them (n , V n = Cn+1 ). Note that Vn is the nth weight space, V n is the

representation with highest weight n. The physicists terminology for this uses

not n, but n2 and calls this number the spinof the representation. We have so

far seen the lowest three examples n = 0, 1, 2, or spin s = n2 = 0, 12 , 1, but there

is an infinite class of larger irreducibles, with dim V = n + 1 = 2s + 1.

The argument of the previous section actually only tells us what the possi-

ble irreducible representations are, in particular it assumes the existence of a

highest-weight vector, without telling us anything about how to construct one.

We need an actual construction of the representation, one that will show exis-

tence of a highest-weight vector, as well as giving us a way to explicitly construct

an irreducible (n , V n ) for each possible highest weight n. There are several pos-

sible constructions, but perhaps the simplest one is the following, which gives

a representation of highest weight n by looking at polynomials in two complex

variables, homogeneous of degree n.

Recall from our early discussion of representations that if one has an action

of a group on a space M , one can get a representation on functions f on M by

taking

((g)f )(x) = f (g 1 x)

acting on column vectors), and we look at a specific class of functions on this

space, the polynomials. We can break up the infinite-dimensional space of

polynomials on C2 into finite-dimensional subspaces as follows:

neous polynomials of degree m in two complex variables z1 , z2 is the space of

functions on C2 of the form

Using the action of SU (2) on C2 , we will see that this space of functions is

exactly the representation space V n that we need Explicitly, for

g= , g 1 =

86

we can construct the representation as follows:

z

(n (g)f )(z1 , z2 ) =f (g 1 1 )

z2

=f (z1 z2 , z1 + z2 )

Xn

= ak (z1 z2 )nk (z1 + z2 )k

k=0

0 d tX d tX z1

n (X)f = n (e )f|t=0 = f (e )

dt dt z2 |t=0

By the chain rule this is

f f d tX z1

n0 (X)f =( , )( e )

z1 z2 dt z2 |t=0

f f

= (X11 z1 + X12 z2 ) (X21 z1 + X22 z2 )

z1 z2

where the Xij are the components of the matrix X.

Computing what happens for X = S3 , S+ , S , we get

1 f f

(n0 (S3 )f )(z1 , z2 ) = ( z1 + z2 )

2 z1 z2

so

1

n0 (S3 ) = (z1 + z2 )

2 z1 z2

and similarly

n0 (S+ ) = z2 , n0 (S ) = z1

z1 z2

The z1k z2nk are eigenvectors for S3 with eigenvalue 12 (n 2k) since

1 1

n0 (S3 )z1k z2nk = (kz1k z2nk + (n k)z1k z2nk ) = (n 2k)z1k z2nk

2 2

z2n will be an explicit highest weight vector for the representation (n , V n ).

An important thing to note here is that the formulas we have found for 0

are not in terms of matrices. Instead we have seen that when we construct our

representations using functions on C2 , for any X su(2) (or its complexification

sl(2, C)), n0 (X) is given by a differential operator. Note that these differential

operators are independent of n: one gets the same operator 0 (X) on all the

V n . This is because the original definition of the representation

((g)f )(x) = f (g 1 x)

this space is infinite-dimensional, issues of analysis dont really come into play

87

here, since polynomial functions are essentially an algebraic construction. Later

on in the course we will need to work with function spaces that require much

more serious consideration of issues in analysis.

Restricting the differential operators 0 (X) to a finite dimensional irreducible

subspace V n , the homogeneous polynomials of degree n, if one chooses a basis

of V n , then the linear operator 0 (X) will be given by a n + 1 by n + 1 matrix.

Clearly though, the expression as a simple first-order differential operator is

much easier to work with. In the examples we will be studying in much of the

rest of the course, the representations under consideration will also be on func-

tion spaces, with Lie algebra representations appearing as differential operators.

Instead of using linear algebra techniques to find eigenvalues and eigenvectors,

the eigenvector equation will be a partial differential equation, and much of the

rest of the course will focus on using Lie groups and their representation theory

to solve such equations.

One issue we havent addressed yet is that of unitarity of the representation.

We need Hermitian inner products on the spaces V n , inner products that will be

preserved by the action of SU (2) that we have defined on these spaces. Up to an

overall scale, there is only one SU (2) invariant inner product on V n (making it

a unitary representation). A standard way to define a Hermitian inner product

on functions on a space M is to define them using an integral: for f , g functions

on M , take their inner product to be

Z

< f, g >= fg

M

is useless for f, g polynomial, since such integrals diverge. What one can do in

this case is define an inner product on polynomial functions on C2 by

Z

1 2 2

< f, g >= 2 f (z1 , z2 )g(z1 , z2 )e(|z1 | +|z2 | ) dx1 dy1 dx2 dy2

C2

Here z1 = x1 + iy1 , z2 = x2 + iy2 . One can do integrals of this kind fairly

easily since they factorize into separate integrals over z1 and z2 , each of which

can be treated using polar coordinates and standard calculus methods. The

polynomials

zj zk

1 2

j!k!

will be an orthornormal basis of the space of polynomial functions with re-

spect to this inner product, and the operators 0 (X), X su(2) will be skew,

Hermitian, something you will be asked to check in the next problem set.

Working out what happens for the first few examples of irreducible SU (2)

representations, one finds orthonormal bases for the representation spaces V n

of homogeneous polynomials as follows

For n = s = 0

1

88

1

For n = 1, s = 2

z1 , z2

For n = 2, s = 1

1 1

z12 , z1 z2 , z22

2 2

3

For n = 3, s = 2

1 1 1 1

z13 , z12 z2 , z1 z22 , z23

6 2 2 6

monics

We would like to now use the classification and construction of representations

of SU (2) to study the representations of the closely related group SO(3). To any

representation (, V ) of SO(3), we can use the double-covering homomorphism

: SU (2) SO(3) to get a representation

=

of SU (2). It can be shown that if is irreducible, will be too, so we must

have = = n , one of the irreducible representations of SU (2) found in

the last section. Using the fact that (1) = 1, we see that

n (1) = (1) = 1

From knowing that the weights of n are n, n + 2, , n 2, n, we know that

in

i e 0 0

e 0 0 ei(n2) 0

n (1) = n = =1

0 ei

0 0 ein

which will only be true for n even, not for n odd. Since the Lie algebra of SO(3)

is isomorphic to the Lie algebra of SU (2), the same Lie algebra argument using

raising and lowering operators as in the last section also applies. The irreducible

representations of SO(3) will be (l , V = C2l+1 ) for l = 0, 1, 2, , of dimension

2l + 1 and satisfying

l = 2l

Just like in the case of SU (2), we can explicitly construct these representa-

tions using functions on a space with an SO(3) action. The obvious space to

choose is R3 , with SO(3) matrices acting on x R3 as column vectors, by the

formula we have repeatedly used

x1

((g)f )(x) = f (g 1 x) = f (g 1 x2 )

x3

89

Taking the derivative, the Lie algebra representation is given by

x1

d d

0 (X)f = (etX )f|t=0 = f (etX x2 )|t=0

dt dt

x3

0 0 0 0 0 1 0 1 0

l1 = 0 0 1 l2 = 0 0 0 l3 = 1 0 0

0 1 0 1 0 0 0 0 0

Im using the notation lj for the real basis of the Lie algebra so(3) = su(2).

For a unitary representation , the 0 (lj ) will be skew-Hermitian linear oper-

ators. For consistency with the physics literature, Ill use the notation Lj =

i0 (lj ) for the self-adjoint version of the linear operator corresponding to lj in

this representation on functions. The Lj satisfy the commutation relations

Well also use elements l = l1 il2 of the complexified Lie algebra to create

raising and lowering operators L = i0 (l ).

As with the SU (2) case, I wont include a factor of ~ as is usual in physics

(e.g. the usual convention is La = i~0 (la )), since for considerations of action

of the rotation group it would just cancel out (physicists define rotations using

i

e ~ La ). The factor of ~ is only of significance when one wants to relate things

to the momentum operator, something that will come up later in the course.

In the SU (2) case, the 0 (Sj ) had half-integral eigenvalues, with the eigen-

values of 0 (2S3 ) the integral weights of the representation. Here the La will

have integer eigenvalues, the weights will be the eigenvalues of 2L3 , which will

be even integers.

90

0 0

0

1

0 0

t x1

d 0 1

0 x )

0 (l1 )f = f (e 2 |t=0

dt

x3

0 0 0 x1

d

= f (0 cos t sin t x2 )|t=0

dt

0 sin t cos t x3

0

d

= f ( x2 cos t + x3 sin t )|t=0

dt

x2 sin t + x3 cos t

0

f f f

=( , , ) x3

x1 x2 x3

x2

f f

=x3 x2

x2 x3

so

0 (l1 ) = x3 x2

x2 x3

and similar calculations give

0 (l2 ) = x1 x3 , 0 (l3 ) = x2 x1

x3 x1 x1 x2

The space of all functions on R3 is much too big: it will give us an infinity of

copies of each finite dimensional representation that we want. Notice that when

SO(3) acts on R3 , it leaves the distance to the origin invariant. If we work in

spherical coordinates (r, , ) (see picture)

91

we will have

x1 =r sin cos

x2 =r sin sin

x3 =r cos

, . It turns out that we can cut down the space of functions to something that

will only contain one copy of the representation we want in various ways. One

way to do this is to restrict our functions to the unit sphere, i.e. just look at

functions f (, ). We will see that the representations we are looking for can

be found in simple trigonmetric functions of these two angular variables.

We can construct our irreducible representations l by explicitly constructing

a function we will call Yll (, ) that will be a highest weight vector of weight

l. The weight l condition and the highest weight condition give two differential

equations for Yll (, ):

L3 Yll = lYll , L+ Yll = 0

These will turn out to have a unique solution (up to scalars).

We first need to change coordinates from rectangular to spherical in our

expressions for L3 , L . Using the chain rule to compute expressions like

f (x1 (r, , ), x2 (r, , ), x3 (r, , ))

r

we find

r sin cos sin sin cos x1

= r cos cos r cos sin sin x 2

r sin sin r sin cos 0 x 3

so

r sin cos sin sin cos x1

1 = cos cos cos sin

sin x

r 2

1

r sin sin cos 0 x

3

This is an orthogonal matrix, so one can invert it by taking its transpose, to get

x1 sin cos cos cos sin r

= sin sin cos sin cos 1

x2 r

1

x

cos sin 0 r sin

3

So we finally have

L1 = i0 (l1 ) = i(x3 x2 ) = i(sin + cot cos )

x2 x3

L2 = i0 (l2 ) = i(x1 x3 ) = i( cos + cot sin )

x3 x1

92

L3 = i0 (l3 ) = i(x1 x3 ) = i

x3 x1

and

L+ = i0 (l+ ) = ei ( + i cot ), L = i0 (l ) = ei ( + i cot )

Now that we have expressions for the action of the Lie algebra on functions in

spherical coordinates, our two differential equations saying our function Yll (, )

is of weight l and in the highest-weight space are

l

L3 Yll (, ) = i Y (, ) = lYll (, )

l

and

L+ Yll (, ) = ei ( + i cot )Yll (, ) = 0

The first of these tells us that

Yll (, ) = eil Fl ()

( l cot )Fl ()

with solution

Fl () = Cll sinl

for arbitrary constant Cll . Finally

2l + 1 dimensional irreducible representation of SO(3). To get functions which

give vectors spanning the rest of the weight spaces, one just repeatedly applies

the lowering operator L , getting functions

=Clm (ei ( + i cot ))lm eil sinl

for m = l, l 1, l 2 , l + 1, l

The functions Ylm (, ) are called spherical harmonics, and they span the

space of complex functions on the sphere in much the same way that the ein

span the space of complex valued functions on the circle (or, equivalently, with

period 2). Unlike the case of polynomials on C2 , for functions on the sphere,

one gets finite numbers by integrating such functions of the sphere, so can define

93

an inner product on these representations for which they are unitary by simply

setting

Z Z 2 Z

< f, g >= f g sin dd = f (, )g(, ) sin dd

S2 =0 =0

The Ylm (, ) are all orthogonal with respect to this inner product. One can de-

rive various general formulas for the Ylm (, ) in terms of Legendre polynomials,

but here well just compute the first few examples, with the proper constants

that give them norm 1 with respect to the chosen inner product.

For the l = 0 representation

r

0 1

Y0 (, ) =

4

For the l = 1 representation

r r r

3 3 1 3

1

Y1 = i 0

sin e , Y1 = cos , Y1 = sin ei

8 4 8

(one can easily see that these have the correct eigenvalues for 0 (L3 ) = i

).

For the l = 2 representation one has

r r

2 15 2 i2 1 15

Y2 = sin e , Y2 = sin cos ei

32 8

r

0 5

Y2 = (3 cos2 1)

16

r r

1 15 i 2 15

Y2 = sin cos e , Y2 = sin2 ei2

8 32

We will see later that these functions of the angular variables in spherical

coordinates are exactly the functions that give the angular dependence of wave

functions for the physical system of a particle in a spherically symmetric po-

tential. In such a case the SO(3) symmetry of the system implies that state

space (the wave functions) will provide a unitary representation of SO(3), and

the action of the Hamiltonian operator H will commute with the action of the

operators L3 , L . As a result all of the states in an irreducible representation

component of will have the same energy. States are thus organized into or-

bitals, single states called s orbitals (l = 0), triplet states called p orbitals

(l = 1), multiplicity 5 states called d orbitals (l = 2), etc.

For both SU (2) and SO(3), we have found that all representations can be

constructed out of function spaces, with the Lie algebra acting as first-order

differential operators. It turns out that there is also a very interesting second-

order differential operator that comes from these Lie algebra representations,

known as the Casimir operator. For the case of SO(3)

94

Definition (Casimir operator for SO(3)). The Casimir operator for the repre-

sentation of SO(3) on functions on S 2 is the second-order differential operator

[L2 , 0 (X)] = 0

for any X so(3). Knowing this, a version of Schurs lemma says that L2 will act

on an irreducible representation as a scalar (i.e. all vectors in the representation

are eigenvectors of L2 , with the same eigenvalue). This eigenvalue can be used

to characterize the irreducible representation.

The easiest way to compute this eigenvalue turns out to be to act with L2 on

a highest weight vector. First one rewrites L2 in terms of raising and lowering

operators (note that the following calculation is like the first expression for L2

above and needs reinterpretation as an equation for operators in an arbitrary

representation, but this is not important for the calculation we are doing).

=L21 + L22 + i[L1 , L2 ]

=L21 + L22 L3

so

L2 = L21 + L22 + L23 = L L+ + L3 + L23

For the representation of SO(3) on functions on S 2 constructed above,

we know that on a highest weight vector of the irreducible representation l

(restriction of to the 2l + 1 dimensional irreducible subspace of functions that

are linear combinations of the Ylm (, )), we have the two eigenvalue equations

L+ f = 0, L3 f = lf

with solution the functions proportional to Yll (, ). Just from these conditions

and our expression for L2 we can immediately find the scalar eigenvalue of L2

since

L2 f = L L+ f + (L3 + L23 )f = 0 + l + l2 = l(l + 1)

We have thus shown that our irreducible representation l can be characterized

as the representation in which L2 acts by the scalar l(l + 1).

In summary, we have two different sets of partial differential equations whose

solutions give the irreducible representation l :

L+ f = 0, L3 f = lf

which are first order equations, with the first using complexification and some-

thing like a Cauchy-Riemann equation, and

L2 f = l(l + 1), L3 f = lf

95

where the first equation is a second order equation, something like a Laplace

equation.

One can compute the explicit second order differential operator L2 in the

representation on functions, it is

=(i(sin + cot cos ))2 + (i( cos + cot sin ))2 + (i )2

1 1 2

=( (sin ) + )

sin sin2 2

We will re-encounter this operator later on in the course as the angular part of

the Laplace operator on R3 .

For the group SU (2) we can go through the same procedure, using the

representation on functions on C2 . In that case, the differential equation

point of view is less useful, since the solutions we are looking for are just the

homogeneous polynomials, which are easily found and manipulated. In the

problem set, youll be asked to find an expression for the Casimir operator

in this representation and check that it behaves in the same way as in the SO(3)

case.

The classification of SU (2) representations is a standard topic in all textbooks

that deal with Lie group representations. A good example is [25], which covers

this material well, and from which the discussion here of the construction of

representations as homogeneous polynomials is drawn (see pages 77-79). The

calculation of the Lj and the derivation of expressions for spherical harmonics

as Lie algebra representations of so(3) appears in most quantum mechanics

textbooks in one form or another (for example, see Chapter 12 of [40]). Another

source used here for the explicit constructions of representations is [12], Chapters

27-30.

96

Chapter 9

Tensor Products,

Entanglement, and

Addition of Spin

If one has two independent quantum systems, with state spaces H1 and H2 ,

the combined quantum system has a description that exploits the mathematical

notion of a tensor product, with the combined state space the tensor product

H1 H2 . Because of the ability to take linear combinations of states, this

combined state space will contain much more than just products of independent

states, including states that are described as entangled, and responsible for

some of the most counter-intuitive behavior of quantum physical system.

This same tensor product construction is a basic one in representation the-

ory, allowing one to construct a new representation (1 2 , W1 W2 ) out of

representations (1 , W1 ) and (2 , W2 ). When we take the tensor product of

states corresponding to two irreducible representations of SU (2) of spins s1 , s2 ,

we will get a new representation (2s1 2s2 , V 2s1 V 2s2 ). It will be reducible,

a direct sum of representations of various spins, a situation we will analyze in

detail.

Starting with a quantum system with state space H that describes a single

particle, one can describe a system of N particles by taking an N -fold tensor

product HN = H H H. A deep fact about the physical world

is that for identical particles, we dont get the full tensor product space, but

only the subspaces either symmetric or antisymmetric under the action of the

permutation group by permutations of the factors, depending on whether our

particles are bosons or fermions. An even deeper fact is that elementary

particles of half-integral spin s must behave as fermions, those of integral spin,

bosons.

Digression. When physicists refer to tensors, they generally mean the ten-

sor fields used in general relativity or other geometry-based parts of physics.

97

The relation to what is being discussed here is only tangential (a pun I realize

after writing that...) A tensor field is a function on a manifold, taking values

in some tensor product of copies of the tangent space and and its dual space.

The simplest tensor fields are just vector fields, functions taking values in the

tangent space. A more non-trivial example is the metric tensor, which takes

values in the dual of the tensor product of two copies of the tangent space.

Given two vector spaces V and W , (assume over C, but one could instead take

R or any other field) one can easily construct the direct sum vector space V W ,

just by taking pairs of elements (v, w) for v V, w W , and giving them a

vector space structure by the obvious addition and multiplication by scalars.

This space will have dimension

dim(V W ) = dim V + dim W

If {e1 , e2 , . . . , edim V } are a basis of V , and {f1 , f2 , . . . , fdim W } a basis of W , the

{(e1 , 0), (e2 , 0), . . . , (edim V , 0), (0, f1 ), (0, f2 ), . . . , (0, fdim W )}

will be a basis of V W .

A much less obvious construction is the tensor product of the vector spaces

V and W . This will be a new vector space called V W , of dimension

dim(V W ) = (dim V )(dim W )

One way to motivate the tensor product is to think of vector spaces as vector

spaces of functions. Elements

v = v1 e1 + v2 e2 + + vdim V edim V V

can be thought of as functions on the dim V points ei , taking values vi at ei . If

one takes functions on the union of the sets {ei } and {fj } one gets elements of

V W . The tensor product V W will be what one gets by taking all functions

on not the union, but the product of the sets {ei } and {fj }. This will be the

set with (dim V )(dim W ) elements, which we will write ei fj , and elements

of V W will be functions on this set, or equivalently, linear combinations of

these basis vectors.

This sort of definition is less than satisfactory, since it is tied to an explicit

choice of bases for V and W . We wont however pursue more details of this

question or a better definition here. For this, one can consult pretty much any

advanced undergraduate text in abstract algebra, but here we will take as given

the following properties of the tensor product that we will need:

Given vectors v V, w W we get an element v w V W , satisfying

bilinearity conditions (c1 , c2 C)

v (c1 w1 + c2 w2 ) = c1 (v w1 ) + c2 (v w2 )

(c1 v1 + c2 v2 ) w = c1 (v1 w) + c2 (v2 w)

98

There are natural isomorphisms

C V ' V, V W ' W V

and

U (V W ) ' (U V ) W

for vector spaces U, V, W

Given a linear operator A on V and another linear operator B on W , we

can define a linear operator A B on V W by

(A B)(v w) = Av Bw

for v V, w W .

With respect to the bases ei , fj of V and W , A will be a (dim V ) by (dim V )

matrix, B will be a (dim W ) by (dim W ) and A B will be a (dim V )(dim W )

by (dim V )(dim W ) matrix (which one can think of as a (dim V ) by (dim V )

matrix of blocks of size (dim W )).

One often wants to consider tensor products of vector spaces and dual vector

spaces. An important fact is that there is an isomorphism between the tensor

product V W and linear maps from V to W given by identifying lw (l V )

with the linear map

v V l(v)w W

ucts

Consider two quantum systems, one defined by a state space H1 and a set of

operators O1 on it, the second given by a state space H2 and set of operators O2 .

One can describe the composite quantum system corresponding to considering

the two quantum systems as a single one, with no interaction between them, by

just taking as a new state space

HT = H1 H 2

A Id + Id B

the state space HT , but with a more general class of operators.

If H is the state space of a quantum system, we can think of this as describing

a single particle, and then to describe a system of N such particles, one uses

the multiple tensor product

HN = H H H H

| {z }

N times

99

The symmetric group SN acts on this state space, and one has a repre-

sentation (, HN ) of SN as follows. For SN a permutation of the set

{1, 2, . . . , N } of N elements, with || the minimal number of transpositions that

combine to give , on a tensor product of vectors one has

various components with different irreducible representations of the group SN .

A fundamental axiom of quantum mechanics is that if HN describes N iden-

tical particles, then all physical states occur as one-dimensional representations

of SN , which are either symmetric (bosons) or anti-symmetric (fermions)

where

Definition. A state v HN is called

symmetric, or bosonic if SN

()v = v

anti-symmetric, or fermionic if SN

()v = (1)|| v

Note that in the fermionic case, one cannot have non-zero states in HN

describing two identical particles in the same state w H, a fact that is known

as the Pauli Principle. This has to be the case, since for a transposition

interchanging the two particles, the antisymetric representation acts on the

factor H H by interchanging vectors, taking

ww HH

to itself, while the antisymmetry requires that this state go to its negative, so

the state cannot be non-zero.

Next semester we will see that when one uses quantum fields to describe

systems of identical particles, then this simple behavior under the symmetric

group occurs naturally, but within the purely quantum mechanical formalism it

is a separate physical axiom.

If one is given a function f on a space X and a function g on a space Y , one

can form a product function f g on the product space X Y by taking (for

x X, y Y )

(f g)(x, y) = f (x)g(y)

However, most functions on X Y are not decomposable in this manner. Sim-

ilarly, for a tensor product of vector spaces, one has:

100

Definition (Decomposable and indecomposable vectors). A vector in V W

is called decomposable if it is of the form v w for some v V, w W . If it

cannot be put in this form it is called indecomposable.

Note that our basis vectors of V W are all decomposable since they are

products of basis vectors of V and W . Linear combinations of these basis vectors

however are in general indecomposable. If we think of an element of V W

as a dim V by dim W matrix, with entries the coordinates with respect to our

basis vectors for V W , then for decomposable vectors we get a special class

of matrices, those of rank one.

In the physics context, the language used is:

state space HT = H1 H2 is called an entangled state.

ing and subtle aspects of quantum mechanical system. The Einstein-Podolsky-

Rosen paradox concerns the behavior of an entangled state of two quantum

systems, when one moves them far apart. Then performing a measurement on

one system can give one information about what will happen if you perform

a measurement on the far removed system, introducing a sort of unexpected

non-locality. Measurement theory itself involves crucially an entanglement be-

tween the state of a system being measured, and the state of the measurement

apparatus, thought of as a quantum system in its own right. For much more

about this, a recommended reading is Chapter 2 of [38].

Given two representations of a group, one can define a new representation, the

tensor product representation, by

representations of a group G, one has a tensor product representation (V W , V

W ) defined by

(V W (g))(v w) = V (g)v W (g)w

To see what happens for the corresponding Lie algebra representation, one

computes (for X in the Lie algebra)

d

V0 W (X)(v w) = V W (etX )(v w)t=0

dt

d

= (V (etX )v W (etX )w)t=0

dt

d d

=(( V (etX )v) W (etX )w)t=0 + (V (etX )v ( W (etX )w))t=0

dt dt

=(V0 (X)v) w + v (W 0

(X)w)

101

which could also be written

0

(X))

Given two representations (V , V ) and (W , W ) of a group G, we can decom-

pose each into irreducibles. To do the same for the tensor product of the two

representations, we need to know how to decompose the tensor product of two

irreducibles. This is a fundamental non-trivial problem for a group G, with the

answer for G = SU (2) as follows:

V n2 ) decomposes into irreducibles as

Proof. Outline

One way to prove this result is to use highest-weight theory, raising and

lowering operators, and the formula for the Casimir operator. We will not try

and show the details of how this works out, but a little later use a very different

and quicker argument using characters. However, in outline (for more details,

see for instance section 5.2 of [36]), heres how one could proceed:

One starts by noting that if vn1 Vn1 , vn2 Vn2 are highest weight vectors

for the two representations, vn1 vn2 will be a highest weight vector in the tensor

product representation (i.e. annihilated by n0 1 +n2 (S+ )), of weight n1 + n2 .

So (n1 +n2 , V n1 +n2 ) will occur in the decomposition. Applying n0 1 +n2 (S ) to

vn1 vn2 one gets a basis of the rest of the vectors in (n1 +n2 , V n1 +n2 ). However,

at weight n1 +n2 2 one can find another kind of vector, a highest-weight vector

orthogonal to the vectors in (n1 +n2 , V n1 +n2 ). Applying the lowering operator

to this gives (n1 +n2 2 , V n1 +n2 2 ). As before, at weight n1 + n2 4 one find

another, orthogonal highest weight vector, and gets another representation, with

this process only terminating at weight |n1 n2 |.

A much more straightforward proof using character theory is given in the

next section.

A standard tool for dealing with representations that we have ignored so far is

that of associating to a representation an invariant called its character. This

will be a conjugation-invariant function on the group that only depends on the

equivalence class of the representation. Given two representations constructed

in very different ways, one can often check whether they are isomorphic just

by seeing if their character functions match. The problem of identifying the

possible irreducible representations of a group can be attacked by analyzing

the possible character functions of irreducible representations. We will not try

and get into the general theory of characters here, but will just see what the

102

characters of irreducible representations are for the case of G = SU (2), and use

this to get a simple argument for the Clebsch-Gordan decomposition of the ten-

sor product of SU (2) representations. For this we dont need general theorems

about the relations of characters and representations, but can directly check

that the irreducible representations of SU (2) correspond to distinct character

functions which are easily evaluated.

Definition (Character). The character of a representation (, V ) of a group G

is the function on G given by

V (g) = T r((g))

be a complex valued, conjugation-invariant function on G. One can easily check

that it will satisfy the relations

V W = V + W , V W = V W

subgroup U (1) of diagonal matrices. Knowing the weights of the irreducible

representations (n , V n ) of SU (2), we know the characters to be the functions

i

e 0

V n ( ) = ein + ei(n2) + + ei(n2) + ein

0 ei

Theorem (Weyl character formula).

ei(n+1) ei(n+1)

V n =

ei ei

Proof.

character of the tensor product using the Weyl character formula for the second

factor (ordering things so that n2 > n1 )

V n1 V n2 =V n1 V n2

ei(n2 +1) ei(n2 +1)

=(ein1 + ei(n1 2) + + ei(n1 2) + ein1 )

ei ei

i(n1 +n2 +1) i(n2 n1 +1)

(e i(n1 +n2 +1)

e ) + + (e ei(n2 n1 +1) )

=

ei ei

=V n1 +n2 + V n1 +n2 2 + + V n2 n1

103

9.4.3 Some examples

Some simple examples of how this works are:

V1V1 =V2V0

This says that the four complex dimensional tensor product of two spinor

representations (which are each two complex dimensional) decomposes

into irreducibles as the sum of a three dimensional vector representation

and a one dimensional trivial (scalar) representation.

1 0

Using the basis , for V 1 , the tensor product V 1 V 1 has a basis

0 1

1 1 1 0 0 1 0 0

, , ,

0 0 0 1 1 0 1 1

The vector

1 1 0 0 1

( )V1V1

2 0 1 1 0

One can show (see problem set) that this vector is invariant under SU (2),

by computing either the action of SU (2) or of its Lie algebra su(2). So,

this vector is a basis for the component V 0 in the decomposition into

irreducibles

V1V1 =V2V0

1 1 1 1 0 0 1 0 0

, ( + ),

0 0 2 0 1 1 0 1 1

These three terms are of weights 2, 0, 2 under the U (1) SU (2) gener-

ated by s3 = i3 su(2). They are symmetric under permutation of the

two factors of V 1 V 1 .

We see that if we take two identical quantum systems with H = V 1 = C2

and make a composite system out of them, if they were bosons we would

get a three dimensional state space, transforming as a vector (spin one)

under SU (2). If they were fermions, we would get a one-dimensional

system of spin zero (invariant under SU (2)). Note that in this second

case we automatically get an entangled state, one that cannot be written

as a decomposable product.

The spin-statistics relation says that if these really correspond to identical

spin-half particles, we must be in the fermionic case. This however only

104

applies to the much more complicated situation of special relativity and

particles with space-time degrees of freedom, which we will come to later

in this course. The bosonic case also occurs in particle physics, when the

SU (2) symmetry group is not the double cover of the group of rotations

in space, but an internal symmetry group acting on elementary particle

labels (in which case it is called isospin). Pions, which in some sense are

composites of two isospin one-half quarks, come in an isospin one triplet,

with names + , 0 , .

Tensor product of three or more spinors:

V 1 V 1 V 1 = (V 2 V 0 ) V 1 = (V 2 V 1 ) (V 0 V 1 ) = V 3 V 1 V 1

This says that the tensor product of three spinor representations decom-

poses as a four dimensional (spin 3/2) representation plus two copies of

the spinor representation.

One can clearly generalize this and consider N -fold tensor products (V 1 )N

of the spinor representation. Taking N high enough one can get any ir-

reducible representation of SU (2) that one wants this way, giving an al-

ternative to our construction using homogeneous polynomials. Doing this

however gives the irreducible as just one component of something larger,

and one needs a method to project out the component one wants. One

can do this using the action of the symmetric group SN on (V 1 )N and

an understanding of the irreducible representations of SN . This relation-

ship between irreducible representations of SU (2) and those of SN coming

from looking at how both groups act on (V 1 )N is known as Schur-Weyl

duality, and generalizes to the case of SU (n), where one looks at N -fold

tensor products of the defining representation of SU (n) matrices on Cn .

For SU (n) this provides perhaps the most straight-forward construction

of all irreducible representations of the group.

For more about the tensor product and tensor product of representations, see

section 6 of [47], or appendix B of [42]. Almost every quantum mechanics text-

book will contain an extensive discussion of the Clebsch-Gordan decomposition.

105

106

Chapter 10

the Quantum Free Particle

Well now turn to the problem that conventional quantum mechanics courses

generally begin with: that of the quantum system describing a free particle

moving in physical space R3 . The state space H will be a space of complex-

valued functions on R3 , called wave-functions. There is one crucial sort of

observable to understand: the momentum operator. This operator will have

the same relationship to spatial translations as the Hamiltonian does to time

translations. In both cases, the operators are given by the Lie algebra represen-

tation corresponding to the unitary representation on the quantum state space

of groups of translations (translation in the three space and one time directions

respectively).

One way to motivate the quantum theory of a particle is that, whatever it

is, it should have the same sort of behavior as in the classical case under the

translational and rotational symmetries of space-time. Invoking the classical

relationship between energy and momentum used in non-relativistic mechan-

ics relates the Hamiltonian and momentum operators, giving the conventional

Schr odinger differential equation for the wave-function of a free particle. We

will examine the solutions to this equation, beginning with the case of periodic

boundary conditions, where spatial translations in each direction are given by

the compact group U (1) (whose representations we have studied in detail).

lations

We have seen that it is a basic axiom of quantum mechanics that the observ-

able operator responsible for infinitesimal time translations is the Hamiltonian

107

operator H, a fact that is expressed as

d

i~ |i = H|i

dt

When H is time-independent, one can understand this equation as reflecting the

existence of a unitary representation (U (t), H) of the group R of time transla-

tions on the state space H. U (t) is a homomorphism from the group R to the

group of unitary transformations of H. It is of the form

i

U (t) = e ~ tH

the group R of time-translations on the state, which means knowing what U (t)

is. We get a Lie algebra representation of R by taking the time derivative of

U (t), which gives us

d i

U (t)|t=0 = H

dt ~

As the Lie algebra representation coming from taking the derivative of a unitary

representation, ~i H will be skew-adjoint, so H will be self-adjoint. The minus

sign is a convention , and the factor of ~ depends on the choice of units of time

and energy. Once one has chosen a unit of either time or energy, it would be

natural to choose the other unit so as to make ~ = 1.

If we have a group R and an action of this group on the space M = R by

translations, with a R acting on x M by

xax=x+a

(g)f (x) = f (g 1 x)

we get

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

The Lie algebra representation will just be the derivative of this

d

0 (1)f = f (x ) =0 = f 0 (x)

d

Note that here the Lie algebra and the Lie group are not related by the same

sort of exponentiation used for matrix groups (where the group action was mul-

tiplicative, not additive like here). For a case where this kind of translation can

be understood using matrix multiplication and the usual exponential map, see

the next section. We wont try and more carefully define the notion of expo-

nentiation used here, but just note that for sufficiently well-behaved functions,

Taylors formula tells us that one gets finite translations by exponentiating the

differentiation operator

df a2 d2 f d

f (x + a) = f (x) + a + 2

+ = ea dx f (x)

dx 2! dx

108

d

In this sense dx is the infinitesimal generator of spatial translations and the Lie

algebra representation of R on functions on R is given by the linear operator

d

dx .

Since we now want to describe quantum systems that depend not just on

time, but on space variables q = (q1 , q2 , q3 ), we will have an action by unitary

transformations of not just the group R of time translations, but also the group

R3 of spatial translations. We will define the corresponding Lie algebra rep-

resentations using self-adjoint operators P1 , P2 , P3 that play the same role for

spatial translations that the Hamiltonian plays for time translations:

Definition (Momentum operators). For a quantum system with state space

H given by complex valued functions of position variables q1 , q2 , q3 , momentum

operators P1 , P2 , P3 are defined by

P1 = i~ , P2 = i~ , P3 = i~

q1 q2 q3

These are given the name momentum operators since we will see that their

eigenvalues have an intepretation as the components of the momentum vector

for the system, just as the eigenvalues of the Hamiltonian have an interpretation

as the energy. Note that while in the case of the Hamiltonian the factor of ~ kept

track of the relative normalization of energy and time units, here it plays the

same role for momentum and length units. It can be set to one if appropriate

choices of units are made.

The differentiation operator is skew-adjoint since, using integration by parts

one has for H

Z + Z + Z +

d d d d

( )dq = ( () ( ))dq = ( )dq

dq dq dq dq

The factor of i in the definition of the Pj is just the same convention used earlier

to multiply the skew-adjoint operator of a unitary Lie algebra representation by

i to get the self-adjoint operator favored by physicists.

Note that the convention for the sign choice here is the opposite from the case

of the Hamiltonian. This means that the sign choice for the Hamiltonian makes

it minus the generator of translations in the time direction. The reason for this

comes from considerations of special relativity, where a relative sign appears in

the treatment of space and time dimensions. We will review this subject in a

later chapter but for now we just need the relationship special relativity gives

between energy and momentum. Space and time are put together in Minkowski

space, which is R4 with indefinite inner product

Energy and momentum are the components of a Minkowski space vector (p0 =

E, p1 , p2 , p3 ) with norm-squared given by the mass-squared:

109

This is the formula for a choice of space and time units such that the speed of

light is 1. Putting in factors of the speed of light c to get the units right one

has

E 2 |p|2 c2 = m2 c4

Two special cases of this are:

For photons, m = 0, and one has the energy momentum relation E = |p|c

For velocities v small compared to c (and thus momenta |p| small com-

pared to mc), one has

p p c|p|2 |p|2

E= |p|2 c2 + m2 c4 = c |p|2 + m2 c2 =

2mc 2m

This is the non-relativistic limit, and we use this energy-momentum rela-

tion to describe particles with velocities small compared to c.

well as other possible ways of constructing quantum systems for relativistic

particles. For now though, we will stick to the non-relativistic case. To describe

a quantum non-relativistic particle, we write the Hamiltonian operator in terms

of the components of the momentum operator, in such a way that the eigenvalues

of the operators (the energy and momentum) will satisfy the classical equation

2

E = |p|

2m :

1 1 ~2 2 2 2

H= (P12 + P22 + P32 ) = |P|2 = ( 2 + 2 + 2)

2m 2m 2m q1 q2 q3

The Schr

odinger equation then becomes:

~2 2 2 2 ~2 2

i~ (q, t) = ( 2 + 2 + 2 )(x, t) = (q, t)

t 2m q1 q2 q3 2m

ential equation. One method of solution is to separate out the time-dependence,

by first finding solutions E to the time-independent equation

~2 2

HE (q) = E (q) = EE (q)

2m

with eigenvalue E for the Hamiltonian operator and then using the fact that

i

(q, t) = E (q)e ~ tE

i~ (q, t) = H(q, t)

t

110

The solutions E (q) to the time-independent equation are just complex ex-

ponentials proportional to

satisfying

~2 ~2 |k|2

(i)2 |k|2 = =E

2m 2m

We see that the solutions to the Schrodinger equation are all linear combinations

of states |ki labeled by a vector k which are eigenstates of the momentum and

Hamiltonian operators with

~2

Pa |ki = ~ka |ki, H|ki = |k|2 |ki

2m

These are states with well-defined momentum and energy

|p|2

pa = ~ka , E =

2m

so satisfy exactly the same energy momentum relations as those for a classical

non-relativistic particle.

While the quantum mechanical state space H contains states with the clas-

sical energy-momentum relation, it also contains much, much more. A general

state will be a linear combination of such states. At t = 0 one has

X

|i = ck eikq

k

where ck are complex numbers, and the general time-dependent state will be

X |k|2

|(t)i = ck eikq eit~ 2m

k

2

i i |p|

X

|(t)i = cp e ~ pq e ~ 2m t

U (1)

We have not yet discussed the inner product on our space of states when they

are given as wave-functions on R3 , and there is a significant problem with doing

this. To get unitary representations of translations, we need to use a translation

invariant, Hermitian inner product on wave-functions, and this will have to be

of the form Z

< 1 , 2 >= C 1 (q)2 (q)dq

R3

111

for some constant C. But if we try and compute the norm-squared of one of

our basis states |ki we find

Z Z

hk|ki = C (eikq )(eikq )dq = C 1 dq =

R3 R3

As a result, while our basis states |ki are orthogonal for different values of k,

there is no value of C which will make them of norm-squared one.

In the finite dimensional case, a linear algebra theorem assure us that given a

self-adjoint operator, we can find an orthonormal basis of its eigenvectors. In this

infinite dimensional case this is no longer true, and a much more sophisticated

formalism (the spectral theorem for self-adjoint operators) is needed to replace

the linear algebra theorem. This is a standard topic in treatments of quantum

mechanics aimed at mathematicians emphasizing analysis, but we will not try

and enter into this here. One place to find such a discussion is section 2.1 of

[45].

One way to deal with the normalization problem is to replace the non-

compact space by one of finite volume. Well show how this works for the

simplified case of a single spatial dimension, since once one sees how this works

for one dimension, treating the others the same way is straight-forward. In this

one dimensional case, one replaces R by the circle S 1 . This is equivalent to

the physicists method of imposing periodic boundary conditions, meaning

to define the theory on an interval, and then identify the ends of the interval.

One can then think of the position variable q as an angle and define the inner

product as Z 2

1

< 1 , 2 >= 1 ()2 ()d

2 0

and take as state space

H = L2 (S 1 )

the space of complex-valued square-integrable functions on the circle.

Instead of the translation group R, we have the standard action of the

group SO(2) on the circle. Elements g() of the group are rotations of the circle

counterclockwise by an angle , or if we parametrize the circle by an angle ,

just shifts

+

Recall that in general we can construct a representation on functions from a

group action on a space by

(g)f (x) = f (g 1 x)

(g())() = ( )

taking

the circle as the unit

2 0 1

circle in R , rotations 2 by 2 matrices, X = , g() = eX ) then the

1 0

112

Lie algebra representation is given by taking the derivative

d

0 (X)f () = f ( )|=0 = f 0 ()

d

so we have

d

0 (X) =

d

This operator is defined on a dense subspace of H = L2 (S 1 ) and is skew-adjoint,

since (using integration by parts)

Z 2

1 d

< 1 , 20 >= 1 2 d

2 0 d

Z 2

1 d d

= ( (1 2 ) ( 1 )2 )d

2 0 d d

= < 10 , 2 >

The eigenfunctions of 0 (X) are just the ein , for n Z, which we will also

write as state vectors |ni. These are also a basis for the space L2 (S 1 ), a basis

that corresponds to the decomposition into irreducibles of

L2 (S 1 )

(, L2 (S 1 )) = nZ (n , C)

n (g()) = ein

The theory of Fourier series for functions on S 1 says that one can expand any

function L2 (S 1 ) in terms of this basis, i.e.

+

X +

X

|i = () = cn ein = cn |ni

n= n=

+

X

|cn |2 <

n=

Z 2

1

cn = hn|i = ein ()d

2 0

113

The Lie algebra of the group S 1 is the same as that of the group (R, +),

and the 0 (X) we have found for the S 1 action on functions is related to the

momentum operator in the same way as in the R case (up to a sign, which is a

convention anyway). So, we can use the same momentum operator

d

P = i~

d

which satisfies

P |ni = ~n|ni

By changing space to the compact S 1 we now have momenta that instead of

taking on any real value, can only be integral numbers times ~. Solving the

Schr

odinger equation

P2 ~2 2

i~ (, t) = (, t) = (, t)

t 2m 2m 2

as before, we find

~2 d2

EE () = E ()

2m d2

an eigenvector equation, which has solutions |ni, with

~2 n 2

E=

2m

The general solution to the Schrodinger equaton is

+

~n2

X

(, t) = cn ein ei 2m t

n=

with the cn determined from the initial condition of knowing the wave-function

at time t = 0, according to the Fourier coefficient formula

Z 2

1

cn = ein (, 0)d

2 0

To get something more realistic, we need to take our circle to have an ar-

bitrary circumference L, and we can study our original problem by considering

the limit L . To do this, we just need to change variables from to L ,

where

L

L =

2

The momentum operator will now be

d

P = i~

dL

2~

and its eigenvalues will be quantized in units of L . The energy eigenvalues

will be

2 2 ~2 n2

E=

mL2

114

10.3 The group R and the Fourier transform

In the previous section, we used the trick of replacing the group R of translations

by a compact group S 1 , and then used the fact that unitary representations of

this group are labeled by integers. This made the analysis rather easy, with H =

L2 (S 1 ) and the self-adjoint operator P = i~ behaving much the same as in

the finite-dimensional case: the eigenvectors of P give a countable orthonormal

basis of H. If one wants to, one can think of P as an infinite-dimensional matrix.

Unfortunately, in order to understand many aspects of quantum mechanics,

we cant get away with this trick, but need to work with R itself. One reason

for this is that the unitary representations of R are labeled by the same group,

R, and we will find it very important to exploit this and treat positions and

momenta on the same footing (see the discussion of the Heisenberg group in

the next chapter). What plays the role here of |ni = ein , n Z will be the

|ki = eikq , k R. These are functions on R that are irreducible representations

under the translation action ((a) acts by translation by a)

We can try and mimic the Fourier series decomposition, with the coefficients

cn that depend on the labels of the irreducibles replaced by a function fe(k)

depending on the label k of the irreducible representation of R.

Definition (Fourier transform). The Fourier transform of a function is given

by Z

1

F = (k) =

e eikq (q)dq

2

The definition makes sense for L1 (R), Lebesgue integrable functions on

R. For the following, it is convenient to instead restrict to the Schwartz space

S(R) of functions such that the function and its derivatives fall off faster than

any power at infinity (which is a dense subspace of L2 (R)). For more details

about the analysis and proofs of the theorems quoted here, one can refer to a

standard textbook such as [44].

Given the Fourier transform of , one can recover itself:

Theorem (Fourier Inversion). For e S(R) the Fourier transform of a func-

tion S(R), one has

Z +

1

(q) = Fee = eikq (k)dk

e

2

Note that Fe is the same linear operator as F, with a change in sign of the

argument of the function it is applied to. Note also that we are choosing one

of various popular ways of normalizing the definition of the Fourier transform.

In others, the factor of 2 may appear instead in the exponent of the complex

exponential, or just in one of F or Fe and not the other.

The operators F and Fe are thus inverses of each other on S(R). One has

115

Theorem (Plancherel). F and Fe extend to unitary isomorphisms of L2 (R)

with itself. In other words

Z Z

2 2

|(q)| dq = |(k)|

e dk

Z

< 1 , 2 >= 1 (q)2 (q)dq

An important example is the case of Gaussian functions where

Z +

q2 1 q2

Fe 2 = eikq e 2 dq

2

Z +

1 2k 2 ik 2

= e 2 ((q+i ) ( ) ) dq

2

1 k2 + q0 2 0

Z

= e 2 e 2 dq

2

1 k2

= e 2

A crucial property of the unitary operator F on H is that it diagonalizes

the differentiation operator and thus the momentum operator P . Under Fourier

transform, differential operators become just multiplication by a polynomial,

giving a powerful technique for solving differential equations. Computing the

Fourier transform of the differentiation operator using integration by parts, we

find

Z +

d

g 1 d

= eikq dq

dq 2 dq

Z +

1 d d

= ( (eikq ) ( eikq ))dq

2 dq dx

Z +

1

=ik eikq dq

2

=ik (k)

e

This is the infinitesimal version of the fact that translation becomes multiplica-

tion by a phase under Fourier transform. If a (q) = (q + a), one has

Z +

fa (k) = 1

eikq (q + a)dq

2

Z +

1 0

= eik(q a) (q 0 )dq 0

2

ika e

=e (k)

116

Since p = ~k, we can easily change variables and work with p instead of k,

and often will do this from now on. As with the factors of 2, theres a choice

of where to put the factors of ~ in the normalization of the Fourier transform.

Well make the following choices, to preserve symmetry between the formulas

for Fourier transform and inverse Fourier transform:

Z +

1 pq

(p)

e = ei ~ dq

2~

Z +

1 pq

(q) = ei ~ dp

2~

Note that in this case we have lost an important property that we had for

finite dimensional H and had managed to preserve by using S 1 rather than

R as our space. If we take H = L2 (R), the eigenvectors for the operator P

(the functions eikq ) are not square-integrable, so not in H. The operator P

is an unbounded operator and we no longer have a theorem saying that its

eigenvectors give an orthornormal basis of H. As mentioned earlier, one way

to deal with this uses a general spectral theorem for self-adjoint operators on a

Hilbert space, for more details see Chapter 2 of [45].

One would like to think of the eigenvectors of the operator P as in some sense

continuing to provide an orthonormal basis for H. A possible approach to this

is to work with generalized functions, known as distributions. We wont try and

develop here this kind formalism in a rigorous form, but will just explain the

non-rigorous form in which it is often used in physics.

Given any function g(q) on R, one can try and define an element of the dual

space of the space of functions on R by integration, i.e by the linear operator

Z +

f g(q)f (q)dq

(we wont try and specify which condition on functions f or g is chosen to make

sense of this). There are however some other very obvious linear functionals on

such a function space, for instance the one given by evaluating the function at

q = c:

f f (c)

Such linear functionals correspond to generalized functions, objects which when

fed into the formula for integration over R give the desired linear functional.

The most well-known of these is the one that gives this evaluation at q = c, it

is known as the delta function and written as (q c). It is the object which,

if it were a function, would satisfy

Z +

(q c)f (q)dq = f (c)

117

To make sense of such an object, one can take it to be a limit of actual functions.

For the -function, consider the limit as 0 of

1 (qc)2

g = e 2

2

which satisfy Z +

g (q)dq = 1

for all > 0 (one way to see this is to use the formula given earlier for the

Fourier transform of a Gaussian).

Heuristically (ignoring obvious problems of interchange of integrals that

dont make sense), one can write the Fourier inversion formula as follows

Z +

1

(x) = eikq (k)dk

e

2

Z + Z +

1 1 0

= eikq ( eikq (q 0 )dq 0 )dk

2 2

Z + Z +

1 0

= ( eik(qq ) (q 0 )dk)dq 0

2

Z +

= (q 0 q)(q 0 )dq 0

one can interpret the above calculation as justifying the formula

Z +

1 0

(q q 0 ) = eik(qq ) dk

2

One then goes on to consider the eigenvectors

1

|ki = eikq

2

of the momentum operator as satisfying a replacement for the finite-dimensional

orthonormality relation, with the -function replacing the ij :

Z + Z +

1 1 1 0

hk 0 |ki = ( eik0 q )( eikq )dq = ei(kk )q dq = (k k 0 )

2 2 2

As mentioned before, we will usually work with the variable p = ~k, in which

case we have

1 pq

|pi = ei ~

2~

and Z +

1 (pp0 )q

(p p0 ) = ei ~ dq

2~

For a more mathematically legitimate version of this calculation, one place

to look is Lecture 6 in the notes on physics by Dolgachev[11].

118

10.4 For further reading

Every book about quantum mechanics covers this example of the free quantum

particle somewhere very early on, in detail. Our discussion here is unusual

just in emphasizing the role of the spatial translation groups and its unitary

representations. Discussions of quantum mechanics for mathematicians (such

as [45]) typically emphasize the development of the functional analysis needed

for a proper description of the Hilbert space H and of the properties of general

self-adjoint operators on this state space. In this class were restricting attention

to a quite limited set of such operators coming from Lie algebra representations,

so will avoid the general theory.

119

120

Chapter 11

the Schr

odinger

Representation

In our discussion of the free particle, we used just the actions of the groups R3 of

spatial translation and the group R of time translation, finding corresponding

observables, the self-adjoint momentum (P ) and Hamiltonian (H) operators.

Weve seen though that the Fourier transform involves a perfectly symmetrical

treatment of position and momentum variables. This allows us to introduce a

position operator Q acting on our state space H. We will analyze in detail in

this chapter the implications of extending the algebra of observable operators in

this way, most of the time restricting to the case of a single spatial dimension,

since the physical case of three dimensions is an easy generalization.

The P and Q operators generate an algebra named the Heisenberg algebra,

since Werner Heisenberg and collaborators used it in the earliest work on a full

quantum-mechanical formalism in 1925. It was quickly recognized by Hermann

Weyl that this algebra comes from a Lie algebra representation, with a cor-

responding group (called the Heisenberg group by mathematicians, the Weyl

group by physicists). The state space of a quantum particle, either free or mov-

ing in a potential, will be a unitary representation of this group, with the group

of spatial translations a subgroup. Note that this particular use of a group

and its representation theory in quantum mechanics is both at the core of the

standard axioms and much more general than the usual characterization of the

significance of groups as symmetry groups. While the group of spatial trans-

lations is a symmetry of the theory of a free particle, the action of the larger

Heisenberg group on the state space H does not commute with any non-zero

Hamiltonian operator. The Heisenberg group does not in any sense correspond

to a group of invariances of the physical situation, but rather plays a much

deeper role.

121

11.1 The position operator and the Heisenberg

Lie algebra

In the description of the state space H as functions of a position variable q, the

momentum operator is

d

P = i~

dq

The Fourier transform F provides a unitary transformation to a description of

H as functions of a momentum variable p in which the momentum operator P

is just multiplication by p. Exchanging the role of p and q, one gets a position

operator Q that acts as

d

Q = i~

dp

when states are functions of p (the sign difference comes from the sign change

in Fe vs. F), or as multiplication by q when states are functions of q.

In the position space representation, taking as position variable q 0 , one has

normalized eigenfunctions describing a free particle of momentum p

1 pq 0

|pi = ei ~

2~

which satisfy

d 1 pq 0 1 pq 0

P |pi = i~ 0

( ei ~ ) = p( ei ~ ) = p|pi

dq 2~ 2~

The operator Q in this representation is just the multiplication operator

Q(q 0 ) = q 0 (q 0 )

that multiplies a function of the position variable q 0 by q 0 . The eigenvectors |qi

of this operator will be the -functions (q 0 q) since

Q|qi = q 0 (q 0 q) = q(q 0 q)

A standard convention in physics is to think of a state written in the notation

|i as being representation independent. The wave-function in the position

space representation is then taken to be the coefficient of |i in the expansion

of a state in Q eigenfunctions |qi, so

Z +

hq|i = (q q 0 )(q 0 )dq 0 = (q)

and in particular

1 pq

hq|pi = ei ~

2~

122

11.1.2 Momentum space representation

In the momentum space description of H as functions of p0 , the state is the

Fourier transform of the state in the position space representation, so one has

Z + Z +

1 i pq

0

1 0 0

i p ~q i pq

0

0 1 pp0 0

|pi = F( e ~ )= e e ~ dq = ei ~ q dq 0 = (pp0 )

2~ 2~ 2~

in this representation

Z +

0 1 p0 q 0 1 p0 q

|qi = F((q q )) = ei ~ (q q 0 )dq 0 = ei ~

2~ 2~

The position operator is

d

Q = i~

dp

and |qi is an eigenvector with eigenvalue q

d 1 p0 q 1 p0 q

Q|qi = i~ 0

( ei ~ ) = q( ei ~ ) = q|qi

dp 2~ 2~

Another way to see that this is the correct operator is to use the unitary trans-

formation F and its inverse Fe that relate the position and momentum space

representations. Going from position space to momentum space one has

Q FQFe

and one can check that this transformed Q operator will act as i~ dpd 0 on functions

of p0 .

Using the representation independent notation, one has

Z +

1 pq 0

hp|i = ( ei ~ )(q 0 )dq 0 = (p)

e

2~

With now both momentum and position operators on H, we have the standard

set-up for describing a non-relativistic quantum particle that is discussed exten-

sively early on in any quantum mechanics textbook, and one of these should be

consulted for more details and for explanations of the physical interpretation of

this quantum system. The classically observable quantity corresponding to the

operator P is the momentum, and eigenvectors of P are the states that have well-

defined values for this, values that will have the correct non-relativistic energy

123

momentum relationship. Note that for the free particle P commutes with the

P2

Hamiltonian H = 2m so there is a conservation law: states with a well-defined

momentum at one time always have the same momentum. This corresponds to

an obvious physical symmetry, the symmetry under spatial translations.

The operator Q on the other hand does not correspond to a physical sym-

metry, since it does not commute with the Hamiltonian. We will see that it

does generate a group action, and from the momentum space picture we can

see that this is a shift in the momentum, but such shifts are not symmetries of

the physics and there is no conservation law for Q. The states in which Q has a

well-defined numerical value are the ones such that the position wave-function

is a delta-function. If one prepares such a state at a given time, it will not

remain a delta-function, but quickly evolve into a wave-function that spreads

out in space.

Since the eigenfunctions of P and Q are non-normalizable, one needs a

slightly different formulation of the measurement theory principle used for finite

dimensional H. In this case, the probability of observing a position of a particle

with wave function (q) in the interval [q1 , q2 ] will be

R q2

q1

(q)(q)dq

R +

(q)(q)dq

This will make sense for states |i L2 (R), which we will normalize to have

norm-squared one when discussing their physical interpretation. Then the sta-

tistical expectation value for the measured position variable will be

h|Q|i

which can be computed with the same result in either the position or momentum

space representation.

Similarly, the probability of observing a momentum of a particle with momentum-

space wave function (q)

e in the interval [p1 , p2 ] will be

R p2

p1

(p)

e (p)dp

e

R +

(p)

e (p)dp

e

and for normalized states the statistical expectation value of the measured mo-

mentum is

h|P |i

Note that states with a well-defined position (the delta-function states in

the position-space representation) are equally likely to have any momentum

whatsoever. Physically this is why such states quickly spread out. States with

a well-defined momentum are equally likely to have any possible position. The

properties of the Fourier transform imply the so-called Heisenberg uncertainty

principle that gives a lower bound on the product of a measure of uncertainty

in position times the same measure of uncertainty in momentum. Examples

124

of this that take on the lower bound are the Gaussian shaped functions whose

Fourier transforms were computed earlier.

For much more about these questions, again most quantum mechanics text-

books will contain an extensive discussion.

In either the position or monentum space representation the operators P and

Q satisfy the relation

[Q, P ] = i~1

Soon after this commutation relation appeared in early work on quantum me-

chanics, Weyl realized that it can be interpreted as the relation between oper-

ators one would get from a representation of a three-dimensional Lie algebra,

now called the Heisenberg Lie algebra.

Definition (Heisenberg Lie algebra). The Heisenberg Lie algebra h3 is the vec-

tor space R3 with the Lie bracket defined by its values on a basis (X, Y, Z) by

Lie bracket is given by

Note that this is a non-abelian Lie algebra, but only minimally so. All Lie

brackets of Z are zero. All Lie brackets of Lie brackets are also zero (as a result,

this is an example of what is known as a nilpotent Lie algebra).

The Heisenberg Lie algebra is isomorphic to the Lie algebra of 3 by 3 strictly

upper triangular real matrices, with Lie bracket the matrix commutator, by the

following isomorphism:

0 1 0 0 0 0 0 0 1

X 0 0 0 , X 0 0 1 , Z 0 0 0

0 0 0 0 0 0 0 0 0

0 x z

xX + yY + zZ 0 0 y

0 0 0

since one has

x0 z0 xy 0 x0 y

0 x z 0 0 0

[0 0 y , 0 0 y 0 ] = 0 0 0

0 0 0 0 0 0 0 0 0

125

For a higher-dimensional generalization of this, one just replaces x and y by

n-dimensional vectors x and y, giving a Lie algebra h2n+1 . The physical case is

n = 3, where elements of the Heisenberg Lie algebra can be written

0 x1 x2 x3 z

0 0 0 0 y3

0 0 0 0 y2

0 0 0 0 y1

0 0 0 0 0

One can easily see that exponentiating matrices in h3 gives

1 x z + 12 xy

0 x z

exp 0 0 y = 0 1 y

0 0 0 0 0 1

so the group with Lie algebra h3 will be the group of upper triangular 3 by 3 real

matrices with ones on the diagonal, and this group will be the Heisenberg group

H3 . For our purposes though, it is better to work in exponential coordinates (i.e.

identifying a group element with the Lie algebra element that exponentiates

to it). Matrix exponentials in general satisfy the Baker-Campbell-Hausdorff

formula, which says

1

eA eB = eA+B+ 2 [A,B]+

where the higher terms all involve commutators of commutators (one can check

this term by term by expanding the exponentials, for a proof, see chapter 3

of [24]). For the case of the Heisenberg Lie algebra, since commutators of

commutators vanish, we get the simpler formula for exponentials of elements of

h3

1

eA eB = eA+B+ 2 [A,B]

and we can use this to explicitly write the group law in exponential coordinates.

the group law

1

(x, y, z) (x0 , y 0 , z 0 ) = (x + x0 , y + y 0 , z + z 0 + (xy 0 yx0 ))

2

Note that the Lie algebra basis elements X, Y, Z each generate subgroups

of H3 isomorphic to R. Elements of the first two of these subgroups generate

the full group, and elements of the third subgroup are central, meaning they

commute with all group elements. Also notice that the non-commutativity of

the Lie algebra or group depends purely on the factor xy 0 yx0 , which for the

generalization to h2n+1 of H2n+1 can be written as x y0 y x0 .

126

11.4 The Schr

odinger representation

Since it can be defined in terms of 3 by 3 matrices, the Heisenberg group H3

has an obvious representation on C3 , but this representation is not unitary and

not of physical interest. What is of great interest is the representation of the

Lie algebra h3 on functions of q given by the Q and P operators:

Definition (Schrodinger representation). The Schr odinger representation of the

Heisenberg Lie algebra h3 is the representation (~ , L2 (R)) satisfying

d

0~ (X)(q) = iQ(q) = iq(q), 0~ (Y )(q) = iP (q) = ~ (q)

dq

0~ (Z)(q) = i~(q)

Factors of i have been chosen to make these operators skew-Hermitian. They

can be exponentiated, giving the unitary representation (~ , L2 (R)) of the group

H3 . Elements of H3 of the form g = exX give one subgroup R H3 and act by

multiplication by a phase depending on q

Those of the form g = eyY give another subset R H3 and act by translation

in q

d

~ (eyY )(q) = eyiP (q) = ey~ dq (q) = (q y~)

while those of the form g = ezZ act by multiplication by a phase

The group analog of the Heisenberg commutation relations (often called the

Weyl form of the commutation relations) is the relation

mula)

1 xy~

~ (exX )~ (eyY ) = eixQ eiyP = ei(xQ+yP )+ 2 [ixQ,iyP ]) = ei 2 ei(xQ+yP )

as well as the same product in the opposite order, and then comparing the

results.

We have seen that the Fourier transform F takes the Schrodinger represen-

tation to a unitarily equivalent representation of H3 , in terms of functions of p

(the momentum space representation). The operators change as

~ (g) F ~ (g)Fe

when one makes the unitary transformation to the momentum space represen-

tation.

127

In typical physics quantum mechanics textbooks, one often sees calculations

made just using the Heisenberg commutation relations, without picking a spe-

cific representation of the operators that satisfy these relations. This turns

out to be justified by the remarkable fact that, for the Heisenberg group, once

one picks the value of ~ with which Z acts, all irreducible representations are

unitarily equivalent. In a sense, the representation theory of the Heisenberg

group is very simple: theres just one irreducible representation. This is very

different than the theory for even the simplest compact Lie groups (U (1) and

SU (2)) which have an infinity of inequivalent irreducibles labeled by weight or

by spin. Representations of a Heisenberg group will appear in different guises

(weve seen two, will see another in the discussion of the harmonic oscillator,

and there are yet others that appear in the theory of theta-functions), but they

are all unitarily equivalent. This statement is known as the Stone-von Neumann

theorem.

So far weve been modestly cavalier about the rigorous analysis needed to

make precise statements about the Schrodinger representation. In order to prove

a theorem like the Stone-von Neumann theorem, which tries to say something

about all possible representations of a group, one needs to invoke a great deal

of analysis. Much of this part of analysis was developed precisely to be able to

deal with general quantum mechanical systems and prove theorems about them.

The Heisenberg group, Lie algebra and its representations are treated in detail

in many expositions of quantum mechanics for mathematicians. Some excellent

references for this material are [45], and [26]. In depth discussions devoted to

the mathematics of the Heisenberg group and its representations can be found

in [30] and [18].

In these references can be found a proof of (not difficult)

Theorem. The Schr

odinger representation ~ described above is irreducible.

and the much more difficult

Theorem (Stone-von Neumann). Any irreducible representation of the group

H3 on a Hilbert space, satisfying

0 (Z) = i~1

odinger representation (~ , L2 (R))

is unitarily equivalent to the Schr

Because of the unitary equivalence ensured by Stone-von Neumann, we will

sometimes be cavalier and use the symbol ~ to refer to any representation

in the unitary equivalence class, without specifying which one we mean (e.g.

Schrodinger representation as a function of q or as a function of p).

Note that all of this can easily be generalized to the case of n spatial di-

mensions, with the Heisenberg group now H2n+1 and the Stone-von Neumann

theorem still true. If one tries however to take n , which is what happens

in quantum field theory, the Stone-von Neumann theorem breaks down. In the

case of quantum field theory one has an infinity of inequivalent irreducible repre-

sentations of the commutation relations to consider, one source of the difficulties

of the subject.

128

It is also important to note that the Stone-von Neumann theorem is for-

mulated for Heisenberg group representations, not for Heisenberg Lie algebra

representations. For infinite-dimensional representations in cases like this, there

are representations of the Lie algebra that are non-integrable: they arent the

derivatives of Lie group representations. For representations of the Heisen-

berg Lie algebra, i.e. the Heisenberg commutator relations, there are counter-

examples to the Stone von-Neumann theorem. It is only for integrable rep-

resentations that the theorem holds and one has a unique sort of irreducible

representation.

For a lot more detail about the mathematics of the Heisenberg group, its Lie

algebra and the Schr odinger representation, see [7], [30], [18] and [46]. An ex-

cellent historical overview of the Stone-von Neumann theorem[37] by Jonathan

Rosenberg is well worth reading.

129

130

Chapter 12

Poisson Bracket

We have seen that the quantum theory of a free particle corresponds to the con-

struction of a representation of the Heisenberg Lie algebra in terms of operators

Q and P . One would like to use this to produce quantum systems with a similar

relation to more non-trivial classical mechanical systems than the free particle.

From the earliest days of quantum mechanics it was recognized that the commu-

tation relations of the Q and P operators somehow corresponded to the Poisson

bracket relations between the position and momentum variables in classical me-

chanics in the Hamiltonian formalism. In this chapter well begin by reviewing

the topic of Hamiltonian mechanics and the Poisson bracket, and go on to study

the Lie algebra structure determined by the Poisson bracket. The Lie algebra

here is infinite dimensional, so not the sort of finite dimensional Lie algebra

given by matrices that we have studied so far (although, historically, it is the

sort of Lie algebra that motivated the discovery of the theory of Lie groups and

Lie algebras by Sophus Lie during the 1870s). It does have certain important

finite-dimensional subalgebras which we will study, one of which is isomorphic

to the Heisenberg Lie algebra h2d+1 seen in the last chapter, the other of which

is something new: the Lie algebra of the symplectic group Sp(2d, R).

In classical mechanics in the Hamiltonian formalism, the space M = R2d

that one gets by putting together positions and the corresponding momenta

is known as phase space. Points in phase space can be thought of as uniquely

parametrizing possible initial conditions for classical trajectories, so another in-

terpretation of phase space is that it is the space that parametrizes solutions

of the equations of motion of a given classical mechanical system. Functions

on phase space carry the following structure (for simplicity well start by just

writing things for d = 1):

131

Definition (Poisson bracket). There is a bilinear operation on functions on the

phase space R2 (with coordinates (p, q)) called the Poisson bracket, given by

f g f g

(f, g) {f, g} =

q p p q

by a point of phase space. Observables are functions f on phase space, and

their time dependence is determined by a distinguished one, the Hamiltonian

function h. An observable f varies along the classical trajectories determined

by a Hamiltonian f according to

df

= {f, h}

dt

Taking f to be just the functions p and q, one gets Hamiltons equations for

classical trajectories in phase space

h

p = {p, h} =

q

h

q = {q, h} =

p

p2

For a non-relativistic free particle, h = 2m and these equations become

p

p = 0, q =

m

which just says that the momentum is the mass times the velocity, and is con-

served. For a particle subject to a potential V (q) one has

p2

h= + V (q)

2m

and the trajectories are the solutions to

V p

p = , q =

q m

V

F = = ma = m

q

q

One can easily check that the Poisson bracket has the properties

Anti-symmetry

{f, g} = {g, f }

132

Jacobi identity

These two properties, together with the bilinearity, show that the Poisson

bracket fits the definition of a Lie bracket, making the space of functions on

phase space into an infinite dimensional Lie algebra. There is a corresponding

group, known to physicists as the group of canonical transformations, but it

is also infinite dimensional, so not a matrix group of the sort we have been

considering so far. This Lie algebra is responsible for much of the structure of

the subject of Hamiltonian mechanics, and it was historically the first sort of

Lie algebra to be studied.

The role of symmetries in classical mechanics is best seen using this Lie

algebra. From the fundamental dynamical equation

df

= {f, h}

dt

we see that

df

{f, h} = 0 = =0

dt

and in this case the function f is called a conserved quantity, since it does

not change under time evolution. Note that if we have two functions f and g

on phase space such that

{f, h} = 0, {g, h} = 0

Another fundamental property of the Poisson bracket is the

Leibniz rule

This property says that taking Poisson bracket with a function h acts on a

product of functions in a way that satisfies the Leibniz rule for what happens

when you take the derivative of a product. This is exactly what one expects,

since one can think of h as a Hamiltonian function and then the dynamical

equation

df

{f, h} =

dt

says that the operation of Poisson bracketing a function by h is a derivative: the

time derivative along the classical trajectory corresponding to the Hamiltonian

h.

133

Taking Poisson brackets with the linear functions p and q gives the deriva-

tives

f f

{f, p} = , {f, q} =

q p

and the Leibniz rule allows one to compute recursively what will happen when

one takes Poisson bracket by any polynomial in p and q. This will always give

some first-order partial differerential operator. A simple example would be the

action of pq which would be

f f

{f, pq} = p{f, q} + {f, p}q = p +q

p q

Notice that taking Poisson bracket with p acts on functions of q as an infinites-

imal translation, so corresponds to the symmetry of spatial translation. Other

symmetries of a Hamiltonian system will be infinitesimally generated by func-

tions in a way similar to the case of the function p , with the Poisson bracket

between such functions giving the Lie bracket relation in the Lie algebra of the

symmetry group.

Digression. If you are familiar with differential manifolds, vector fields and

differential forms, a more general notion of a phase space than just M = R2d

is the following sort of manifold:

Definition (Symplectic manifold). A symplectic manifold M is a manifold with

a two-form (, ) (called a symplectic two-form) satisfying the conditions

is non-degenerate, i.e. for a nowhere zero vector field X, (X, ) is a

nowhere zero one-form.

d = 0, in which case is said to be closed.

The cotangent bundle T N of a manifold N (i.e. the space of pairs of a

point on N together with a linear function on the tangent space at that point)

provides one class of symplectic manifolds, generalizing the linear case N = Rd ,

and corresponding physically to a particle moving on N . A simple example that

is neither linear nor a cotangent bundle is the sphere M = S 2 , with the area

two-form.

Note that there is no assumption here that M has a metric (i.e. it may

not be a Riemannian manifold). A symplectic two-form is a structure on a

manifold analogous to a metric but with opposite symmetry properties. Whereas

a metric is a symmetric non-degenerate bilinear form on the tangent space at

each point, a symplectic form is an anti-symmetric non-degenerate bilinear form

on the tangent space.

Using the symplectic form , for any symplectic manifold M , one has a

generalization of Hamiltons equations. This is the following equality of one-

forms, giving the relation between a Hamiltonian function h and the vector field

Xh that determines time evolution on M

iXh = (Xh , ) = dh

134

The Poisson bracket in this context can be defined as

{f, g} = (Xf , Xg )

and the fact that the Poisson bracket gives a Lie bracket structure on functions

corresponds to the existence of a Lie bracket on vector fields, satisfying

[Xf , Xg ] = X{f,g}

(the minus sign is due to the fact that vector fields are a Lie algebra repre-

sentation on the space of functions on M , with the standard inverse appearing

when going from a group action on M to the corresponding representation on

functions on M ).

The condition d = 0 is then equivalent to the condition that the Poisson

bracket satisfies the Jacobi identity. The infinite dimensional Lie algebra of

functions on M with Lie bracket the Poisson bracket is the Lie algebra for an

infinite dimensional group, the group of all diffeomorphisms of M that preserve

. For much more about symplectic geometry and Hamiltonian mechanics, a

good reference is [7].

From the point of view of symplectic geometry, at each point the fundamental

structure is the anti-symmetric bilinear form on tangent vectors given by .

In discussions of quantization however, what is most important is the Poisson

bracket which, in the linear approximation near each point, is an anti-symmetric

bilinear form on linear combinations of the local coordinate functions (qi , pi ).

These lie not in the tangent space but in its dual space, and it is important to

keep that in mind as we go on to exploit symmetries based on the Poisson bracket,

rather than the closely related symplectic form. Our discussion of quantization

will rely crucially on having a linear structure on phase space, so will not apply

to general symplectic manifolds.

als

The Poisson bracket of any two polynomial functions can be computed by using

the Leibniz rule and anti-symmetry, reducing the problem down to the case of

Poisson brackets of linear functions, which are determined by the relations

so it is these simple relations (together with the Leibniz rule) which fundamen-

tally determine the entire infinite dimensional Lie algebra structure. The vector

space of linear polynomials on M is three dimensional, with basis elements q, p, 1

and this space is closed under Poisson bracket, so gives a three dimensional Lie

algebra. This Lie algebra is isomorphic to the Heisenberg Lie algebra h3 with

the isomorphism given by

X q, Y p, Z 1

135

This isomorphism preserves the Lie bracket relations since

[X, Y ] = Z {q, p} = 1

The homogeneous quadratic polynomials in p and q form another three-

dimensional sub-Lie algebra of the Lie algebra of functions on phase space,

since the non-zero Poisson bracket relations between them are

q 2 p2

{ , } = qp {qp, p2 } = 2p2 {qp, q 2 } = 2q 2

2 2

One can identify this Lie algebra with a Lie algebra of 2 by 2 matrices as

follows

p2 q2

0 1 0 0 1 0

E= F = qp G =

2 0 0 2 1 0 0 1

The commutation relations amongst these matrices are

[E, F ] = G [G, E] = 2E [G, F ] = 2F

which are the same as the Poisson bracket relations between the corresponding

quadratic polynomials. Mathematicians should note that what we have called

G here is more usually denoted H, but in this context using that symbol would

just be too confusing due to our use of it to denote the Hamiltonian.

The matrices E, F, G give a basis of the real 2 by 2 matrices of trace zero,

which span the Lie algebra sl(2, R). So our Lie algebra of quadratic polynomials

is the Lie algebra of the group SL(2, R) of 2 by 2 real matrices of determinant

1. This three-dimensional Lie group is non-compact and quite unlike SU (2). In

particular, all of its non-trivial irreducible unitary representations are infinite-

dimensional, forming an important topic in mathematics, but one that is beyond

the scope of this course.

Two important subgroups of SL(2, R) are

The subgroup of elements one gets by exponentiating g, which is isomor-

phic to the multiplicative group of positive real numbers

t

e 0

etG =

0 et

Here one can explicitly see that this group has elements going off to infinity.

Exponentiating the Lie algebra element E F gives rotations of the plane

cos sin

e(EF ) =

sin cos

Note that the Lie algebra element being exponentiated here is

1 2

EF (p + q 2 )

2

which we will later reencounter as the Hamiltonian function for the har-

monic oscillator.

136

We have found two three-dimensional Lie algebras (h3 and sl(2, R)) as subal-

gebras of the infinite dimensional Lie algebra of functions on phase space. They

do not commute with each other, since one has the following non-zero Poisson

bracket relations between them

{pq, p} = p, {pq, q} = q

p2 q2

{ , q} = p, { , p} = q

2 2

Taking all quadratic polynomials together, we get a six-dimensional Lie algebra

with basis elements 1, q, p, qp, q 2 , p2 . This is not the product of h3 and sl(2, R),

but an example of a structure that we will study in detail later on, called a

semi-direct product and written

h3 o sl(2, R)

Heisenberg Lie algebra

Our construction of the Lie algebra h3 as the linear functions on phase space

M = R2 with Poisson bracket as Lie bracket was based on a specific choice of

a basis q, p of M = M . We could equally well have chosen any other basis, as

long as changing basis did not change the Poisson brackets, or equivalently, the

Lie bracket relation for h3 . Restricted to M, the Poisson bracket determines an

antisymmetric bilinear form by

=cq c0p cp c0q

0

0 1 cq

= cq cp

1 0 c0p

cq a b cq

cp c d cp

T

a b 0 1 a b 0 1

=

c d 1 0 c d 1 0

or

0 ad bc 0 1

=

ad + bc 0 1 0

137

so

a b

det = ad bc = 1

c d

which say that the change of basis matrix must be an element of SL(2, R). The

infinitesimal version of this SL(2, R) action on M is given by considering the

action of elements of sl(2, R) on the basis elements q, p. Identifying elements

of sl(2, R) with linear combinations of q 2 , p2 , pq, this action is given by taking

Poisson brackets of q, p elements of sl(2, R) using the Poisson bracket relations

{pq, p} = p, {pq, q} = q

p2 q2

{ , q} = p, { , p} = q

2 2

In higher dimensions, choosing a basis for the phase space M = R2d gives

a basis {q1 , p1 , qd , pd } for the dual space M = M consisting of the linear

coefficient functions of vectors in M . Taking as an additional basis element the

constant function 1, we have a 2d + 1 dimensional space

MR

with basis {q1 , p1 , qd , pd , 1}, and the Poisson bracket turns this space into a

Lie algebra, the Heisenberg Lie algebra h2d+1 .

This Lie bracket on h2d+1 is explicitly given on linear combinations of the

basis elements by an antisymmetric, non-degenerate bilinear form as follows.

If (u, c) and (u0 , c0 ) are elements of h2d+1 = M R, with

u = cq1 q1 + cp1 p1 + + cqd qd + cpd pd M, c R

u0 = c0q1 q1 + c0p1 p1 + + c0qd qd + c0pd pd M, c0 R

then

[(u, c), (u0 , c0 )] = (u, u0 )1

where

(u, u0 ) ={cq1 q1 + cp1 p1 + + cqd qd + cpd pd , c0q1 q1 + c0p1 p1 + + c0qd qd + c0pd pd }

=cq1 c0p1 cp1 c0q1 + cqd c0pd cpd c0qd

0

0 1 ... 0 0 cq1

1 0 ... 0 0 c0p

1

= cq1 cp1 . . . cqd cpd ... .. .. .. ..

. . . .

0 0 ... 0 1 c0qd

0 0 ... 1 0 c0pd

The group of linear transformations of M that leave invariant (i.e. linear

transformations g such that

(gu, gu0 ) = (u, u0 )

for all u M) is called the symplectic group, so we have seen that it can be

defined explicitly as

138

Definition (Symplectic group Sp(2d, R)). The group Sp(2d, R) is the group of

real 2d by 2d matrices g that satisfy

0 1 ... 0 0 0 1 ... 0 0

1 0 . . . 0 0 1 0 . . . 0 0

T .. .

.. .

.. .

.. g = . .. .. ..

g . ..

. . .

0 0 . . . 0 1 0 0 . . . 0 1

0 0 . . . 1 0 0 0 . . . 1 0

For d = 1 we have Sp(2, R) = SL(2, R), but for higher d the symplectic

groups have no such simple characterization. The Lie algebra of Sp(2d, R) is

denoted sp(2d, R) and it is isomorphic to the Lie algebra of order two homoge-

neous polynomials on M = R2d , with Poisson bracket as Lie bracket.

Our definition of the Poisson bracket determined a specific bilinear form

, but we can define more generally the class of bilinear forms that have the

properties we need:

Definition (Symplectic form). A symplectic form on a vector space V is a

bilinear map

:V V R

such that

is antisymmetric, i.e. (v1 , v2 ) = (v2 , v1 )

is nondegenerate, i.e. if v 6= 0, then (v, ) V is non-zero.

A vector space V with a symplectic form is called a symplectic vector

space. One can show that such a vector space must be even dimensional, and

a basis can always be found such that has exactly the matrix expression we

used earlier. Our interest will be in symplectic vector spaces V that are the

dual vector spaces M for some phase space M = R2d . It is for these that

Poisson brackets correspond to symplectic forms, and for these that we have an

identification with the Lie algebra h2d+1 .

To summarize more abstractly this section, we have shown explicitly that

the group Sp(2d, R) acts on the Lie algebra h2d+1 by automorphisms (maps

from the Lie algebra to itself, preserving the Lie bracket). Using the exponen-

tial map, this gives an action of Sp(2d, R) on the Heisenberg group H2d+1 by

automorphisms (maps from H2d+1 to itself, preserving the group law). When

we study semi-direct products we will see that this is exactly the information

needed to define a semi-direct product group H2d+1 o Sp(2d, R).

algebra

We have been careful here to keep track of the difference between phase space

M = R2d and its dual M = M , since it is M R that is given the structure

139

of a Lie algebra (in this case h2d+1 ) by the Poisson bracket, and it is this Lie

algebra we want to use in the next chapter when we define a quantization of

the classical system. From a purely classical point of view, one can generalize

the notion of phase space and Poisson bracket, defining them purely in terms

of Lie algebras, by starting not with h2d+1 , but with a general Lie algebra g.

Since the dual of a dual is the original vector space, one can think of this g as

the dual of g , i.e.

g = (g )

and then the Lie bracket on g has an interpretation as a Poisson bracket on

linear functions on g . So, for any Lie algebra g, its dual g is a space with a

Poisson bracket, and one can think of this as a generalized classical phase space.

Such a generalized classical phase space is best thought of as a family of

conventional phase spaces. For g = h2d+1 one has

h2d+1 = M R

phase space.

Digression. In the context of the earlier discussion of generalizing phase spaces

to arbitrary symplectic manifolds, g will be a family of such symplectic mani-

folds. For a simple example, consider g = su(2), where

su(2) = R3

sphere is an example of a symplectic manifold, with symplectic form the area

two-form.

The book An Introduction to Symplectic Geometry by Rolf Berndt[7] contains

a detailed discussion of the geometrical formulation of Hamiltonian mechanics

using symplectic geometry. See also [23], especially chapter 14.

140

Chapter 13

Quantization

Given any Hamiltonian classical mechanical system with phase space R2d , physics

textbooks have a standard recipe for producing a quantum system, by a method

known as canonical quantization. We will see that this is just the construc-

tion we have already seen of a unitary representation 0 of the Heisenberg Lie

algebra, the Schr odinger representation, and the Stone-von Neumann theorem

assures us that this is the unique such construction, up to unitary equivalence.

We will also see that this recipe can only ever be partially successful, with the

Schrodinger representation giving us a representation of a sub-algebra of the al-

gebra of all functions on phase space (the polynomials of degree two and below),

and a no-go theorem showing that this cannot be extended to a representation

of the full infinite dimensional Lie algebra. Recipes for quantizing higher-order

polynomials will alway suffer from a lack of uniqueness, a phenomenon known

to physicists as the existence of operator ordering ambiguities.

In later chapters we will see that this quantization prescription does give

unique quantum systems corresponding to some Hamiltonian systems (in par-

ticular the harmonic oscillator and the hydrogen atom), and does so in a manner

that allows a description of the quantum system purely in terms of representa-

tion theory.

Very early on in the history of quantum mechanics, when Dirac first saw the

Heisenberg commutation relations, he noticed an analogy with the Poisson

bracket. One has

i

{q, p} = 1 and [Q, P ] = 1

~

as well as

df d i

= {f, h} and O(t) = [O, H]

dt dt ~

where the last of these equations is the equation for the time dependence of a

Heisenberg picture observable O(t) in quantum mechanics. Diracs suggestion

141

was that given any classical Hamiltonian system, one could quantize it by

finding a rule that associates to a function f on phase space a self-adjoint

operator Of (in particular Oh = H) acting on a state space H such that

i

O{f,g} = [Of , Og ]

~

This is completely equivalent to asking for a unitary representation ( 0 , H)

of the infinite dimensional Lie algebra of functions on phase space (with the

Poisson bracket as Lie bracket). To see this, note that one can choose units

for momentum p and position q such that ~ = 1. Then, as usual getting a

skew-adjoint Lie algebra representation operator by multiplying a self-adjoint

operator by i, setting

0 (f ) = iOf

the Lie algebra homomorphism property

corresponds to

iO{f,g} = [iOf , iOg ] = [Of , Og ]

so one has Diracs suggested relation.

Recall that the Heisenberg Lie algebra is isomorphic to the three-dimensional

sub-algebra of functions on phase space given by linear combinations of the con-

stant function, the function q and the function p. The Schrodinger representa-

tion ~=1 = provides a unitary representation not of the Lie algebra of all

functions on phase space, but of these polynomials of degree at most one, as

follows

O1 = 1, Oq = Q, Op = P

so

d

0 (1) = i1, 0 (q) = iQ = iq, 0 (p) = iP =

dq

Moving on to quadratic polynomials, these can also be quantized using the

Schr

odinger representation, as follows

P2 p2 P2 i d2

O p2 = , 0 ( ) = i =

2 2 2 2 2 dq 2

Q2 q2 Q2 i

O q2 = , 0 ( ) = i = q2

2 2 2 2 2

For the function pq one can no longer just replace p by P and q by Q since

the operators P and Q dont commute, and P Q or QP is not self-adjoint. What

does work, satisfying all the conditions to give a Lie algebra homomorphism is

1 i

Opq = (P Q + QP ), 0 (pq) = (P Q + QP )

2 2

142

This shows that the Schr odinger representation 0 that was defined as a

representation of the Heisenberg Lie algebra h3 extends to a unitary Lie algebra

representation of a larger Lie algebra, that of all quadratic polynomials on phase

space. Restricting to just the homogeneous order two polynomials, we get a

representation of sl(2, R), called the metaplectic representation.

As a Heisenberg Lie algebra representation, the Schrodinger representation

0 exponentiates to give a representation of the corresponding Heisenberg Lie

group. As an sl(2, R) representation however, one can show that 0 has the same

sort of problem as the spinor representation of su(2) = so(3), which was not a

representation of SO(3), but only of its double cover SU (2) = Spin(3). To get

a group representation, one must go to a double cover of the group SL(2, R),

which will be called the metaplectic group and denoted M p(2, R). We will

return to this point later, since the calculation necessary to see what is going on

is best performed not in the Schr odinger representation but in a different but

unitarily equivalent one that we will find when studying the harmonic oscillator.

One should keep in mind though that 0 is not a representation of the product

of h3 and sl(2, R), but of something more non-trivial called the semi-direct

product h3 o sl(2, R). In a later chapter we will study the general theory

of such semi-direct products for Lie groups in more detail. Using that general

theory, the group we want is the semi-direct product of H3 and SL(2, R), written

H3 o SL(2, R). To construct that semi-direct product, one uses an action of

SL(2, R) on H3 , and it is this action that is at issue here.

The Schr odinger representation will be an infinite dimensional unitary

representation not of this larger group, but of a double cover of it. We will see

this more clearly later on when we work with a different but unitarily equivalent

representation based on the quantization of the harmonic oscillator. On general

grounds though, one can argue that acting by g Sp(2d, R) on elements X

h2d+1 takes 0 to a different representation 0g defined by

0g (X) = 0 (g X)

unitary operators U (g) satisfying

0g = U (g)0 U (g)1

These U (g) should then give a unitary representation of the group Sp(2d, R)

on our state space H, and we should have what we expect: a symmetry of the

classical phase space under quantization becomes a unitary representation on

the quantum state space. The problem with this argument is that the U (g) are

not uniquely defined. Schurs lemma tells us that since the representation on H

is irreducible, the operators commuting with the representation operators will

just be the complex scalars, but these complex scalars give a phase ambiguity

in how we define the U (g). So it turns out that the U (g) only have to give

a representation of Sp(2d, R) on H up to a phase. The more detailed explicit

information we have about the representation turns out to show that this phase

ambiguity is really just a sign ambiguity, and this is what is responsible for the

143

necessity of introducing a double cover of Sp(2d, R) called M p(2d, R) if we want

a true representation. For infinite dimensional phase spaces, the ambiguity may

not just be a sign ambiguity, a situation that is described as an anomaly, where

a classical symmetry does not survive intact as a symmetry of the quantum state

space.

If one wants to quantize polynomial functions on phase space of degree greater

than two, it quickly becomes clear that the problem of operator ordering am-

biguities is a significant one. Different prescriptions involving different ways

of ordering the P and Q operators lead to different Of for the same function

f , with physically different observables (although the differences involve the

commutator of P and Q, so higher-order terms in ~).

When physicists first tried to find a consistent prescription for producing an

operator Of corresponding to a polynomial function on phase space of degree

greater than two, they found that there was no possible way to do this consistent

with the relation

i

O{f,g} = [Of , Og ]

~

for polynomials of degree greater than two. Whatever method one devises for

quantizing higher degree polynomials, it can only satisfy that relation to lowest

order in ~, and there will be higher order corrections, which depend upon ones

choice of quantization scheme. Equivalently, it is only for the six-dimensional

Lie algebra of polynomials of degree up to two for which the Schroodinger repre-

sentation gives one a Lie algebra representation, and this cannot be consistently

extended to a representation of a larger subalgebra of the functions on phase

space. This problem is made precise by the following no-go theorem

on R2 to self-adjoint operators on L2 (R) satisfying

i

O{f,g} = [Of , Og ]

~

and

Op = P, Oq = Q

for any Lie subalgebra of the functions on R2 larger than the subalgebra of

polynomials of degree less than or equal to two.

Proof. For a detailed proof, see section 5.4 of [7], section 4.4 of [18], or chapter

16 of [23]. In outline, the proof begins by showing that taking Poisson brack-

ets of polynomials of degree three leads to higher order polynomials, and that

furthermore for degree three and above there will be no finite-dimensional sub-

algebras of polynomials of bounded degree. The assumptions of the theorem

force certain specific operator ordering choices in degree three. These are then

144

used to get a contradiction in degree four, using the fact that the same degree

four polynomial has two different expressions as a Poisson bracket:

1 2 1

q 2 p2 = {q p, p2 q} = {q 3 , p3 }

3 9

Just about all quantum mechanics textbooks contain some version of the discus-

sion here of canonical quantization starting with classical mechanical systems

in Hamiltonian form. For discussions of quantization from the point of view of

representation theory, see [7] and chapters 14-16 of [23]. For a detailed discus-

sion of the Heisenberg group and Lie algebra, together with their representation

theory, see chapter 2 of [30].

145

146

Chapter 14

Coulomb Potential, and the

Hydrogen Atom

We will now turn to the physical case of 3-dimensional space, and will be working

odinger representation, so we have H = L2 (R3 ). Using the action of

in the Schr

SO(3) on R3 we get an induced representation on H by

q1

(g)(q) = (g 1 q) = (g 1 q2 )

q3

This representation was studied in detail in chapter 8, where we found that the

Lie algebra representation

q1

d d

0 (X)(q) = (etX )t=0 (q) = (etX q2 )t=0

dt dt

q3

0 (l1 ) = q3 q2

q2 q3

0 (l2 ) = q1 q3

q3 q1

0 (l3 ) = q2 q1

q1 q2

In our general study of the action of the Heisenberg group H2n+1 and the

symplectic group Sp(2n, R) on the state space H, we saw that, at least at the

147

Lie algebra level, this came from just using the operators

Pj = i~ , Qj = qj , j = 1, 2, 3

qj

rotation group acts on both position and momentum variables, preserving inner

products, so it will also preserve the symplectic form on the dual phase space

M = R6 , since this is defined just using the inner products of momentum and

position vectors. We thus we have SO(3) Sp(6, R), the group whose Lie

algebra acts on H by quadratic polynomials in the Qj , Pj . We will find that

certain specific quadratic polynomials (the angular momenta) give the so(3)

action, recovering the same Lie algebra representation we found earlier just

from the action of SO(3) on functions on R3 .

When the Hamiltonian function is invariant under rotations, we then expect

eigenfunctions of the corresponding Hamiltonian operator to carry representa-

tions of SO(3). The eigenfunctions of a given energy break up into irreducible

representations of SO(3), and we have seen that these are labeled by a inte-

ger l = 0, 1, 2, . . .. One can use this to find properties of the solutions of the

Schrodinger equation whenever one has a rotation-invariant potential energy,

and we will work out what happens for the case of the Coulomb potential de-

scribing the hydrogen atom. We will see that this specific case is exactly solvable

because it has a second not-so-obvious SO(3) symmetry, in addition to the one

coming from rotations in R3 .

In our study of the Lie algebra so(3) we used the following basis elements

0 0 0 0 0 1 0 1 0

l1 = 0 0 1 l2 = 0 0 0 l3 = 1 0 0

0 1 0 1 0 0 0 0 0

these relations as Poisson bracket relations. Then quantization will give us a

unitary representation 0 of so(3) on H = L2 (R3 ). From elementary classical

physics, one knows that the angular momentum vector is defined to be l = qp,

which has components

l1 = q2 p3 q3 p2 , l2 = q3 p1 q1 p3 , l3 = q1 p2 q2 p1

One can easilly check that these satisfy the Poisson bracket relations

148

Quantization gives us corresponding self-adjoint operators on L2 (R3 )

L1 = i0~ (l1 ) = Q2 P3 Q3 P2 = i~(q2 q3 )

q3 q2

L2 = i0~ (l2 ) = Q3 P1 Q1 P3 = i~(q3 q1 )

q1 q3

L3 = i0~ (l3 ) = Q1 P2 Q2 P1 = i~(q1 q2 )

q2 q1

which satisfy

One can check that these operators Lj on H = L2 (R3 ) are (up to a factor

of ~) precisely the same ones

Lj = i0 (lj )

derived earlier using nothing about quantization and the representation , but

just the infinitesimal action of rotations on functions on R3 (the representation

). The factor of ~ appears in the quantization version since now eigenvalues of

the Lj have a correspondence to the classical angular momentum and so depend

on the relative units use for momentum and length.

It is the operators

0 i

e (lj ) = e ~ Lj

that give the operator corresponding to a rotation by about the j-axis. From

the point of view of quantum mechanics, this story is independent of ~, with

the eigenvalues of observables just the integer weights one finds in the action of

SO(3) on its representations. It is only if one wants to interpret these eigenvalues

in terms of the classical notion of angular monentum that one needs to introduce

~ and then one can characterize the eigenvalues as integer multiples of ~.

Note that the spatial geometry of R3 provides us with two symmetry groups.

There is a group R3 of translations which well call T3 , with Lie algebra t3 ,

and a group SO(3) of rotations, with Lie algebra so(3). Since rotations act

on translations, what we have is not the product of these two groups, but a

semi-direct product. This is the so-called Euclidean group

E(3) = T3 o SO(3)

and we will study the group and its representations in more detail in a later

chapter. This group is six dimensional, and its Lie algebra structure can be seen

from the Poisson brackets

momentum operators

149

14.2 Quantum particle in a central potential

In classical physics, to describe not free particles, but particles experiencing

some sort of force, one just needs to add a potential energy term to the

kinetic energy term in the expression for the energy (the Hamiltonian function).

In one dimension, for potential energies that just depend on position, one has

p2

h= + V (q)

2m

for some function V (q). In the physical case of three dimensions, this will be

1 2

h= (p + p22 + p23 ) + V (q1 , q2 , q3 )

2m 1

Quantizing and using the Schrodinger representation, the Hamiltonian op-

erator for a particle moving in a potential V (q1 , q2 , q3 ) will be

1

H= (P 2 + P22 + P33 ) + V (Q1 , Q2 , Q3 )

2m 1

~2 2 2 2

= ( 2 + 2 + 2 ) + V (q1 , q2 , q3 )

2m q1 q2 q3

2

~

= + V (q1 , q2 , q3 )

2m

We will be interested in so-called central potentials, potential functions that

are functions only of q12 + q22 + q32 , and thus only depend upon r, the radial

distance to the origin. For such V , both terms in the Hamiltonian will be

SO(3) invariant, and eigenfunctions of H will be representations of SO(3).

Using the expressions for the angular momentum operators in spherical coor-

dinates derived in chapter 8, one can show that the Laplacian has the following

expression in spherical coordinates

2 2 1

= 2

+ 2 L2

r r r r

where L2 is the Casimir operator in the representation , which we have shown

has eigenvalues l(l+1) on irreducible representations of dimension 2l+1 (integral

spin l). So, restricted to such an irreducible representation, we have

d2 2 d l(l + 1)

= 2

+

dr r dr r2

To solve the Schr odinger equation, we need to find the eigenfunctions of H.

These will be irreducible representations of SO(3) acting on H, which we have

seen can be explicitly expressed in terms of the spherical harmonic functions

Ylm (, ) in the angular variables. So, to find eigenfunctions of the Hamiltonian

~2

H= + V (r)

2m

150

we need to find functions glE (r) depending on l = 0, 1, 2, . . . and the energy

eigenvalue E satisfying

~2 d2 2 d l(l + 1)

( ( 2+ ) + V (r))glE (r) = EglE (r)

2m dr r dr r2

Given such a glE (r) we will have

and the

(r, , ) = glE (r)Ylm (, )

will be a 2l + 1 dimensional (since m = l, l + 1, . . . , l 1, l) space of energy

eigenfunctions for H of eigenvalue E.

For a general potential function V (r), exact solutions for the eigenvalues E

and corresponding functions glE (r) cannot be found in closed form. One special

case where we can find such solutions is for the three-dimensional harmonic

oscillator, where V (r) = 21 m 2 r2 . These are much more easily found though

using the creation and annihilation operator techniques discussed in the last

chapter.

The other well-known and physically very important case is the case of a 1r

potential, called the Coulomb potential. This describes a light charged particle

moving in the potential due to the electric field of a much heavier charged

particle, a situation that corresponds closely to that of a hydrogen atom. In

this case we have

e2

V =

r

where e is the charge of the electron, so we are looking for solutions to

~2 d2 2 d l(l + 1) e2

( ( 2+ 2

) )glE (r) = EglE (r)

2m dr r dr r r

Since having

d2

(rg) = Erg

dr2

is equivalent to

d2 2

( + )g = Eg

dr2 r

for any function g, glE (r) will satisfy

~2 d2 l(l + 1) e2

( ( 2 2

) )rglE (r) = ErglE (r)

2m dr r r

The solutions to this equation can be found through a rather elaborate pro-

cess. For E 0 there are non-normalizable solutions that describe scattering

phenomena that we wont study here. For E < 0 solutions correspond to an

151

integer n = 1, 2, 3, . . ., with n l + 1. So, for each n we get n solutions, with

l = 0, 1, 2, . . . , n 1, all with the same energy

me4

En =

2~2 n2

The unexpected degeneracy in the energy values leads one to suspect that

there is some extra group action in the problem commuting with the Hamilto-

nian, so energy eigenfunctions will come in irreducible representations of some

larger group than SO(3), one such that when one restricts to the SO(3) sub-

group, the representation is reducible, giving n copies of our SO(3) representa-

tion of spin l. In the next section we will see that this is the case, and there use

representation theory to derive the above formula for En .

We wont go through the process of showing how to explicitly find the func-

152

tions glEn (r) but just quote the result. Setting

~2

a0 =

me2

(this has dimensions of length and is known as the Bohr radius), and defining

gnl (r) = glEn (r) the solutions are of the form

r 2r l 2l+1 2r

gnl (r) e na0 ( ) Ln+l ( )

na0 na0

n+l are certain polynomials known as associated Laguerre polyno-

mials.

So, finally, the energy eigenfunctions will be

for

n = 1, 2, . . .

l = 0, 1, . . . , n 1

m = l, l + 1, . . . , l 1, l

1 r

100 = p 3 e a0

a0

1 r r

200 = p (1 )e 2a0

3

8a0 2a0

l = 1 ) states with basis elements

1 r r

211 = p 3 e 2a0 sin ei

8 a0 a0

1 r 2ar

211 = p e 0 cos

4 2a0 a0

3

1 r r

211 = p 3 e 2a0 sin ei

a

8 a0 0

153

14.3 so(4) symmetry and the Coulomb potential

The Coulomb potential problem is very special in that it has an additional

symmetry, of a non-obvious kind. This symmetry appears even in the classi-

cal problem, where it is responsible for the relatively simple solution one can

find to the essentially identical Kepler problem. This is the problem of finding

the classical trajectories for bodies orbiting around a central object exerting a

gravitational force, which also has a 1r potential. Keplers second law for such

motion comes from conservation of angular momentum, which corresponds to

the Poisson bracket relation

{lj , h} = 0

Here well take the Coulomb version of the Hamiltonian that we need for the

hydrogen atom problem

1 e2

h= |p|2

2m r

One can read the relation {lj , h} = 0 in two ways: it says both that the hamil-

tonian h is invariant under the action of the group (SO(3)) whose infinitesimal

generators are lj , and that the components of the angular momentum (lj ) are

invariant under the action of the group (R of time translations) whose infinites-

imal generator is h.

Keplers first and third laws have a different origin, coming from the existence

of a new conserved quantity for this special choice of Hamiltonian. This quantity

is, like the angular momentum, a vector, often called the Lenz (or sometimes

Runge-Lenz, or even Laplace-Runge-Lenz) vector.

Definition (Lenz vector). The Lenz vector is the vector-valued function on the

phase space R6 given by

1 q

w= (p l) e2 2

m |q|

lw =0

We wont here explicitly calculate the various Poisson brackets involving the

components wa of w, since this is a long and unilluminating calculation, but

will just quote the results, which are

{wj , h} = 0

This says that, like the angular momentum, the vector wa is a conserved

quantity under time evolution of the system, and its components generate

symmetries of the classical system.

154

{wj , lk } = jkl wl

These relations say that the generators of the SO(3) symmetry act on wj

the way you would expect, since wj is a vector.

2h

{wj , wk } = jkl ll ( )

m

This is the most surprising relation, and it has no simple geometrical

explanation (although one can change variables in the problem to try and

give it one). It expresses a highly non-trivial relationship between the two

sets of symmetries generated by the vectors l, w and the Hamiltonian h.

The wj are cubic in the q and p variables, so one would expect that the

Groenewold-van Hove no-go theorem would tell one that there is no consistent

way to quantize this system by finding operators Wj corresponding to the wj

that would satisfy the commutation relations corresponding to these Poisson

brackets. It turns out though that this can be done, although not over the

entire phase-space. One gets around the no-go theorem by doing something

that only works where the Hamiltonian h is positive (well be taking a square

root of h).

The choice of operator Wa that works is

1 Q

W= (P L L P) e2

2m |Q|2

which (by elaborate and unenlightening computation) one can show satisfy

[Wj , H] = 0

2

[Wj , Lk ] = i~jkl Ll ( H)

m

[Wj , Lk ] = i~jkl Wl

as well as

LW =WL=0

Most importantly, one has the following relation between W 2 , H and the Casimir

operator L2

2

W 2 = e4 + H(L2 + ~2 )

m

and it is this which will allow us to find the eigenvalues of H, since we know

those for L2 , and can find those of W 2 by changing variables to identify a second

so(3) Lie algebra.

To do this, first change normalization by defining

r

m

K= W

2E

155

where E is the eigenvalue of the Hamiltonian that we are trying to solve for.

Note that it is at this point that we violate the conditions of the no-go theorem,

since we must have E < 0 to get a K with the right properties, and this restricts

the validity of our calculations to a subset of phase space. For E > 0 one can

proceed in a similar way, but the Lie algebra one gets is different (so(3, 1) instead

of so(4)).

One then has the following relation between operators

2H(K 2 + L2 + ~2 ) = me4

[Lj , Lk ] = i~jkl Ll

[Kj , Lk ] = i~jkl Kl

[Kj , Kk ] = i~jkl Ll

Defining

1 1

M= (L + K), N = (L K)

2 2

one has

[Mj , Mk ] = i~jkl Ml

[Nj , Nk ] = i~jkl Nl

[Mj , Nk ] = 0

which is just two commuting copies of so(3).

Recall from our discussion of rotations in three dimensions that representa-

tions of so(3) = su(2) correspond to representations of Spin(3) = SU (2), the

double cover of SO(3) and the irreducible ones have dimension 2l + 1, with l

half-integral. Only for l integral does one get representations of SO(3), and it

is these that occur in the SO(3) representation on functions on R3 . For four di-

mensions, we found that Spin(4), the double cover of SO(4), is SU (2) SU (2),

and one thus has spin(4) = so(4) = su(2)su(2) = so(3)so(3). This is exactly

the Lie algebra we have found here, so one can think of the Coulomb problem

as having an so(4) symmetry. The representations that will occur will include

the half-integral ones, since for this larger symmetry one of our so(3) factors no

longer corresponds to the action of physical rotations on functions on R3 .

Using the fact that

LK=KL=0

one finds that

M2 = N2

so on our energy eigenstates the values of the Casimirs for the two so(3) factors

must match. The relation between the Hamiltonian and these Casimirs is

156

On irreducible representations, we will have

M 2 = ( + 1)

for some half-integral , so we get the following equation for the energy eigen-

values

me4 me4

E= 2 = 2

2~ (4( + 1) + 1) 2~ (2 + 1)2

Letting n = 2 + 1, for = 0, 21 , 1, . . . we get n = 1, 2, 3, . . . and precisely the

same equation for the eigenvalues described earlier

me4

En =

2~2 n2

The irreducible representations of a product like so(3) so(3) are just tensor

products of irreducibles, and in this case the two factors of the product are

identical due to the equality of the Casimirs M 2 = N 2 . The dimension of the

so(3) so(3) irreducibles is thus (2 + 1)2 = n2 , explaining the multiplicity of

states one finds at energy eigenvalue En .

The Coulomb potential problem provides a good description of the quantum

physics of the hydrogen atom, but it is missing an important feature of that

system, the fact that electrons are spin 12 systems. To describe this, one really

needs to take as space of states

H = L2 (R3 ) C2

The action of the rotation group is now an SU (2) = Spin(3) action on a tensor

product. It is given by the homomorphism to SO(3) and the rotation group

action of SO(3) on the first factor of the tensor product. On the second factor

it is the spinor representation. Physicists generally write the tensor product Lie

algebra representation of the SU (2) operators on H as

J=L+S

where the first term acts just on L2 (R3 ) by the rotation action, the second on

C2 as the spinor representation, so should perhaps be written as

J=L1+1S

+ (q)

|i =

(q)

where the + (q) is thought of as the wave-function for the electron to be in the

spin up state (+ 21 eigenvalue of S3 ), (q) the wave-function for the spin

157

down state ( 12 eigenvalue of S3 ). The Hamiltonian operator is just the identity

on the spinor factor, so the only effect of this factor is to double the number of

energy eigenstates at each energy. Electrons are fermions, so anti-symmetry of

multi-particle wave functions implies the Pauli principle that states can only be

occupied by a single particle. As a result, one finds that when adding electrons

to an atom described by the Coulomb potential problem, the first two fill up the

lowest Coulomb energy eigenstate (the 100 or 1S state at n = 1), the next eight

fill up the n = 2 states ( two each for 200 , 211 , 210 , 211 ), etc. This goes a

long ways towards explaining the structure of the periodic table of elements.

When one puts a hydrogen atom in a constant magnetic field B, the Hamil-

tonian then acts non-trivially on the spinor factor of H, through a term

e

BS

mc

This is exactly the sort of Hamiltonian we began our study of quantum mechan-

ics with for a simple two-state system. It causes a shift in energy eigenvalues

proportional to |B| depending on the eigenvalue of S3 (the Zeeman effect), giving

different energies to the states in + (q) and (q).

This is a standard topic in all quantum mechanics books. For example, see

chapters 12 and 13 of [40]. The so(4) calculation is not in [40], but is in some of

the other such textbooks, a good example is chapter 7 of [4]. For an extensive

discussion of the symmetries of the 1r potential problem, see [22].

158

Chapter 15

In this chapter well begin the study of the most important exactly solvable

physical system, the harmonic oscillator. Later chapters will discuss extensions

of the methods developed here to the case of fermionic oscillators, as well as free

quantum field theories, which are harmonic oscillator systems with an infinite

number of degrees of freedom.

rem tells us that there is essentially just one way to non-trivially represent the

(exponentiated) Heisenberg commutation relations as operators on a quantum

mechanical state space. We have seen two unitarily equivalent constructions

of these operators: the Schr odinger representation in terms of functions on ei-

ther coordinate space or momentum space. It turns out that there is another

class of quite different constructions of these operators, one that depends upon

introducing complex coordinates on phase space and then using properties of

holomorphic functions. Well refer to this as the Bargmann-Fock representation,

although quite a few mathematicians have had their name attached to it for one

good reason or another (some of the other names one sees are Friedrichs, Segal,

Shale, Weil, as well as the descriptive terms holomorphic and oscillator).

Hamiltonian operator for a fundamental sort of quantum system: the harmonic

oscillator. In the Bargmann-Fock representation the energy eigenstates of such

a system are the simplest states and energy eigenvalues are just integers. These

integers label the irreducible representations of the U (1) symmetry generated

by the Hamiltonian, and they can be interpreted as counting the number of

quanta in the system. It is the ubiquity of this example that justifies the

quantum in quantum mechanics. The operators on the state space can be

simply understood in terms of basic operators called annihilation and creation

operators which increase or decrease by one the number of quanta.

159

15.1 The harmonic oscillator with one degree of

freedom

An even simpler case of a particle in a potential than the Coulomb potential of

the last chapter is the case of V (q) quadratic in q. This is also the lowest-order

approximation when one studies motion near a local minimum of an arbitrary

V (q), expanding V (q) in a power series around this point. Well write this as

p2 1

h= + m 2 q 2

2m 2

with coefficients chosen so as to make the angular frequency of periodic motion

of the classical trajectories. These satisfy Hamiltons equations

V p

p = = m 2 q, q =

q m

so

q = 2 q

which will have solutions with periodic motion of angular frequency . These

solutions can be written as

solutions of the equation of motion is thus two-dimensional, and abstractly one

can think of this as the phase space of the system.

More conventionally, one can parametrize the phase space by initial values

that determine the classical trajectories, for instance by the position q(0) and

momentum p(0) at an initial time t(0). Since

we have

so

1 1

c+ = q(0) + i p(0)

2 2m

The classical phase space trajectories are

1 1 1 1

q(t) = ( q(0) + i p(0))eit + ( q(0) i p(0))eit

2 2m 2 2m

im 1 im 1

p(t) = ( q(0) p(0))eit + ( q(0) + p(0))eit

2 2 2 2

160

Instead of using two real coordinates to describe points in the phase space

(and having to introduce a reality condition when using complex exponentials),

one can instead use a single complex coordinate

1 i

z(t) = (q(t) + p(t))

2 m

Then the equation of motion is a first-order rather than second-order differential

equation

z = iz

with solutions

z(t) = z(0)ei(t)

The classical trajectories are then realized as complex functions of t, and paramet-

rized by the complex number

1 i

z(0) = (q(0) + p(0))

2 m

Since the Hamiltonian is just quadratic in the p and q, we have seen that we

can construct the corresponding quantum operator uniquely using the Schroding-

er representation. For H = L2 (R) we have a Hamiltonian operator

P2 1 ~2 d2 1

H= + m 2 Q2 = + m 2 q 2

2m 2 2m dq 2 2

To find solutions of the Schr odinger equation, as with the free particle, one

proceeds by first solving for eigenvectors of H with eigenvalue E, which means

finding solutions to

~2 d2 1

HE = ( 2

+ m 2 q 2 )E = EE

2m dq 2

Solutions to the Schr

odinger equation will then be linear combinations of the

functions

i

E (q)e ~ Et

Standard but somewhat intricate methods for solving differential equations

like this show that one gets solutions for E = En = (n + 21 )~, n a non-negative

integer, and the normalized solution for a given n (which well denote n ) will

be r

m 1 m m q2

n (q) = ( 2n 2

) 4 Hn ( q)e 2~

~2 (n!) ~

where Hn is a family of polynomials called the Hermite polynomials. The

n provide an orthonormal basis for H (one does not need to consider non-

normalizable wave-functions as in the free particle case), so any initial wave-

function (q, 0) can be written in the form

X

(q, 0) = cn n (q)

n=0

161

with Z +

cn = n (q)(q, 0)dq

(note that the n are real-valued). At later times, the wavefunction will be

i 1

X X

(q, t) = cn n (q)e ~ En t = cn n (q)ei(n+ 2 )t

n=0 n=0

It turns out that there is a quite easy method to solve for eigenfunctions of

the harmonic oscillator Hamiltonian, which also leads to a new representation

of the Heisenberg group (of course unitarily equivalent to the Schrodinger one

by the Stone-von Neumann theorem). Instead of working with the self-adjoint

operators Q and P that satisfy the commutation relation

[Q, P ] = i~1

we define

r r r r

m 1 m 1

a= Q+i P, a = Qi P

2~ 2m~ 2~ 2m~

which satisfy the commutation relation

[a, a ] = 1

corresponds to a specific choice of units for energy, distance and time, and the

general case of arbitrary constants can be recovered by rescaling the results of

our calculations. So, now

1 1

a = (Q + iP ), a = (Q iP )

2 2

and

1 1

Q = (a + a ), P = (a a )

2 i 2

and the Hamiltonian operator is

1 2 1 1 1

H= (Q + P 2 ) = ( (a + a )2 (a a )2 )

2 2 2 2

1

= (aa + a a)

2

1

=a a +

2

The fact that the eigenvalues of H will be (n + 12 ) (in units of ~) corresponds

to

162

Theorem. The number operator N = a a has eigenvalues n = 0, 1, 2, . . . on

H = L2 (R).

Since

|a|ci|2 = hc|a a|ci = hc|N |ci = chc|ci

if |ci has norm 1 then

1

a|ci = |c 1i

c

will be normalized.

In order to have non-negative eigenvalues for N , the sequence one gets by

repeatedly applying a to a N eigenvector must terminate. The only way for this

to happen is for there to be an eigenvector |0i = a|1i for which we will have

a|0i = 0

ultimately lead to states with negative values of c.

Similarly, one can show that

a |ci = c + 1|c + 1i

N.

a|0i = 0

for |0i (the lowest energy or vacuum state) which will have energy eigenvalue

1 1

2 , then acting by a n-times on |0i to get states with energy eigenvalue n + 2 .

The equation for |0i is thus

1 1 d

a|0i = (Q + iP )0 (q) = (q + )0 (q) = 0

2 2 dq

One can check that solutions to this are all of the form

q2

0 (q) = Ce 2

163

Choosing C to make this of norm 1 gives

1 q2

0 (q) = 1 e 2

a a a 1 d q2

|ni = |0i = 1 n (q )n e 2

n 2 1 4 2 2 n! dq

In the physical interpretation of this quantum system, the state |ni, with

energy ~(n + 21 ) is thought of as a state describing n quanta. The state

|0i is the vacuum state with zero quanta, but still carrying a zero-point

energy of 12 ~. The operators a and a have somewhat similar properties to

the raising and lowering operators we used for SU (2) but their commutator

is different (just the identity operator), leading to simpler behavior. In this

case they are called creation and annihilation operators respectively, due

to the way they change the number of quanta. The relation of such quanta to

physical particles like the photon is that quantization of the electromagnetic field

involves quantization of an infinite collection of oscillators, with the quantum

of an oscillator corresponding physically to a photon with a specific momentum

and polarization. This leads to a well-known problem of how to handle the

infinite vacuum energy corresponding to adding up 21 ~ for each oscillator.

The first few eigenfunctions are plotted below. The lowest energy eigenstate

is a Gaussian centered at q = 0, with a Fourier transform that is also a Gaussian

centered at p = 0. Classically the lowest energy solution is an oscillator at rest at

its equilibrium point (q = p = 0), but for a quantum oscillator one cannot have

such a state with a well-defined position and momentum. Note that the plot

gives the wave functions, which in this case are real and can be negative. The

square of this function is what has an intepretation as the probability density

for measuring a given position.

164

15.3 The Bargmann-Fock representation

Working with the operators a and a and their commutation relation

[a, a ] = 1

makes it clear that there is a simpler way to represent these operators than

the Schr

odinger representation as operators on position space functions that we

have been using, while the Stone-von Neumann theorem assures us that this will

be unitarily equivalent to the Schrodinger representation. This representation

appears in the literature under a large number of different names, depending on

the context, all of which refer to the same representation:

Definition (Bargmann-Fock or oscillator or holomorphic or Segal-Shale-Weil

representation). The Bargmann-Fock (etc.) representation is given by taking

H = C[w]

(this is the vector space of polynomials of a complex variable z), and operators

acting on this space

d

a= , a = w

dw

165

The inner product on H is

Z

1 2

h1 |2 i = 1 (w)2 (w)e|w| dudv

C

where w = u + iv.

One has

d d n

[a, a ]z n = (wwn ) w w = (n + 1 n)wn = wn

dw dw

so this commutator is the identity operator on polynomials

[a, a ] = 1

and

The elements

wn

n!

of H for n = 0, 1, 2, . . . are orthornormal.

The operators a and a are adjoints with respect to the given inner product

on H.

The basis

wn

n!

of H for n = 0, 1, 2, . . . is complete.

Proof. The proofs of the above statements are not difficult, in outline they are

Z

2

wm wn e|w| dudv

C

in polar coordinates.

d

To show that w and dw are adjoint operators, use integration by parts.

For completeness, assume hn| = 0 for all n. The expression for the |ni

as Hermite polynomials times a Gaussian implies that

Z

q2

F (q)e 2 (q)dq = 0

166

q2

for all polynomials F (q). Computing the Fourier transform of (q)e 2

gives

(ikq)j q2

Z

q2 X

eikq e 2 (q)dq = e 2 (q)dq = 0

j=0

j!

q2

So (q)e 2 has Fourier transform 0 and must be 0 itself. Alternatively,

one can invoke the spectral theorem for the self-adjoint operator H, which

guarantees that its eigenvectors form a complete and orthonormal set.

The operators a and a+ take one from any energy eigenstate |ni to any

other, so linear combinations in H form an irreducible representation. By

completeness there are no states |i orthogonal to all the |ni. Alterna-

tively, one can argue that if there was such a |i, by the argument that

showed that the eigenvalues of N are natural numbers, the same argu-

ment would imply that |i would be in a chain of states terminating with

a state |0i0 such that a|0i0 = 0, with |0i0 6= |0i. But we saw that there

was only one solution for a|0i = 0. In general, given a representation of

the Heisenberg algebra, one can check for irreducibility by seeing if there

is a unique solution to a|0i = 0.

d n

N wn = w w = nwn

dw

the monomials in z diagonalize the number and energy operators, so one has

wn

|ni =

n!

for the normalized energy eigenstate of energy ~(n + 21 ).

Up until now we have been working with the simple case of one physical degree of

freedom, i.e. one pair (Q, P ) of position and momentum operators satisfying the

Heisenberg relation [Q, P ] = i1, or one pair of adjoint operators a, a satisfying

[a, a ] = 1. We can easily extend this to any number d of degrees of freedom by

taking tensor products of our state space Hd=1 , and d copies of our operators,

acting on each factor of the tensor product. Our new state space will be

H = Hd=1 Hd=1

| {z }

d times

Qj , Pj , j = 1, . . . d

167

satisfying

[Qj , Pk ] = ijk 1, [Qj , Qk ] = [Pj , Pk ] = 0

where Qj and Pj just act on the jth term of the tensor product in the usual

way. Similarly, we will have

1 1

aj = (Qj + iPj ), aj = (Qj iPj ), j = 1, . . . d

2 2

satisfying

[aj , ak ] = jk 1, [aj , ak ] = [aj , ak ] = 0

These sets of commutation relations are known as the canonical commutation

relations (often abbreviated CCR).

Using the fact that tensor products of function spaces correspond to func-

tions on the product space, in the Schrodinger representation we have

H = L2 (Rd )

H = C[w1 , , wd ]

The harmonic oscillator Hamiltonian for d degrees of freedom will be

d d

1X 2 X 1

H= (Pj + Q2j ) = (aj aj + )

2 j=1 j=1

2

where one should keep in mind that one can rescale each degree of freedom

separately, allowing different parameters j for the different degrees of freedom.

The energy and number operator eigenstates will be written

|n1 , , nd i

where

aj aj |n1 , , nd i = Nj |n1 , , nd i = nj |n1 , , nd i

Either the Pj , Qk or the aj , ak together with the identity operator will give

a representation of the Heisenberg Lie algebra h2d+1 on H, and by exponentia-

tion a representation of the Heisenberg group H2d+1 . Quadratic combinations

of these operators will give a representation of sp(2d, R), the Lie algebra of

Sp(2d, R). In the next two chapters we will study these and other aspects of

the quantum harmonic oscillator as a unitary representation.

All quantum mechanics books should have a similar discussion of the harmonic

oscillator, with a good example the detailed one in chapter 7 of Shankar[40].

168

Chapter 16

a Representation of the

Heisenberg Group

provides new insight into the infinite dimensional representation of the Heisen-

berg and metaplectic groups, using the existence of not just the Schrodinger

version of , but the unitarily equivalent Bargmann-Fock version. In this chap-

ter well examine various aspects of the Heisenberg group H2d+1 part of this

story that the formalism of annihilation and creation operators illuminates, go-

ing beyond the understanding of this representation one gets from the physics

of the free particle and the Schrodinger representation.

The Schr odinger representation of H2d+1 (which well now call S ) uses a

specific choice of extra structure on classical phase space, a decomposition of

its coordinates into positions qj and momenta pj . For the unitarily equivalent

Bargmann-Fock representation (which well denote BF ), a different sort of ex-

tra structure is needed, a decomposition into complex coordinates and their

complex conjugates. This is called a complex structure: we must first com-

plexify our phase space R2d , then make a specific choice of complex coordinates

and their complex conjugates, corresponding to how we choose to define an-

nihilation and creation operators. This choice has physical significance, it is

equivalent to a specification of the lowest energy state |0i H. Equivalently, it

corresponds to a choice of Hamiltonian operator H.

The Heisenberg group action on H = C[w1 , . . . , wd ] given by BF does not

commute with the Hamiltonian H, so it is not what is normally described as

a symmetry, which would take states to other states with the same energy.

While it doesnt commute with the Hamiltonian, it does have physically inter-

esting aspects, since it takes the state |0i to a distinguished set of states known

as coherent states. These states are labeled by the phase space R2d and pro-

vide the closest analog possible in the quantum system of classical states (i.e.

169

those with a well-defined value of position and momentum variables).

Quantization of phase space M = R2d using the Schrodinger representation

gives a representation 0S which takes the qj and pj coordinate functions on

phase space to operators Qj and Pj on HS = L2 (Rd ). To deal with the har-

monic oscillator, it is far more convenient to define annihilation and creation

operators aj , aj and work with the Bargmann-Fock representation 0BF of these

on polynomials HBF = C[w1 , . . . , wd ]. In the Schrodinger case we needed to

make a choice, that of taking states to be functions of the qj , or of the pj . A

different kind of choice needs to be made in the Bargmann-Fock case: certain

complex linear combinations of the Qj , Pj are taken as annihilation operators,

and complex conjugate ones are taken to be creation operators. Once we decide

which are the annihilation operators, that specifies the state |0i as the state

they all annihilate.

To understand what the possible consistent such choices are, we need to

begin by introducing the following definition:

ator

J : R2d R2d

such that

J 2 = 1

The point of this is that, abstractly, to make a real vector space a complex

vector space, we need to know how to multiply vectors by complex numbers. We

already know how to multiply them by real numbers, so just need to specify what

happens when we multiply by the imaginary unit i. The choice of a complex

structure J provides exactly that piece of information.

Given such a pair (V = R2d , J), we can now break up complex linear com-

binations of vectors in V into those on which J acts as i and those on which it

acts as i (since J 2 = 1, its eigenvalues must be i). We have

V C = VJ+ VJ

eigenspace. Complex conjugation takes elements of VJ+ to VJ and vice-versa.

We can construct a generalization of the Bargmann-Fock representation by

choosing a complex structure J on the dual phase space M = M , which gives

a decomposition

M C = M+ J MJ

h2d+1 = M R

170

with Lie bracket determined by the anti-symmetric bilinear form

Taking complex linear combinations of h2d+1 elements to get a new Lie algebra,

the complexification h2d+1 C, we just use the same formula for the Lie bracket,

with u, u0 now complex vectors in M C, and c, c0 complex numbers.

The choice of a complex structure J on M allows one to write elements of

h2d+1 C as

(u = u+ + u , c) h2d+1 C = (M C) C = M+

J MJ C

where

u+ M+

J , u MJ

, satisfying

(Ju, Jv) = (u, v)

Then, if u+ , v + M+

J , u , v MJ we have

(u , v ) = 0

so

(u+ + u , v + + v ) = (u+ , v ) + (u , v + )

When we quantize elements of M+ J as annihilation operators, these will com-

mute among themselves (similarly for elements of M J which will be creation

operators), and the only non-zero commutation relations will come from the

case of commuting annihilation operators with creation operators.

To define a generalization of the Bargmann-Fock representation for each J,

we need to find a unitary Lie algebra representation 0BF,J of h2d+1 such that,

when we extend by complex linearity to linear operators

we have

operator, and 0BF,J (0, c) = ic1. The first two components can be explicitly

realized as differentiation and multiplication operators on a state space HBF,J of

complex polynomials in d variables, the third is just multiplication by a scalar.

As a technical aside, an additional positivity condition on J is required here

to make the analog of the standard Bargmann-Fock inner product well-defined,

giving a unitary representation.

171

The standard Bargmann-Fock representation 0BF corresponds to a standard

choice of complex structure J on M, one such that on basis elements

Jqj = pj , Jpj = qj

J(cq q + cp p) = cp q cq p

so J in matrix form is

c 0 1 cq cp

J q = =

cp 1 0 cp cq

It is compatible with since

0

0 1 cq T 0 1 0 1 cq

(J(cq q + cp p), J(c0q q + c0p p)) =( ) ( )

1 0 cp 1 0 1 0 c0p

0

0 1 0 1 0 1 cq

= cq cp

1 0 1 0 1 0 c0p

0

0 1 cq

= cq cp

1 0 c0p

=(cq q + cp p, c0q q + c0p p)

Basis elements of M+

J are then

1

zj = (qj + ipj )

2

since one has

1

Jzj = (pj + iqj ) = izj

2

Basis elements of M

J are the complex conjugates

1

z j = (qj ipj )

2

The definition of Poisson brackets extends to functions of complex coordi-

nates with the conjugate variable to zj becoming iz j since

1 1

{zj , z k } = { (qj + ipj ), (qj ipj )} = ijk

2 2

or

{zj , iz k } = jk

These Poisson bracket relations are just the Lie bracket relations on the Lie

algebra h2d+1 C, since as we have seen, a choice of complex structure J has

been made allowing us to write

h2d+1 C = M+

J MJ C

172

and the only non-zero Lie brackets in this Lie algebra are the ones between

elements of M+J and MJ given by the Poisson bracket relation above.

Quantization takes

zj aj , z j aj , 1 1

and these operators can be realized on the state space C[w1 , . . . , wd ]. The

aj , aj are neither self-adjoint nor skew-adjoint, but if we follow our standard

convention of multiplying by i to get the Lie algebra representation, we have

0BF (zj , 0) = iaj = i , 0BF (z j , 0) = iaj = iwj , 0BF (0, 1) = i1

wj

=jk (i)(i1) = jk 0BF (0, i)

h2d+1 C

The Stone von-Neumann theorem implies the existence of

There is a unitary map called the Bargmann transform

B : HBF HS

satisfying the relation

0BF (X) = B 1 0S (X)B

for X h2d+1 .

In practice, knowing B explicitly is often not needed, since one can use the

representation independent relation

1

aj = (Qj + iPj )

2

expression

aj = , aj = wj

wj

173

in the Bargmann-Fock representation, or purely in terms of Qj and Pj which

have a simple expression

Qj = qj , Pj = i

qj

in the Schrodinger representation.

To give an idea of what the Bargmann transform looks like explicitly, well

just give the formula for the d = 1 case here, without proof. If (q) is a state

in HS = L2 (R), then

Z +

1 2 q2

(B)(z) = 1 ez 2 +2qz (q)dq

4

One can check this equation for the case of the lowest energy state in the

Schrodinger representation, where

1 q2

|0i = (q) = 1 e 2

and

Z +

1 2 2

q2 +2qz 1 q2

B(z) = 1 ez 1 e 2 dq

4 4

Z +

1 2

q 2 +2qz

= 1 ez dq

2

Z +

1 2

= 1 e(qz) dq

2

Z +

1 2

= 1 eq dq

2

=1

action

Since the Hamiltonian for the harmonic oscillator does not commute with the

operators aj or aj which give the representation of the Lie algebra h2d+1 on

the state space HBF , the Heisenberg Lie group and its Lie algebra are not

symmetries of the system in the conventional sense. Energy eigenstates do

not break up into irreducible representations of the group but rather the entire

state space makes up such an irreducible representation. The state space for the

harmonic oscillator does however have a distinguished state, the lowest energy

state |0i, and one can ask what happens to this state under the Heisenberg

group action. Well study this question for the simplest case of d = 1.

Considering the basis of operators for the Lie algebra representation 1, a, a ,

we see that the first acts as a constant on |0i, generating a phase tranformation

174

of the state, while the second annihilates |0i, so generates group transformations

that leave the state invariant. It is only the third operator a , that takes |0i to

distinct states, and one could consider the family of states

ea |0i

for C. The transformations ea are not unitary since a is not skew ad-

joint. It is better to fix this by replacing a with the skew-adjoint combination

a a, defining

Definition (Coherent states). The coherent states in H are the states

a

|i = ea |0i

where C.

Since ea a is unitary, the |i will be a family of distinct normalized states

in H, with = 0 corresponding to the lowest energy state |0i. These are, up

to phase transformation, precisely the states one gets by acting on |0i with

arbitrary elements of the Heisenberg group H3 .

Using the Baker-Campbell-Hausdorff formula one has

||2

a

|i = ea |0i = ea ea e 2 |0i

||2 ||2 X n

|i = e 2 ea |0i = e 2 |ni

n=0 n!

Since a|ni = n|n 1i one has

||2 X n

a|i = e 2 p |n 1i = |i

n=0 (n 1)!

Note that coherent states are superpositions of different states |ni, so are

not eigenvectors of the number operator N . They are eigenvectors of

1

a = (Q + iP )

2

with eigenvalue so one can try and think of as a complex number whose

real part gives the position and imaginary part the momentum. This does not

lead to a violation of the Heisenberg uncertainly principle since this is not a

self-adjoint operator, and thus not an observable. Such states are however very

useful for describing certain sorts of physical phenomena, for instance the state

of a laser beam, where one does not have a definite number of photons, but does

have a definite amplitude and phase.

175

One thing coherent states do provide is an alternate complete set of norm

one vectors in H, so any state can be written in terms of them. However, these

states are not orthogonal (they are eigenvectors of a non-self-adjoint operator so

the spectral theorem for self-adjoint operators does not apply). One can easily

compute that

2

|h|i|2 = e||

Digression (Spin coherent states). One can perform a similar construction

replacing the group H3 by the group SU (2), and the state |0i by a highest weight

vector of an irreducible representation (n , V n = Cn+1 ) of spin n2 . Writing | n2 i

for a highest weight vector, we have

n n n

n0 (S3 )| i = , n0 (S+ )| i = 0

2 2 2

and we can create a family of spin coherent states by acting on | n2 i by elements

of SU (2). If we identify states in this family that differ just by a phase, the

states are parametrized by a sphere.

By analogy with the Heisenberg group coherent states, with n0 (S+ ) playing

the role of the annihilation operator a and n0 (S ) playing the role of the creation

operator a , we can define a skew-adjoint transformation

1 i 0 1

e n (S ) ei n0 (S+ )

2 2

and exponentiate to get a family of unitary transformations parametrized by

(, ). Acting on the highest weight state we get a definition of the family of

spin coherent states as

1 i 0

(S ) 21 ei n

0 n

|, i = e 2 e n (S+ )

| i

2

One can show that the SU (2) group element used here corresponds, in terms of

its action on vectors, to a rotation by an angle about the axis (sin , cos , 0),

so one can associate the state |, i to the unit vector along the z-axis, rotated

by this transformation.

Some references for more details about the Bargmann transform and the Bargmann-

Fock representation as a representation of h2d+1 are [18] and [7]. Coherent states

and spin coherent states are discussed in chapter 21 of [40].

176

Chapter 17

a Representation of the

Metaplectic Group

In the last chapter we examined those aspects of the harmonic oscillator quan-

tum system and the Bargmann-Fock representation that correspond to quan-

tization of phase space functions of order less than or equal to one, giving a

unitary representation BF of the Heisenberg group H2d+1 . Well now turn

to what happens when one also includes quadratic functions, extending BF

to a representation of a larger group, the semi-direct product of H2d+1 and

M p(2d, R) (a double cover of Sp(2d, R)).

As a representation of M p(2d, R), BF is unitarily equivalent (by the Barg-

mann transform) to S , the Schr odinger representation version studied earlier.

The Bargmann-Fock version makes some aspects of the representation easier to

study; in particular it makes clear the nature of the double-cover that appears.

This construction involves making a specific choice of complex structure J, a

choice which corresponds to a a choice of Hamiltonian operator H. This choice

also picks out a subgroup U (d) Sp(2d, R) of transformations which commute

with J. The Hamiltonian gives the action of the U (1) diagonal subgroup of

U (d), which commutes with the U (d) action, so energy eigenstates are sets of

irreducible representations of U (d).

The representation BF thus carries an action of several different sorts of

operators: linear combinations of the Qj , Pj studied in the last chapter, one

quadratic combination that gives the Hamiltonian H, a set of quadratic com-

binations that give an SU (d) action on the energy eigenstates, and yet other

quadratic combinations that dont commute with the Hamiltonian, and change

the lowest energy state |0i (these are called Bogoliubov transformations).

177

17.1 The metaplectic representation for d = 1

In the last chapter we saw that introducing annihilation and creation operators

corresponded to choosing a complex structure on the dual phase space M =

R2d , allowing one to use complex coordinates zj , z j instead of real coordinates

qj , pj . After quantization, instead of operators Qj , Pj , we could work with

operators aj , aj , and restricting to first order operators these give the Bargmann-

Fock representation 0BF of h2d+1 on H = C[w1 , . . . , wd ], which corresponds to

the quantum harmonic oscillator system in d degrees of freedom.

Recall from our discussion of the Schrodinger representation 0S that we can

extend this quantization to include quadratic combinations of the qj , pj , getting

a unitary representation of the semi-direct product h2d+1 o sp(2d, R). Restrict-

ing attention to the sp(2d, R) factor, we get the metaplectic representation, and

it is this that we will construct explicitly using 0BF in this chapter, beginning

in this section with the case d = 1, where sp(2, R) = sl(2, R).

One can readily compute the Poisson brackets of quadratic combinations of

z and z using the basic relation {z, z} = i and the Leibniz rule, finding the

following for the non-zero cases

{zz, z 2 } = 2iz 2 , {zz, z 2 } = 2iz 2 , {z 2 , z 2 } = 4izz

In the case of the Schrodinger representation, our quadratic combinations of p

and q were real, and we could identify the Lie algebra they generated with the

Lie algebra sl(2, R) of traceless 2 by 2 real matrices with basis

0 1 0 0 1 0

E= , F = , G=

0 0 1 0 0 1

Since we have complexified, our quadratic combinations of z and z are in the

complexification of sl(2, R), the Lie algebra sl(2, C) of traceless 2 by 2 complex

matrices. We can take as a basis of sl(2, C) over the complex numbers

1

Z = E F, X = (G i(E + F ))

2

which satisfy

[Z, X+ ] = 2iX+ , [Z, X ] = 2iX , [X+ , X ] = iZ

and then use as our isomorphism between quadratics in z, z and sl(2, C)

z2 z2

X+ , X , zz Z

2 2

The element

1 0 1

zz = (p2 + q 2 ) Z =

2 1 0

exponentiates to give a SO(2) = U (1) subgroup of SL(2, R) with elements of

the form

cos sin

eZ =

sin cos

178

Note that h = 12 (p2 + q 2 ) = zz is the classical Hamiltonian function for the

harmonic oscillator.

We can now quantize quadratics in z and z using annihilation and creation

operators acting on H = C[w]. There is no operator ordering ambiguity for

d2

z 2 a2 = , z 2 (a )2 = w2

dw2

For the case of zz (which is real), in order to get the sl(2, R) commutation

relations to come out right, we must take the symmetric combination

1 1 d 1

zz (aa + a a) = a a + = w +

2 2 dw 2

(which of course is just the standard harmonic oscillator quantization of the

Hamiltonian).

Multiplying as usual by i to represent real elements of the Lie algebra as

the skew-Hermitian operator of a unitary group representation, one can now

define

Definition (Metaplectic representation of sl(2, R)). The representation 0BF

on H = C[w] given by

i i 1

0BF (X+ ) = a2 , 0BF (X ) = (a )2 , 0BF (Z) = i (a a + aa )

2 2 2

is called the metaplectic representation of sl(2, R).

One can check that we have made the right choice of 0BF (Z) to get a sl(2, R)

representation by computing

i i

[0BF (X+ ), 0BF (X )] =[ a2 , (a )2 ]

2 2

1

= (aa + a a)

2

= i0BF (Z)

Note that we have actually defined the Lie algebra homomorphism on a basis of

sl(2, C), so we have a Lie algebra representation of this larger Lie algebra, but

it is only on the real sl(2, R) subalgebra that it is a unitary representation.

This representation 0BF will be unitarily equivalent to the Schrodinger ver-

sion 0S found earlier when quantizing q 2 , p2 , pq as operators on H = L2 (R).

It is however much easier to work with since it can be studied as the state

space of the quantum harmonic oscillator, with the Lie algebra acting simply

by quadratics in the annihilation and creation operators.

One thing that can easily now be seen is that this representation 0BF does

not integrate to give a representation of the group SL(2, R). If the Lie algebra

representation 0BF comes from a Lie group representation BF of SL(2, R), we

have 0

BF (eZ ) = eBF (Z)

179

where

1 1

0BF (Z) = i(a a + ) = i(N + )

2 2

so

1

(eZ )|ni = ei(n+ 2 ) |ni

Taking = 2, this gives an inconsistency

BF (1)|ni = |ni

which has its origin in the physical phenomenon that the energy of the lowest

energy eigenstate |0i is 21 rather than 0, so not an integer.

This is precisely the same sort of problem we found when studying the

spinor representation of the Lie algebra so(3). Just as in that case, the problem

indicates that we need to consider not the group SL(2, R), but a double cover,

often called the metaplectic group and denoted M p(2, R). The behavior here

is quite a bit more subtle than in the Spin(3) double cover case, where Spin(3)

was just the group SU (2), and topologically the only non-trivial cover of SO(3)

was the Spin(3) one. Here one has 1 (SL(2, Z)) = Z, and each time one goes

around the U (1) subgroup we are looking at one gets a topologically different

non-contractible loop in the group. As a result, SL(2, R) has lots of non-trivial

covering groups, only one of which is of interest here.

Digression. This group M p(2, R) is quite unusual in that it is a finite-dim-

ensional Lie group, but does not have any sort of description as a group of

finite-dimensional matrices. This is related to the fact that its only interesting

irreducible representation is the infinite-dimensional one we are studying. The

lack of any significant irreducible finite-dimensional representations corresponds

to its not having a matrix description, which would give such a representation.

Note that the lack of a matrix description means that this is a case where the

definition we gave of a Lie algebra in terms of the matrix exponential does not

apply. The more general geometric definition of the Lie algebra of a group in

terms of the tangent space at the identity of the group does apply, although to do

this one really needs a construction of the double cover M p(2, R), which is quite

non-trivial. This is not actually a problem for purely Lie algebra calculations,

since the Lie algebras of M p(2, R) and SL(2, R) can be identified.

Another aspect of the metaplectic representation that is relatively easy to

see in the Bargmann-Fock construction is that the state space H = C[w] is not

an irreducible representation, but is the sum of two irreducible representations

H = Heven Hodd

where Heven consists of the even-degree polynomials, Hodd the odd degree poly-

nomials. Since the generators of the Lie algebra representation are degree two

combinations of annihilation and creation operators, they will not change the

parity of a polynomial, taking even to even and odd to odd. The separate irre-

ducibility of these two pieces is due to the fact that (when n = m(2)), one can

get from state |ni to any another |mi by repeated application of the Lie algebra

representation operators.

180

17.2 The metaplectic representation and the choice

of complex structure

Our construction of the metaplectic representation depends on using a choice

of a basis q, p of M = R2 to get an irreducible unitary representation of the

Heisenberg algebra, and the group SL(2, R) acts on M in a way that takes one

such construction to another unitarily equivalent one. Elements of this group

are given by matrices

a b

c d

satisfying ac bd = 1. The Bargmann-Fock construction depends on a choice

of complex structure J to provide the splitting M C = M+ M into i

eigenspaces of J (and thus into annihilation and creation operators). We have

so far used the standard complex structure

0 1

J=

1 0

and one can ask how the SL(2, R) action interacts with this choice. Only a

subgroup of SL(2, R) will respect the splitting provided by J, it will be the

subgroup of matrices commuting with J. Acting with such matrices on M (and

extending by complex linearity to an action on M C), vectors that are J

eigenvectors with eigenvalue i with transform into other J eigenvectors with

eigenvalue i, and similarly for the i eigenspace. This commutativity condition

is explicitly

a b 0 1 0 1 a b

=

c d 1 0 1 0 c d

so

b a c d

=

d c a b

which implies b = c and a = d. The elements of SL(2, R) that we want will

be of the form

a b

b a

with unit determinant, so a2 + b2 = 1. This subgroup is the SO(2) = U (1)

subgroup of matrices of the form

cos sin

sin cos

To see how this works more explicitly, recall that what is happening here

is that SL(2, R) is acting by automorphisms on the Lie algebra h3 . The in-

finitesimal version of this action is given by the Poisson brackets between the

order-two polynomials of sl(2, R) and the order-one or zero polynomials of h3 .

These are

{pq, p} = p, {pq, q} = q

181

p2 q2

{

, q} = p, { , p} = q

2 2

Once we intoduce a complex structure and variables z, z, these Poisson bracket

relations break up into two sorts, the first of which is

This expresses the fact that the Hamiltonian function h = zz is the generator

of the U (1) action on M C that commutes with J, and it acts with weight 1

on the M+ which has basis z, and with weight 1 on M which has basis z.

Quantizing, one has the following operator relations

1 1

[ (aa + a a), a] = a, [ (aa + a a), a ] = a

2 2

which are the infinitesimal versions of

+a a)

+a a)

ei 2 (aa aei 2 (aa = ei a

and

+a a) i 2 (aa +a a)

ei 2 (aa a e = ei a

We see that, on operators, conjugation by the action of the U (1) subgroup of

SL(2, R) does not mix creation and annihilation operators. On states the U (1)

subgroup acts by

1

ei 2 (aa +a a) |ni = ei(n+ 2 ) |ni

and in particular

+a a)

ei 2 (aa |0i = ei 2 |0i

leaving the lowest energy state |0i invariant (up to a phase).

The second sort of Poisson bracket relation one has are these

{z 2 , z} = 2iz, {z 2 , z} = 0, {z 2 , z} = 2iz, {z 2 , z} = 0

i(a2 (a )2 )

will give the action of the metaplectic representation in the other two direc-

tions than the U (1) direction corresponding to zz. Such transformations act

non-trivially on the chosen complex structure J, taking it to a different one

J 0 , and giving a new set of annihilation a0 and creation operators (a )0 that

mixes the original ones. This sort of transformation is known to physicists as

a Bogoliubov transformation. The lowest energy state will be the state |0iJ 0

satisfying

a0 |0iJ 0 = 0

182

(this determines the state up to phase). The different Js (or equivalently, differ-

ent possible |0iJ up to phase) will be parametrized by the space SL(2, R)/U (1).

This is a space well-known to mathematicians as having a two-dimensional hy-

perbolic geometry, with two specific models the Poincare upper half plane model

and the Poincare disk model. Acting with both SL(2, R) and the Heisenberg

group on |0i we get a generalization of our construction of coherent states in

the previous chapter, with these new coherent states parametrized not just by

an element of M, but also by an element of SL(2, R)/U (1).

Whenever one has a product involving both z and z that one would like to

quantize, the non-trivial commutation relation of a and a means that one has

different inequivalent possibilities, depending on the order one chooses for the a

and a . In this case, we chose to quantize zz as 21 (aa + a a) = a a + 21 because

it then satisfied the commutation relations for sl(2, R). We could instead have

chosen to use a a, which is an example of a normal-ordered product.

Definition. Normal-ordered product

Given any product P of a and a , the normal ordered product of P , written

: P : is given by re-ordering the product so that all factors a are on the left, all

factors a on the right, for example

: a2 a a(a )3 := (a )4 a3

The advantage of working with the normal-ordered a a instead of 21 (aa +

a a) is that it acts trivially on |0i and has integer eigenvalues on all the states.

There is then no need to invoke a double-covering. The disadvantage is that one

loses the relation to sl(2, R). One also needs to keep in mind that the definition

of normal-ordering depends upon the choice of J, or equivalently, the choice

of lowest energy state |0iJ . A better notation would be something like : P :J

rather than just : P :.

grees of freedom

Generalizing to the case of d degrees of freedom, we can consider the polynomials

of order two in coordinates qj , pj on phase space. These are a basis for the Lie

algebra sp(2d, R), using the Poisson bracket as Lie bracket. The quantization

of qj and pj as operators Qj , Pj in the Schrodinger representation gives the

representation 0S of this Lie algebra on H = L2 (Rd ). To get the Bargmann-

Fock representation 0BF , recall that we need to introduce annihilation and

creation operators, which can be done with the standard choice

1 1

aj = (Qj + iPj ), aj = (Qj iPj ), j = 1, . . . d

2 2

and then realize these operators as differentiation and multiplication operators

aj = , aj = wj

wj

183

on the state space

H = C[w1 , , wd ]

of polynomials in d complex variables. One can easily generalize the Bargmann-

Fock inner product defined using the Gaussian from the d = 1 case to this one.

One can think of this as the quantization of the dual phase space M = R2d

given by choosing the standard complex structure J on R2d and using it to

decompose

M C = M+ M

where zj , j = 1, d are basis elements in M+ while z j , j = 1, d are

basis elements in M . As in the d = 1 case, one can ask what subgroup of

Sp(2d, R) preserves the decomposition. We wont show this here, but it turns

out to be a subgroup isomorphic to the unitary group U (d). The analog of

the d = 1 parametrization of possible |0iJ by SL(2, R)/U (1) here would be a

parametrization of such states (or, equivalently, of possible choices of J) by the

space Sp(2d, R)/U (d), known as the Siegel upper half-space.

The situation at the Lie algebra level is a bit easier to work out explicitly.

The choice of J gives a decomposition of the complexified Lie algebra sp(2d, C)

into three sub-algebras as follows:

A Lie subalgebra with basis elements zj zk (as usual, the Lie bracket is the

Poisson bracket). There are 12 (d2 + d) distinct such basis elements. This is

a commutative Lie subalgebra, since the Poisson bracket of any two basis

elements is zero.

A Lie subalgebra with basis elements z j z k . Again, it is has dimension

1 2

2 (d + d) and is a commutative Lie subalgebra.

puting Poisson brackets one finds

{zj z k , zl z m } =zj {z k , zl z m } + z k {zj , zl z m }

=zj z m ikl zl z k ijm

This can be identified with the Lie algebra gl(n, C) of all n by n complex

matrices, since if Mjk is the matrix with 1 at the jth row and kth column,

zeros elsewhere, one has

[Mjk , Mlm ] = Mjm kl Mlk jm

and these provide a basis of gl(n, C). Identifying bases by

izj z k Mjk

gives the isomorphism of Lie algebras.

The standard Hamiltonian

d

X

h= zj z j

j=1

184

lies in this sub-algebra, and its Poisson brackets with the rest of the sub-

algebra are zero since

This Lie subalgebra is the complexification of the Lie algebra u(d), with

h corresponding to the one-dimensional u(1) subalgebra that commutes

with everything.

Quantization takes

zj zk aj ak , z j z k aj ak

zj z k aj ak = ak aj

is to take

1

zj z j (aj aj + aj aj )

2

which will have the proper sp(2d, R) commutation relations but require going

to a double cover to get a true representation of the group. The other is to use

the normal-ordered version

zj z j aj aj

and then the state space H provides a representation of U (d), with no need for

a double cover.

The normal-ordered quantization of zj z k uses a property of the annihilation

and creation operators that is quite useful:

[A, B] = C

[aj , ak ] = jk

X X X

A= aj Ajk ak , B = aj Bjk ak , C = aj Cjk ak

j,k j,k j,k

[A, B] = C

185

Proof.

X

[A, B] =

j,k,l,m

X

=

j,k,l,m

X

=

j,k,l,m

the theorem says that

X

0 : A 0 (A) = aj Ajk ak

j,k

is a Lie algebra representation of u(d) on the space H that the aj , aj act on.

This is exactly the Lie algebra representation 0BF , restricted to the subalgebra

u(d) sp(2d, R) and normal-ordered. It exponentiates to give a representation

of U (d) on H = C[w1 , . . . , wd ], one that commutes with the Hamiltonian. From

the physics point of view, this is useful, as it provides a decomposition of energy

eigenstates into irreducible representations of U (d). From the mathematics

point of view, the quantum harmonic oscillator provides a construction of a

large class of representations of U (d) (the energy eigenstates). With no normal-

ordering, the representation 0BF itself also provides a representation of the Lie

algebra u(d), but this is one that does not exponentiate to U (d), requiring a

double-cover of U (d) to get a true representation.

If we specialize the above to the case d = 2 of two degrees of freedom, we

will reencounter our earlier construction of SU (2) representations in terms of

homogeneous polynomials, in a new context. This use of the energy eigenstates

of a two-dimensional harmonic oscillator appears in the physics literature as the

Schwinger boson method for studying representations of SU (2).

The state space for the d = 2 Bargmann-Fock representation is

H = C[w1 , w2 ]

the polynomials in two complex variables w1 , w2 . Recall from our SU (2) dis-

cussion that it was useful to organize these polynomials into finite dimensional

sets of homogeneous polynomials of degree n for n = 0, 1, 2, . . .

H = H0 H1 H2

186

There are four annihilation or creation operators

a1 = , a = w1 , a2 = , a = w2

w1 1 w2 2

acting on H. These are the quantizations of complex phase space coordinates

z1 , z2 , z 1 , z 2 with quantization just the Bargmann-Fock Lie algebra representa-

tion of the Lie algebra these span, with Lie bracket the Poisson bracket, so

Our original dual phase space was M = R4 , with a group Sp(4, R) acting

on it, preserving the Poisson bracket. When picking the coordinates z1 , z2 , we

have made a standard choice of complex structure J, which breaks up

M C = M+ M = C2 C2

of J picks out a distinguished subgroup U (2) Sp(4, R).

The quadratic combinations of the creation and annihilation operators give

representations on H of three subalgebras of the complexification of sp(4, R):

A three dimensional commutative Lie sub-algebra spanned by z1 z2 , z12 , z22 ,

with quantization

with quantization

z1 z 1 , z2 z 2 , z1 z 2 , z 1 z2

and quantization

i i

0BF (z1 z 1 ) = (a1 a1 + a1 a1 ), 0BF (z2 z 2 ) = (a2 a2 + a2 a2 )

2 2

0BF (z 1 z2 ) = ia1 a2 , 0BF (z1 z 2 ) = ia2 a1

The real elements (the ones that are in sp(4, R)) are a basis for u(2).

Inside this last subalgebra, there is a distinguished element h = z1 z 1 +

z2 z 2 that commutes withe the rest of the subalgebra, which give the quantum

Hamiltonian

1 1 1

H= (a1 a1 + a1 a1 + a2 a2 + a2 a2 ) = N1 + + N2 + = w1 + w2 +1

2 2 2 w1 w2

187

This operator will just multiply a homogeneous polynomial by its degree plus

one, so it acts just by multiplication by n + 1 on Hn . Exponentiating this

operator (multiplied by i) one gets a representation of a U (1) subgroup of the

metaplectic cover M p(4, R). Taking instead the normal-ordered version

: H := a1 a1 + a2 a2 = N1 + N2 = w1 + w2

w1 w2

one gets a representation of the U (1) Sp(4, R) that acts with opposite weight

on M+ and M . In either case, H does not commute with operators coming

from quantization of the first two subalgebras. These change the eigenvalue of

H or : H : by 2 so take

Hn Hn2

in particular taking |0i to 0 or a state in H2 .

Besides h, which gives the u(1) in u(2) = u(1) su(2), a basis for the su(2)

part, with its correspondence to our basis Xj = i 2j in terms of 2 by 2 matrices

is

1 i 1

X1 (z 1 z2 + z 2 z1 ), X2 (z 2 z1 z 1 z2 ), X3 (z 1 z1 z2 z2 )

2 2 2

This relates two different but isomorphic ways of describing su(2): as 2 by 2

matrices with Lie bracket the commutator, or as quadratic polynomials, with

Lie bracket the Poisson bracket.

Quantization will give a representation of su(2) on H

i 1

0BF (X1 ) = (a2 a1 + a1 a2 ), 0BF (X2 ) = (a2 a1 a1 a2 )

2 2

i

0BF (X3 ) = (a1 a1 a2 a2 )

2

Note that another way to get this result is by the theorem of the last section,

computing the quadratic operators as

a1

a1 a2 Xj

a2

that these are the same, up to a minus sign due to a relative inverse in the

definition used there, so

0BF (X) = 0 (X)

Also note that the inner product on the space of polynomials in two variables

studied in one of the problem sets is exactly the Bargmann-Fock inner product,

restricted to homogeneous polynomials.

We see that, up to this inverse, and the slight difference between 0BF and its

normal-ordered version, which only affects the u(1) factor (the Hamiltonian), the

Bargmann-Fock representation on polynomials and the SU (2) representation

188

on homogeneous polynomials are the same. The Bargmann-Fock representation

does extend as a unitary representation to a larger group (H5 o M p(4, R)).

The fact that we have an SU (2) group acting on the state space of the d = 2

harmonic oscillator and commuting with the action of the Hamiltonian H means

that energy eigenstates will can be organized as irreducible representations of

SU (2). In particular, one sees that the space Hn of energy eigenstates of energy

n+1 will be a single irreducible SU (2) representation, the spin n2 representation

of dimension n + 1 (so n + 1 will be the multiplicity of energy eigenstates of that

energy).

The case d = 3 corresponds physically to the so-called isotropic quantum

harmonic oscillator system, and it is an example of the sort of central po-

tential problem we have already studied (since the potential just depends on

r2 = q12 + q22 + q32 ). For such problems, we saw that since the classical Hamilto-

nian is rotationally invariant, the quantum Hamiltonian will commute with the

action of SO(3) on wavefunctions and energy eigenstates can be decomposed

into irreducible representations of SO(3).

Here the Bargmann-Fock representation gives an action of H7 o M p(6, R)

on the state space, with a U (3) subgroup commuting with the Hamiltonian

(actually one has a double cover of U (3), but by normal-ordering one can get an

actual U (3)). The eigenvalue of the U (1) corresponding to the Hamiltonian gives

the energy of a state, and states of a given energy will be sums of irreducible

representations of SU (3). This works much like in the d = 2 case, although here

our irreducible representations are the spaces Hn of homogeneous polynomials

of degree n in three variables rather than two. These spaces have dimension

1

2 (n + 1)(n + 2). A difference with the SU (2) case is that one does not get all

irreducible representations of SU (3) this way.

The rotation group SO(3) will be a subgroup of this U (3) and one can ask

how the SU (3) irreducible Hn decomposes into a sum of irreducibles of the

subgroup (which will be characterized by l = 0, 1, 2, ). One can show that

for even n we get all even values of l from 0 to n, and for odd n we get all odd

values of l from 1 to n. A derivation can be found in many quantum mechanics

textbooks, see for example pgs. 351-2 of [40].

The angular momentum operators L1 , L2 , L3 will be examples of the quadratic

combinations of annihilation and creation operators discussed in the general

case. For example, one can simply calculate

L3 =Q1 P2 Q2 P1

i

= ((a1 + a1 )(a2 a2 ) (a2 + a2 )(a1 a1 ))

2

i

= (a1 a2 a1 a2 a2 a1 + a2 a1 )

2

=i(a2 a1 a1 a2 )

189

One could also use the fact that in the matrix representation of the Lie alge-

bra so(3) acting on vectors, infinitesimal rotations are given by antisymmetric

matrices, and use our theorem from earlier in this section to find the quadratic

operator corresponding to a given antisymmetric matrix. For instance, for in-

finitesimal rotations about the 3 axis,

0 1 0

l3 = 1 0 0

0 0 0

X

aj (l3 )jk ak = a1 a2 + a2 a1

j,k

The metaplectic representation is not usually mentioned in the physics litera-

ture, and the discussions in the mathematical literature tend to be aimed at

an advanced audience. Two good examples of such detailed discussions can be

found in [18] and chapters 1 and 11 of [46].

Remarkably, the metaplectic representation plays a significant role in num-

ber theory, in which context the fundamental theorem of quadratic reciprocity

becomes a description of the nature of the non-trivial metaplectic cover of the

symplectic group. This is also closely related to the theory of theta functions.

This relationship was first studied by Andre Weil in the early sixties [50]. For

some (rather advanced and technical) expositions of this material, see [6] and

[32].

190

Chapter 18

The quantum state spaces describing fundamental physical systems that occur

in the real world often are fermionic: they have a property of anti-symmetry

under the exchange of two states. It turns out that the entire quantum me-

chanical formalism we have developed so far has a parallel version in which such

anti-symmetry is built in from the beginning. Besides providing further exam-

ples of the connections between representation theory and quantum mechanics,

this new formalism encompasses in a surprising way some mathematics of great

importance in geometry.

In this chapter well consider in detail the basic example of how this works,

as a simple variation on techniques we used to study the harmonic oscillator,

leaving for later chapters a discussion of the new mathematics and new general

formalism embodied in the example.

bosonic oscillator

Recall that the Hamiltonian for the quantum harmonic oscillator system in d

degrees of freedom (setting ~ = m = = 1) is

d

X 1

H= (Q2j + Pj2 )

j=1

2

1 1

aj = (Qj + iPj ), aj = (Qj iPj )

2 2

191

This is done by noting that the operators

Nj = aj aj

nj = 0, 1, 2, . . .

and that

d d

1X 1

(aj aj + aj aj ) =

X

H= (Nj + )

2 j=1 j=1

2

d

X 1

En1 ,n2 , ,nd = (nj + )

j=1

2

Putting back in factors of ~, and noting that one can choose to be

a different number j for each degree of freedom, one finds that the energy

eigenvalues are

d

X 1

En1 ,n2 , ,nd = (nj + )~j

j=1

2

We will see later in this course that collections of identical particles can be

described using this formalism, when states are symmetric under interchange

of two particle (and are then called bosons). The state |n1 , n2 , , nN i is

interpreted as describing n1 bosons of one momentum, n2 of another momentum,

etc. The lowest energy state |0, 0, , 0i describes the vacuum state with no

particles, but it carries an energy

d

X 1

~j

j=1

2

Because of this ability to describe bosonic particles, well often call the harmonic

oscillator the bosonic oscillator.

the fermionic oscillator

The simple change in the harmonic oscillator problem that takes one from bosons

to fermions is the replacement of the bosonic annihilation and creation operators

192

(which well now denote aB and aB ) by fermionic annihilation and creation

operators called aF and aF , which satisfy a variant of the bosonic commutation

relations

[aF , aF ]+ = 1, [aF , aF ]+ = 0, [aF , aF ]+ = 0

The change from the bosonic case is just the replacement of the commutator

[A, B] AB BA

[A, B]+ AB + BA

NF = aF aF

now satisfies

2

NF2 = aF aF aF aF = aF (1 aF aF )aF = NF aF a2F = NF

2

(using the fact that a2F = aF = 0). So one has

NF2 NF = NF (NF 1) = 0

which implies that the eigenvalues of NF are just 0 and 1. Well denote eigen-

vectors with such eigenvalues by |0i and |1i. The simplest representation of the

operators aF and aF on a complex vector space HF will be on C2 , and choosing

the basis

0 1

|0i = , |1i =

1 0

the operators are represented as

0 0 0 1 1 0

aF = , aF = , NF =

1 0 0 0 0 0

Since

1

H= (a aF + aF aF )

2 F

is just 12 the identity operator, to get a non-trivial quantum system, instead we

make a sign change and set

1

1 1 0

H = (aF aF aF aF ) = NF 1 = 2

2 2 0 21

The energies of the energy eigenstates |0i and |1i will then be 12 since

1 1

H|0i = |0i, H|1i = |1i

2 2

193

Note that the quantum system we have constructed here is nothing but our

old friend the qubit. Taking complex linear combinations of the operators

aF , aF , NF , 1

tation of the algebra of these operators). The relation to the Pauli matrices is

just

1 1 1

aF = (1 + i2 ), aF = (1 i2 ), H = 3

2 2 2

For the case of d degrees of freedom, one has this variant of the canonical

commutation relations (CCR) amongst the bosonic annihilation and creation

operators aB j and aB j :

1, . . . , d is said to satisfy the canonical anti-commutation relations (CAR) when

one has

In this case the state space is the tensor product of N copies of the single

fermionic oscillator state space

HF = (C2 )d = C2 C2 C2

| {z }

d times

and aF j can be explicitly given by

0 0

aF j = 3 3 3 1 1

| {z } 1 0

j1 times

0 1

aF j = 3 3 3 1 1

| {z } 0 0

j1 times

are satisfied for j 6= k since in these cases one will get a factor in the tensor

product of

0 0 0 1

[3 , ]+ = 0 or [3 , ] =0

1 0 0 0 +

194

While this sort of tensor product construction is useful for discussing the

physics of multiple qubits, in general it is easier to not work with large tensor

products, and the Clifford algebra formalism we will describe in the next chapter

avoids this.

The number operators will be

NF j = aF j aF j

with eigenvalues nj = 0, 1. One can take as an orthonormal basis of HF the 2d

states

|n1 , n2 , , nd i

As an example, for the case d = 3 the pattern of states and their energy

levels for the bosonic and fermionic cases looks like this

In the bosonic case the lowest energy state is at positive energy and there are

an infinite number of states of ever increasing energy. In the fermionic case the

195

lowest energy state is at negative energy, with the pattern of energy eigenvalues

of the finite number of states symmetric about the zero energy level.

No longer setting ~ = = 1 and allowing different for different degrees of

freedom, the Hamiltonian will be

d

1X

H= ~j (aF j aF j aF j aF j )

2 j=1

|n1 , n2 , , nd i

d

X 1

En1 ,n2 , ,nd = (nj )~j

j=1

2

Most quantum field theory books and a few quantum mechanics books contain

some sort of discussion of the fermionic oscillator, see for example Chapter 21.3

of [40] or Chapter 5 of [9]. The standard discussion often starts with considering

a form of classical analog using anti-commuting fermionic variables and then

quantizes to get the fermionic oscillator. Here we are doing things in the opposite

order, starting in this chapter with the quantized oscillator, then considering

the classical analog in a later chapter.

196

Chapter 19

harmonic oscillator quantum system to a very different one (the fermionic os-

cillator), with this new system having in many ways a parallel structure. It

turns out that this parallelism goes much deeper, with every aspect of the har-

monic oscillator story having a fermionic analog. Well begin in this chapter by

studying the operators of the corresponding quantum systems.

In mathematics, a ring is a set with addition and multiplication laws that are

associative and distributive (but not necessarily commutative), and an algebra

is a ring that is also a vector space over some field of scalars. The canonical

commutation and anti-commutation relations define interesting algebras, called

the Weyl and Clifford algebras respectively. The case of complex numbers as

scalars is simplest, so well start with that, before moving on to the real number

case.

Starting with the one degree of freedom case (corresponding to two operators

Q, P , which is why the notation will have a 2) we can define

Definition (Complex Weyl algebra, one degree of freedom). The complex Weyl

algebra in the one degree of freedom case is the algebra Weyl(2, C) generated by

the elements 1, aB , aB , satisfying the canonical commutation relations:

In other words, Weyl(2, C) is the algebra one gets by taking arbitrary prod-

ucts and complex linear combinations of the generators. By repeated use of the

commutation relation

aB aB = 1 + aB aB

197

any element of this algebra can be written as a sum of elements in normal order,

of the form

cl,m (aB )l am

B

with all annihilation operators aB on the right, for some complex constants cl,m .

As a vector space over C, Weyl(2, C) is infinite-dimensional, with a basis

aB = w, aB =

w

one sees that Weyl(2, C) can be identified with the algebra of polynomial coeffi-

cient differential operators on functions of a complex variable w. As a complex

vector space, the algebra is infinite dimensional, with a basis of elements

m

wl

wm

In our study of quantization and the harmonic oscillator we saw that the

subset of such operators consisting of linear combinations of

2

1, w, , w2 , , w

w w2 w

is closed under commutators, so it forms a Lie algebra of dimension 6. This Lie

algebra includes as subalgebras the Heisenberg Lie algebra (first three elements)

and the Lie algebra of SL(2, R) (last three elements). Note that here we are

allowing complex linear combinations, so we are getting the complexification of

the real six-dimensional Lie algebra that appeared in our study of quantization.

Since the aB and aB are defined in terms of P and Q, one could of course

also define the Weyl algebra as the one generated by 1, P, Q, with the Heisenberg

commutation relations, taking complex linear combinations of all products of

these operators.

Changing commutators to anti-commutators, one gets a different algebra, the

Clifford algebra

Clifford algebra in the one degree of freedom case is the algebra Cliff(2, C) gen-

erated by the elements 1, aF , aF , subject to the canonical anti-commutation re-

lations

[aF , aF ]+ = 1, [aF , aF ]+ = [aF , aF ]+ = 0

198

This algebra is a four dimensional algebra over C, with basis

1, aF , aF , aF aF

since higher powers of the operators vanish, and one can use the anti-commutation

relation betwee aF and aF to normal order and put factors of aF on the right.

We saw in the last chapter that this algebra is isomorphic with the algebra

M (2, C) of 2 by 2 complex matrices, using

1 0 0 0 0 1 1 0

1 , aF , aF , aF aF

0 1 1 0 0 0 0 0

We will see later on that there is also a way of identifying this algebra with

differential operators in fermionic variables, analogous to what happens in

the bosonic (Weyl algebra) case.

Recall that the bosonic annihilation and creation operators were originally

defined in term of the P and Q operators by

1 1

aB = (Q + iP ), aB = (Q iP )

2 2

Looking for the fermionic analogs of the operators Q and P , we use a slightly

different normalization, and set

1 1

aF = (1 + i2 ), aF = (1 i2 )

2 2

so

1

1 = aF + aF , 2 = (aF aF )

i

and the CAR imply that the operators j satisfy the anti-commutation relations

[1 , 1 ]+ = [aF + aF , aF + aF ]+ = 2

[2 , 2 ]+ = [aF aF , aF aF ]+ = 2

1

[1 , 2 ]+ = [aF + aF , aF aF ]+ = 0

i

From this we see that

One could alternatively have defined Cliff(2, C) as the algebra generated

by 1, 1 , 2 , subject to the relations

[j , k ]+ = 2jk

One can just drop mention of 2 , and get an algebra Cliff(1, C), generated

by 1, , with the relation

2 = 1

This is a two-dimensional complex algebra, isomorphic to C C.

199

19.1.3 Multiple degrees of freedom

For a larger number of degrees of freedom, one can generalize the above and

define Weyl and Clifford algebras as follows:

Definition (Complex Weyl algebras). The complex Weyl algebra for d degrees

of freedom is the algebra Weyl(2d, C) generated by the elements 1, aB j , aB j ,

j = 1, . . . , d satisfying the CCR

ferential operators in m complex variables w1 , w2 , . . . , wd . The subspace of

complex linear combinations of the elements

2

1, wj , , wj wk , , wj

wj wj wk wk

algebra h2d+1 o sp(2d, R) built out of the Heisenberg Lie algebra in 2d variables

and the Lie algebra of the symplectic group Sp(2d, R). Recall that this is the

Lie algebra of polynomials of degree at most 2 on the phase space R2d , with the

Poisson bracket as Lie bracket.

One could also define the complex Weyl algebra by taking complex linear

combinations of products of generators 1, Pj , Qj , subject to the Heisenberg com-

mutation relations.

For Clifford algebras one has

erators). The complex Clifford algebra for d degrees of freedom is the algebra

Cliff(2d, C) generated by 1, aF j , aF j for j = 1, 2, . . . , d satisfying the CAR

or, alternatively, one has the following more general definition that also works

in the odd-dimensional case

ables is the algebra Cliff(n, C) generated by 1, j for j = 1, 2, . . . , n satisfying

the relations

[j , k ]+ = 2jk

We wont try and prove this here, but one can show that, abstractly as

algebras, the complex Clifford algebras are something well-known. Generalizing

the case d = 1 where we saw that Cliff(2, C) was isomorphic to the algebra of 2

by 2 complex matrices, one has isomorphisms

Cliff(2d, C) M (2d , C)

200

in the even-dimensional case, and in the odd-dimensional case

1, j , j k , j k l , . . . , 1 2 3 n1 n

consider all products of the generators, and use the commutation relations

for the j to identify any such product with an element of this basis. The

relation j2 = 1 shows that one can remove repeated occurrences of a

j . The relation j k = k j can then be used to put elements of the

product in the order of a basis element as above.

this is to consider the product

(1 + 1 )(1 + 2 ) (1 + n )

which will have 2n terms that are exactly those of the basis listed above.

We can define real Clifford algebras Cliff(n, R) just as for the complex case, by

taking only real linear combinations:

the algebra Cliff(n, R) generated over the real numbers by 1, j for j = 1, 2, . . . , n

satisfying the relations

[j , k ]+ = 2jk

For reasons that will be explained in the next chapter, it turns out that a

more general definition is useful. We write the number of variables as n = r + s,

for r, s non-negative integers, and now vary not just r + s, but also r s, the

so-called signature.

Definition (Real Clifford algebras, arbitrary signature). The real Clifford al-

gebra in n = r + s variables is the algebra Cliff(r, s, R) over the real numbers

generated by 1, j for j = 1, 2, . . . , n satisfying the relations

[j , k ]+ = 2jk 1

r + 1, . . . , n.

201

In other words, as in the complex case different j anti-commute, but only

the first r of them satisfy j2 = 1, with the other s of them satisfying j2 = 1.

Working out some of the low-dimensional examples, one finds:

Cliff(0, 1, R). This has generators 1 and 1 , satisfying

12 = 1

Taking real linear combinations of these two generators, the algebra one

getsis just the algebra C of complex numbers, with 1 playing the role of

i = 1.

Cliff(0, 2, R). This has generators 1, 1 , 2 and a basis

1, 1 , 2 , 1 2

with

12 = 1, 22 = 1, (1 2 )2 = 1 2 1 2 = 12 22 = 1

This four-dimensional algebra over the real numbers can be identified with

the algebra H of quaternions by taking

1 i, 2 j, 1 2 k

possible identification as follows

1 0 0 1 0 1 1 0

1 , 1 , 2 , 1 2

0 1 1 0 1 0 0 1

Note that one can construct this using the aF , aF for the complex case

Cliff(2, C)

1 = aF + aF , 2 = aF aF

since these are represented as real matrices.

It turns out that Cliff(r, s, R) is always one or two copies of matrices of real,

complex or quaternionic elements, of dimension a power of 2, but this requires

a rather intricate algebraic arguments that we will not enter into here. For the

details of this and the resulting pattern of algebras one gets, see for instance

[31]. One special case where the pattern is relatively simple is when one has

r = s. Then n = 2r is even-dimensional and one finds

Cliff(r, r, R) = M (2r , R)

We will see in the next chapter that just as quadratic elements of the Weyl

algebra give a basis of the Lie algebra of the symplectic group, quadratic ele-

ments of the Clifford algebra give a basis of the Lie algebra of the orthogonal

group.

202

19.3 For further reading

A good source for more details about Clifford algebras and spinors is Chapter

12 of the representation theory textbook [46]. For the details of what what

happens for all Cliff(r, s, R), another good source is Chapter 1 of [31].

203

204

Chapter 20

Geometry

The definitions given in last chapter of Weyl and Clifford algebras were purely

algebraic, based on a choice of generators. These definitions do though have a

more geometrical formulation, with the definition in terms of generators corre-

sponding to a specific choice of coordinates. For the Weyl algebra, the geometry

involved is known as symplectic geometry, and we have already seen that in the

bosonic case quantization of a phase space R2d depends on the choice of a

non-degenerate anti-symmetric bilinear form which determines the Poisson

brackets and thus the Heisenberg commutation relations. Such a also deter-

mines a group Sp(2d, R), which is the group of linear transformations of R2d

preserving . The Clifford algebra also has a coordinate invariant definition,

based on a more well-known structure on a vector space Rn , that of a non-

degenerate symmetric bilinear form, i.e. an inner product. In this case the

group that preserves the inner product is an orthogonal group. In the symplec-

tic case anti-symmetric forms require an even number of dimensions, but this is

not true for symmetric forms, which also exist in odd dimensions.

In the case of M = R2d , the dual phase space, for two vectors u, u0 M

explicitly by

205

(u, u0 ) =cq1 c0p1 cp1 c0q1 + cqd c0pd cpd c0qd

0

0 1 ... 0 0 cq1

1 0 ... 0 0 c0p

1

= cq1 cp1 . . . cqd cpd ... .. .. .. ..

. . . .

0 0 ... 0 1 c0qd

0 0 ... 1 0 c0pd

0 1 ... 0 0 0 1 ... 0 0

1 0 ... 0 0 1 0 ... 0 0

g T ... .. .. .. g = .. .. .. ..

. . . .

. . .

0 0 ... 0 1 0 0 ... 0 1

0 0 ... 1 0 0 0 ... 1 0

This choice of is much less arbitrary than it looks. One can show that

given any non-degenerate anti-symmetric bilinear form on R2d a basis can be

found with respect to which it will be the given here (for a proof, see [7]).

This is also true if one complexifies, taking (q, p) C2d and using the same

formula for , which is now a bilinear form on C2d . In the real case the group

that preserves is called Sp(2d, R), in the complex case Sp(2d, C).

To get a fermionic analog of this, it turns out that all we need to do is replace

non-degenerate anti-symmetric bilinear form (, ) with non-degenerate sym-

metric bilinear form < , >. Such a symmetric bilinear form is actually some-

thing much more familiar from geometry than the anti-symmetric case: it is

just a notion of inner product. Two things are different in the symmetric case

The underlying vector space does not have to be even dimensional, one

can take M = Rn for any n, including n odd. To get a detailed analog of

the bosonic case though, we will mostly consider the even case n = 2d.

For a given dimension n, there is not just one possible choice of < , >

up to change of basis, but one possible choice for each pair of integers r, s

such that r + s = n. Given r, s, any choice of < , > can be put in the

206

form

0

1 0 ... 0 0 u1

0 1 . . . 0 0 u02

= u1 . . . un ... ... .. .. ..

. . .

0 0 . . . 1 0 u0n1

0 0 . . . 0 1 u0n

| {z }

r + signs, s - signs

We can thus extend our definition of the orthogonal group as the group of

transformations g preserving an inner product

Definition (Orthogonal group O(r, s, R)). The group O(r, s, R) is the group of

real r + s by r + s matrices g that satisfy

1 0 ... 0 0 1 0 ... 0 0

0 1 . . . 0 0 0 1 ... 0 0

g T ... ... .. .. g = .. .. .. ..

. .

.

. . .

0 0 . . . 1 0 0 0 . . . 1 0

0 0 . . . 0 1 0 0 ... 0 1

| {z } | {z }

r + signs, s - signs r + signs, s - signs

formula for < , >, one can change basis by multiplying the s basis elements

by a factor of i, and in this new basis all basis vectors ej satisfy < ej , ej >= 1.

One finds that on Cn , as in the symplectic case, up to change of basis there

is only one non-degenerate bilinear form. The group preserving this is called

O(n, C). Note that on Cn < , > is not the Hermitian inner product (which is

anti-linear on the first variable), and it is not positive definite.

As defined by generators in the last chapter, Clifford algebras have no obvious

geometrical significance. It turns out however that they are powerful tools in

the study of the geometry of linear spaces with an inner product, including

especially the study of linear transformations that preserve the inner product,

207

i.e. rotations. To see the relation between Clifford algebras and geometry,

consider first the positive definite case Cliff(n, R). To an arbitrary vector

v = (v1 , v2 , . . . , vn ) Rn

v = v 1 1 + v2 2 + + vn n

Using the Clifford algebra relations for the j , given two vectors v, w the

product of their associated Clifford algebra elements satisfies

vw + wv = [v1 1 + v2 2 + + vn n , w1 1 + w2 2 + + wn n ]+

= 2(v1 w1 + v2 w2 + + vn wn )

= 2 < v, w >

where < , > is the symmetric bilinear form on Rn corresponding to the stan-

dard inner product of vectors. Note that taking v = w one has

nations of the generators j , and one can think of it as a sort of enhancement of

the vector space Rn that encodes information about the inner product. In this

larger structure one can multiply as well as add vectors, with the multiplication

determined by the inner product.

In general one can define a Clifford algebra whenever one has a vector space

with a symmetric bilinear form:

V with a symmetric bilinear form < , >, the Clifford algebra Cliff(V, < , >)

is the algebra generated by 1 and elements of V , with the relations

vw + wv = 2 < v, w >

vw + wv = 2 < v, w >

another common choice. One also sees variants without the factor of 2.

For n dimensional vector spaces over C, we have seen that for any non-

degenerate symmetric bilinear form a basis can be found such that < , > has

the standard form

< z, w >= z1 w1 + z2 w2 + + zn wn

As a result, there is just one complex Clifford algebra in dimension n, the one

we defined as Cliff(n, C).

208

For n dimensional vector spaces over R with a non-degenerate symmetric

bilinear forms of type r, s such that r+s = n, the corresponding Clifford algebras

Cliff(r, s, R) are the ones defined in terms of generators in the last chapter.

In special relativity, space-time is a real 4-dimensional vector space with an

indefinite inner product corresponding to (depending on ones choice of conven-

tion) either the case r = 1, s = 3 or the case s = 1, r = 3. The group of linear

transformations preserving this inner product is called the Lorentz group, and

its orientation preserving component is written as SO(3, 1) or SO(1, 3) depend-

ing on the choice of convention. Later on in the class we will consider what

happens to quantum mechanics in the relativistic case, and there encounter the

corresponding Clifford algebras Cliff(3, 1, R) or Cliff(1, 3, R). The generators

j of such a Clifford algebra are well-known in the subject as the Dirac -

matrices.

For now though, we will restrict attention to the positive definite case, so

just will be considering Cliff(n, R)and seeing how it is used to study the group

O(n) of n-dimensional rotations in Rn .

Well consider two different ways of seeing the relationship between the Clifford

algebra Cliff(n, R) and the group O(n) of rotations in Rn . The first is based

upon the geometrical fact (known as the Cartan-Dieudonne theorem) that one

can get any rotation by doing multiple orthogonal reflections in different hy-

perplanes. Orthogonal reflection in the hyperplane perpendicular to a vector w

takes a vector v to the vector

< v, w >

v0 = v 2 w

< w, w >

209

the reflection in w transformation is

< v, w >

v v 0 =v 2 w

< w, w >

w

=v (vw + wv)

< w, w >

but since

w < w, w >

w = =1

< w, w > < w, w >

we have for w 6= 0

w

w1 =

< w, w >

and the reflection transformation is just conjugation by w times a minus sign

v v 0 = v v wvw1 = wvw1

So, thinking of vectors as lying in the Clifford algebra, the orthogonal trans-

formation that is the result of one reflection is just a conjugation (with a minus

sign). These lie in the group O(n) since they change orientation. The result of

two reflections in different hyperplanes orthogonal to w1 , w2 will be

group SO(n).

This construction not only gives an efficient way of representing rotations (as

conjugations in the Clifford algebra), but it also provides a construction of the

group Spin(n) in arbitrary dimension n. The spin group is generated from unit

vectors in Rn by taking products of the corresponding elements of the Clifford

algebra. This construction generalizes to arbitrary n the one we gave in chapter

6 of Spin(3) in terms of unit length elements of the quaternion algebra H. One

can see the characteristic fact that there are two elements of the Spin(n) group

giving the same rotation in SO(n) by noticing that changing the sign of the

Clifford algebra element does not change the conjugation action, where signs

cancel.

elements of the Clifford algebra

For a second approach to understanding rotations in arbitrary dimension, one

can use the fact that these are generated by taking products of rotations in the

coordinate planes. A rotation by an angle in the j k coordinate plane (j < k)

will be given by

v eLjk v

where Ljk is an n by n matrix with only two non-zero entries: jk entry 1 and

kj entry +1. Restricting attention to the j k plane, eLjk acts as the standard

210

rotation matrix in the plane

vj cos sin vj

vk sin cos vk

In the SO(3) case we saw that there were three of these matrices

providing a basis of the Lie algebra so(3), in n dimensions there will be 21 (n2 n)

of them, providing a basis of the Lie algebra so(n).

Just as in the case of SO(3) where unit length quaternions were used, we can

use elements of the Clifford algebra to get these same rotation transformations,

but as conjugations in the Clifford algebra. To see how this works, consider the

quadratic Clifford algebra element j k and notice that

1 1 1 1

( j k )2 = (j k j k ) = j2 k2 =

2 4 4 4

so one has

(/2)2

e 2 j k =(1 + ) + j k (/2 )

2!

= cos( ) + j k sin( )

2 2

Conjugating a vector vj j +vk k in the j k plane by this, one finds (working

out the details here will be a problem on the problem set)

e 2 j k (vj j + vk k )e 2 j k = (vj cos vk sin )j + (vj sin + vk cos )k

This shows that, just as in our earlier calculations in three dimensions, one

gets a double cover of the group of rotations, with here the elements e 2 j k of

the Clifford algebra giving a double cover of the group of rotations in the j k

plane. General elements of the spin group can be constructed by multiplying

these for different angles in different coordinate planes. One sees that the Lie

algebra spin(n) can be identified with the Lie algebra so(n) by

1

Ljk j k

2

Yet another way to see this would be to compute the commutators of the 21 j k

for different values of j, k and show that they satisfy the same commutation

relations as the corresponding matrices Ljk .

Recall that in the bosonic case we found that quadratic combinations of the

Qj , Pk (or of the aB j , aB j ) gave operators satisfying the commutation relations

of the Lie algebra sp(2n, R). This is the Lie algebra of the group Sp(2n, R),

the group preserving the non-degenerate anti-symmetric bilinear form (, ) on

211

the phase space R2n . The fermionic case is precisely analogous, with the role

of the anti-symmetric bilinear form (, ) replaced by the symmetric bilinear

form < , > and the Lie algebra sp(2n, R) replaced by so(n) = spin(n).

In the bosonic case the linear functions of the Qj , Pj satisfied the commuta-

tion relations of another Lie algebra, the Heisenberg algebra, but in the fermionic

case this is not true for the j . In the next chapter we will see that one can

define a notion of a Lie superalgebra that restores the parallelism.

Some more detail about the relationship between geometry and Clifford algebras

can be found in [31], and an exhaustive reference is [35].

212

Chapter 21

Anti-commuting Variables

and Pseudo-classical

Mechanics

The analogy between the algebras of operators in the bosonic (Weyl algebra) and

fermionic (Clifford algebra) cases can be extended by introducing a fermionic

analog of phase space and the Poisson bracket. This gives a fermionic ana-

log of classical mechanics, sometimes called pseudo-classical mechanics, the

quantization of which gives the Clifford algebra as operators, and spinors as

state spaces. In this chapter well intoduce anti-commuting variables j that

will be the fermionic analogs of the variables qj , pj . These objects will become

generators of the Clifford algebra under quantization, and will later be used in

the construction of fermionic state spaces, by analogy with the Schrodinger and

Bargmann-Fock constructions in the bosonic case.

anti-commuting generators

Given a phase space V = Rn , one gets classical observables by taking polynomial

functions on V . These are generated by the linear functions qj , pj , j = 1, . . . , d

for n = 2d, which lie in the dual space V . It is also possible to start with the

same space V of linear functions on V (with n not necessarily even), and pick

a different notion of multiplication, one that is anti-commutative on elements

of V . Using such a multiplication, we can generate an analog of the algebra of

polynomials on V , sometimes called the Grassmann algebra:

Definition (Grassmann algebra). The algebra over the real numbers generated

by j , j = 1, . . . , n, satisfying the relations

j k + k j = 0

213

is called the Grassmann algebra, and denoted (Rn ).

Note that these relations imply that generators satisfy j2 = 0. Also note that

sometimes the product in the Grassmann algebra is called the wedge product

and the product of j and k is denoted j k . We will not use a different

symbol for the product in the Grassmann algebra, relying on the notation for

generators to keep straight what is a generator of a conventional polynomial

algebra (e.g. qj or pj ) and what is a generator of a Grassman algebra (e.g. j ).

The Grassmann algebra behaves in many ways like the polynomial algebra

on Rn , but it is finite dimensional, with basis

1, j , j k , j k l , , 1 2 n

monomials are characterized by a degree (number of generators in the product),

which takes values from 0 to n. k (Rn ) is the subspace of (Rn ) of linear

combinations of monomials of degree k.

exterior algebra, and we will examine its relationship to anti-symmetric elements

of the tensor algebra later on. Readers may have already seen the exterior al-

gebra in the context of differential forms on Rn . These are known to physicists

as anti-symmetric tensor fields, and given by taking elements of the exte-

rior algebra (Rn ) with coefficients not constants, but functions on Rn . This

construction is important in the theory of manifolds, where at a point x in a

manifold M , one has a tangent space Tx M and its dual space (Tx M ) . A set of

local coordinates xj on M gives basis elements of (Tx M ) denoted by dxj and

differential forms locally can be written as sums of terms of the form

a space is in terms of the functions on it. One can think of what we have done

here as creating a new kind of space out of Rn , where the algebra of functions

on the space is (Rn ), generated by coordinate functions j with respect to a

basis of Rn . The enlargement of conventional geometry to include new kinds

of spaces such that this makes sense is known as supergeometry, but we will

not attempt to pursue this subject here. Spaces with this new kind of geometry

have functions on them, but do not have conventional points: one cant ask

what the value of an anti-commuting function at a point is.

Remarkably, one can do calculus on such unconventional spaces, introducing

analogs of the derivative and integral for anti-commuting functions. For the case

n = 1, an arbitrary function is

F () = c0 + c1

214

and one can take

F = c1

For larger values of n, an arbitrary function can be written as

F (1 , 2 , . . . , n ) = FA + j FB

where FA , FB are functions that do not depend on the chosen j (one gets FB by

using the anti-commutation relations to move j all the way to the left). Then

one can define

F = FB

j

This derivative operator has many of the same properties as the conventional

derivative, although there are unconventional signs one must keep track of. An

unusual property of this derivative that is easy to see is that one has

=0

j j

Taking the derivative of a product one finds this version of the Leibniz rule

for monomials F and G

(F G) = ( F )G + (1)|F | F ( G)

j j j

A notion of integration (often called the Berezin integral) with many of the

usual properties of an integral can also be defined. It has the peculiar feature

of being the same operation as differentiation, defined in the n = 1 case by

Z

(c0 + c1 )d = c1

Z

F (1 , 2 , , n )d1 d2 dn = F = cn

n n1 1

where cn is the coefficient of the basis element 1 2 n in the expression of F

in terms of basis elements.

This notion of integration is a linear operator on functions, and it satisfies

an analog of integration by parts, since if one has

F = G

j

then Z

F dj = F = G=0

j j j

using the fact that repeated derivatives give zero.

215

21.2 Pseudo-classical mechanics

The basic structure of Hamiltonian classical mechanics depends on an even

dimensional space phase space R2d with a Poisson bracket {, } on functions on

this space. Time evolution of a function f on phase space is determined by

d

f = {f, h}

dt

for some Hamiltonian function h. This says that taking the derivative of any

function in the direction of the velocity vector of a classical trajectory is the

linear map

f {f, h}

on functions. As we saw in chapter 12, since this linear map is a derivative, the

Poisson bracket will have the derivation property, satisfying the Leibniz rule

for arbitrary functions f1 , f2 , f3 on phase space. Using the Leibniz rule and

anti-symmetry, one can calculate Poisson brackets for any polynomials, just

from knowing the Poisson bracket on generators qj , pj , which we chose to be

bracket is anti-symmetric.

To get pseudo-classical mechanics, we think of the Grassmann algebra (Rn )

as our algebra of classical observables, an algebra we can think of as functions

on a fermionic phase space Rn (note that in the fermionic case, the phase

space does not need to be even dimensional). We want to find an appropriate

notion of fermionic Poisson bracket operation on this algebra, and it turns out

that this can be done. While the standard Poisson bracket is an antisymmetric

bilinear form on linear functions, the fermionic Poisson bracket will be based on

a choice of symmetric bilinear form on linear functions, equivalently a notion of

inner product.

Denoting the fermionic Poisson bracket by {, }+ , for a multiplication anti-

commutative on generators one has to adjust signs in the Leibniz rule, and the

derivation property analogous to the derivation property of the usual Poisson

bracket is

where |F1 | and |F2 | are the degrees of F1 and F2 . It will also have the symmetry

property

{F1 , F2 }+ = (1)|F1 ||F2 | {F2 , F1 }+

and one can use these properties to compute the fermionic Poisson bracket for

arbitrary functions in terms of the relation for generators.

216

One can think of the j as the coordinate functions with respect to a

basis ei of Rn , and we have seen that the symmetric bilinear forms on Rn are

classified by a choice of positive signs for some basis vectors, negative signs for

the others. So, on generators j one can choose

{j , k }+ = jk

corresponding to the possible inequivalent choices of non-degenerate symmetric

bilinear forms.

Taking the case of a positive-definite inner product for simplicity, one can

calculate explicitly the fermionic Poisson brackets for linear and quadratic com-

binations of the generators. One finds

{j , k l }+ = {j , k }+ l k {j , l }+ = jk l jl k

and

{j k , l m }+ ={j k , l }+ m + l {j k , m }+

= {l , j k }+ m l {m , j k }+

= jl k m + lk j m mj l k + mk l j

The second of these equations shows that the quadratic combinations of the

generators j satisfy the relations of the Lie algebra of the group of rotations in

n dimensions (so(n) = spin(n)). The first shows that the k l acts on the j as

infinitesimal rotations in the k l plane.

Notice that this is again exactly parallel to what happens in the bosonic

case, where the quadratic combinations of the qj , pj satisfy the relations of the

Lie algebra sp(2d, R) and act on the generators qj , pj as infinitesimal Sp(2d, R)

transformations of the phase space R2d .

In the case of the conventional Poisson bracket, the anti-symmetry of the bracket

and the fact that it satisfies the Jacobi identity implies that it is a Lie bracket

determining a Lie algebra (the infinite dimensional Lie algebra of functions on

a phase space R2d ). The fermionic Poisson bracket provides an example of a

something called a Lie superalgebra. These can be defined for vector spaces

with some usual and some fermionic coordinates:

plex vector space V is given by a Lie superbracket [, ], a bilinear map on V

which on generators X, Y, Z (which may be usual coordinates or fermionic ones)

satisfies

[X, Y ] = (1)|X||Y | [Y, X]

217

and a super Jacobi identity

where |X| takes value 0 for a usual coordinate, 1 for a fermionic coordinate.

As in the bosonic case, on functions of generators of order less than or equal

to two, the Fermionic Poisson bracket {, , }+ is a Lie superbracket, giving a

Lie superalgebra of dimension 1 + n + 12 (n2 n) (since there is one constant, n

linear terms j and 21 (n2 n) quadratic terms j k ).

We will see in the next chapter that quantization will give a representation

of this Lie superalgebra, with a parallel construction to that of the bosonic case.

In pseudo-classical mechanics, the dynamics will be determined by choosing a

hamiltonian h in (Rn ). Observables will be other functions F (Rn ), and

they will satisfy the analog of Hamiltons equations

d

F = {F, h}+

dt

Well consider two of the simplest possible examples.

Using pseudo-classical mechanics, one can find a classical analog of something

that is quintessentially quantum: the degree of freedom that appears in the qubit

or spin 1/2 system that we have seen repeatedly in this course. Taking R3 with

the standard inner product as fermionic phase space, we have three generators

1 , 2 , 3 satisfying the relations

{j , k }+ = jk

1, 1 , 2 , 3 , 1 2 , 1 3 , 2 3 , 1 2 3

have to be a linear combination

d

j (t) = {j , h}+

dt

but since h is a quadratic combination of the generators, recall from the earlier

discussion of quadratic combinations of the j that the right hand side is just

218

an infinitesimal rotation. So the solution to the classical equations of motion

will be a time-dependent rotation of the j in the plane perpendicular to

B = (B23 , B13 , B12 )

at a constant speed proportional to |B|.

We have already studied the fermionic oscillator as a quantum system, and one

can ask whether there is a corresponding pseudo-classical system. Such a system

is given by taking an even dimensional fermionic phase space R2d , with a basis

of coordinate functions 1 , 2d that generate (R2d ). On generators the

fermionic Poisson bracket relations come from the standard choice of positive

defnite symmetric bilinear form

{j , k }+ = jk

As discussed earlier, for this choice the quadratic products j k act on the gen-

erators by infinitesimal rotations in the j k plane, and satisfy the commutation

relations of so(2d).

To get a classical system that will quantize to the fermionic oscillator one

makes the choice

Xd

h=i 2j1 j

j=1

As in the bosonic case, we can make the standard choice of complex structure

J on R2d and get a decomposition

R2d C = Cd Cd

into eigenspaces of J of eigenvalue i. This is done by defining

1 1

j = (2j1 + i2j ), j = (2j1 i2j )

2 2

for j = 1, . . . , d. These satisfy the fermionic Poisson bracket relations

{j , k }+ = {j , k }+ = 0, {j , k }+ = jk

In the bosonic case we could use any complex structure J that satisfied

(Ju, Ju0 ) = (u, u0 ). For the Fermionic case we need instead < Ju, Ju0 >=<

u, u0 >. The subgroup of the SO(2d) transformations that preserve < , > that

also commute with J will be a U (d), just like in the bosonic case where a U (d)

Sp(2d, R) commuted with J, and possible choices of J will be parametrized by

the space SO(2d)/U (d).

In terms of the j , the Hamiltonian is

d

X

h= j j

j=1

219

Using the derivation property of {, }+ one finds

d

X d

X

{j , h}+ = {j , k k }+ = ({j , k }+ k k {j , k }+ ) = j

k=1 k=1

and, similarly,

{j , h}+ = j

so one sees that h is just the generator of U (1) U (d) phase rotations on the

variables j .

For more details on pseudo-classical mechanics, a very readable original ref-

erence is [5], and there is a detailed discussion in the textbook [45], chapter

7.

220

Chapter 22

Spinors

In this chapter well begin by investigating the fermionic analog of the notion

of quantization, which takes functions of anti-commuting variables on a phase

space with symmetric bilinear form < , > and gives an algebra of operators

satisfying the relations of the corresponding Clifford algebra. We will then

consider analogs of the constructions used in the bosonic case which there gave

us an essentially unique representation 0 of the corresponding Weyl algebra on

a space of states.

We know that for a fermionic oscillator with d degrees of freedom, the alge-

bra of operators will be Cliff(2d, C), the algebra generated by annihilation and

creation operators aF j , aF j . These operators will act on HF , a complex vec-

tor space of dimension 2d , and this will be our fermionic analog of the bosonic

0 . Since the spin group consists of invertible elements of the Clifford algebra,

it has a representation on HF . This is known as the spinor representation,

and it can be constructed by analogy with the Bargmann-Fock construction in

the bosonic case. Well also consider the analog in the fermionic case of the

Schrodinger representation, which turns out to have a problem with unitarity,

but finds a use in physics as ghost degrees of freedom.

In the bosonic case, quantization was based on finding a representation of the

Heisenberg Lie algebra of linear functions on phase space, or more explicitly,

for basis elements qj , pj of this Lie algebra finding operators Qj , Pj satisfying

the Heisenberg commutation relations. In the Fermionic case, the analog of

the Heisenberg Lie algebra is not a Lie algebra, but a Lie superalgebra, with

basis elements 1, j , j = 1, . . . , n and a Lie superbracket given by the fermionic

221

Poisson bracket, which on basis elements is

classical system) will be given by finding a linear map + that takes basis

elements j to operators + (j ) satisfying the anti-commutation relations

[+ (j ), + (k )]+ = jk 1

identity operator.

These relations are, up to a factor of 2, exactly the Clifford algebra rela-

tions, since if

j = 2+ (j )

then

[j , k ]+ = 2jk

We have seen that these real Clifford algebras Cliff(p, q, R) are isomorphic to

either one or two copies of the matrix algebras M (2l , R), M (2l , C), or M (2l , H),

with the power l depending on p, q. The irreducible representations of such a

matrix algebra are just the column vectors of dimension 2l , and there will be

either one or two such irreducible representations for Cliff(p, q, R) depending on

the number of copies of the matrix algebra.

Note that here we are not introducing the factors of i into the definition of

quantization that in the bosonic case were necessary to get a unitary represen-

tation of the Lie group corresponding to the real Heisenberg Lie algebra h2d+1 .

In the bosonic case we worked with all complex linear combinations of powers

of the Qj , Pj (the complex Weyl algebra Weyl(2d, C)), and thus had to identify

the specific complex linear combinations of these that gave unitary represen-

tations of the Lie algebra h2d+1 o sp(2d, R). Here we are not complexifying

for now, but working with the real Clifford algebra Cliff(p, q, R), and it is the

irreducible representations of this algebra that provide an analog of the unitary

representation 0 of h2d+1 .

As in the bosonic case, we can extend this quantization from linear anti-

commuting functions on Rn to quadratic anti-commuting functions on Rn . A

basis of these is given by j k for j < k and taking

+ (j k ) = + (j )+ (k )

the Clifford algebra elements

1

j k

2

which are exactly the commutation relations of the Lie algebra so(p, q, R).

So, we have seen that quantization of a fermionic phase space Rn with a

symmetric bilinear form of signature p, q gives a representation of the Clifford

222

algebra Cliff(p, q, R), with the quadratic elements providing a representation of

so(p, q, R). The explicit construction of these representations will show that,

as in the bosonic case, the irreducible representations are representations of a

double-cover, in this case the double cover Spin(p, q, R) of SO(p, q, R).

odinger representation for fermions:

ghosts

We would like to construct representations of Cliff(p, q, R) and thus fermionic

state spaces by using analogous constructions to the Schrodinger and Bargmann-

Fock ones in the bosonic case. The Schrodinger construction took the state space

H to be a space of functions on a subspace of the classical phase space which had

the property that the basis coordinate-functions Poisson-commuted. Two exam-

ples of this are the position coordinates qj , since {qj , qk } = 0, or the momentum

coordinates pj , since {pj , pk } = 0.. Unfortunately, for symmetric bilinear forms

< , > of definite sign, such as the positive definite case Cliff(n, R), the only

subspace the bilinear form is zero on is the zero subspace.

To get an analog of the bosonic situation, one needs to take the case of

signature d, d. The fermionic phase space will then be 2d dimensional, with d-

dimensional subspaces on which < , > and thus the fermionic Poisson bracket

is zero. Quantization will give the Clifford algebra

Clif f (d, d, R) = M (2d , R)

d

which has just one irreducible representation, R2 . One can complexify this to

get a complex state space

d

H F = C2

This state space will come with a representation of Spin(d, d, R) from exponen-

tiating quadratic combinations of the generators of Clif f (d, d, R). However,

this is a non-compact group, and one can show that on general grounds it can-

not have unitary finite-dimensional representations, so there must be a problem

with unitarity.

To see what happens explicitly, consider the simplest case d = 1 of one degree

of freedom. In the bosonic case the classical phase space is R2 , and quantization

gives operators Q, P which in the Schrodinger representation act on functions

of q, with Q = q and P = i q . In the fermionic case with signature 1, 1, basis

coordinate functions on phase space are 1 , 2 , with

{1 , 1 }+ = 1, {2 , 2 }+ = 1, {1 , 2 }+ = {2 , 1 }+ = 0

Defining

1 1

= (1 + 2 ), = (1 2 )

2 2

we get objects with fermionic Poisson bracket analogous to those of q and p

{, }+ = {, }+ = 0, {, }+ = 1

223

Quantizing, we get analogs of the Q, P operators

1 1

= (+ (1 ) + + (2 )),

= (+ (1 ) + (2 ))

2 2

which satisfy anti-commutation relations

2 =

2 = 0,

+

= 1

variable as

= multiplication by ,

=

This state space is two complex dimensional, with an arbitrary state

f () = c1 1 + c2

with cj complex numbers. The inner product on this space given by the

fermionic integral Z

(f1 (), f2 ()) = f1 ()f2 ()d

with

f () = c1 1 + c2

With respect to this inner product, one has

(1, 1) = (, ) = 0, (1, ) = (, 1) = 1

This inner product is indefinite and can take on negative values, since

(1 , 1 ) = 2

as a physical system, since this negative number is supposed to the probability of

finding the system in this state. Such quantum systems are called ghosts, and

do have applications in the description of various quantum systems, but only

when a mechanism exists for the negative-norm states to cancel or otherwise be

removed from the physical state space of the theory.

tion

While the fermionic analog of the Schrodinger construction does not give a

unitary representation, it turns out that the fermionic analog of the Bargmann-

Fock construction does, giving exactly the fermionic oscillator quantum system

that we have already studied. This will work for the case of a positive definite

symmetric bilinear form < , >. Note though that n must be even since this

will involve choosing a complex structure on the fermionic phase space Rn .

224

The corresponding pseudo-classical system will be the classical fermionic

oscillator studied in the last chapter. Recall that this uses a choice of complex

structure J on the fermionic phase space R2d , with the standard choice given

by defining

1 1

j = (2j1 + i2j ), j = (2j1 i2j )

2 2

for j = 1, . . . , d. Here < , > is positive-definite, and the j are coordinates with

respect to an orthonormal basis, so we have the standard relation {j , k }+ = jk

and the j , j satisfy

{j , k }+ = {j , k }+ = 0, {j , k }+ = jk

satisfy

[+ (j ), + (k )]+ = [+ (j ), + (k )]+ = 0

[+ (j ), + (k )]+ = jk 1

but these are just the CAR satisfied by fermionic annihilation and creation

operators. We can choose

+ (j ) = aF j , + (j ) = aF j

aF j = , aF j = multiplication by j

j

This is a complex vector space of dimension 2d , isomorphic with the state space

HF of the fermionic oscillator in d degrees of freedom, with the isomorphism

given by

If one defines a Hermitian inner product (, ) on HF by taking these basis

elements to be orthonormal, the operators aF j and aF j will be adjoints with

respect to this inner product. This same inner product can also be defined

using fermionic integration by analogy with the Bargmann-Fock definition in

the bosonic case as

Z Pd

(f1 (1 , , d ), f2 (1 , , d )) = f1 f2 e j=1 j j d1 dd d1 dd

where f1 and f2 are complex linear combinations of the powers of the anticom-

muting variables j . For the details of the construction of this inner product,

see chapter 7.2 of [45].

225

We saw in chapter 20 that quadratic combinations of the generators j of the

Clifford algebra satisfy the commutation relations of the Lie algebra so(2d) =

spin(2d). The fermionic quantization given here provides an explicit realization

of a representation of the Clifford algebra

Cliff(2d, R) on the complex vector

space HF since we know that the 2+ (j ) will satisfy the Clifford algebra

relations, and we have

1 1

2j1 = (j + j ), 2j = (j j )

2 i 2

As a result, the following linear combinations of annihilation and creation op-

erators

aj + aj , i(aj aj )

will satisfy the same Clifford algebra relations as 2j1 , 2j and taking quadratic

combinations of these operators on HF provides a representation of the Lie

algebra spin(2d).

This representation exponentiates to a representation not of the group SO(2d),

but of its double-cover Spin(2d). The representation that we have constructed

here on the fermionic oscillator state space HF is called the spinor representa-

tion of Spin(2d), and we will sometimes denote HF with this group action as

S.

In the bosonic case, HB is an irreducible representation of the Heisenberg

group, but as a representation of M p(2d, R), it has two irreducible components,

corresponding to even and odd polynomials. The fermionic analog is that HF

is irreducible under the action of the Clifford algebra Cliff(2d, C). To show

this, note that Cliff(2d, C) is isomorphic to the matrix algebra M (2d , C) and its

d

action on HF = C2 is isomorphic to the action of matrices on column vectors.

While HF is irreducible as a representation of the Clifford algebra, it is the

sum of two irreducible representations of Spin(2d), the so-called half-spinor

representations . Spin(2d) is generated by quadratic combinations of the Clif-

ford algebra generators, so these will preserve the subspaces

S+ = span{|0iF , aF j aF k |0iF , } S = HF

and

S = span{aF j |0iF , aF j aF k aF l |0iF , } S = HF

the number of creation operators used to get an element of S by action on |0iF .

The construction of the spinor representation here has involved making a spe-

cific choice of relation between the Clifford algebra generators and the fermionic

annihilation and creation operators . This is determined by our choice of com-

plex structure J, which, as in bosonic case, splits the complexification of the real

phase space R2d with its coordinates j into a complex vector space Cd with

coordinates j and a conjugate vector space with coordinates j . Again as in the

bosonic case, this choice is reflected in the existence of a distinguished direction

226

in the spinor space, |0iF HF which (up to a scalar factor), is determined by

the condition that the quantized j annihilate the state

+ (j )|0iF = aF j |0iF = 0

transformations that commute with J. Just as in the bosonic case, the Lie

algebra u(d) of this U (d) subgroup is given by the quadratic combinations of

annihilation and creation operators involving one annihilation operator and one

creation operator. The fermionic oscillator Hamiltonian

d

X

h= j j

j=1

d

1X

H= (aF j aF j aF j aF j )

2 j=1

which has half-integral eigenvalues, corresponding to the fact that the spinor

representation HF = S will be a true representation only of the double-cover

Spin(2d) of SO(2d). In parallel with the bosonic case, one can work with

normal-ordered products of the annihilation and creation operators, getting a

true representation of U (d) on HF , but when does this one works with a u(1)

that is different than the u(1) that lies in the so(2d) = spin(2d) determined by

the fermionic Poisson bracket on quadratic combinations of the j .

To summarize much of the material we have covered, it may be useful to con-

sider the following table, which explicitly gives the correspondence between the

parallel constructions we have studied in the bosonic and fermionic cases.

Bosonic Fermionic

Non-degenerate anti-symmetric Non-degenerate symmetric

bilinear form (, ) on M bilinear form < , > on M

Poisson bracket {, } = (, ) Poisson bracket {, }+ =< , >

on functions on M = R2d on anti-commuting functions on R2d

Coordinates qj , pj , basis of M Coordinates j , basis of M

Weyl algebra Weyl(2d, C) Clifford algebra Cliff(2d, C)

Momentum, position operators Pj , Qj Clifford algebra generators j

227

Sp(2d, R) preserves (, ) SO(2d, R) preserves < , >

Quadratics in Pj , Qj Quadratics in j

Representation of sp(2d, R) Representation of so(2d)

J : J 2 = 1, (Ju, Jv) = (u, v) J : J 2 = 1, < Ju, Jv >=< u, v >

M C = M+ M M C = M+ M

U (d) Sp(2d, R) commutes with J U (d) SO(2d, R) commutes with J

M p(2d, R) double-cover of Sp(2d, R) Spin(2d) double-cover of SO(2d)

Metaplectic representation Spinor representation

aj , aj satisfying CCR aj F , aj F satisfying CAR

aj |0i = 0, |0i depends on J aj F |0i = 0, |0i depends on J

For more about pseudo-classical mechanics and quantization, see [45] Chapter

7. For another discussion of the spinor representation from a similar point of

view to the one here, see [46] Chapter 12.

228

Chapter 23

Supersymmetry, Some

Simple Examples

If one considers fermionic and bosonic quantum system that each separately

have operators coming from Lie algebra or superalgebra representations on their

state spaces, when one combines the systems by taking the tensor product, these

operators will continue to act on the combined system. In certain special cases

new operators with remarkable properties will appear that mix the fermionic and

bosonic systems. These are generically known as supersymmetries. In such

supersymmetrical systems, important conventional mathematical objects often

appear in a new light. This has been one of the most fruitful areas of interaction

between mathematics and physics in recent decades, beginning with Edward

Wittens highly influential 1982 paper Supersymmetry and Morse theory [54]. A

huge array of different supersymmetric quantum field theories have been studied

over the years, although the goal of using these to develop a successful unified

theory remains still out of reach. In this chapter well examine in detail some

of the simplest such quantum systems, examples of superymmetric quantum

mechanics. These have the characteristic feature that the Hamiltonian operator

is a square.

In the previous chapters we discussed in detail

HB generated by applying N creation operators aB j to a lowest energy

state |0iB . The Hamiltonian is

N N

1 X 1

(aB j aB j + aB j aB j ) =

X

H= ~ (NB j + )~

2 j=1 j=1

2

229

where NB j is the number operator for the jth degree of freedom, with

eigenvalues nB j = 0, 1, 2, .

space HF generated by applying N creation operators aF j to a lowest

energy state |0iF . The Hamiltonian is

N N

1 X 1

(aF j aF j aF j aF j ) =

X

H= ~ (NF j )~

2 j=1 j=1

2

where NF j is the number operator for the jth degree of freedom, with

eigenvalues nF j = 0, 1.

Putting these two systems together we get a new quantum system with state

space

H = HB HF

and Hamiltonian

N

X

H= (NB j + NF j )~

j=1

Notice that the lowest energy state |0i for the combined system has energy 0,

due to cancellation between the bosonic and fermionic degrees of freedom.

For now, taking for simplicity the case N = 1 of one degree of freedom, the

Hamiltonian is

H = (NB + NF )~

with eigenvectors |nB , nF i satisfying

H|nB , nF i = (nB + nF )~

Notice that while there is a unique lowest energy state |0, 0i of zero energy, all

non-zero energy states come in pairs, with two states

|n, 0i and |n 1, 1i

This kind of degeneracy of energy eigenvalues usually indicates the existence

of some new symmetry operators commuting with the Hamiltonian operator.

We are looking for operators that will take |n, 0i to |n 1, 1i and vice-versa,

and the obvious choice is the two operators

Q+ = aB aF , Q = aB aF

which are not self adjoint, but are each others adjoints ((Q ) = Q+ ).

The pattern of energy eigenstates looks like

230

Computing anti-commutators using the CCR and CAR for the bosonic and

fermionic operators (and the fact that the bosonic operators commute with the

fermionic ones since they act on different factors of the tensor product), one

finds that

Q2+ = Q2 = 0

and

(Q+ + Q )2 = [Q+ , Q ]+ = H

1

Q1 = Q+ + Q , Q2 = (Q+ Q )

i

which satisfy

[Q1 , Q2 ]+ = 0, Q21 = Q22 = H

Q , and this fact alone tells us that the energy eigenvalues will be non-negative.

It also tells us that energy eigenstates of non-zero energy will come in pairs

231

with the same energy. To find states of zero energy, instead of trying to solve

the equation H|0i = 0 for |0i, one can look for solutions to

Q1 |0i = 0 or Q2 |0i = 0

These operators dont correspond to a Lie algebra representation as H does,

but do come from a Lie superalgebra representation, so are described as gen-

erators of a supersymmetry transformation. In more general theories with

operators like this with the same relation to the Hamiltonian, one may or may

not have solutions to

Q1 |0i = 0 or Q2 |0i = 0

If such solutions exist, the lowest energy state has zero energy and is described

as invariant under the supersymmetry. If no such solutions exist, the lowest

energy state will have a non-zero, positive energy, and satisfy

Q1 |0i =

6 0 or Q2 |0i =

6 0

In this case one says that the supersymmetry is spontaneously broken, since

the lowest energy state is not invariant under supersymmetry.

There is an example of a physical quantum mechanical system that has ex-

actly the behavior of this supersymmetric oscillator. A charged particle confined

to a plane, coupled to a magnetic field perpendicular to the plane, can be de-

scribed by a Hamiltonian that can be put in the bosonic oscillator form (to

show this, we need to know how to couple quantum systems to electromagnetic

fields, which we will come to later in the course). The equally spaced energy

levels are known as Landau levels. If the particle has spin one-half, there will

be an additional term in the Hamiltonian coupling the spin and the magnetic

field, exactly the one we have seen in our study of the two-state system. This

additional term is precisely the Hamiltonian of a fermionic oscillator. For the

case of gyromagnetic ratio g = 2, the coefficients match up so that we have ex-

actly the supersymmetric oscillator described above , with exactly the pattern

of energy levels seen there.

a superpotential

The supersymmetric oscillator system can be generalized to a much wider class

of potentials, while still preserving the supersymmetry of the system. In this

section well introduce a so-called superpotential W (q), with the harmonic

oscillator the special case

q2

W (q) =

2

For simplicity, we will here choose constants ~ = = 1

Recall that our bosonic annihilation and creation operators were defined by

1 1

aB = (Q + iP ), aB = (Q iP )

2 2

232

We will change this definition to introduce an arbitrary superpotential W (q):

1 1

aB = (W 0 (Q) + iP ), aB = (W 0 (Q) iP )

2 2

but keep our definition of the operators

Q+ = aB aF , Q = aB aF

These satisfy

Q2+ = Q2 = 0

for the same reason as in the oscillator case: repeated factors of aF or aF vanish.

Taking as the Hamiltonian the same square as before, we find

H =(Q+ + Q )2

1 1

= (W 0 (Q) + iP )(W 0 (Q) iP )aF aF + (W 0 (Q) iP )(W 0 (Q) + iP )aF aF

2 2

1 1

= (W (Q) + P )(aF aF + aF aF ) + (i[P, W 0 (Q)])(aF aF aF aF )

0 2 2

2 2

1 0 2 2 1 0

= (W (Q) + P ) + (i[P, W (Q)])3

2 2

But iP is the operator corresponding to infinitesimal translations in Q, so we

have

i[P, W 0 (Q)] = W 00 (Q)

and

1 1 00

H= (W 0 (Q)2 + P 2 ) + W (Q)3

2 2

which gives a large class of quantum systems, all with state space

H = HB HF = L2 (R) C2

odinger representation for the bosonic factor).

The energy eigenvalues will be non-negative, and energy eigenvectors with

positive energy will occur in pairs

.

There may or may not be a state with zero energy, depending on whether

or not one can find a solution to the equation

If such a solution does exist, thinking in terms of super Lie algebras, one calls

Q1 the generator of the action of a supersymmetry on the state space, and

describes the ground state |0i as invariant under supersymmetry. If no such

233

solution exists, one has a theory with a Hamiltonian that is invariant under su-

persymmetry, but with a ground state that isnt. In this situation one describes

the supersymmetry as spontaneously broken. The question of whether a given

supersymmetric theory has its supersymmetry spontaneously broken or not is

one that has become of great interest in the case of much more sophisticated su-

persymmetric quantum field theories. There, hopes (so far unrealized) of making

contact with the real world rely on finding theories where the supersymmetry

is spontaneously broken.

In this simple quantum mechanical system, one can try and explicitly solve

the equation Q1 |i = 0. States can be written as two-component complex

functions

+ (q)

|i =

(q)

and the equation to be solved is

1 + (q)

(Q+ + Q )|i = ((W 0 (Q) + iP )aF + (W 0 (Q) iP )aF )

2 (q)

1 d 0 1 d 0 0 + (q)

= ((W 0 (Q) + ) + (W 0 (Q) ) )

2 dq 0 0 dq 1 0 (q)

1 0 1 d 0 1 + (q)

= (W 0 (Q) + )

2 1 0 dq 1 0 (q)

1 0 1 d + (q)

= ( W 0 (Q)3 ) =0

2 1 0 dq (q)

which has general solution

c+ eW (q)

+ (q) c

= eW (q)3 + =

(q) c c eW (q

for complex constants c+ , c . Such solutions will only be normalizable if

c+ = 0, lim W (q) = +

q

or

c = 0, lim W (q) =

q

If, for example, W (q) is an odd polynomial, one will not be able to satisfy either

of these conditions, so there will be no solution, and the supersymmetry will be

spontaneously broken.

differential forms

If one considers supersymmetric quantum mechanics in the case of d degrees of

freedom and in the Schrodinger representation, one has

H = L2 (Rd ) (Rd )

234

the tensor product of complex-valued functions on Rd (acted on by the Weyl

algebra Weyl(2d, C)) and anti-commuting functions on Rd (acted on by the

Clifford algebra Cliff(2d, C)). There are two operators Q+ and Q , adjoints

of each other and of square zero. If one has studied differential forms, this

should look familiar. This space H is well-known to mathematicians, as the

complex valued differential forms on Rd , often written (Rd ), where here the

d-forms). In the theory of differential forms, it is well known that one has an

operator d on (Rd ) with square zero, called the de Rham differential. Using

the inner product on Rd , one can put a Hermitian inner product on (Rd )

by integration, and then d has an adjoint , also of square zero. The Laplacian

operator on differential forms is

= (d + )2

precisely to this, once one conjugates d, as follows

fact that one can construct them not just on Rd , but on a general differentiable

manifold M , with a corresponding construction of d, , operators. In Hodge

theory, one studies solutions of

= 0

(these are called harmonic forms) and finds that the dimension of the space

of such solutions can be used to get topological invariants of the manifold M .

Wittens famous 1982 paper on supersymmetry and Morse theory[54] first ex-

posed these connections, using them both to give new ways of thinking about

the mathematics involved, as well as applications of topology to the question

of deciding whether supersymmetry was spontaneously broken in various super-

symmetric models.

the Dirac operator

So far we have been considering supersymmetric quantum mechanical systems

where both the bosonic and fermionic variables that get quantized take values

in an even dimensional phase space R2d . We get two supersymmetry operators

Q1 and Q2 , so this is sometimes called N = 2 supersymmetry (in an alternate

normalization, counting complex variables, it is called N = 1). It turns out

however that there are very interesting supersymmetrical quantum mechanics

systems that one can get by quantizing bosonic variables in R2d , but fermionic

variables in Rd . Our operators will be given by the tensor product of the Weyl

235

algebra in 2d variables and the Clifford algebra in d variables. The state space

will be

H = L2 (Rd ) S

where S is the spinor space that the Clifford algebra in d variables acts on

irreducibly.

If our Hamiltonian operator is the Laplacian on Rd , we have

d

X

H= Pj2

j=1

and one can make this act not just on functions L2 (Rd ) but on H, the spinor-

valued functions on Rd . Doing this, we find again that there is a supersymmetry

Q, just one this time (so called N = 1, or alternately N = 12 supersymmetry),

and H is a square. The operator Q is given by

d

X

j P j

j=1

where j are the generators of the Clifford algebra. Using the defining relations

for the j , one can easily see that

Q2 = H

This operator Q is called the Dirac operator, and it plays a critical role in

physics, which we will study later in the course when we study relativistic quan-

tum systems. It also plays an important role in mathematics, since like the de

Rham differential it can be constructed for not just Rn , but arbitrary manifolds

with metrics, where it can be used to study their geometry and topology.

For a reference at the level of these notes, see [20]. For more details about

supersymmetric quantum mechanics and its relationship to the Dirac operator

and the index theorem, see the graduate level textbook of Tahktajan[45], and

lectures by Orlando Alvarez[1]. These last two sources also describe the formal-

ism one can get by putting together the standard Hamiltonian mechanics and

its fermionic analog, consitently describing both classical system with com-

muting and anti-commuting variables and its quantization to get the quantum

systems described in this chapter.

236

Chapter 24

the Path Integral

In this chapter well give a rapid survey of a different starting point for devel-

oping quantum mechanics, based on the Lagrangian rather than Hamiltonian

classical formalism. Lagrangian methods have quite different strengths and

weaknesses than those of the Hamiltonian formalism, and well try and point

these out, while referring to standard physics texts for more detail about these

methods.

The Lagrangian formalism leads naturally to an apparently very different

notion of quantization, one based upon formulating quantum theory in terms of

infinite-dimensional integrals known as path integrals. A serious investigation

of these would require another and very different volume, so again well have to

restrict ourselves to outlining how path integrals work, describing their strengths

and weaknesses, and giving references to standard texts for the details.

In the Lagrangian formalism, instead of a phase space R2d of positions qj and

momenta pj , one considers just the position space Rd . Instead of a Hamiltonian

function h(q, p), one has

Definition. Lagrangian

The Lagrangian L is a function

L : (q, v) Rd Rd L(q, v) R

: t [t1 , t2 ] Rd

237

which we will write in terms of their position and velocity vectors as

(t) = (q(t), q(t))

Definition. Action

The action S for a path is

Z t2

S[] =

L(q(t), q(t))dt

t1

ism is that classical trajectories are given by critical points of the action func-

tional. These may correspond to minima of the action (so this is sometimes

called the principle of least action), but one gets classical trajectories also for

critical points that are not minima of the action. One can define the appropriate

notion of critical point as follows

A path is a critical point of the functional S[] if

d

S() S(s )|s=0 = 0

ds

when

s : [t1 , t2 ] Rd

is a smooth family of paths parametrized by an interval s (, ), with 0 = .

Well now ignore analytical details and adopt the physicists interpretation

of this as the first-order change in S due to an infinitesimal change =

(q(t), q(t)) in the path.

When (q(t), q(t)) satisfy a certain differential equation, the path will be a

critical point and thus a classical trajectory:

One has

S[] = 0

for all variations of with endpoints (t1 ) and (t2 ) fixed if

L d L

(q(t), q(t)) (

(q(t), q(t))) =0

qj dt qj

Proof. Ignoring analytical details, the Euler-Lagrange equations follow from the

following calculations, which well just do for d = 1, with the generalization to

238

higher d straightfoward. We are calculating the first-order change in S due to

an infinitesimal change = (q(t), q(t))

Z t2

S[] = L(q(t), q(t))dt

t1

Z t2

L L

= ( (q(t), q(t))q(t)

+ (q(t), q(t))

q(t))dt

t1 q q

But

d

q(t)

= q(t)

dt

and, using integration by parts

L d L d L

q(t)

= ( q) ( )q

q dt q dt q

so

Z t2

L d L d L

S[] = (( )q ( q))dt

t1 q dt q dt q

Z t2

L d L L L

= ( )qdt ( q)(t2 ) + ( q)(t1 )

t1 q dt q q q

L d L

(q(t), q(t))

( (q(t), q(t)))

=0

q dt q

the integral will be zero for arbitrary variations q.

As an example, a particle moving in a potential V (q) will be described by a

Lagrangian

d

1 X 2

= m

L(q, q) q V (q)

2 j=1 j

V d

= (mqj )

qj dt

This is just Newtons second law which says that the force coming from the

potential is equal to the mass times the acceleration of the particle.

The derivation of the Euler-Lagrange equations can also be used to study

the implications of Lie group symmetries of a Lagrangian system. When a Lie

group G acts on the space of paths, preserving the action S, it will take classical

trajectories to classical trajectories, so we have a Lie group action on the space

of solutions to the equations of motion (the Euler-Lagrange equations). In

239

good cases, this space of solutions is just the phase space of the Hamiltonian

formalism. On this space of solutions, we have, from the calculation above

L L

S[] = ( q(X))(t1 ) ( q(X))(t2 )

q q

where now q(X) is the infinitesimal change in a classical trajectory coming

from the infinitesimal group action by an element X in the Lie algebra of G.

From invariance of the action S under G we must have S=0, so

L L

( q(X))(t2 ) = ( q(X))(t1 )

q q

This is an example of a more general result known as Noethers theorem.

It says that given a Lie group action on a Lagrangian system that leaves the

action invariant, for each element X of the Lie algebra we will have a conserved

quantity

L

q(X)

q

which is independent of time along the trajectory. A basic example is when

the Lagrangian is independent of the position variables qj , depending only on

the velocities qj , for example in the case of a free particle, when V (q) = 0. In

such a case one has invariance under the Lie group Rd of space-translations.

An infinitesimal transformation in the j-direction is qiven by

qj (t) =

L

qj

For the case of the free particle, this will be

L

= mqj

qj

and the conservation law is conservation of momentum.

Given a Lagrangian classical mechanical system, one would like to be able

to find a corresponding Hamiltonian system. To do this, we proceed by defining

the momenta pj as above, as the conserved quantities corresponding to space-

translations, so

L

pj =

qj

Then, instead of working with trajectories characterized at time t by

(q(t), q(t)) R2d

240

where pj = Lqj and this R

2d

is the Hamiltonian phase space with the conven-

tional Poisson bracket.

This transformation between position-velocity and phase space is known as

the Legendre transform, and in good cases (for instance when L is quadratic

in all the velocities) it is an isomorphism. In general though, this is not an

isomorphism, with the Legendre transform often taking position-velocity space

to a lower-dimensional subspace of phase space. Such cases require a much more

elaborate version of Hamiltonian formalism, known as constrained Hamiltonian

dynamics and are not unusual: one example we will see later is that of the

equations of motion of a free electromagnetic field (Maxwells equations).

Besides a phase space, for a Hamiltonian system one needs a Hamiltonian

function. Choosing

Xd

h= pj qj L(q, q)

j=1

L

pj =

qj

to solve for the velocities qj and express them in terms of the momentum vari-

ables. In that case, computing the differential of h one finds (for d = 1, the

generalization to higher d is straightforward)

L L

dh =pdq + qdp dq dq

q q

L

=qdp dq

q

So one has

h h L

= q,

=

p q q

but these are precisely Hamiltons equations since the Euler-Lagrange equations

imply

L d L

= = p

q dt q

The Lagrangian formalism has the advantage of depending concisely just on

the choice of action functional, which does not distinguish time in the same

way that the Hamiltonian formalism does by its dependence on a choice of

Hamiltonian function h. This makes the Lagrangian formalism quite useful in

the case of relativistic quantum field theories, where one would like to exploit the

full set of space-time symmetries, which can mix space and time directions. On

the other hand, one loses the infinite dimensional group of symmetries of phase

space for which the Poisson bracket is the Lie bracket (the so-call canonical

transformations). In the Hamiltonian formalism we saw that the harmonic

oscillator could be best understood using such symmetries, in particular the

241

U (1) symmetry generated by the Hamiltonian function. The harmonic oscillator

is a more difficult problem in the Lagrangian formalism, where this symmetry

is not manifest.

After the Legendre transform of a Lagrangian classical system to phase space

and thus to Hamiltonian form, one can then apply the method of quantization

we have discussed extensively earlier (known to physicists as canonical quan-

tization. There is however a very different approach to quantization, which

completely bypasses the Hamiltonian formalism. This is the path integral for-

malism, which is based upon a method for calculating matrix elements of the

time-evolution operator

i

hqT |e ~ HT |q0 i

in the position eigenstate basis in terms of an integral over the space of paths

that go from q0 to q1 in time T . Here |q0 i is an eigenstate of Q with eigenvalue

q0 (a delta-function at q0 in the position space representation), and |qT i has Q

eigenvalue qT (as in many cases, well stick to d = 1 for this discussion). This

matrix element has a physical interpretation as the amplitude for a particle

starting at q0 at t = 0 to end up at qT at time T , with its norm-squared giving

the probability density for observing the particle at position qT .

To try and derive a path-integral expression for this, one breaks up the

interval [0, T ] into N equal-sized sub-intervals and calculates

i

hqT |(e N ~ HT )N |q0 i

that

i i i

hqT |e ~ HT |q0 i = lim hqT |(e N ~ KT e N ~ V T )N |q0 i

N

If K(P ) can be chosen to depend only on the momentum operator P and V (Q)

depends only on the operator Q then one can insert alternate copies of the

identity operator in the forms

Z Z

|qihq|dq = 1, |pihp|dp = 1

i i

hqj |e N ~ K(P )T |pj ihpj |e N ~ V (Q)T |qj1 i

where the index j goes from 1 to N and the pj , qj variable will be integrated

over. Such a term can be evaluated as

i i

hqj |pj ihpj |qj1 ie N ~ K(pj )T e N ~ V (qj1 )T

242

1 i 1 i i i

= e ~ q j pj e ~ qj1 jpj e N ~ K(pj )T e N ~ V (qj1 )T

2~ 2~

1 i pj (qj qj1 ) i (K(pj )+V (qj1 ))T

= e~ e N~

2~

1 N

The N factors of this kind give an overall factor of ( 2~ ) times something

which is a discretized approximation to

i

RT

(pqh(q(t),p(t)))dt

e~ 0

where the phase in the exponential is just the action. Taking into account the

integrations over qj and pj one has

N Z

1 NY

Z RT

i i

hqT |e ~ HT |q0 i = lim ( ) dpj dqj e ~ 0 (pqh(q(t),p(t)))dt

N 2~ j=1

and one can interpret the integration as an integral over the space of paths in

phase space, thus a phase space path integral.

This is an extremely simple and seductive expression, apparently saying that,

once the action S is specified, a quantum system is defined just by considering

integrals Z

i

D e ~ S[]

over paths in phase space, where D is some sort of measure on this space of

paths. Since the integration just involves factors of dpdq and the exponential

just pdq and h, this formalism seems to share the same sort of invariance under

the infinite-dimensional group of canonical transformations (transformations of

the phase space preserving the Poisson bracket) as the classical Hamiltonian

formalism. It also appears to solve our problem with operator ordering am-

biguities, since introducing products of P s and Qs at various times will just

give a phase space path integral with the corresponding p and q factors in the

integrand, but these commute.

Unfortunately, we know from the Groenewold-van Hove theorem that this is

too good to be true. This expression cannot give a unitary representation of the

full group of canonical transformations, at least not one that is irreducible and

restricts to what we want on transformations generated by linear functions q

and p. Another way to see the problem is that a simple argument shows that by

canonical transformations one can transform any Hamiltonian into free-particle

Hamiltonian, so all quantum systems would just be free particles in some choice

of variables. For the details of these arguments and a careful examination of

what goes wrong, see chapter 31 of [39]. One aspect of the problem is that, as a

measure on the discrete sets of points qj , pj , points in phase space for successive

values of j are not likely to be close together, so thinking of the integral as an

integral over paths is not justified.

When the Hamiltonian h is quadratic in the momentum p, the pj integrals

will be Gaussian integrals that can be performed exactly. Equivalently, the

243

kinetic energy part K of the Hamiltonian operator will have a kernel in position

space that can be computed exactly. Using one of these, the pj integrals can be

eliminated, leaving just integrals over the qj that one might hope to interpret as

a path integral over paths not in phase space, but in position space. One finds,

P2

if K = 2m

i

hqT |e ~ HT |q0 i =

N Z

m Y m(qj qj1 )2

r

i2~T N i

PN T

lim ( ) 2 dqj e ~ j=1 ( 2T /N V (qj ) N )

N N m i2~T j=1

Z T

1

S() = dt( m(q2 ) V (q(t)))

0 2

One can try and properly normalize things so that this limit becomes an integral

Z

i

D e ~ S[]

where now the paths (t) are paths in the position space.

An especially attractive aspect of this expression is that it provides a simple

understanding of how classical behavior emerges in the classical limit as ~ 0.

The stationary phase approximation method for oscillatory integrals says that,

for a function f with a single critical point at x = xc (i.e. f 0 (xc ) = 0) and for a

small parameter , one has

Z +

1 1

dx eif / = p eif (xc )/ (1 + O())

i2 f 00 (c)

Using the same principle for the infinite-dimensional path integral, with f = S

the action functional on paths, and = ~, one finds that for ~ 0 the path

integral will simplify to something that just depends on the classical trajectory,

since by the principle of least action, this is the critical point of S.

Such position-space path integrals do not have the problems of principle

of phase space path integrals coming from the Groenewold-van Hove theorem,

but they still have serious analytical problems since they involve an attempt to

integrate a wildly oscillating phase over an infinite-dimensional space. One does

not naturally get a unitary result for the time evolution operator, and it is not

clear that whatever results one gets will be independent of the details of how

one takes the limit to define the infinite-dimensional integral.

Such path integrals though are closely related to integrals that are known

to make sense, ones that occur in the theory of random walks. There, a well-

defined measure on paths does exist, Wiener measure. In some sense Wiener

measure is what one gets in the case of the path integral for a free particle, but

taking the time variable t to be complex and analytically continuing

t it

244

So, one can use Wiener measure techniques to define the path integral, getting

results that need to be analytically continued back to the physical time variable.

In summary, the path integral method has the following advantages:

Study of the classical limit and semi-classical effects (quantum effects

at small ~) is straightforward.

Calculations for free particles and for series expansions about the free

particle limit can be done just using Gaussian integrals, and these are rela-

tively easy to evaluate and make sense of, despite the infinite-dimensionality

of the space of paths.

ing Wiener measure techniques, and often evaluated numerically even in

cases where no exact solution is known.

On the other hand, there are disadvantages:

Some path integrals such as phase space path integrals do not at all have

the properties one might expect, so great care is required in any use of

them.

How to get unitary results can be quite unclear. The analytic continua-

tion necessary to make path integrals well-defined can make their physical

interpretation obscure.

just symmetries of configuration space are difficult to see using the con-

figuration space path integral, with the harmonic oscillator providing a

good example. One can see such symmetries using the phase-space path

integral, but this is not reliable.

with the bosonic case, using the notion of fermionic integration discussed earlier.

For much more about Lagrangian mechanics and its relation to the Hamiltonian

formalism, see [2]. More along the lines of the discussion here can be found in

most quantum mechanics and quantum field theory textbooks. For the path

integral, Feynmans original paper [13] or his book [14] are quite readable. A

typical textbook discussion is the one in chapter 8 of Shankar [40]. The book by

Schulman [39] has quite a bit more detail, both about applications and about

the problems of phase-space path integrals. Yet another fairly comprehensive

treatment, including the fermionic case, is the book by Zinn-Justin.

245

246

Chapter 25

Non-relativistic Quantum

Fields and Field

Quantization

The quantum mechanical systems we have studied so far describe a finite num-

ber of degrees of freedom, which may be of a bosonic or fermionic nature. In

particular we have seen how to describe a quantized free particle moving in

three-dimensional space. By use of the notion of tensor product, we can then

describe any particular fixed number of such particles. We would however, like

a formalism capable of conveniently describing an arbitrary number of parti-

cles. From very early on in the history of quantum mechanics, it was clear

that at least certain kinds of particles, photons, were most naturally described

not one by one, but by thinking of them as quantized excitations of a classical

system with an infinite number of degrees of freedom: the electromagnetic field.

In our modern understanding of fundamental physics not just photons, but all

elementary particles are best described in this way.

Conventional textbooks on quantum field theory often begin with relativis-

tic systems, but well start instead with the non-relativistic case. Well study a

simple quantum field theory that extends the conventional single-particle quan-

tum systems we have dealt with so far to deal with multi-particle systems. This

version of quantum field theory is what gets used in condensed matter physics,

and is in many ways simpler than the relativistic case, which well take up in a

later chapter.

Quantum field theory is a large and complicated subject, suitable for a full-

year course at an advanced level. Well be giving only a very basic introduc-

tion, mostly just considering free fields, which correspond to systems of non-

interacting particles. Most of the complexity of the subject only appears when

one tries to construct quantum field theories of interacting particles. A remark-

able aspect of the theory of free quantum fields is that in many ways it is little

more than something we have already discussed in great detail, the quantum

247

harmonic oscillator problem. However, the classical harmonic oscillator phase

space that is getting quantized in this case is an infinite dimensional one, the

space of solutions to the free particle Schrodinger equation. To describe multiple

non-interacting fermions, we just need to use fermionic oscillators.

For simplicity well set ~ = 1 and start with the case of a single spatial

dimension. Well also begin using x to denote a spatial variable instead of the q

conventional when this is the coordinate variable in a finite-dimensional phase

space.

of a harmonic oscillator

It turns out that quantum systems of identical particles are best understood by

thinking of such particles as quanta of a harmonic oscillator system. We will

begin with the bosonic case, then later consider the fermionic case, which uses

the fermionic oscillator system.

A fundamental postulate of quantum mechanics is that given a space of states

H1 describing a bosonic single particle, a collection of N particles is described

by

(H1 H1 )S

| {z }

N times

where the superscript S means we take elements of the tensor product invariant

under the action of the group SN by permutation of the factors. We want to

consider state spaces containing an arbitrary number of particles, so we define

Definition (Bosonic Fock space, the symmetric algebra). Given a complex vec-

tor space V , the symmetric Fock space is defined as

F S (V ) = C V (V V )S (V V V )S

S N (V ) = (V V )S

| {z }

N times

consisting of linear combinations of states with N quanta (i.e. states one gets

by applying N creation operators to the lowest energy state), for N = 0, 1, 2, . . ..

We have seen in our discussion of the quantization of the harmonic oscillator that

in the Bargmann-Fock representation, the state space is just C[z1 , z2 , . . . , zd ],

the space of polynomials in d complex variables.

248

The part of the state space with N quanta has dimension

N +d1

N

which grows with N . This is just the binomial coefficient giving the number

of d-variable monomials of degree N The quanta of a harmonic oscillator are

indistinguishable, which corresponds to the fact that the space of states with

N quanta can be identified with the symmetric part of the tensor product of N

copies of Cd .

More precisely, what one has is

the symmetric algebra S (V ) (where V is the dual vector space to V ) and the

algebra C[V ] of polynomial functions on V .

We wont try and give a detailed proof of this here, but one can exhibit the

isomorphism explicitly on generators. If zj V are the coordinate functions

with respect to a basis ej of V (i.e. zj (ek ) = jk ), they give a basis of V which

is also a basis of the linear polynomial functions on V . For higher degrees, one

makes the identification

zj zj S N (V ) zjN

| {z }

N times

algebra.

Besides describing states of multiple identical quantum systems as polyno-

mials or as symmetric parts of tensor products, a third description is useful.

This is the so-called occupation number representation, where states are la-

beled by d non-negative integers nj = 0, 1, 2, , with an identification with the

polynomial description as follows:

n

|n1 , n2 , . . . , nj , . . . , nd i z1n1 z2n2 zj j zdnd

So, starting with a single particle state space H1 , we have three equivalent

descriptions of the Fock space describing multi-particle bosonic states

H1 , the dual vector space to H1 .

As polynomial functions on H1 .

0, 1, 2, . . ..

operators aj , aj , satisfying the canonical commutation relations, and these will

generate an algebra of operators on the Fock space.

249

25.1.2 Fermions and the fermionic oscillator

For the case of fermionic particles, theres an analogous Fock space construction

using tensor products:

space V , the Fermionic Fock space is

F A (V ) = C V (V V )A (V V V )A

where the superscript A means the subspace of the tensor product that just

changes sign under interchange of two factors. This is known to mathemati-

cians as the exterior algebra (V ), with

N (V ) = (V V )A

| {z }

N times

now polynomials in anti-commuting variables.

For the case of fermionic particles with single particle state space H1 , one

can again define the multi-particle state space in three equivalent ways:

also think of such polynomials as anti-symmetric multilinear functions on

the product of copies of H1 .

nj = 0, 1.

For each of these descriptions, one can define d annihilation or creation oper-

ators aj , aj satisfying the canonical anti-commutation relations, and these will

generate an algebra of operators on the Fock space.

odinger

equation

To describe multi-particle quantum systems we will take as our single-particle

state space H1 the space of solutions of the free-particle Schrodinger equation,

as already studied in chapter 10. As we saw in that chapter, one can use a finite

box normalization, which gives a discrete, countable basis for this space and

then try and take the limit of infinite box size. To fully exploit the symmetries of

phase space though, we will need the continuum normalization, which requires

considering not just functions but distributions.

250

25.2.1 Box normalization

Recall that for a free particle in one dimension the state space H consists of

complex-valued functions on R, with observables the self-adjoint operators for

momentum

d

P = i

dx

and energy (the Hamiltonian)

P2 1 d2

H= =

2m 2m dx2

Eigenfunctions for both P and H are the functions of the form

p (x) eipx

2

p

for p R, with eigenvalues p for P and 2m for H.

Note that these eigenfunctions are not normalizable, and thus not in the

conventional choice of state space as L2 (R). One way to deal with this issue is

to do what physicists sometimes refer to as putting the system in a box, by

imposing periodic boundary conditions

(x + L) = (x)

ered to those on an interval of length L. For our eigenfunctions, this condition

is

eip(x+L) = eipx

so we must have

eipL = 1

which implies that

2

p= l

L

for l an integer. Then p will take on a countable number of discrete values

corresponding to the l Z, and

1 1 2l

|li = l (x) = eipx = ei L x

L L

will be orthornormal eigenfunctions satisfying

This box normalization is one form of what physicists call an infrared cutoff,

a way of removing degrees of freedom that correspond to arbitrarily large sizes,

in order to make a problem well-defined. To get a well-defined problem one

starts with a fixed value of L, then one studies the limit L .

251

The number of degrees of freedom is now countable, but still infinite. In order

to get a completely well-defined problem, one typically needs to first make the

number of degrees of freedom finite. This can be done with an additional cutoff,

an ultraviolet cutoff, which means restricting attention to |p| for some

finite , or equivalently |l| < L

2 . This makes the state space finite dimensional

and one then studies the limit.

For finite L and our single-particle state space H1 is finite dimensional,

with orthonormal basis elements l (x). An arbitrary solution to the Schrodinger

equation is then given by

+ L

2

X 2l 4 2 l2

(x, t) = l ei L x ei 2mL2 t

l= L

2

initial value at t = 0

+ L

2

2l

X

(x, 0) = (l)ei L x

l= L

2

these coordinates are in the dual space H1 .

Multi-particle states are now described by the Fock spaces F S (H1 ) or F A (H1 ),

depending on whether the particles are bosons or fermions. In the occupation

number representation of the Fock space, orthonormal basis elements are

where the subscript j indexes the possible values of the momentum p (which are

discretized in units of 2

L , and in the interval [, ]). The occupation number

npj is the number of particles in the state with momentum pj . In the bosonic

case it takes values 0, 1, 2, , , in the fermionic case it takes values 0 or 1.

The state with all occupation numbers equal to zero is denoted

| , 0, 0, 0, i = |0i

For each pj we have annihilation and creation operators apj and apj . These

satisfy the commutation relations

[apj , apk ] = jk

apj | , npj1 , npj , npj+1 , i = npj | , npj1 , npj 1, npj+1 , i

p

npj + 1| , npj1 , npj + 1, npj+1 , i

Observables one can build out of these operators include

252

The number operator

X

N

b= apk apk

k

X

b | , np , np , np , i = (

N npk )| , npj1 , npj , npj+1 , i

j1 j j+1

k

X

Pb = pk apk apk

k

X

Pb| , npj1 , npj , npj+1 , i = ( nk pk )| , npj1 , npj , npj+1 i

k

The Hamiltonian

X p2

k

H

b = a ap

2m pk k

k

b , np , np , np , i = (

X p2k

H| j1 j j+1

n k )| , npj1 , npj , npj+1 , i

2m

k

With ultraviolet and infrared cutoffs in place, the possible values of pj are

finite in number, H1 is finite dimensional and this is nothing but the standard

quantized harmonic oscillator (with a Hamiltonian that has different frequencies

p2j

(pj ) =

2m

for different values of j). In the limit as one or both cutoffs are removed, H1

becomes infinite dimensional and the Stone-von Neumann theorem no longer

applies. This implies that not all choices of state space are unitarily equivalent.

In particular, state spaces with different choices of vacuum state |0i can be

unitarily inequivalent, with not just the dynamics of states in the state space

dependent on the Hamiltonian, but the state space itself depending on the

Hamiltonian (through the characterization of |0i as lowest energy state). Even

for the free particle, we have here defined the Hamiltonian as the normal-ordered

version, which for finite dimensional H1 differs from the non-normal-ordered one

just by a constant, but as cut-offs are removed this constant becomes infinite,

requiring careful treatment of the limit.

253

25.2.2 Continuum normalization

A significant problem introduced by using cutoffs such as the box normalization

is that these ruin some of the space-time symmetries of the system. The one-

particle space with an infrared cutoff is a space of functions on a discrete set of

points, and this set of points will not have the same symmetries as the usual

continuous momentum space (for instance in three dimensions it will not carry

an action of the rotation group SO(3)). In our study of quantum field theory

we would like to exploit the action of space-time symmetry groups on the state

space of the theory, so need a formalism that preserves such symmetries.

In our earlier discussion of the free particle, we saw that physicists often

work with a continuum normalization such that

1

|pi = p (x) = eipx , hp0 |pi = (p p0 )

2

where formulas such as the second one need to be interpreted in terms of dis-

tributions. In quantum field theory we want to be able to think of each value

of p as corresponding to a classical degree of freedom that gets quantized, and

this continuum normalization will then correspond to an uncountable number

of degrees of freedom, requiring great care when working in such a formalism.

This will however allow us to readily see the action of space-time symmetries

on the states of the quantum field theory and to exploit the duality between

position and momentum space embodied in the Fourier transform.

In the continuum normalization, an arbitrary solution to the free-particle

Schrodinger equation is given by

Z

1 p2

(x, t) = (p)eipx ei 2m t dp

2

solutions are in one-to-one correspondence with initial data

Z

1

(x, 0) = (p)eipx dp

2

^

terms of its Fourier transform (x, 0)(p) = (p). Note that we want to con-

sider not just square integrable functions (p), but non-integrable functions

like (p) = 1 (which corresponds to (x, 0) = (x)), and distributions such as

(p) = (p), which corresponds to (x, 0) = 1.

We will generally work with this continuum normalization, taking as our

single-particle space H1 the space of complex valued functions (x, 0) on R.

One can think of the |pi as an orthornomal basis of H1 , with (p) the coor-

dinate function for the |pi basis vector. Quantization will give corresponding

annihilation and creation operators a(p), a (p), which need to be thought of as

operator-valued distributions: what really is a well-defined operator is not a(p),

254

but Z +

f (p)a(p)dp

Observable operators of the theory will include the number operator

Z +

Nb= a(p) a(p)dp

Z +

Pb = pa(p) a(p)dp

+

p2

Z

H

b = a(p) a(p)dp

2m

treatment (something we will not enter into here) for their proper definition.

The formalism developed so far works well to describe states of multiple free

particles, but does so purely in terms of states with well-defined momenta, with

no information at all about their position. To get operators that know about

position, one can Fourier transform the annihilation and creation operators for

momentum eigenstates as follows (well begin with the box normalization):

Definition (Quantum field operator). The quantum field operators for the free

particle system are

X X 1

(x)

b = p (x)ap = eipx ap

p p L

X 1

p (x)ak = eikx ak

X

b (x) =

p k

L

2j

(where k takes the discrete values kj = L ).

Note that these are not self-adjoint operators, and thus not themselves ob-

servables. To get some idea of their behavior, one can calculate what they do

to the vacuum state. One has

(x)|0i

b =0

255

1 X ipx

b (x)|0i = e | , 0, np = 1, 0, i

L p

While this sum makes sense as long as it is finite, when cutoffs are removed it is

clear that (x) will have a rather singular limit as an infinite sum of operators.

It can be interpreted as the operator that creates a particle localized precisely

at x.

The field operators allow one to recover conventional wave-functions, for

single and multiple-particle states. One sees by orthonormality of the occupation

number basis states that

1

h , 0, np = 1, 0, |b (x)|0i = eipx = p (x)

L

the complex conjugate wave function of the single-particle state of momentum

p. An arbitrary one particle state |1 i with wave function (x) is a linear

combination of such states, and taking complex conjugates one finds

h0|(x)|

b 1 i = (x)

Similarly, for a two-particle state of identical particles with momenta pj1 and

pj2 one finds

h0|(x b 2 )| , 0, np = 1, 0, , 0, np = 1, 0, i = p ,p (x1 , x2 )

b 1 )(x

j1 j2 j1 j2

where

pj1 ,pj2 (x1 , x2 )

is the wavefunction (symmetric under interchange of x1 and x2 for bosons) for

this two particle state. For a general two-particle state |2 i with wavefunction

(x1 , x2 ) one has

h0|(x b 2 )|2 i = (x1 , x2 )

b 1 )(x

and one can easily generalize this to see how field operators are related to wave-

functions for an arbitrary number of particles.

Cutoffs ruin translational invariance and calculations with them quickly be-

come difficult. Well now adopt the physicists convention of working directly

in the continuous case with no cutoff, at the price of having formulas that only

make sense as distributions. One needs to be aware that the correct interpreta-

tion of such formulas may require going back to the cutoff version.

In the continuum normalization we take as normalized eigenfunctions for the

free particle

1

|pi = p (x) = eipx

2

with Z

1 0

hp0 |pi = ei(pp )x dx = (p p0 )

2

The annihilation and creation operators satisfy

[a(p), a (p0 )] = (p p0 )

256

The field operators are then

Z

1

(x)

b = eipx a(p)dx

2

Z

1

b (x) = eipx a (p)dx

2

[(x),

b b 0 )] = [b (x), b (x0 )] = 0

(x

Z Z

1 0 0

[(x),

b b (x0 )] = eipx eip x [a(p), a (p0 )]dpdp0

2

Z Z

1 0 0

= eipx eip x (p p0 )dpdp0

2

Z

1 0

= eip(xx ) dp

2

=(x x0 )

operator-valued distributions, with the distributions defined on some appropri-

ately chosen function space, such as a Schwartz space of functions such that the

function and its derivatives fall off faster than any power at infinity. Given

such functions f, g, one gets operators

Z Z

(f

b )= f (x)(x)dx,

b b (g) = g(x)b (x)dx

Z Z Z

0 0 0

[(f ), (g)] =

b b f (x)g(x )(x x )dxdx = f (x)g(x)dx

of distributions are not necessarily defined, so one has trouble making sense of

rather innocuous looking expressions like

4

((x))

b

There are observables that one can define simply using field operators. These

include:

. One can define a number density operator

The number operator N

b(x) = b (x)(x)

n b

257

and integrate it to get an operator with eigenvalues the total number of

particles in a state

Z

N=

b n

b(x)dx

Z Z Z

1 0 1

= eip x a (p0 ) eipx a(p)dpdp0 dx

2 2

Z Z

= (p p0 )a (p0 )a(p)dpdp0

Z

= a (p)a(p)dp

The total momentum operator Pb. This can be defined in terms of field

operators as

Z

d b

P =

b b (x)(i )(x)dx

dx

Z Z Z

1 0 1

= eip x a (p0 )(i)(ip) eipx a(p)dpdp0 dx

2 2

Z Z

= (p p0 )pa (p0 )a(p)dpdp0

Z

= pa (p)a(p)dp

The Hamiltonian H.

b This can be defined much like the momentum, just

changing

d 1 d2

i

dx 2m dx2

to find

Z Z 2

1 d2 b p

H=

b (x)(

b

2

)(x)dx = a (p)a(p)dp

2m dx 2m

observable O

b corresponding to any one-particle quantum mechanical observable

O Z

O

b= b (x)O(x)dx

b

potential V (x), one takes the Hamiltonian to be

1 d2

Z

H

b = b (x)( + V (x))(x)dx

b

2m dx2

258

If one can solve the one-particle Schrodinger equation for a complete set of

orthonormal wave functions n (x), one can describe this quantum system using

the same techniques as for the free particle. A creation-annihilation operator

pair an , an is associated to each eigenfunction, and quantum fields are defined

by X X

(x)

b = n (x)an , b (x) = n (x)an

n n

For Hamiltonians just quadratic in the quantum fields, quantum field the-

ories are quite tractable objects. They are in some sense just free quantum

oscillator systems, with all of their symmetry structure intact, but taking the

number of degrees of freedom to infinity. Higher order terms though make

quantum field theory a difficult and complicated subject, one that requires a

year-long graduate level course to master basic computational techniques, and

one that to this day resists mathematicians attempts to prove that many ex-

amples of such theories have even the basic expected properties. In the theory

of charged particles interacting with an electromagnetic field, when the electro-

magnetic field is treated classically one still has a Hamiltonian quadratic in the

field operators for the particles. But if the electromagnetic field is treated as a

quantum system, it acquires its own field operators, and the Hamiltonian is no

longer quadratic in the fields, a vastly more complicated situation described as

aninteracting quantum field theory.

Even if one restricts attention to the quantum fields describing one kind of

particles, there may be interactions between particles that add terms to the

Hamiltonian, and these will be higher order than quadratic. For instance, if

there is an interaction between such particles described by an interaction energy

v(y x), this can be described by adding the following quartic term to the

Hamiltonian

1 b

Z Z

(x)b (y)v(y x)(y)

b (x)dxdy

b

2

major topic in condensed matter physics.

One can easily extend the above to three spatial dimensions, getting field op-

erators (x)

b and b (x), defined by integrals over three-dimensional momentum

space. For instance, in the continuum normalization

Z

1 3

(x) =

b ( ) 2 eipx a(p)d3 p

R3 2

and the Hamiltonian for the free field is

Z

1 2 b

Hb = b (x)( )(x)d3 x

R3 2m

More remarkably, one can also very easily write down theories of quantum

systems with an arbitrary number of fermionic particles, just by changing com-

mutators to anti-commutators for the creation-annihilation operators and using

259

fermionic instead of bosonic oscillators. One gets fermionic fields that satisfy

anti-commutation relations

[(x),

b b (x0 )]+ = (x x0 )

Instead of motivating the definition of quantum fields by starting with anni-

hilation and creation operators for free particle of fixed momentum, one can

more simply just define them as what one gets by taking the space H1 of solu-

tions of the free single particle Schrodinger equation as a classical phase space,

and quantizing to get a unitary representation (of a Heisenberg algebra that is

now infinite-dimensional). This procedure is sometimes called second quan-

tization, with first quantization what was done when one started with the

classical phase space for a single particle and quantized to get the space H1 of

wavefunctions.

Some things to note about this are

The Schr odinger equation is first-order in time, so solutions are determined

by the initial values (x, 0) of the wave-function at t = 0 and H1 is thus

just a space of complex functions, the wave-functions at t = 0. This is

unlike other classical systems we have quantized, where the equation of

motion is second-order in time, so solutions are parametrized by two pieces

of initial-value data, the coordinates and momenta (since one needs initial

velocities as well as positions).

Since wave-functions are complex valued, H1 is already a complex vector

space, and we can quantize by the Bargmann-Fock method using this com-

plex structure. This is quite unlike our previous examples of quantization,

where we started with a real phase space and needed to choose a complex

structure (to get annihilation and creation operators).

Using Fourier transforms we can think of H1 either as a space of functions

of position x, or as a space of functions of momentum p. This corresponds

to two possible choices of orthonormal bases of the function space H1 : the

|pi (plane waves of momentum p) or the |xi (delta-functions at position

x). In the finite dimensional case it is the coordinate functions qj , pj on

phase space, which lie in the dual phase space M, that get mapped to

operators Qj , Pj under quantization. Here one can quantize either the

(p) (coordinates with respect to the |pi basis) as annihilation operators

a(p), or the (x) (field value at x, coordinates with respect to the |xi

basis) as field operators (x).

b

H1 , one should just recall that in the Bargmann-Fock quantization we found

260

that, choosing a complex structure and complex coordinates zj on phase space,

the non-zero Poisson bracket relations were

{zj , iz l } = jl

If we use the |pi basis for H1 , our complex coordinates will be (p), (p)

with Poisson bracket

{(p), i(p0 )} = (p p0 )

and under quantization we have

The fact that one gets a unitary Heisenberg Lie algebra representation 0 when

multiplying these operators by i implies the commutator relations

[a(p), a (p0 )] = (p p0 )1

Using instead the |xi basis for H1 , our complex coordinates will be (x), (x)

with Poisson brackets

{(x), i(x0 )} = (x x0 )

and quantization takes

(x) (x),

b (x) b (x), 1 1

relations

[(x),

b b (x0 )] = (x x0 )1

Pretty much exactly the same formalism works to describe fermions, with

the same H1 and the same choice of bases. The only difference is that the coor-

dinate functions are now taken to be anti-commuting, satisfying the fermionic

Poisson bracket relations of a super Lie algebra rather than a Lie algebra. After

quantization, the fields (x),

b b (x) satisfy anticommutation relations and gen-

erate an infinite-dimensional Clifford algebra, rather than the Weyl algebra of

the bosonic case.

This material is discussed in essentially any quantum field theory textbook.

Many do not explicitly discuss the non-relativistic case, two that do are [21]

and [29]. Two books aimed at mathematicians that cover the subject much

more carefully than those for physicists are [19] and [10].

A good source for learning about quantum field theory from the point of view

of non-relativistic many-body theory is Feynmans lecture notes on statistical

mechanics[17].

261

262

Chapter 26

for Non-relativistic

Quantum Fields

In our study of the harmonic oscillator we found that the symmetries of the sys-

tem could be studied using quadratic functions on the phase space. Classically

these gave a Lie algebra under the Poisson bracket, and quantization provided a

unitary representation 0 of the Lie algebra, with quadratic functions becoming

quadratic operators. In the case of quantum fields, the same pattern holds, with

the phase space now an infinite dimensional space, the single particle Hilbert

space. Certain specific quadratic functions of the fields will provide a Lie al-

gebra under the Poisson bracket, with quantization then providing a unitary

representation of the Lie algebra in terms of quadratic field operators.

Well begin with the most fundamental symmetry in quantum mechanics,

time translation symmetry, which determines the dynamics of the theory. For

the case of a free particle, the field theory Hamiltonian will be a quadratic

function of the fields, providing a basic example of how such functions generate

a unitary representation on the states of the quantum theory.

In classical Hamiltonian mechanics, the Hamiltonian function h determines how

an observable f evolves in time by the differential equation

d

f = {f, h}

dt

Multiplying this by i gives a skew-adjoint operator that exponentiates (well

263

assume here H time-independent) to the unitary operator

U (t) = eiHt

that determines how states (in the Schrodinger picture) evolve under time trans-

lation. In the Heisenberg picture states stay the same and operators evolve, with

their time evolution given by

d b

f = [fb, iH]

dt

which is the quantization of the classical dynamical equation. This has the

solution

ally easier to work with the Heisenberg picture (in which the time dependence

is in the quantum field operators) than the Schrodinger picture (in which the

time dependence is in the states). This is especially true in relativistic systems

where one wants to as much as possible treat space and time on the same foot-

ing. It is however also true in non-relativistic cases due to the complexity of

the description of the states (inherent since one is trying to describe arbitrary

numbers of particles) versus the description of the operators, which are built

simply out of the quantum fields.

The classical phase space to be quantized is the space H1 of solutions of

the free particle Schrodinger equation, parametrized by the initial data of a

complex-valued wavefunction (x, 0) (x), with Poisson bracket

{(x), i(x0 )} = (x x0 )

Time translation on this space is given by the Schrodinger equation, which says

that wavefunctions will evolve with time dependence given by

i 2

(x, t) = (x, t)

t 2m x2

+

1 2

Z

h= (x) (x)dx

2m x2

then we will get the single-particle Schrodinger equation from the Hamiltonian

264

dynamics, since

t

Z +

1 2

={(x, t), (x0 , t) 02

(x0 , t)dx0 }

2m x

1 + 2

Z

= ({(x, t), (x0 , t)} 0 2 (x0 , t)+

2m x

2

(x0 , t){(x, t), 0 2 (x0 , t)})dx0 )

x

1 + 2

2

Z

= (i(x x0 ) 0 2 (x0 , t) + (x0 , t) 0 2 {(x, t), (x0 , t)})dx0 )

2m x x

i 2

= (x, t)

2m x2

Here we have used the derivation property of the Poisson bracket and the lin-

2

earity of the operator x 02 .

Note that there are other forms of the same Hamiltonian function, related

to the one we chose by integration by parts. One has

d2 d d d

(x) 2

(x) = ((x) (x)) | (x)|2

dx dx dx dx

d d d d2

= ((x) (x) ( (x))(x)) + ( 2 (x))(x)

dx dx dx dx

so neglecting integrals of derivatives (assuming boundary terms go to zero at

infinity), one could have used

Z +

1 + d2

Z

1 d

h= | (x)|2 dx or h = ( (x))(x)dx

2m dx 2m dx2

Instead of working with position space fields (x, t) we could work with

their momentum space components. Recall that we can write solutions to the

Schrodinger equation as

Z

1 p2

(x, t) = (p, t)eipx ei 2m t dp

2

where

p2

(p, t) = (p)ei 2m t

Using these as our coordinates on the H1 , dynamics is given by

p2

(p, t) = {(p, t), h} = i (p, t)

t 2m

and one can easily see that one can choose

Z 2

p

h= |(p)|2 dp

2m

265

as Hamiltonian function in momentum space coordinates. This is the same

expression one would get by substituting the expression for in terms of and

calculating h from its formula as a quadratic polynomial in the fields.

In momentum space, quantization is simply given by

Z 2

p

(p) a(p), h H b = a (p)a(p)dp

2m

In position space the expression for the Hamiltonian operator (again normal-

ordered) will be:

Z +

1 2 b

Hb = b (x) (x)dx

2m x2

Using this quantized form, essentially the same calculation as before (now with

operators and commutators instead of functions and Poisson brackets) shows

that the quantum field dynamical equation

d b

(x, t) = i[(x,

b t), H]

b

dt

becomes

b i 2 b

(x, t) = (x, t)

t 2m x2

The field operator (x,

b t) satisfies the Schrodinger equation which now ap-

pears as a differential equation for operators rather than for wavefunctions.

One can explicitly solve such a differential equation just as for wavefunctions,

by Fourier transforming and turning differentiation into multiplication. If the

operator (x,

b t) is related to the operator a(p, t) by

Z

b t) = 1

(x, eipx a(p, t)dp

2

then the Schr

odinger equation for the a(p, t) will be

ip2

a(p, t) = a(p, t)

t 2m

with solution

p2

a(p, t) = ei 2m t a(p, 0)

The solution for the field will then be

Z

b t) = 1

(x,

p2

eipx ei 2m t a(p)dp

2

where the operators a(p) a(p, 0) are the initial values.

We will not enter into the important topic of how to compute observables

in quantum field theory that can be connected to experimentally important

quantities such as scattering cross-sections. A crucial role in such calculations

is played by the following observables:

266

Definition (Greens function or propagator). The Greens function or propa-

gator for a quantum field theory is the amplitude, for t > t0

G(x, t, x0 , t0 ) = h0|(x,

b t)b (x0 , t0 )|0i

The physical interpretation of these functions is that they describe the am-

plitude for a process in which a one-particle state localized at x is created at time

t0 , propagates for a time t t0 , and then its wave-function is compared to that

of a one-particle state localized at x. Using the solution for the time-dependent

field operator given earlier we find

Z + Z +

1 p2 0 p0 2 0

G(x, t, x0 , t0 ) = h0|eipx ei 2m t a(p)eip x ei 2m t a (p0 )|0idpdp0

2

Z + Z +

p2 0 0 p0 2 0

= eipx ei 2m t eip x ei 2m t (p p0 )dpdp0

Z+

0 p2 0

= eip(xx ) ei 2m (tt ) dp

One can evaluate this integral, finding

im 3 im

(x0 x)2

G(x0 , t0 , x, t) = ( 0

) 2 e 2(t0 t)

2(t t)

and that

lim G(x0 , t0 , x, t) = (x0 x)

tt0

While we have worked purely in the Hamiltonian formalism, one might won-

der what the Lagrangian for this system is. A Lagrangian that will give the

Schr

odinger equation as an Euler-Lagrange equation is

1 2

h = i +

L = i

t t 2m x2

or, using integration by parts to get an alternate form of h mentioned earlier

1 2

L = i | |

t 2m x

If one tries to define a canonical momentum for as L

on just gets i.

This justifies the Poisson bracket relation

{(x), i(x0 )} = (x x0 )

but, as expected for a case where the equation of motion is first-order in time,

such a canonical momentum is not independent of and the space of the wave-

functions is already a phase space. One could try and quantize this system

by path integral methods, for instance computing the propagator by doing the

integral Z

i

D e ~ S[]

over paths from (x, t) to (x0 , t0 ). However one needs to keep in mind the

warnings given earlier about path integrals over phase space, since that is what

one has here.

267

26.2 Spatial symmetries

The specification of the Hamiltonian gives the dynamics of the theory and de-

termines how states and operators transform under time translation. The action

of translations and rotations of the physical space R3 on a quantum field theory

is of a different nature. It is determined from the fact that the phase space H1

is a space of functions on this R3 , so any group action

xgx

(x) (g 1 x)

When this group action preserves the Poisson bracket on H1 , we expect to get

a unitary representation of the group on the quantum field theory state space,

as well as an action on the field operators by conjugation.

This will be the case for the group R3 of spatial translations and the group

SO(3) of spatial rotations, which (since rotations act non-trivially on transla-

tions) fit together into a group E(3) which is a semi-direct product R3 o SO(3),

something that we will study in detail in a later chapter. Since the action of

E(3) preserves Poisson brackets, it will have a unitary representation on the

quantum field theory, which we will study in this section.

If the action of E(3) also commutes with the group action on H1 generated

by the Hamiltonian h, then we will not only have a representation of E(3) on

the state space, but energy eigenstates of each energy will give representations

of E(3). This will be the case for the free particle, since it has a Hamiltonian h

that is invariant under rotations and translations. If one takes not free particles,

but particles in an external potential V , then one will still have a representation

of E(3) on the state space, but only the subgroup of E(3) that preserves the

potential (for instance, SO(3) for a central potential) will preserve the energy

eigenspaces.

For reasons that we will see later when we discuss special relativity, in the

single-particle theory the momentum operators P that generate space transla-

tions are taken to act oppositely to the Hamiltonian operator H (iH = t , and

iP = x ). For time translations we found that the action on field operators

was infinitesimally

d b b t), iH]

(x, t) = [(x, b

dt

which exponentiates to

b t)eia0 Hb

b t + a0 ) = eia0 Hb (x,

(x,

momentum operators P b for the infinitesimal action of such translations, but

because of the sign change we will have

b + a, t) = eiaP

(x (x, t)eiaP

b b b

268

The momentum operator P b in the quantum field theory can be constructed in

terms of quadratic operators in the fields in the same way as the Hamiltonian,

just replacing the single-particle Hamiltonian operator by the single-particle

momentum operator P = i. So one has

Z

P=

b b (x)(i)(x)d

b 3

x

R3

In the last chapter we saw that, in terms of annihilation and creation operators,

this operator is just Z

P=

b p a (p)a(p)d3 p

R3

which is just the integral over momentum space of the momentum times the

number-density operator in momentum space.

For spatial rotations, we found in chapter 14 that these had generators the

angular momentum operators

L = X P = X (i)

momentum operators in the quantum field theory as quadratic field operators

by Z

L=

b b (x)(x i)(x)d

b 3

x

R3

These will generate the action of rotations on the field operators. For instance,

if R() is a rotation about the x3 axis by angle , we will have

b b b b

(check sign).

Since our phase space H1 is a space of complex functions, perhaps the simplest

group that acts on this space is the group U (1) of phase transformations. Such

a group action that acts trivially on the spatial coordinates x, but non-trivially

on the values of the function (x) is called an internal symmetry. If our

functions are complex valued, only U (1) can act unitarily on the value, but

we will see later that if we introduce functions that take values in Cm , we will

have an action of the larger group U (m).

In an early chapter we saw that the fact that irreducible representations of

U (1) are labeled by integers is what is responsible for the term quantization:

since quantum states are representations of this group, they break up into states

269

characterized by integers, with these integers counting the number of quanta.

In the non-relativistic quantum field theory, the integer will just be the total

particle number.

On fields the U (1) group action is given by

to recall the simple case of a harmonic oscillator in one variable, identifying the

phase space R2 with C so the coordinates are z, z, with a U (1) action

z ei z, z ei z

{z, iz} = 1

which implies

{zz, z} = iz, {zz, z} = iz

So, it is zz that infinitesimally generates this U (1) action on phase space.

Quantizing takes z a, z a and we need to multiply by i to get a

skew-adjoint operator that will generate the U (1) representation. We have two

choices here:

i

zz (a a + aa )

2

This will have eigenvalues i(n + 21 ), n = 0, 1, 2 . . . .

zz ia a

This is the normal-ordered form, with eigenvalues in.

With either choice, we get a number operator

1

N= (a a + aa ), or N = a a

2

In both cases we have

[N, a] = a, [N, a ] = a

so

eiN aeiN = ei a, eiN a eiN = ei a

Either choice of N will give the same action on operators. Hoever, on states

only the normal-ordered one will have the desirable feature that

and since we now want to treat fields, adding together an infinite number of

such 2d phase spaces, we will need the normal-ordered version in order to not

get 21 as the particle number eigenvalue for the vacuum state.

270

In momentum space, we simply do the above for each value of p and sum,

getting Z +

N=

b a (p)a(p)dp

where one needs to keep in mind that a (p)a(p) is really an operator valued

distribution, which must be integrated against something (here the identity

function) to get an operator.

Instead of working with a(p), the quantization of the Fourier transform of

(x), one could work with (x) directly, do the above for each value of x and

sum, getting Z +

N

b= b (x)(x)dx

b

b . b (x)(x)

with the Fourier transform relating the two formulas for N b is also an

operator valued distribution, with the interpretation of measuring the number

density at x.

On field operators, Nb satisfies

[N b = ,

b , ] b , b ] = b

b [N

states by increasing the eigenvalue of N by one. Exponentiating, one has

b b i N

eiN e = ei ,

b eiNb b eiNb = ei b

b

which are the quantum version of the U (1) symmetry on the phase space we

began our discussion with.

An important property of N b that can be straightforwardly checked is that

Z +

1 2 b

[N , H] = [N ,

b b b b (x) (x)dx] = 0

2m x2

This implies that particle number is a conserved quantity: if we start out with

a state with a definite particle number, this will remain constant. Note that the

origin of this conservation law comes from the fact that N b is the quantized gen-

erator of the U (1) symmetry of phase transformations on complex-valued fields

. If we start with any hamiltonian function h on H1 that is invariant under the

U (1) (i.e. built out of terms with an equal number of s and s), then for such

a theory N b will commute with H b and particle number will be conserved. Note

though that one needs to take some care with arguments like this, which assume

that symmetries of the classical phase space give rise to unitary representations

in the quantum theory. The need to normal-order operator products, working

with operators that differ from the most straightforward quantization by an in-

finite constant, can cause a failure of symmetries to be realized as expected in

the quantum theory, a phenomenon known as an anomaly in the symmetry.

In quantum field theories, due to the infinite number of degrees of freedom,

the Stone-von Neumann theorem does not apply, and one can have unitarily

271

inequivalent representations of the algebra generated by the field operators,

leading to new kinds of behavior not seen in finite dimensional quantum systems.

In particular, one can have a space of states where the lowest energy state |0i

does not have the property

b |0i = 0, eiNb |0i = |0i

N

b to some other state, with

N 6 0, eiN |0i |i =

b |0i = 6 |0i (for 6= 0)

b

In this case, the vacuum state is not an eigenstate of N b so does not have a

well-defined particle number. If [N , H] = 0, the states |i will all have the same

b b

energy as |0i and there will be a multiplicity of different vacuum states, labeled

by . In such a case the U (1) symmetry is said to be spontaneously broken.

This phenomenon occurs when non-relativistic quantum field theory is used to

describe a superconductor. There the ground state will be a state without a

definite particle number, with electrons pairing up in a way that allows them to

lower their energy, condensing in the lowest energy state.

By taking fields with values in Cm , or, equivalently, m different species of

complex-valued field b , b = 1, 2, . . . , m, one can easily construct quantum field

theories with larger internal symmetry groups than U (1). Taking as Hamilto-

nian function Z + X m

1 2

h= b (x) b (x)dx

2m x2

b=1

one can see that this will be invariant not just under U (1) phase transformations,

but also under transformations

1 1

2 2

.. U ..

. .

m m

As in the U (1) case, one can begin by considering the case of one particular

value of k or of x, for which the phase space is Cm , with coordinates zb , z b .

One can check that the m2 quadratic combinations zb z c for b = 1, . . . , m, c =

1, . . . , m will generalize the role played by zz in the m = 1 case, with their

Poisson bracket relations exactly the Lie bracket relations of the Lie algebra

u(m).

272

After quantization, these quadratic combinations become quadratic combi-

nations in annihilation and creation operators ab , ab satisfying

[ab , ac ] = bc

m m m

ab Xbc ac , ab Ybc ac ] = ab [X, Y ]bc ac

X X X

[

b,c=1 b,c=1 b,c=1

So, for each X u(m) quantization will give us a representation of u(m) where

X acts as the operator

m

ab Xbc ac

X

b,c=1

As in the U (1) case, one gets an operator in the quantum field theory just by

summing over either the a(k) in momentum space, or the fields in configuration

space, finding for each X u(m) an operator

Z + m

bb (x)Xbc bc (x)dx

X

X

b=

b,c=1

that provides a representaion of u(m) on the quantum field theory state space.

When, as for the free-particle h we chose, the Hamiltonian is invariant under

U (m) transformations of the fields b , then we will have

[X,

b H]

b

In this case, if |0i is invariant under the U (m) symmetry, then energy eigenstates

of the quantum field theory will break up into irreducible representations of

U (m) and can be labeled accordingly. As in the U (1) case, the U (m) symmetry

may be spontaneously broken, with

X|0i

b 6 0

=

for some directions X in u(m). When this happens, just as in the U (1) case

states did not have well-defined particle number, now they will not carry well-

defined irreducible U (m) representation labels.

26.4 Fermions

Everything that was done in this chapter carries over straightforwardly to the

fermionic case. Our field operators now generate an infinite-dimensional Clifford

algebra and the quantum state space is now an infinite-dimensional version of

spinors. All the symmetries considered here have Lie algebra representations

constructed using quadratic combinations of the field operators in just the same

273

way as in the bosonic case. The method of annihilation and creation operators

used to take matrices with given commutation relations and produce quadratic

field operators with the same commutation relations works in the same way,

since if one checks the proof of this in chapter 17 one finds it continues to work

when one uses the CAR relations instead of the CCR relations.

The material of this chapter is often developed in conventional quantum field

theory texts in the context of relativistic rather than non-relativistic quantum

field theory. Symmetry generators are also more often derived via Lagrangian

methods (Noethers theorem) rather than the Hamiltonian methods used here.

For an example of a detailed discussion relatively close to this one, see [21].

274

Chapter 27

Lorentz Group

For the case of non-relativistic quantum mechanics, we saw that systems with

an arbitrary number of particles, bosons or fermions, could be described by

taking as the Hamiltonian phase space the state space H1 of the single-particle

quantum theory (e.g. the space of complex-valued wave-functions on R3 in the

bosonic case). This phase space is infinite-dimensional, but it is linear and it

can be quantized using the same techniques that work for the finite-dimensional

harmonic oscillator. This is an example of a quantum field theory since it is a

space of functions (fields, to physicists) that is being quantized.

We would like to find some similar way to proceed for the case of rela-

tivistic systems, finding relativistic quantum field theories capable of describ-

ing arbitrary numbers of particles, with the energy-momentum relationship

E 2 = |p|2 c2 + m2 c4 characteristic of special relativity, not the non-relativistic

2

limit |p| mc where E = |p| 2m . In general, a phase space can be thought of as

the space of initial conditions for an equation of motion, or equivalently, as the

space of solutions of the equation of motion. In the non-relativistic field theory,

the equation of motion is the first-order in time Schrodinger equation, and the

phase space is the space of fields (wave-functions) at a specified initial time,

say t = 0. This space carries a representation of the time-translation group R,

the space-translation group R3 and the rotation group SO(3). To construct a

relativistic quantum field theory, we want to find an analog of this space. It will

be some sort of linear space of functions satisfying an equation of motion, and

we will then quantize by applying harmonic oscillator methods.

Just as in the non-relativistic case, the space of solutions to the equation

of motion provides a representation of the group of space-time symmetries of

the theory. This group will now be the Poincare group, a ten-dimensional

group which includes a four-dimensional subgroup of translations in space-time,

and a six-dimensional subgroup (the Lorentz group), which combines spatial

rotations and boosts (transformations mixing spatial and time coordinates).

275

The representation of the Poincare group on the solutions to the relativistic

wave equation will in general be reducible. Irreducible such representations will

be the objects corresponding to elementary particles. Our first goal will be to

understand the Lorentz group, in later sections we will find representations of

this group, then move on to the Poincare group and its representations.

Special relativity is based on the principle that one should consider consider

space and time together, and take them to be a four-dimensional space R4 with

an indefinite inner product:

Definition (Minkowski space). Minkowski space M 4 is the vector space R4

with an indefinite inner product given by

(x, y) x y = x0 y0 x1 y1 x2 y2 x3 y3

nates of y R4 .

This inner product will also sometimes be written using the matrix

1 0 0 0

0 1 0 0

= 0 0 1 0

0 0 0 1

as

3

X

xy = x x

,=0

The choice of overall sign of is a convention, with the choice we are making

the most common one in discussions of relativistic quantum field theories.

Digression (Upper and lower indices). In many physics texts it is conventional

in discussions of special relativity to write formulas using both upper and lower

indices, related by

X3

x = x = x

=0

with the last form of this using the Einstein summation convention.

One motivation for introducing both upper and lower indices is that special

relativity is a limiting case of general relativity, which is a fully geometrical

theory based on taking space-time to be a manifold M with a metric g that

varies from point to point. In such a theory it is important to distinguish between

elements of the tangent space Tx (M ) at a point x M and elements of its dual,

the co-tangent space Tx (M ), while using the fact that the metric g provides an

inner product on Tx (M ) and thus an isomorphism Tx (M ) ' Tx (M ). In the

276

special relativity case, this distinction between Tx (M ) and Tx (M ) just comes

down to an issue of signs, but the upper and lower index notation is useful for

keeping track of those.

A second motivation is that position and momenta naturally live in dual

vector spaces, so one would like to distinguish between the vector space M 4 of

positions and the dual vector space of momenta. In the case though of a vector

space like M 4 which comes with a fixed inner product , this inner product

gives a fixed identification of M 4 and its dual, an identification that is also an

identification as representations of the Lorentz group. For simplicity, we will

not here try and distinguish by notation whether a vector is in M 4 or its dual,

so will just use lower indices, not both upper and lower indices.

coordinate x0 is a time coordinate, related to the conventional time coordinate

t with respect to chosen units of time and distance by x0 = ct where c is the

speed of light. Mostly we will assume units of time and distance have been

chosen so that c = 1.

Vectors v M 4 such that |v|2 = v v < 0 are called spacelike, those with

|v| > 0 time-like and those with |v|2 = 0 are said to lie on the light-cone.

2

space looks like this:

We can take Fourier transforms with respect to the four space-time variables,

which will take functions of x0 , x1 , x2 , x3 to functions of the Fourier transform

variables p0 , p1 , p2 , p3 . The definition we will use for this Fourier transform will

277

be

Z

1

fe(p) = eipx f (x)d4 x

(2)2 M 4

Z

1

= ei(p0 x0 p1 x1 p2 x2 p3 x3 ) f (x)dx0 d3 x

(2)2 M 4

and the Fourier inversion formula is

Z

1

f (x) = eipx fe(p)d4 p

(2)2 M4

Note that our definition puts one factor of 12 with each Fourier (or inverse

Fourier) transform with respect to a single variable. A common alternate con-

vention among physicists is to put all factors of 2 with the p integrals (and

thus in the inverse Fourier transform), none in the definition of fe(p), the Fourier

transform itself.

The reason why one conventionally defines the Hamiltonian operator as i t

but the momentum operator with components i x j

is due to the sign change

between the time and space variables that occurs in this Fourier transform in

the exponent of the exponential.

Recall that in 3 dimensions the group of linear transformations of R3 pre-

serving the standard inner product was the group O(3) of 3 by 3 orthogonal

matrices. This group has two disconnected components: SO(3), the subgroup

of orientation preserving (determinant +1) transformations, and a component

of orientation reversing (determinant 1) transformations. In Minkowksi space,

one has

Definition (Lorentz group). The Lorentz group O(1, 3) is the group of linear

transformations preserving the Minkowski space inner product on R4 .

In terms of matrices, the condition for a 4 by 4 matrix to be in O(1, 3)

will be

1 0 0 0 1 0 0 0

0 1 0 0 0 1 0 0

T

0 0 1 0 = 0 0 1 0

0 0 0 1 0 0 0 1

The Lorentz group has four components, with the component of the iden-

tity a subgroup called SO(1, 3) (which some call SO+ (1, 3)). The other three

components arise by multiplication of elements in SO(1, 3) by P, T, P T where

1 0 0 0

0 1 0 0

P = 0 0 1 0

0 0 0 1

278

is called the parity transformation, reversing the orientation of the spatial

variables, and

1 0 0 0

0 1 0 0

T =0

0 1 0

0 0 0 1

reverses the time orientation.

The Lorentz group has a subgroup SO(3) of transformations that just act

on the spatial components, given by matrices of the form

1 0 0 0

0

=0

0

where is in SO(3). For each pair j, k of spatial directions one has the usual

SO(2) subgroup of rotations in the j k plane, but now in addition for each pair

0, j of the time direction with a spatial direction, one has SO(1, 1) subgroups

of matrices of transformations called boosts in the j direction. For example,

for j = 1, one has the subgroup of SO(1, 3) of matrices of the form

cosh sinh 0 0

sinh cosh 0 0

=

0

0 1 0

0 0 0 1

for R.

The Lorentz group is six-dimensional. For a basis of its Lie algebra one can

take six matrices J for , 0, 1, 2, 3 and j < k. For the spatial indices, these

are

0 0 0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0 1 0 0 0 0

J12 = 0 1 0 0 , J13 = 0 0 0 0 , J23 = 0 0 0 1

0 0 0 0 0 1 0 0 0 0 1 0

which correspond to the basis elements of the Lie algebra of SO(3) that we saw

in an earlier chapter. One can rename these as

279

Taking the first index 0, one gets three elements corresponding to infinitesi-

mal boosts in the three spatial directions

0 1 0 0 0 0 1 0 0 0 0 1

1 0 0 0 0 0 0 0 0 0 0 0

J01 = , J = , J =

0 02 1 0 03 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 1 0 0 0

One can easily calculate the commutation relations between the Kj and Jj ,

which show that the Kj transform as a vector under infinitesimal rotations. For

instance, for infinitesimal rotations about the x1 axis, one finds

1 1

Aj = (Jj + iKj ), Bj = (Jj iKj )

2 2

one finds

[A1 , A2 ] = A3 , [A3 , A1 ] = A2 , [A2 , A3 ] = A1

and

[B1 , B2 ] = B3 , [B3 , B1 ] = B2 , [B2 , B3 ] = B1

linear combinations of basis elements) the Lie algebra so(1, 3) of SO(1, 3) and

work with the complex Lie algebra so(3, 1) C. It shows that this Lie algebra

splits into a product of two sub Lie algebras, which are each copies of the

(complexified) Lie algebra of SO(3), so(3) C. Since

we have

so(1, 3) C = sl(2, C) sl(2, C)

In the next section well see the origin of this phenomenon at the group level.

280

27.3 Spin and the Lorentz group

Just as the groups SO(n) have double covers Spin(n), the group SO(1, 3) has a

double cover, which we will show can be identified with the group SL(2, C) of

2 by 2 complex matrices with unit determinant. This group will have the same

Lie algebra as the SO(1, 3), and we will sometimes refer to either group as the

Lorentz group.

Recall that for SO(3) the spin double cover Spin(3) can be identified with

either Sp(1) (the unit quaternions) or SU (2), and then the action of Spin(3)

as SO(3) rotations of R3 was given by conjugation of imaginary quaternions or

certain 2 by 2 complex matrices respectively. In the SU (2) case this was done

explicitly by identifying

x3 x1 ix2

(x1 , x2 , x3 )

x1 + ix2 x3

and then showing that conjugating this matrix by an element of SU (2) was a

linear map leaving invariant

x3 x1 ix2

det = x21 + x22 + x23

x1 + ix2 x3

The same sort of thing works for the Lorentz group case. Now we identify

R4 with the space of 2 by 2 complex self-adjoint matrices by

x0 + x3 x1 ix2

(x0 , x1 , x2 , x3 )

x1 + ix2 x0 x3

x0 + x3 x1 ix2

det = x20 x21 x22 x23

x1 + ix2 x0 x3

This provides a very useful way to think of Minkowski space: as complex self-

adjoint 2 by 2 matrices, with norm-squared the determinant of the matrix.

The linear transformation

x0 + x3 x1 ix2 x0 + x3 x1 ix2

x1 + ix2 x0 x3 x1 + ix2 x0 x3

for

x0 + x3 x1 ix2 det x0 + x3 x1 ix2 )

det( ) =(det ) (det

x1 + ix2 x0 x3 x1 + ix2 x0 x3

=x20 x21 x22 x23

281

It also takes self-adjoint matrices to self-adjoints, and thus R4 to R4 , since

x0 + x3 x1 ix2 x0 + x3 x1 ix2

( ) =( )

x1 + ix2 x0 x3 x1 + ix2 x0 x3

= x0 + x3 x1 ix2

x1 + ix2 x0 x3

Note that both and give the same linear transformation when they act by

conjugation like this. One can show that all elements of SO(3, 1) arise as such

conjugation maps, by finding appropriate that give rotations or boosts in the

planes, since these generate the group.

Recall that the double covering map

: SU (2) SO(3)

SO(3)

x3 x1 ix2 x3 x1 ix2

1

x1 + ix2 x3 x1 + ix2 x3

We have found an extension of this map to a double covering map from SL(2, C)

to SO(1, 3). This restricts to on the subgroup SU (2) of SL(2, C) matrices

=

satisfying 1 .

Digression (The complex group Spin(4, C) and its real forms). Recall that

we found that Spin(4) = Sp(1) Sp(1), with the corresponding SO(4) trans-

formation given by identifying R4 with the quaternions H and taking not just

conjugations by unit quaternions, but both left and right multiplication by dis-

tinct unit quaternions. Rewriting this in terms of complex matrices instead of

quaternions, we have Spin(4) = SU (2) SU (2), and a pair 1 , 2 of SU (2)

matrices acts as an SO(4) rotation by

x0 ix3 x2 ix1 x0 ix3 x2 ix1

1 2

x2 ix1 x0 + ix3 x2 ix1 x0 + ix3

For another example, consider the identification of R4 with 2 by 2 real ma-

trices given by

x0 + x3 x2 + x1

(x0 , x1 , x2 , x3 )

x2 x1 x0 x3

Given a pair of matrices 1 , 2 in SL(2, R), the linear transformation

x0 + x3 x2 + x1 x0 + x3 x2 + x1

1 2

x2 x1 x0 x3 x2 x1 x0 x3

preserves the reality condition on the matrix, and preserves

x + x3 x2 + x1

det 0 = x20 + x21 x22 x23

x2 x1 x0 x3

282

so gives an element of SO(2, 2) and we see that Spin(2, 2) = SL(2, R)

SL(2, R).

The three different examples

and

Spin(2, 2) = SL(2, R) SL(2, R)

that we have seen are all so-called real forms of a fact about complex groups

that one can get by complexifying any of the examples, i.e. considering elements

(x0 , x1 , x2 , x3 ) C4 , not just in R4 . For instance, in the Spin(4) case, taking

the x0 , x1 , x2 , x3 in the matrix

x0 ix3 x2 ix1

x2 ix1 x0 + ix3

matrices, and the transformation

z0 iz3 z2 iz1 z iz3 z2 iz1

1 0 2

z2 iz1 z0 + iz3 z2 iz1 z0 + iz3

preserves this space as well as the determinant (z02 + z12 + z22 + z32 ) for 1 and 2

not just in SU (2), but in the larger group SL(2, C). So we find that the group

SO(4, C) of complex orthogonal transformations of C4 has spin double cover

and one gets from this the three different real forms described above by picking

various choices of real subspaces R4 C4 .

Since spin(4, C) = so(1, 3) C, this relation between complex Lie groups

corresponds to the Lie algebra relation

of generators Jj and Kj of so(1, 3) we could find generators Aj and Bj of two

different sl(2, C) sub-algebras.

Those not familiar with special relativity should consult a textbook on the

subject for the physics background necessary to appreciate the significance of

Minkowski space and its Lorentz group of invariances. An example of a suitable

such book aimed at mathematics students is Woodhouses Special Relativity[55].

Most quantum field theory textbook have some sort of discussion of the

Lorentz group and its Lie algebra, although the issue of its complexification is

283

often not treated. A typical example is Peskin-Schroeder[33], see the beginning

of Chapter 3. Another example is Quantum Field Theory in a Nutshell by

Tony Zee, see Chapter II.3 [56] (and test your understanding by interpreting

properly some of the statements included there such as The mathematically

sophisticated say the algebra SO(3, 1) is isomorphic to SU (2) SU (2)).

284

Chapter 28

Representations of the

Lorentz Group

sentations of the rotation group SO(3) and its double cover Spin(3) = SU (2)

one would like to also understand the representations of the Lorentz group. Well

consider this question for the double cover SL(2, C). As in the three-dimensional

case, only some of these will also be representations of SO(1, 3). One difference

from the SO(3) case is that these will be non-unitary representations, so do not

by themselves provide physically sensible state spaces. All finite-dimensional

irreducible representations of the Lorentz group are non-unitary (except for the

trivial representation). The Lorentz group does have unitary irreducible repre-

sentations, but these are infinite-dimensional and a topic we will not cover.

Having seen the importance in quantum mechanics of understanding the repre-

sentations of the rotation group SO(3) and its double cover Spin(3) = SU (2)

one would like to also understand the representations of the Lorentz group.

Well consider this question for the double cover SL(2, C). As in the three-

dimensional case, only some of these will also be representations of SO(1, 3).

In the SU (2) case we found irreducible unitary representations (n , V n ) of

dimension n + 1 for n = 0, 1, 2, . . .. These could also be labeled by s = n2 .

called the spin of the representation, and we will do that from now on. These

representations can be realized explicitly as homogeneous polynomials of degree

n = 2s in two complex variables z1 , z2 . For the case of Spin(4) = SU (2)SU (2),

the irreducible representation will just be tensor products

V s1 V s2

of SU (2) irreducibles, with the first SU (2) acting on the first factor, the second

on the second factor. The case s1 = s2 = 0 is the trivial representation, s1 =

285

1

2 , s2= 0 is one of the half-spinor representations of Spin(4) on C2 , s1 = 0, s2 =

1

2 is the other, and s1 = s2 = 21 is the representation on four-dimensional

(complexified) vectors.

Turning now to Spin(1, 3) = SL(2, C), can take the SU (2) matrices acting

on homogeneous polynomials to instead be SL(2, C) matrices and still have

irreducible representations of dimension 2s + 1 for s = 0, 12 , 1, . . .. These will

now be representations (s , V s ) of SL(2, C). There are several things that are

different though about these representations:

They are not unitary (except in the case of the trivial representation).

1

For example, for the defining representation V 2 on C2 and the Hermitian

inner product < , >

0

10

1 1

< , >= 1 2 = 1 10 + 2 20

2 20 20

0 0

1 1 1

< , >= 1 2

2 20 20

The condition that matrices SL(2, C) do satisfy is that they have

determinant 1. It turns out that having determinant one is equivalent

to the condition that preserves the anti-symmetric bilinear form on C2 ,

i.e.

T 0 1 0 1

=

1 0 1 0

To see this, take

=

and calculate

0 1 0 0 1

T = = det

1 0 0 1 0

1

As a result, on representations V 2 of SL(2, C) we do have a non-degenerate

bilinear form

0 0

1 0 1 1

( 1 , ) 1 2 = 1 20 2 10

2 20 1 0 20

1

that is invariant under the SL(2, C) action on V 2 and can be used to

identify the representation and its dual.

Such a non-degenerate bilinear form is called a symplectic form, and we

have already made extensive use of these as a fundamental structure that

286

occurs on a Hamiltonian phase space. For the simplest case of the phase

space R2 for one degree of freedom, such a form is given with respect to

a basis as the same matrix

0 1

=

1 0

of SL(2, C). In the phase space case, everything was real, and the invari-

ance group of was the real symplectic group Sp(2, R) = SL(2, R). What

occurs here is just the complexification of this story, with the symplectic

form now on C2 , and the invariance group now SL(2, C).

In the case of SU (2) representations, the complex conjugate representa-

tion one gets by taking as representation matrices (g) instead of (g) is

equivalent to the original representation (the same representation, with a

different basis choice, so matrices changed by a conjugation). To see this

for the spin- 12 representation, note that SU (2) matrices are of the form

=

and one has

1

0 1 0 1

=

1 0 1 0

so the matrix

0 1

1 0

is the change of basis matrix relating the representation and its complex

conjugate.

This is no longer true for SL(2, C). One cannot complex conjugate arbi-

trary 2 by 2 complex matrices of unit determinant by a change of basis,

and representations will not be equivalent to their complex conjugates.

The classification of irreducible finite dimensional SU (2) representation was

done earlier in this course by considering its Lie algebra su(2), complexified to

give us raising and lowering operators, and this complexification is sl(2, C). If

you take a look at that argument, you see that it mostly also applies to irre-

ducible finite-dimensional sl(2, C) representations. There is a difference though:

now flipping positive to negative weights (which corresponds to change of sign

of the Lie algebra representation matrices, or conjugation of the Lie group rep-

resentation matrices) no longer takes one to an equivalent representation. It

turns out that to get all irreducibles, one must take both the representations

we already know about and their complex conjugates. Using the fact that the

tensor product of one of each type of irreducible is still an irreducible, one can

show (we wont do this here) that the complete list of irreducible representations

of sl(2, C) is given by

287

Theorem (Classification of finite dimensional sl(2, C) representations). The

irreducible representations of sl(2, C) are labeled by (s1 , s2 ) for sj = 0, 21 , 1, . . ..

These representations are built out of the representations (s , V s ) with the ir-

reducible (s1 , s2 ) given by

(s1 s2 , V s1 V s2 )

and having dimension (2s1 + 1)(2s2 + 1).

All these representations are also representations of the group SL(2, C) and

one has the same classification theorem for the group, although we will not try

and prove this. We will also not try and study these representations in general,

but will restrict attention to the four cases of most physical interest.

(0, 0): The trivial representation on C, also called the spin 0 or scalar

representation.

( 12 , 0): These are called left-handed (for reasons we will see later on) Weyl

spinors. We will often denote the representation space C2 in this case as

SL , and write an element of it as L .

(0, 12 ): These are called right-handed Weyl spinors. We will often denote

the representation space C2 in this case as SR , and write an element of it

as R .

( 21 , 12 ): This is called the vector representation since it is the complexifi-

cation of the action of SL(2, C) as SO(1, 3) transformations of space-time

vectors that we saw earlier. Recall that for SL(2, C) this action was

x0 + x3 x1 ix2 x0 + x3 x1 ix2

x1 + ix2 x0 x3 x1 + ix2 x0 x3

representation SL SR . This representation is on a vector space C4 =

M (2, C), but preserves the subspace of self-adjoint matrices that we have

identified with the Minkowski space R4 .

The reducible 4 complex dimensional representation ( 21 , 0) (0, 12 ) is known as

the representation on Dirac spinors. As explained earlier, of these representa-

tions, only the trivial one is unitary. Only the trivial and vector representations

are representations of SO(1, 3) as well as SL(2, C).

One can manipulate spinors like tensors, distinguishing between a spinor

space and its dual by upper and lower indices, and using the SL(2, C) invariant

bilinear form to raise and lower indices. With complex conjugates and duals,

there are four kinds of irreducible SL(2, C) representations on C2 to keep track

of

SL : This is the standard defining representation of SL(2, C) on C2 , with

SL(2, C) acting on L SL by

L L

288

A standard index notation for such things is called the van der Waer-

den notation. It uses a lower index taking values 1, 2 to label the

components

1

L = =

2

and in this notation acts by

= ei 2 3

that acts on vectors by a rotation by an angle around the z-axis acts on

SL by

1 i 2 3 1

e

2 2

acting on L SL by

L (1 )T L

representation ((g), V ), the pairing between V and its dual V is pre-

served by acting on V by matrices ((g)1 )T , and these provide a repre-

sentation (((g)1 )T , V ). In van der Waerden notation, one uses upper

indices and writes

((1 )T )

Writing elements of the dual as row vectors, our example above of a par-

ticular acts by

1 2 1 2 ei 2 3

SL and SL , given in index notation by

=

where

0 1

=

1 0

acting on R SR by

R R

The van der Waerden notation uses a separate set of dotted indices for

these, writing this as

289

Another common notation among physicists puts a bar over the to

denote that the vector is in this representation, but well reserve that

notation for complex conjugation. The corresponding to a rotation

about the z-axis acts as

1 1

ei 2 3

2 2

SR : This is the dual representation to SR , with SL(2, C) acting on

R SR by

1

R ( )T R

and the index notation uses raised dotted indices

1 T

(( ) )

2 ei 2 3

1 2 1

Another copy of

0 1

=

1 0

gives the isomorphism of SR and SR as representations, by

=

are unitary, and equivalent. As SL(2, C) representations, they are not unitary,

and while the representations are equivalent to their duals, SL and SR are

inequivalent.

In our discussion of the fermionic version of the harmonic oscillator, we defined

the Clifford algebra Cliff(r,s) and found that elements quadratic in its generators

gave a basis for the Lie algebra of so(r, s) = spin(r, s). Exponentiating these

gave an explicit construction of the group Spin(r, s). We can apply that general

theory to the case of Cliff(1,3) and this will give us explicitly the representations

( 21 , 0) and (0, 12 ) (the Aj and Bj we constructed using the Jj and Kj were 4 by

4 matrices for the complexification of the vector representation ( 12 , 12 ), now we

want matrices for something different, the spinor representations).

If we complexify our R4 , then its Clifford algebra becomes just the algebra

of 4 by 4 complex matrices

290

We will represent elements of Cliff(1, 3) as such 4 by 4 matrices, but should

keep in mind that we are working in the complexification of the Clifford algebra

that corresponds to the Lorentz group, so there is some sort of condition on

the matrices that should be kept track of. There are several different choices of

how to explicitly represent these matrices, and for different purposes, different

ones are most convenient. The one we will begin with and mostly use is some-

times called the chiral or Weyl representation, and is the most convenient for

discussing massless charged particles. We will try and follow the conventions

used for this representation in the standard quantum field theory textbook of

Peskin and Schroeder[33].

Writing 4 by 4 matrices in 2 by 2 block form and using the Pauli matrices

j we assign the following matrices to Clifford algebra generators

0 1 0 1 0 2 0 3

0 = , 1 = , 2 = , 3 =

1 0 1 0 2 0 3 0

One can easily check that these satisfy the Clifford algebra relations for gener-

ators of Cliff(1, 3): they anti-commute with each other and

02 = 1, 12 = 1, 22 = 1, 32 = 1

The quadratic Clifford algebra elements 12 j k for k < j satisfy the commu-

tation relations of so(1, 3). These are explicitly

1 i 3 0 1 i 2 0 1 i 1 0

1 2 = , 1 3 = , 2 3 =

2 2 0 3 2 2 0 2 2 2 0 1

and

1 1 1 0 1 1 2 0 1 1 3 0

0 1 = , 0 2 = , 0 3 =

2 2 0 1 2 2 0 2 2 2 0 3

1 1 1

0 (J1 ) = 2 3 , 0 (J2 ) = 1 3 , 0 (J3 ) = 1 2

2 2 2

and

1 1 1

0 (K1 ) =

0 1 , 0 (K2 ) = 0 2 , 0 (K3 ) = 0 3

2 2 2

On the two commuting SL(2, C) subalgebras of so(1, 3) C with bases

1 1

Aj = (Jj + iKj ), Bj = (Jj iKj )

2 2

this representation is

i 1 0 i 2 0 i 3 0

0 (A1 ) = , 0 (A2 ) = , 0 (A3 ) =

2 0 0 2 0 0 2 0 0

291

and

0 i 0 0 0 i 0 0 0 i 0 0

(B1 ) = , (B2 ) = , (B3 ) =

2 0 1 2 0 2 2 0 3

We see explicitly that the action of the quadratic elements of the Clifford

algebra on the spinor representation C4 is reducible, decomposing as the direct

sum SL SR of two inequivalent representations on C2

L

=

R

tions on the components. The Aj act just on SL , the Bj just on SR . An

alternative standard notation to the two-component van der Waerden notation

is to use the four components of C4 with the action of the matrices. The

relation between the two notations is given by

=

where the index on the left takes values 1, 2, 3, 4 and the indices , on the

right each take values 1, 2.

An important element of the Clifford algebra is the one you get by multi-

plying all of the basis elements together. Physicists traditionally multiply this

by i to make it self-adjoint and define

1 0

5 = i0 1 2 3 =

0 1

This can be used to produce projection operators from the Dirac spinors onto

the left and right-handed Weyl spinors

1 1

(1 5 ) = L , (1 + 5 ) = R

2 2

There are two other commonly used representations of the Clifford algebra

relations, related to the one above by a change of basis. The Dirac representation

is useful to describe massive charged particles, especially in the non-relativistic

limit. Generators are given by

1 0 0 1 0 2 0 3

0D = , 1D = , 2D = , 3D =

0 1 1 0 2 0 3 0

and the projection operators for Weyl spinors are no longer diagonal, since

0 1

5D =

1 0

292

A third representation, the Majorana representation, is given by

1 0 0 0

0 0 0 1

|p|2

0 0 1 0 ipx i 2m t

1 0

0M , 1M = i e e 0 0

= i

0 1 0 0 0 0 1 0

1 0 0 0 0 0 0 1

0 0 0 1 0 1 0 0

0 0 1 0 1 0 0 0

2M , 3M = i

= i

0 1 0 0 0 0 0 1

1 0 0 0 0 0 1 0

with

0 1 0 0

1 0 0 0

5M = i

0

0 0 1

0 0 1 0

possibility of having spinors that are not complex numbers, but real numbers,

since one sees that (up to a factor of i), the Clifford algebra matrices can be

chosen to be real. The factor of i indicates that if one wants to do this, one

really should adopt the opposite sign convention for the Minkowski metric and

work with Cliff(3, 1) rather than Cliff(1, 3) since one has Cliff(3, 1) = M (4, R)

and can use purely real matrices.

Removing the is and working with this opposite sign convention, one has

0 1 0 0

1 0 0 0

0 1 2 3 =

0

0 0 1

0 0 1 0

and

(0 1 2 3 )2 = 1

operator on this space with square 1, so it provides a complex structure on

SM . Recall that a complex structure on a real vector space gives a splitting of

the complexification of the real vector space into a sum of two complex vector

spaces, related by complex conjugation. In this case this corresponds to

SM C = SL SR

the fact that complexifying Majorana spinors gives the two kinds of Weyl

spinors.

293

28.3 For further reading

Most quantum field theory textbook have extensive discussions of spinor rep-

resentations of the Lorentz group and gamma matrices. Typical examples are

Peskin-Schroeder[33] and Quantum Field Theory in a Nutshell by Tony Zee, see

Chapter II.3 and Appendix E[56].

294

Chapter 29

their Representations

The Lorentz group and the group of space-time translations fit together into a

larger group, the Poincare group. However, before moving on to discuss that

group and its representations, well first consider what happens in general for

constructions like this, and then examine specifically two simpler such cases, the

groups E(2) and E(3) of Euclidean motions in 2 and 3 dimensions. These raise

many of the same issues as the Poincare group, and the case of E(2) appears as

an important part of the Poincare group story.

Given two groups G0 and G00 , one can form the product group by taking pairs

of elements (g 0 , g 00 ) G0 G00 , but something more subtle happens when G0

and G00 are related by the existence of an action of G00 on G0 . As an example,

consider the case of v2 V and g2 GL(V ) (with V a vector space, thus an

additive group, and GL(V ) the group of invertible linear transformations of V ),

acting on V by

v (v2 , g2 ) v = g2 v + v2

If we then act on the result with (v1 , g1 ) we get

(v1 , g1 ) (g2 v + v2 ) = g1 g2 v + g1 v2 + v1

Note that this is not what we would get if we took the product group law on

V GL(V ), since then the action of (v1 , g1 )(v2 , g2 ) on V would be

v g1 g2 v + v1 + v2

To put a group structure on the set V GL(V ) that makes this a group action,

define the group V o GL(V ) as V GL(V ) with the group law

(v1 , g1 ) (v2 , g2 ) = (g1 v2 + v1 , g1 g2 )

295

This is an example of the following general construction

Definition (Semi-direct product). A semi-direct product

G=N oK

of two groups N and K is a group such that

As a set

G=N K

and any element g G can uniquely be written as g = nk, with n N

and k K.

N is a normal subgroup of G, a condition that is written as

N G

There is a homomorphism

: k K k Aut(N )

such that the group law in N o K is given by

(n1 k1 )(n2 k2 ) = n1 k1 n2 k2 = n1 k1 (n2 )k1 k2

is a normal subgroup of G iff

n N = gng 1 N, n N, g G

In our semi-direct product case, this is only non-trivial for g K, so could be

rewritten

n N = knk 1 N, n N, k K

One can show that N G is equivalent to the existence of a group structure on

the set of N cosets in G. Defining [gN ] for g G to be the equivalence class

where one identifies all elements of G of the form gn, for n N , when N G

one gets a well-defined product on these equivalence classes, giving the quotient

group G/N . One can always define the set G/N , but only when N G is this

set a group.

Aut(N ) is the so-called automorphism group of N , the group of homomor-

phisms from N to itself. Note that the product group N K is the special case

of N o K where k is the identity map on N , for all k K.

The symbol N o K is supposed to be a mixture of the and symbols

(note that some authors define it to point in the other direction). It has the

defect that it does not specify which is being used to construct the semi-direct

product, and there may be multiple possibilities. If one knows the group law on

N o K one can recover since

n1 k1 n2 k2 = n1 k1 (n2 )k1 k2 = k1 (n2 ) = k1 n2 k11

Some examples of semi-direct products are

296

For any vector space V , the group V o GL(V ) described at the beginning

of this section.

Recall that we defined the Heisenberg group H2n+1 as the space R2n+1

with the group law

0

1 p p

(p, q, c) (p0 , q0 , c0 ) = (p + p0 , q + q0 , c + c0 + (( , ))

2 q q0

where p Rn , q Rn , c R and is the symplectic (non-degenerate

antisymmetric bilinear) form

0

p p

( , ) = qp0 p q0

q q0

Taking (p, 0, c) N = Rn+1 H2n+1 and (0, q, 0) K = Rn H2n+1

one has

H2n+1 = Rn+1 o Rn

In this case, since we know the group law, we find that is given by

(0,q,0) (p, 0, c) =(0, q, 0) (p, 0, c) (0, q, 0)1

=(0, q, 0) (p, 0, c) (0, q, 0)

1

=(0, q, 0) (p, q, c + p q)

2

=(p, 0, c + p q)

polynomials in Qj , Pj give (after exponentiation) a representation of the

Heisenberg group H2n+1 , but one can also quantize quadratic polynomials

in Qj , Pj . These give a representation of M p(2n, R), a double cover of

the group Sp(2n, R). The action of M p(2n, R) or Sp(2n, R) preserves the

symplectic form which determines the multiplication law in H2n+1 . As

a result, elements of these groups give automorphisms of H2n+1 , and thus

the map we need to define a semi-direct product

H2n+1 o M p(2n, R)

This semi-direct product group is the group for which our quantization

construction provides a unitary representation, infinite dimensional but

essentially unique by the Stone-von Neumann theorem.

Another class of semi-direct product groups are the Euclidean groups

Definition (Euclidean group). The Euclidean group in dimension n is the semi-

direct product

Rn o SO(n)

of the translation and rotation groups of Rn , with multiplication law

(v1 , R1 ) (v2 , R2 ) = (R1 v2 + v1 , R1 R2 )

(where vj Rn , Rj SO(n))

297

Here our map just associates to an element of SO(n) the corresponding

rotation of Rn (an automorphism of the translation group)

R (v) = Rv

One can similarly construct a double cover Rn oSpin(n) of the Euclidean group,

and defining the multiplication using

taking pairs (v, R)

R (v) = (R)v

E(n) can also be written as a matrix group, taking it to be the subgroup of

GL(n+1, R) of matrices of the form (R is an n by n matrix, v an n-dimensional

column vector)

R v

0 1

One gets the multiplication law for E(n) from matrix multiplication since

R1 v1 R2 v2 R1 R2 R1 v2 + v1

=

0 1 0 1 0 1

Given a representation (, V ) of a semi-direct product N o K, by restriction

to the subgroups N and K, one gets a representation (N , V ) of N and a

representation (K , V ), on the same vector space V . These two representations

determine (, V ), but in order to combine together to give a representation of

N o K, since

k (n) = knk 1

they must satisfy the condition

H2n+1 , K = Sp(2n, R), the Stone-von Neumann theorem uniqueness theorem

implies there is a single irreducible rep of H2n+1 , so H2n+1 (n) and H2n+1 (k (n))

must be equivalent representations. Schurs lemma implies that the Sp(2n,R) (k)

giving the equivalence are unique up to a scalar. We have seen how to explic-

itly construct an appropriate true representation of the double cover M p(2n, R)

(just by exponentiating quadratic combinations of the Qj , Pj operators).

In all the other examples of semi-direct products that we have seen, as well

as the Poincare group case we are interested in, the normal subgroup N of the

semi-direct product is a commutative group, so from now on we will just consider

that case. Things simplify considerably since the irreducible representations of

N are then all one-dimensional and one can define

298

be the set of

Definition (Character group). For N a commutative group, let N

characters of N , i.e. functions

:N C

form a group, with multiplication

The elements of N

We only will actually need the case N = Rn , where we have already seen

that the irreducible representations are one-dimensional and given by

p (x) = eipx

So the character group in this case is N

vector p.

Recall that whenever we have an action of a group G on a space X, we get

an action of G on functions on M by

g f (x) = f (g 1 x)

by

k n = k (n)

and so we get an action of K on the functions on N , and thus on N . We can use

this to study representations of the semi-direct product by proceeding roughly

as follows.

Given a representation on (, V ) of the semi-direct product N o K, we can

consider this as a representation of the subgroup N , and decompose it into

irreducibles. This decomposition can be thought of as

V = V

, i.e.

where V is the subspace on which N acts by the character N

v V = (n)v = (n)v

If V and N were finite dimensional this would be a finite sum. We are interested

in cases where this is not true, but do not want to try and enter into the

necessary analytical machinery to properly deal with this question, so this should

be considered just a motivation for the construction that follows.

If one N occurs in this sum, so will all the other N that one

gets by acting on with elements of K. The set of these is called the K-orbit

O of in N under the action of K. Distinct orbits will correspond to distinct

299

irreducible representations, which suggests that we try and construct irreducible

representations of N o K by taking functions on K-orbits in N .

One example of this phenomenon is the Schrodinger representation of the

Heisenberg group H2n+1 = Rn+1 o R2n on functions on Rn . We will not work

out the details here, but in this case N is the commutative group Rn+1 of

elements of H2n+1 of the form (p, 0, c), and N is just the dual space, also Rn+1 .

n

The group K = R of elements of the form (0, q, 0) acts on this, and the K-

orbit O of any element of N that is non-zero on (0, 0, c) will be a copy of Rn .

Functions on this orbit will give a version of the Schrodinger representation.

With the material of the previous section as motivation, well consider the case

of the Euclidean group E(n) and try and find irreducible representations. Here

elements of the group are given by pairs (x, R) with x N = Rn and R K =

SO(n). Elements of N = Rn can be thought of as the functions

p = eipx

1

eipx eipR x

= ei(Rp)x

orbit of any element p N is the sphere of radius r = |p|.

There are two different cases:

E(n) on which the translation subgroup N acts trivially. Any SO(n)

representation can occur, so the irreducible representations of E(n) we

get this way are just the finite-dimensional representations of SO(n), with

the rest of the group acting trivially.

that are non-zero only on the n 1 dimensional sphere of radius r in

momentum space.

To better understand the second case, where the representations are infinite-

dimensional function spaces, one can consider the case n = 3, and observe that

this is just the space of eigenfunctions of the free-particle Hamiltonian of a

fixed energy, and thus also the space of solutions of the Schrodinger equation

with this energy. This space of solution is an irreducible representation of the

Euclidean group E(3), with the action of the translation subgroup generated

by the momentum operators P, and the rotation subgroup SO(3) generated

by the angular momentum operators (both momentum and angular momentum

operators commute with the Hamiltonian).

300

More explicitly, consider the quantum mechanical state space H of functions

on R3 , describing a free particle. Recall that we would like to work with the

functions

k (x) = eipx = |pi

as a sort of continuous basis of H, even though these are not integrable functions.

The formalism for doing this uses distributions and the orthonormality relation

hp|p0 i = (p p0 )

function (p)

e on momentum space as

ZZZ

1

(x) = eipx (p)d

e 3

x

(2)3/2

The Euclidean group E(3) will act on H, with the action generated by mo-

mentum and angular momentum operators, but this representation will not be

irreducible. To get an irreducible representation, we need to restrict to an ap-

propriate subspace of H, taking only those with Fourier transforms (p)

e that

are non-zero just when |p|2 has a fixed value (these are just the SO(n)-orbits

Op of the rotation group on momentum space). One can also think of this

in terms of the action of the operator |P|2 , which plays the role of a Casimir

operator in this case. Recall that irreducible representations of SO(3) could

be characterized by the eigenvalue with which the Casimir operator L2 acted

on them. Here we want to consider states with a fixed eigenvalue of a Casimir

operator for E(3), the operator |P|2 (the operator L2 is no longer a Casimir in

this case, since it does not commute with the generators P of the translation

part of E(3)).

Instead of the Casimir operator, we can use something proportional to it,

the Hamiltonian H, and think of the subspace we are looking for as a space of

eigenfunctions of H with energy E. We have

1 1 d2 d2 d2 1

H= (P12 + P22 + P32 ) = ( 2 + 2 + 2) =

2m 2m dx1 dx2 dx3 2m

where is the Laplacian. The energy eigenfunctions of energy E form a sub-

space HE of the functions in H, those satisfying

1 d2 d2 d2

H = ( 2 + 2 + 2 ) = E

2m dx1 dx2 dx3

This has solutions

p (x) = eipx = |pi

where p satisfies |p|2 = 2mE. So we get an orbit Op in momentum space

R3 given by a sphere of radius r = 2mE and this corresponds to a single

irreducible representation of E(3) on HE . This representation is infinite di-

mensional and unitary, although we need an appropriate version of the inner

301

product, which we will not write down here, although we will later do this for

the case of the Poincare group.

One could of course do the same thing in two dimensions, getting infinite

dimensional unitary irreducible representations of E(2) corresponding to circles

in R2 of radius r > 0, as well as one dimensional representations labeled by

integers (the representations of SO(2)), corresponding to the orbit of radius 0.

In the case of E(2), this is all of the irreducible representations, but for

E(n), n > 2 there will be other irreducibles, and the outline of their construction

goes as follows. For each orbit Op in N = Rn , it turns out that we can get the

group to act not just on scalar functions on Op , but on certain vector-valued

functions. For each p Rn there is a subgroup of SO(n) of rotations that leave

p invariant. One can see by going to a coordinate system where p points in

the nth direction that rotations that just act on the other n 1 directions will

leave this invariant, and these will form a group SO(n 1). For different p

in an orbit one gets different specific SO(n 1) subgroups of SO(n), but they

are all isomorphic. SO(n 1) is called the little group of the element p of

N , or of the corresponding orbit Op . Taking not scalar functions on the orbit,

but functions that take values in a representation of SO(n 1), it is not hard

to show that one can get a consistent action of the full group E(n) on such

functions, and that this representation is irreducible and unitary. For the case

of E(3), to each orbit of radius r, we will get an irreducible representation for

each choice of a representation of the little group SO(2). These are labeled by

integers m, so our irreducible representations are labeled by pairs (r, m). In the

case of E(2), the little group of each non-zero p is trivial, so irreducibles are

labeled just by the radius r.

While the case of the Poincare group that we will examine in the next chapter

is often covered in physics texts, semi-direct products and their representations

are less commonly discussed in either physics or mathematics textbooks. The

general theory was developed by Mackey during the late 1940s and 1950s. Some

textbooks that do cover the subject include section 3.8 of [42], Chapter 5 of [46],

[8], and [48].

302

Chapter 30

The Poincar

e Group and its

Representations

In the previous chapter we saw that one can take the semi-direct product of

spatial translations and rotations and that the resulting group has infinite-

dimensional unitary representations on the state space of a quantum free parti-

cle. The free particle Hamiltonian plays the role of a Casimir operator: to get

irreducible representations one fixes the eigenvalue of the Hamiltonian (the en-

ergy), and then the representation is on the space of solutions to the Schrodinger

equation with this energy. This is a non-relativistic procedure, treating time and

space (and correspondingly the Hamiltonian and the momenta) differently. For

a relativistic analog, we will use instead the semi-direct product of space-time

translations and Lorentz transformations. Irreducible representations of this

group will be labeled by a continuous parameter (the mass) and a discrete pa-

rameter (the spin or helicity), and these will correspond to possible relativistic

elementary particles.

In the non-relativistic case, the representation occurred as a space of solu-

tions to a differential equation, the Schrodinger equation. There is an analogous

description of the irreducible Poincare group representations as spaces of solu-

tions of relativistic wave equations, but we will put off that story until succeeding

chapters.

e group and its Lie algebra

Definition (Poincare group). The Poincare group is the semi-direct product

P = R4 o SO(1, 3)

with double-cover

P = R4 o SL(2, C)

The action of SO(1, 3) or SL(2, C) on R4 is the action of the Lorentz group on

Minkowski space.

303

We will refer to both of these groups as the Poincare group, with the

double-cover the meaning only when we need it because spinor representations

of the Lorentz group are involved. The two groups have the same Lie algebra,

so the distinction is not needed in discussions that only need the Lie algebra.

Elements of the group P will be written as pairs (a, ), with a R4 and

SO(1, 3) and the group law is

The Lie algebra LieP = LieP has dimension 10, with basis

P0 , P1 , P2 , P3 , J1 , J2 , J3 , K1 , K2 , K3

where the first four elements are a basis of the Lie algebra of the translation

group (note that we are following physicists in using the same notation for Lie

algebra basis elements and the corresponding operators in a representation of

the Lie algebra), and the next six are a basis of so(1, 3), with the Jj giving

the subgroup of spatial rotations, the Kj the boosts. We already know the

commutation relations for the translation subgroup, which is commutative so

[Pj , Pk ] = 0

and that the commutation relations between the Jj and Kj correspond to the

fact that the Kj transform as a vector under spatial rotations, so for example

commuting the Kj with J1 gives an infinitesimal rotation about the 1-axis and

To find the Lie algebra relations between translation and so(1, 3) basis ele-

ments, recall that the Lie algebra commutators are related to the group multi-

plication by

d

[X, Y ] = (etX Y etX )|t=0

dt

Using the semi-direct product relation

mation (0, ) P one finds

=(0, )(a + 1 0, 1 ) = (0, )(a, 1 )

=(a, 1)

304

Taking = etJ1 one finds

1 0 0 0 a0 0

d d

0 1 0 0 a1

0

a|t=0 = =

dt dt 0 0 cos t sin t a2 a3

0 0 sin t cos t a3 |t=0 a2

This is a relation on finite translations (a, 1), but it also hold for infinitesimal

translations, so we have

P0 1 0 0 0 P0 0

P1 d 0 1 0 0 P1

0

P2 ] = dt 0 0 cos t sin t P2

[J1 , =P3

which shows that the Pj transform as vectors under infinitesimal rotations about

the 1-axis, so

For boosts along the 1-axis, one does the same calculation with = etK1 to

find

cosh t sinh t 0 0 a0 a2

d d sinh t cosh t 0 0 a1

a1

a|t=0 = =

dt dt 0 0 1 0 a2 0

0 0 0 1 a3 |t=0 0

so

P0 P1

P1 P0

[K1 ,

P2 ] = 0

P3 0

and in general one will have

e group

We want to find unitary irreducible representations of the Poincare group. These

will be infinite dimensional, so given by operators (g) on a Hilbert space H,

which will have an interpretation as a single-particle relativistic quantum state

space. The standard physics notation for the operators giving the represen-

tation is U (a, ), with the U emphasizing their unitarity. To classify these

representations, we recall from the last chapter that irreducible representations

of semi-direct products N o K are associated with pairs of a K-orbit O in the

space N and an irreducible representation of the corresponding little group K .

305

= R4 is the space of characters of the translation

For the Poincare group, N

group of Minkowski space, so functions

Lorentz group acts on this R4 by

p p

and, restricting attention to the p0 p3 plane, the picture of the orbits looks

like this

Unlike the Euclidean group case, here there are several different kinds of

orbits O . Well examine them and the corresponding stabilizer groups K

each in turn, and see what can be said about the associated representations.

One way to understand the equations describing these orbits is to note that

the different orbits correspond to different eigenvalues of the Poincare group

Casimir operator

P 2 = P02 P12 P22 P32

This operator commutes with all the generators of the Lie algebra of the Poincare

group, so by Schurs lemma it must act as a scalar times the identity on an

306

irreducible representation (recall that the same phenomenon occurs for SU (2)

representations, which can be characterized by the eigenvalue j(j +1) of the Cas-

mir operator J2 for SU (2)). At a point p = (p0 , p1 , p2 , p3 ) in energy-momentum

space, the Pj operators are diagonalized and P 2 will act by the scalar

m. The value of the scalar will be the same everywhere on the orbit, so in

energy-momentum space orbits will satisfy one of the three equations

2

m

2 2 2 2

p0 p1 p2 p3 = m2

0

Note that in this chapter we are just classifying Poincare group representa-

tions, not actually constructing them. It is possible to construct these represen-

tations using the data we will find that classifies them, but this would require

introducing some techniques (for so-called induced representations) that go

beyond the scope of this course. In later chapters we will explicitly construct

these representations in certain specific cases as solutions to certain relativistic

wave equations.

One way to get positive values m2 of the Casimir P 2 is to take the vector

p = (m, 0, 0, 0), m > 0 and generate an orbit Om,0,0,0 by acting on it with

the Lorentz group. This will be the upper, positive energy, hyperboloid of the

hyperboloid of two sheets

so q

p0 = p21 + p22 + p23 + m2

the form

1 0

0

where SO(3), so Km,0,0,0 = SO(3). Irreducible representations are classi-

fied by the spin. For spin 0, points on the hyperboloid can be identified with

positive energy solutions to a wave equation called the Klein-Gordon equation

and functions on the hyperboloid both correspond to the space of all solutions

of this equation and carry an irreducible representation of the Poincare group.

In the next chapter we will study the Klein-Gordon equation, as well as the

quantization of the space of its solutions by quantum field theory methods.

307

We will later study the case of spin 12 , where one must use the double cover

SU (2) of SO(3). The Poincare group representation will be on functions on

the orbit that take values in two copies of the spinor representation of SU (2).

These will correspond to solutions of a wave equation called the massive Dirac

equation.

For choices of higher spin representations of the stabilizer group, one can

again find appropriate wave equations and construct Poincare group represen-

tations on their space of solutions, but we will not enter into this topic.

Starting instead with the energy-momentum vector p = (m, 0, 0, 0), m > 0,

the orbit Om,0,0,0 one gets is the lower, negative energy component of the

hyperboloid

p20 p21 p22 p23 = m2

satisfying q

p0 = p21 + p22 + p23 + m2

Again, one has the same stabilizer group Km,0,0,0 = SO(3) and the same con-

stuctions of wave equations of various spins and Poincare group representations

on their solution spaces as in the positive energy case. Since negative energies

lead to unstable, unphysical theories, we will see that these representations are

treated differently under quantization, corresponding physically not to particles,

but to anti-particles.

One can get negative values m2 of the Casimir P 2 by considering the orbit

O0,0,0,m of the vector p = (0, 0, 0, m). This is a hyperboloid of one sheet,

satisfying the equation

It is not too difficult to see that the stabilizer group of the orbit is K0,0,0,m =

SO(2, 1). This is isomorphic to the group SL(2, R), and it has no finite-

dimensional unitary representations. These orbits correspond physically to

tachyons, particles that move faster than the speed of light, and there is

no known way to consistently incorporate them in a conventional theory.

The simplest case where the Casimir P 2 is zero is the trivial case of a point

p = (0, 0, 0, 0). This is invariant under the full Lorentz group, so the orbit

O0,0,0,0 is just a single point and the stabilizer group K0,0,0,0 is the entire Lorentz

group SO(1, 3). For each finite-dimensional representation of SO(1, 3), one gets

a corresponding finite dimensional representation of the Poincare group, with

308

translations acting trivially. These representations are not unitary, so not usable

for our purposes.

One has P 2 = 0 not only for the zero-vector in momentum space, but for a

three-dimensional set of energy-momentum vectors, called the null-cone. By

the term cone one means that if a vector is in the space, so are all products

of the vector times a positive number. Vectors p = (p0 , p1 , p2 , p3 ) are called

light-like or null when they satisfy

One such vector is p = (1, 0, 0, 1) and the orbit of the vector under the action

of the Lorentz group will be the upper half of the full null-cone, the half with

energy p0 > 0, satisfying q

p0 = p21 + p22 + p23

x3 axis, but also boosts in the other two directions. It is isomorphic to the

Euclidean group E(2). Recall that this is a semi-direct product group, and it

has two sorts of irreducible representations

Representations such that the two translations act trivially. These are

irreducible representations of SO(2), so one-dimensional and characterized

by an integer n (half-integers when one uses the Poincare group double

cover).

a circle of radius r

The first of these two gives irreducible representations of the Poincare group

on certain functions on the positive energy null-cone, labeled by the integer n,

which is called the helicity of the representation. We will in later chapters

consider the cases n = 0 (massless scalars, wave-equation the Klein-Gordon

equation), n = 21 (Weyl spinors, wave equation the Weyl equation), and n =

1 (photons, wave equation the Maxwell equations).

The second sort of representation of E(2) gives representations of the Poincare

group known as continuous spin representations, but these seem not to cor-

respond to any known physical phenomena.

Looking instead at the orbit of p = (1, 0, 0, 1), one gets the negative energy

part of the null-cone. As with the time-like hyperboloids of non-zero mass

m, these will correspond to anti-particles instead of particles, with the same

classification as in the positive energy case.

309

30.3 For further reading

The Poincare group and its Lie algebra is discussed in pretty much any quantum

field theory textbook. Weinberg [51] (Chapter 2) has some discussion of the

representations of the Poincare group on single particle state spaces that we have

classified here. Folland [19] (Chapter 4.4) and Berndt [7] (Chapter 7.5) discuss

the actual construction of these representations using the induced representation

methods that we have chosen not to try and explain here.

310

Chapter 31

and Scalar Quantum Fields

theory describing arbitrary numbers of particle by second quantization of the

standard quantum theory of a free particle. This was done by taking as classical

phase space the space of solutions to the free particle Schrodinger equation, a

space which carries a unitary representation of the Euclidean group E(3). This is

an infinite dimensional space of functions (the space of solutions can be identified

with the space of initial conditions, which is the space of wave-functions at a

fixed time), but one can quantize it using analogous methods to the case of

the finite-dimensional harmonic oscillator (annihilation and creation operators).

After such quantization we get a quantum field theory, with a state space that

describes an arbitrary number of particles. Such a state space provides a unitary

representation of the E(3) group and we saw how to construct the momentum

and angular momentum operators that generate it.

start with an irreducible unitary representation not of E(3), but of the Poincare

group P. In the last chapter we saw that such things were classified by orbits

O of the Lorentz group on momentum space, together with a choice of rep-

resentation of the stabilizer group K of the orbit. The simplest case will be

the orbits Om,0,0,0 , and the choice of the trivial spin-zero representation of the

stabilizer group Km,0,0,0 = SO(3). These orbits are characterized by a positive

real number m, and are hyperboloids in energy-momentum space. Points on

these orbits correspond to solutions of a relativistic analog of the Schrodinger

representation, the Klein-Gordon equation, so we will begin by studying this

equation and its solutions.

311

31.1 The Klein-Gordon equation and its solu-

tions

Recall that a condition characterizing the orbit in momentum space that we

want to study was that the Casimir operator P 2 of the Poincare group acts on

the representation corresponding to the orbit as the scalar m2 . So, we have the

operator equation

P 2 = P02 P12 P22 P32 = m2

characterizing the Poincare group representation we are interested in. Interpret-

ing the Pj as the standard differentiation operators i x j

on a space of wave-

functions, generating the infinitesimal action of the translation group on such

wave-functions, we get the following differential equation for wave-functions:

Definition (Klein-Gordon equation). The Klein-Gordon equation is the second-

order partial differential equation

2 2 2 2

( + 2 + 2 + ) = m2

t2 x1 x2 x23

or

2

( + m2 ) = 0

t2

for functions (x) on Minkowski space (which may be real or complex valued).

This equation is the obvious simplest Lorentz-invariant wave equation to try,

and historically was the one Schrodinger first tried (he then realized it could not

account for atomic spectra and instead used the non-relativistic equation that

bears his name). Taking Fourier transforms

Z

1

(p) =

e d4 xeipx (x)

(2)2

e =0

e that are non-zero only on the hyperboloid

positive and negative energy

p0 = p

where q

p = p21 + p22 + p23 + m2

Ignoring one dimension these look like

312

In the non-relativistic case, a continuous basis of solutions of the Schrodinger

equation labeled by p R3 was given by the functions

|p|2

eipx ei 2m t

e given by

the Fourier inversion formula

Z

1 |p|2

ipx i 2m t 3

(x, t) = 3/2

(p)e

e e d x

(2) R3

e gave coordinates on our single-particle space H1 , and

we have actions on this of the group of time translations (generated by the

Hamiltonian) and the Euclidean group E(3) (generated by momentum and an-

gular momentum).

In the relativistic case we want to study the corresponding single-particle

space H1 of solutions to the Klein-Gordon equation, but parametrized in a way

that makes clear the action of the Poincare group on this space. Coordinates on

the space of such solutions will now be given by complex-valued functions (p)

e on

4

the energy-momentum space R , supported on the two-component hyperboloid

(here p = (p0 , p)). The Fourier inversion formula giving a general solution in

terms of these coordinates will be

Z

1

(x, t) = (p20 p2 )(p)e

e i(pxp0 t) 4

d p

(2)3/2 M 4

313

with the integral over the 3d hyperboloid expressed as a 4d integral over R4

with a delta-function on the hyperboloid in the argument.

The Poincare group action on the coordinates (p)

e on H1 will be given by

e = eipa (

U (a, )(p) e 1 p)

(Fix this and make it more explicit after deal with symmetries later).

The delta function distribution with argument a function f (x) depends only

on the zeros of f , and if f 0 = 0 at such these zeros, one has

X 1

(f (x)) = (f 0 (xj )(x xj )) = 0 (x xj )

|f (xj )|

xj :f (xj )=0

For each p, one can applying this to the case of the function of p0 given by

f = p20 p2

on R4 , and using

d 2

(p p2 ) = 2p0 = 2p

dp0 0

one finds

Z

1 1 i(pxp0 t)

(x, t) = ((p0 p ) + (p0 + p ))(p)e

e dp0 d3 p

(2)3/2 M 4 2p

Z 3

1 e+ (p)eip t + e (p)eip t )eipx d p

= (

(2)3/2 R3 2p

Here

e+ (p) = (

e p , p), e (p) = (

e p , p)

are the values of e on the positive and negative energy hyperboloids. We see

that instead of thinking of the Fourier transforms of solutions as taking values

on energy-momentum hyperboloids, we can think of them as taking values just

on the space R3 of momenta (just as in the non-relativistic case), but we do

have to use both positive and negative energy Fourier components, and to get

a Lorentz invariant measure need to use

d3 p

2p

instead of d3 p.

A general complex-valued solution to the Klein-Gordon equation will be

given by the two complex-valued functions e+ , e , but we could impose the

condition that the solution be real-valued, in which case one can check that the

pair of functions must satisfy the condition

e (p) = e+ (p)

Real-valued solutions of the Klein-Gordon equation thus correspond to arbitrary

complex-valued functions e+ defined on the positive energy hyperboloid, which

fixes the value of the other function e .

314

31.2 Classical relativistic scalar field theory

We would like to set up the Hamiltonian formalism, finding a phase space H1 and

a Hamiltonian function h on it such that Hamiltons equations will give us Klein-

Gordon equation as equation of motion. Such a phase space will be an infinite-

dimensional function space and the Hamiltonian will be a functional. We will

here blithely ignore the analytic difficulties of working with such spaces, and use

physicists methods, with formulas that can be given a legitimate interpretation

by being more careful and using distributions. Note that now we will take the

fields to be real-valued, this is the so-called real scalar field.

Since the Klein-Gordon equation is second order in time, solutions will be

parametrized by initial data which, unlike the non-relativistic case now requires

the specification at t = 0 of not one, but two functions,

(x) = (x, 0), (x) = (x, t)|t=0

t

the values of the field and its first time derivative.

We will take as our phase space H1 the space of pairs of functions (, ),

with coordinates (x), (x) and Poisson brackets

{(x), (x0 )} = (x x0 ), {(x), (x0 )} = {(x), (x0 )} = 0

We want to get the Klein-Gordon equation for (x, t) as the following pair of

first order equations

= , = ( m2 )

t t

which together imply

2

= ( m2 )

t2

To get these as equations of motion, we just need to find a Hamiltonian

function h on the phase space H1 such that

={, h} =

t

={, h} = ( m2 )

t

One can check that two choices of Hamiltonian function that will have this

property are Z

h= H(x)d3 x

R3

where

1 2 1

H= ( + m2 2 ) or H = ( 2 + ()2 + m2 2 )

2 2

where the two different integrands H(x) are related (as in the non-relativistic

case) by integration by parts so these just differ by boundary terms that are

assumed to vanish.

315

(Should I put in here an explicit calculation, following Das).

One could instead have taken as starting point the Lagrangian formalism,

with an action Z

S= L d4 x

M4

where

1 2

L= (( ) ()2 m2 2 )

2 t

This action is a functional now of fields on Minkowski space M 4 and is Lorentz

invariant. The Euler-Lagrange equations give as equation of motion the Klein-

Gordon equation

(2 m2 ) = 0

One recovers the Hamiltonian formalism by seeing that the canonical momentum

for is

L

= =

and the Hamiltonian density is

1

H = L = ( 2 + ()2 + m2 2 )

2

Besides the position-space Hamiltonian formalism, we would like to have one

for the momentum space components of the field, since for a free field it is these

that will decouple into an infinite collection of harmonic oscillators. For a real

solution to the Klein-Gordon equation we have

Z 3

1 e+ (p)eip t + e (p)eip t )eipx d p

(x, t) = ( +

(2)3/2 R3 2p

d3 p

Z

1

= 3/2

(e+ (p)eip t eipx + e+ (p)eip t eipx )

(2) R3 2p

where we have used the symmetry of the integration over p to integrate over

p instead of p.

We can choose a new way of normalizing Fourier coefficients, one that reflects

the fact that the Lorentz-invariant notion is that of integrating over the energy-

momentum hyperboloid rather than momentum space

e+ (p)

(p) = p

2p

d3 p

Z

1

(x, t) = ((p)eip t eipx + (p)eip t eipx ) p

(2)3/2 R3 2p

316

The (p), (p) will have the same sort of Poisson bracket relations as the

z, z for a single harmonic oscillator, or the (p), (p) Fourier coefficients in the

case of the non-relativistic field:

To see this, one can compute the Poisson brackets for the fields as follows. We

have

d3 p

Z

1 ipx ipx

(x) = (x, t)|t=0 = (i p )((p)e (p)e )

(2)3/2 R3

p

t 2p

and

d3 p

Z

1

(x) = ((p)eipx + (p)eipx ) p

(2)3/2 R3 2p

so

Z

0 1 0 0

{(x), (x )} = ({(p), i(p0 )}ei(pxp x )

2(2)3 R3 R3

0 0

{i(p), (p0 )}ei(px+p x ) )d3 pd3 p0

Z

1 0 0 0 0

= 3

3 (p p0 )(ei(pxp x ) + ei(px+p x ) )d3 pd3 p0

2(2) R3 R3

Z

1 0 0

= 3

(eip(xx ) + eip(xx ) )d3 p

2(2) R3

= 3 (x x0 )

coordinates (x), (x)

Z

1 2

h= ( + ()2 + m2 2 )d3 x

R 3 2

and by laborious calculation one can substitute the above expressions for ,

in terms of (p), (p) to find h as a quadratic polynomial in these coordinates

on the momentum space fields. A quicker way to find the correct expression is

to use the fact that different momentum components of the field decouple, and

we know the time-dependence of such components, so just need to find the right

h that generates this.

We can decompose a solution (x, t) of the Klein-Gordon equation as =

+ + , where

Z 3

1 ip t ipx d p

+ (x, t) = (p)e e

(2)3/2 R3

p

2p

and

d3 p

Z

1

(x, t) = (p)eip t eipx p

(2)3/2 R3 2p

317

If, as in the non-relativistic case, we interpret as a single-particle wave-

function, Hamiltons equation of motion says

{, h} =

t

and applying this to the component of + with momentum p, we just get

multiplication by ip . The energy of such a wave-function would be p , the

eigenvalue of i t . These are called positive frequency or positive energy

wave-functions. In the case of momentum components of , the eigenvalue is

p , and one has negative frequency or negative energy wave-functions.

An expression for h in terms of momentum space field coordinates that will

have the right Poisson brackets on + , is

Z

h= p (p)(p)d3 p

R3

and this is the same expression one could have gotten by a long direct calcula-

tion.

In the non-relativistic case, the eigenvalues of the action of i t on the

2

wave-functions were non-negative ( |p| 2m ) so the single-particle states had non-

negative energy. Here we find instead eigenvalues p of both signs, so single-

particle states can have arbitrarily negative energies. This makes a physically

sensible interpretation of H1 as a space of wavefunctions describing a single

relativistic particle difficult if not impossible. We will however see in the next

section that there is a way to quantize this H1 as a phase space, getting a

sensible multi-particle theory with a stable ground state.

Given the description we have found in momentum space of a real scalar field

satisfying the Klein-Gordon equation, it is clear that one can proceed to quan-

tize the theory in exactly the same way as was done with the non-relativistic

Schrodinger equation, taking momentum components of fields to operators by

replacing

(p) a(p), (p) a (p)

where a(p), a (p) are operator valued distributions satisfying the commutation

relations

[a(p), a (p0 )] = 3 (p p0 )

For the Hamiltonian we take the normal-ordered form

Z

H=

b p a (p)a(p)d3 p

R3

Starting with a vacuum state |0i, by applying creation operators one can create

arbitary positive energy multiparticle states of free relativistic particles with

318

single-particle states having the energy momentum relation

p

E(p) = p = |p|2 + m2

in the non-relativistic case we could define a position space complex-valued field

operator by Z

1

(x)

b = a(p)eipx d3 p

(2)3/2 R3

which has an interpretation as an annhilation operator for a particle localized

at x. Solving the dynamics of the theory gave the time-dependence of the field

operator. Z

1 p2

(x,

b t) =

3/2

a(p)ei 2m t eipx d3 p

(2) R3

For the relativistic case we must do something somewhat different, defining

Definition (Real scalar quantum field). The real scalar quantum field operators

are the operator-valued distributions defined by

d3 p

Z

1 ipx ipx

(x)

b = (a(p)e + a (p)e )

(2)3/2 R3

p

2p

d3 p

Z

1 ipx ipx

b(x) = (i p )(a(p)e a (p)e )

(2)3/2 R3

p

2p

By essentially the same computation as for Poisson brackets, one can com-

pute commutators, finding

[(x),

b b(x0 )] = i 3 (x x0 ), [(x),

b b 0 )] = [b

(x b(x0 )] = 0

(x),

berg Lie algebra, now the infinite dimensional Lie algebra corresponding to the

phase space H1 of solutions of the Klein-Gordon equation.

The Hamiltonian operator will be quadratic in the field operators and can

be chosen to be

Z

1

H

b = (x)2 + ((x))

: (b b 2

+ m2 (x)

b 2 ) : d3 x

R 3 2

This operator is normal ordered, and a computation (see for instance [9]) shows

that in terms of momentum space operators this is just

Z

Hb = p a (p)a(p)d3 p

R3

The dynamical equations of the quantum field theory are now

b b iH]

= [, b =

b

t

319

b = ( m2 )b

b = [b , iH]

t

which have as solution the following equation for the time-dependent field op-

erator:

d3 p

Z

1 ip t ipx ip t ipx

(x,

b t) = (a(p)e e + a (p)e e )

(2)3/2 R3

p

2p

of positive energy and creation operators for momentum eigenstates of negative

energy. Note that, unlike the non-relativistic case, here the quantum field oper-

ator is self-adjoint. In the next chapter we will see what happens in the case of

a complex scalar quantum field, where the operator and its adjoint are distinct.

It is a characteristic feature of relativistic field theory that what one is quan-

tizing is not just a space of positive energy wave-functions, but a space that

includes both positive and negative energy wave-functions, assigning creation

operators to one sign of the energy, annihilation operators to the other, and by

this mechanism getting a Hamiltonian operator with spectrum bounded below.

In more complicated quantum field theories there will be other operators (for

instance, a charge operator) that can distinguish between particle states that

correspond to wave-functions of positive energy and anti-particle states that

correspond to wave-functions of negative energy. The real scalar field case is a

bit special in that there are no such operators and one says that here a particle

is its own anti-particle.

To understand the nature of the negative energy solutions in a way indepen-

dent of the annihilation-creation operator formalism, one should realize that

the notion of quantization we are using requires an extra piece of structure not

present in the classical field theory, a choice of complex structure. Recall from

the discussion of quantization of the bosonic harmonic oscillator in chapter 16

that the Bargmann-Fock quantization began with a real phase space M , then

took the dual space M with a choice of complex structure J such that

M C = M+

J MJ

where M+ J is the +i eigenspace of J, MJ the i eigenspace. The quantum

state space will be the symmetric algebra S (M+ J ) or equivalently the space of

polynomials on the dual of M+ J . The choice of J is reflected in the fact that

the Bargmann-Fock construction requires a choice of distinguished state |0iJ ,

the state corresponding to the constant polynomial function 1.

In non-relativistic quantum field theory our classical phase space was just

the space of complex wave-functions (x) at a constant time, and we could

take as J just multiplication by i, so our M+ J was just the space of complex

wave-functions itself, with M J the space of complex conjugated wave-functions.

Elements of M+ J got quantized as operators , and annihilation operators on the

b

Bargmann-Fock state space, while elements of M J were quantized as adjoint

operators b , acting as creation operators on the state space.

320

In relativistic quantum field theory, we must do something quite different.

The real scalar fields satisfying the Klein-Gordon equation form a real phase

space M to be quantized. When we complexify and look at the space M C

of complex-valued solutions, it naturally decomposes as a representation of the

Poincare group into two pieces: M+ , the complex functions on the positive

energy hyperboloid and M , the complex functions on the negative energy

hyperboloid. What we do to quantize is to take the complex structure to be

the operator J that is +i on positive energy wave-functions and i on negative

energy wavefunctions. Complexified classical fields in M+ get quantized as

annihilation operators, those in M as creation operators. Since conjugation

interchanges M+ and M , non-zero real-valued classical fields have components

in both M+ and M since they are their own conjugates. As we have seen,

they end up being a superposition of annihilation and creation operators.

As explained in the non-relativistic case, in quantum field theory explicitly

dealing with states and their time-dependence is awkward, so we work in the

Heisenberg picture, expressing everything in terms of a fixed, unchangeable

state |0i and time-dependent operators. For the free scalar field theory, we have

explicitly solved for the time-dependence of the field operators. A basic quantity

needed for describing the propagation of quanta of a quantum field theory is

the propagator:

function or propagator for a scalar field theory is the amplitude, for t > t0

b t)(x

x x0 , so we can just evaluate the case (x0 , t0 ) = (0, 0), using the formula for

the time dependent field to get

Z

1

G(x, t, 0, 0) = h0|(a(p)eip t eipx + a (p)eip t eipx )

(2)3 R3 R3

d3 p d3 p0

(a(p0 ) + a (p0 ))|0i p p

2p 2p0

d3 p d3 p0

Z

= 3 (p p0 )eip t eipx p p

R3 R3 2p 2p0

d3 p

Z

= eip t eipx

R3 2p

For t > 0, this gives the amplitude for propagation of a particle in time t

from the origin to the point x.

321

31.5 For further reading

Pretty much every quantum field theory textbook has a treatment of the rela-

tivistic scalar field with more details than here, and significantly more physical

motivation. A good example with some detailed versions of the calculations

done here is [9]. See Folland [19], chapter 5 for a mathematically more careful

treatment of the distributional nature of the scalar field operators.

322

Chapter 32

Quantum Fields

323

324

Chapter 33

Coupling to the

Electromagnetic Field

325

326

Chapter 34

Quantization of the

Electromagnetic Field: the

Photon

327

328

Chapter 35

Spin-1/2 Fields

329

330

Chapter 36

An Introduction to the

Standard Model

331

332

Chapter 37

Further Topics

course that arent discussed here due to lack of time in the class and lack of

energy of the author. Two important ones are:

Scattering theory. Here one studies solutions to Schrodingers equation

that in the far past and future correspond to free-particle solutions, with

a localized interaction with a potential occuring at finite time. This is

exactly the situation analyzed experimentally through the study of scat-

tering processes. Use of the representation theory of the Euclidean group,

the semi-direct product of rotations and translations in R3 provides in-

sight into this problem and the various functions that occur, including

spherical Bessel functions.

Perturbation methods. Rarely can one find exact solutions to quantum

mechanical problems, so one needs to have at hand an array of approxima-

tion techniques. The most important is perturbation theory, the study of

how to construct series expansions about exact solutions. This technique

can be applied to a wide variety of situations, as long as the system in

question is not too dramatically of a different nature than one for which

an exact solution exists.

333

334

Appendix A

Conventions

Ive attempted to stay close to the conventions used in the physics literature,

leading to the choices listed here. Units have been chosen so that ~ = 1.

To get from the self-adjoint operators used by physicists as generators of

symmetries, multiply by i to get a skew-adjoint operator in a unitary repre-

sentation of the Lie algebra, for example

The Lie bracket on the space of functions on phase space M is given by

the Poisson bracket, determined by

{q, p} = 1

a unitary representation of the Heisenberg Lie algebra h3 , multiply the

self-adjoint operators by i, so they satisfy

that satisfies

example

l1 = q2 p3 q3 p2

under quantization go to the self-adjoint operator

L1 = Q2 P3 Q3 P2

of the Lie algebra so(3). The three such operators will satisfy the Lie

bracket relations of so(3), for instance

335

For the spin 12 representation, the self-adjoint operators are Sj = 2j ,

the Xj = i 2j give the Lie algebra representation. Unlike the integer

spin representations, this representation does not come from the bosonic

quantization map .

Given a unitary Lie algebra representation 0 (X), the unitary group action

on states is given by

0

|i (eX )|i = e (X) |i

Instead of considering the action on states, one can consider the action on

operators by conjugation

0 0

O O() = e (X) Oe (X)

d

O() = [O, 0 (X)]

d

If a group G acts on a space M , the representation one gets on functions on

M is given by

(g)(f (x)) = f (g 1 x)

Examples include

|i eiaP |i

d d

eia(i dq ) (q) = ea dq (q) = (q a)

d

So, the Lie algebra action is given by the operator iP = dq . On

operators one has

O(a) = eiaP OeiaP

or infinitesimally

d

O(a) = [O, iP ]

da

Time translation (t t a). The convention for the Hamiltonian H is

opposite that for the momentum P , with the Schrodinger equation saying

that

d

iH =

dt

On states, time evolution is translation in the positive time direction, so

states evolve as

|(t)i = eitH |(0)i

336

Operators in the Heisenberg picture satisfy

or infinitesimally

d

O(t) = [O, iH]

dt

which is the quantization of the Poisson bracket relation in Hamiltonian

mechanics

d

f = {f, h}

dt

Conventions for special relativity.

Conventions for representations on field operators.

Conventions for anti-commuting variables. for unitary and odd super Lie

algebra actions.

337

338

Appendix B

Problems

Problem 1:

set of 3 elements. Consider the representation (, C3 ) this gives on the vector

space C3 of complex valued functions on the set of 3 elements (as defined in

class). Choose a basis of this set of functions, and find the matrices (g) for

each element g S 3 .

Is this representation irreducible? If not, can you give its decomposition

into irreducibles, and find a basis in which the representation matrices are block

diagonal?

Problem 2:

Use a similar argument to the one given in class for G = U (1) to classify

the irreducible representations of the group R under the group law of addition.

Which of these are unitary?

Problem 3:

one. What are the complex representations of this group?

There is an obvious representation of SO(2) on R2 given by matrix multi-

plication on real 2-vectors. If I replace the real 2-vectors by complex 2-vectors

I get a 2-complex dimensional representation. How does this decompose as a

direct sum of irreducibles?

Problem 4:

339

Consider a quantum mechanical system with state space H = C3 and Hamil-

tonian operator

0 1 0

H = 1 0 0

0 0 2

odinger equation for this system to find its state vector |(t)i

Solve the Schr

at any time t > 0, given that the state vector at t = 0 was

1

2

3

for i C.

Problem 1: Calculate the exponential etM for

0 0

0 0

0 0 0

by two different methods:

Diagonalize the matrix M (i.e. write as P DP 1 , for D diagonal), then

show that 1

etP DP = P etD P 1

and use this to compute etM .

Calculate etM using the Taylor series expansion for the exponential, as

well as the series expansions for the sine and cosine.

H = Bx 1

(this is the sort of thing that occurs for a spin-1/2 system subjected to a mag-

netic field in the x-direction).

Find the eigenvectors and eigenvalues of H. What are the possible energies

that can occur in this quantum system?

If the system starts out at time t = 0 in the state

1

|(0)i =

0

340

Problem 3: By using the fact that any unitary matrix can be diagonalized

by conjugation by a unitary matrix, show that all unitary matrices can be

written as eX , for X a skew-adjoint matrix in u(n).

By contrast, show that

1 1

A=

0 1

is in the group SL(2, C), but is not of the form eX for any X sl(2, C) (this

Lie algebra is all 2 by 2 matrices with trace zero.

Hint: For 2 by 2 matrices X, one can show (this is the Cayley-Hamilton

theorem: matrices X satisfy their own characteristic equation det(1 X) = 0,

and for 2 by 2 matrices, this equation is 2 tr(X) + det(X) = 0)

X 2 tr(X)X + det(X)1 = 0

For X sl(2, C), tr(X) = 0, so here X 2 = det(X)1. Use this to show that

p

p sin( det(X))

eX = cos( det(X))1 + p X

det(X)

= A and derive a contradiction (taking the trace of the

equation, what is cos( det(X))?)

Problem 4: Show that the Lie algebra u(n) is not a complex vector space.

Problem 1: On the Lie algebras g = su(2) and g = so(3) one can define the

Killing form K(, ) by

1. For both Lie algebras, show that this gives a bilinear, symmetric form,

negative definite, with the basis vectors sa in one case and La in the other

providing an orthogonal basis if one uses K(, ) as an inner product.

Here the Lie algebra adjoint representation (ad, g) gives for each X g a

linear map

ad(X) : R3 R3

and thus a 3 by 3 real matrix. This K 0 is determined by taking the trace

of the product of two such matrices. How are K and K 0 related?

341

Problem 2: Under the homomorphism given in class, what elements of

SO(3) do the quaternions i, j, k (unit length, so elements of Sp(1)) correspond

to?

R , with an inner product such that < v, v >= v02 v12 v22 v32 , where

4

one preserving this inner product is written SO(1, 3) and known as the Lorentz

group. Show that, just as SO(4) has a double-cover Spin(4) = Sp(1) Sp(1),

the Lorentz group has a double cover SL(2, C), with action on vectors given by

identifying R4 with 2 by 2 Hermitian matrices according to

v0 + v3 v1 iv2

(v0 , v1 , v2 , v3 , v4 )

v1 + iv2 v0 v3

and using the conjugation action of SL(2, C) on these matrices. (Hint: use

determinants).

Note that the Lorentz group has a spinor representation, but it is not unitary.

deriving a correct one. The notes refer to a place this is done with another

convention. To do this calculation, note that your result will be a 3 by 3

real matrix, with the first column given by the result of applying this linear

transformation to

1

0

0

then doing the same for the next two columns. Each of these three calculations

is done by conjugating the corresponding basis element in su(2) and expanding

the result in terms of these basis elements.

Problem 1: Using the definition

Z

1 2

+|z2 |2 )

< f, g >= f (z1 , z2 )g(z1 , z2 )e(|z1 | dx1 dy1 dx2 dy2

2 C2

342

Show that the representation on such polynomials given in class (induced

from the SU (2) representation on C2 ) is a unitary representation with

respect to this inner product.

zj zk

1 2

j!k!

are orthonormal with respect to this inner product (break up the integrals

into integrals over the two complex planes, use polar coordinates).

(S+ ) are adjoint of each other.

Problem 2: Using the formulas for the Y1m (, ) and the inner product

weight vector.

Show that this operators commutes with the 0 (X) for all X so(3). Use

this to show that L2 has the same eigenvalue on all vectors in an irreducible

representation of so(3).

eigenfunctions, and calculate the eigenvalue.

343

B.5 Problem Set 5

Problem 1: Consider the action of SU (2) on the tensor product V 1 V 1 of two

spin one-half representations. According to the Clebsch-Gordan decomposition,

this breaks up into irreducibles as V 0 V 2 .

1. Show that

1 1 0 0 1

( )

2 0 1 1 0

is a basis of the V 0 component of the tensor product, by computing first

the action of SU (2) on this vector, and then the action of su(2) on the

vector (i.e. compute the action of 0 (X) on this vector, for the tensor

produ ct representation, and X basis elements of su(2)).

2. Show that

1 1 1 1 0 0 1 0 0

, ( + ),

0 0 2 0 1 1 0 1 1

give a basis for the irreducible representation V 2 , by showing that they are

eigenvectors of 0 (s3 ) with the right eigenvalues (weights), and computing

the action of the raising and lowering operators for su(2) on these vectors.

relativistic particle of mass m in R3 . Using tensor products, how would you

describe a system of two identical such particles? Find the Hamiltonian and

momentum operators. Find a basis f or the energy and momentum eigenstates

for such a system, first under the assumption that the particles are bosons, then

under the assumption that the particles are fermions.

spatial dimension, of size R (the wavefunction satisfies (x, t) = (x + R, t)).

If the wave-function at time t = 0 is given by

6 4

(x, 0) = C(sin( x) + cos( x + ))

R R

where C is a constant and is an angle, find the wave-function for all t. For

what values of C is this a normalized wave-function (|(x, t)|2 = 1)?

Problem 1: Fill in the details of the proof of the Groenewold-van Hove theorem

following the outline given in Chapter 5.4 of Rolf Berndts An Introduction to

Symplectic Geometry.

344

Problem 2: If a quantum harmonic oscillator is in a state

1

(|0i + |1i)

2

at time t = 0, find its position-space wavefunction (q, t) for all t.

the expectation values in the energy eigen-state |ni of the following operators

Q, P, Q2 , P 2

and

Q4

Use these to find the standard deviations in the statistical distributions of ob-

served values of q and p in these states. These are

p p

Q = hn|Q2 |ni hn|Q|ni2 , P = hn|P 2 |ni hn|P |ni2

Problem 1: Show that the Lie algebras of Spin(n) and SO(n) are the same

by showing that the quadratic elements

1

j k

2

for j < k of the Clifford algebra Clif f (n, R) satisfy the same commutation

relations as the Ljk (elementary antisymmetric matrices).

Clifford algebra element of the previous problem gives a rotation in the j k

plane.

cases of the Clifford algebra in 4 or 6 dimensions, corresponding to the fermionic

oscillator in 2 or 3 variables.

The Hamiltonian operator generates a U (1) action on the spinors. What

is it explicitly? This U (1) is a subgroup of the Spin group ( in 4 or 6

dimensions respectively), and so acts not just on spinors, but on vectors

as a rotation. What is the rotation on vectors?

345

For a rotation by an angle in the j k plane in 4 or 6 dimensions, what

are the elements of the Spin group that correspond to this, and how do

they act on the fermionic oscillator states? Do this by expressing things

in terms of annihilation and creation operators and their action on the

spinors, thought of as a fermionic oscillator state space.

Fock construction construct spinors in even dimensions as spaces of functions

of anticommuting variables. Find the inner product on such spinors that is the

analog of the one constructed using an integral in the bosonic case. Show that

the operators aF j and aF j are adjoints with respect to this inner product.

Problem 1: Compute the propagator

0 , t0 ) (x, t)|0i

G(x0 , t0 , x, t) = h0|(x

for the free non-relativistic particle of mass m (for t0 > t). First do this in

momentum space, showing that

2

k

e 0 , t0 , k, t) = ei 2m (t0 t)

G(k (k 0 k)

m 3 m

(x0 x)2

G(x0 , t0 , x, t) = ( 0

) 2 e i2(t0 t)

i2(t t)

Show also that

lim G(x0 , t0 , x, t) = (x0 x)

tt0

including interaction with a potential, show that there is a particle number

operator N that is conserved (commutes with the Hamiltonian) and generates

a U (1) symmetry.

in three dimensions, find the classical angular momentum functions Lj on

the phase space of complex valued functions on R3 . Show that these functions

Poisson-commute with the Hamiltonian function. Find the corresponding quan-

tized operators Lj , show that these commute with the Hamiltonian operator,

and satisfy the Lie algebra commutation relation

1, L

[L 2] = L

3

346

Problem 4: Show that the Lie algebra so(4, C) is sl(2, C)sl(2, C). Within

this Lie algebra, identify the sub-Lie algebras of the groups Spin(4), Spin(1, 3)

and Spin(2, 2).

Problem 1: Compute the following commutators of elements of the Lie algebra

of the Poincare group

[Kl , Pm ]

where the Kl (l = 1, 2, 3) generate boosts in the Lorentz subgroup, and Pm

(m = 0, 1, 2, 3) generate translations.

Problem 2:

For the real scalar field theory, find the momentum operator acting on

the quantum field theory state space, in terms of the scalar quantum

field. Show that this gives the expected expression in terms of a sum of

the number operators for each momentum mode and the corresponding

momentum.

Repeat the same calculation for the complex scalar field theory.

Problem 3: In class we showed that taking two real free scalar fields, one

could make a theory with SO(2) symmetry, and we found the charge operator Q

that gives the action of the Lie algebra of SO(2) on the state space of this theory.

Instead, consider two complex free scalar fields, and show that this theory has

a U (2) symmetry. Find the four operators that give the Lie algebra action for

this symmetry on the state space, in terms of a basis for the Lie algebra of U (2).

Note that this is the field content and symmetry of the Higgs sector of

the standard model (where the difference is that the theory is not free, but

interacting, and has a lowest energy state not invariant under the symmetry).

347

348

Bibliography

[1] Alvarez, O., Lectures on quantum mechanics and the index theorem, in

Geometry and Quantum Field Theory, Freed, D, and Uhlenbeck, K., eds.,

American Mathematical Society, 1995.

[2] Arnold, V., Mathematical Methods of Classical Mechanics, Springer-Verlag,

1978.

[3] Artin, M., Algebra, Prentice-Hall, 1991.

[4] Baym, G., Lectures on Quantum Mechanics, Benjamin, 1969.

[5] Berezin, F., and Marinov, M., Particle Spin Dynamics as the Grassmann

Variant of Classical Mechanics, Annals of Physics 104 (1972) 336-362.

[6] Berg, M., The Fourier-Analytic Proof of Quadratic Reciprocity, Wiley-

Interscience, 2000.

[7] Berndt, R., An Introduction to Symplectic Geometry, AMS, 2001.

[8] Berndt, R., Representations of Linear Groups, Vieweg, 2007.

[9] Das, A., Field Theory, a Path Integral Approach, World Scientific, 1993.

[10] De Faria, E. and De Melo, W., Mathematical Aspects of Quantum Field

Theory, Cambridge University Press, 2010.

[11] Dolgachev, I., Introduction to Physics, Lecture Notes

http://www.math.lsa.umich.edu/~idolga/lecturenotes.html

[12] Fadeev, L.D., Yakubovskii, O.A., Lectures on Quantum Mechanics for

Mathematics Students, AMS, 2009.

[13] Feynman, R., Space-time approach to non-relativistic quantum mechanics,

Reviews of Modern Physics, 20(1948) 367-387.

[14] Feynman, R. and Hibbs, A., Quantum Mechanics and Path Integrals,

McGraw-Hill, 1965.

[15] Feynman, R., Feynman Lectures on Physics,Volume 3, Addison-Wesley,

1965.

349

[16] Feynman, R., The Character of Physical Law, page 129, MIT Press, 1967.

[17] Feynman, R., Statistical Mechanics: A set of lectures, Benjamin, 1972.

[18] Folland, G., Harmonic Analysis in Phase Space, Princeton, 1989.

[19] Folland, G., Quantum Field Theory: A tourist guide for mathematicians,

AMS, 2008.

[20] Gendenshtein, L., and Krive, I., Supersymmetry in quantum mechanics,

Sov. Phys. Usp. 28 (1985) 645-666.

[21] Greiner, W. and Reinhardt, J., Field Quantization Springer, 1996.

[22] Guillemin, V., and Sternberg, S., Variations on a Theme of Kepler, AMS,

1990.

[23] Guillemin, V,., and Sternberg, S. Symplectic Techniques in Physics, Cam-

bridge University Press, 1984.

[24] Hall, B., Lie Groups, Lie Algebras, and Representations: An Elementary

Introduction, Springer-Verlag, 2003.

[25] Hall, B., An Elementary Introduction to Groups and Representations

http://arxiv.org/abs/math-ph/0005032

[26] Hall, B., An Introduction to Quantum Theory for Mathematicians, (to be

published).

[27] Hannabuss, K., An Introduction to Quantum Theory, Oxford University

Press, 1997.

[28] Haroche, S. and Ramond, J-M., Exploring the Quantum: Atoms, Cavities

and Photons, Oxford University Press, 2006.

[29] Hatfield, B., Quantum Field Theory of Point Particles and Strings,

Addison-Wesley, 1992.

[30] Kirillov, A., Lectures on the Orbit Method, AMS, 2004.

[31] Lawson, H. B., and Michelsohn, M-L., Spin Geometry, Princeton University

Press, 1989.

[32] Mumford, D., Tata Lectures on Theta III, Birkhauser, 1991.

[33] Peskin, M., and Schroeder, D., An Introduction to Quantum Field Theory,

Westview Press, 1995.

[34] Presskill, J., Quantum Computing Course Notes,

http://www.theory.caltech.edu/people/preskill/ph229/#lecture

[35] Porteous, I., Clifford Algebras and the Classical Groups Cambridge Univer-

sity Press, 1995.

350

[36] Ramond, P., Group Theory: A physicists survey, Cambridge University

Press, 2010.

[37] Rosenberg, J., A Selective History of the Stone-von Neumann Theorem,

in Operator Algebras, Quantization and Noncommutative Geometry, AMS,

2004.

[38] Schlosshauer, M., Decoherence and the Quantum-to-Classical Transition,

Springer, 2007.

[39] Schulman, L., Techniques and Applications of Path Integration, John Wiley

and Sons, 1981.

[40] Shankar, R., Principles of Quantum Mechanics, 2nd Ed., Springer, 1994.

[41] Singer, S., Linearity, Symmetry, and Prediction in the Hydrogen Atom,

Springer-Verlag, 2005.

[42] Sternberg, S., Group Theory and Physics, Cambridge University Press,

1994.

[43] Stillwell, J., Naive Lie Theory Springer-Verlag, 2010.

http://www.springerlink.com/content/978-0-387-78214-0

[44] Strichartz, R., A Guide to Distribution Theory and Fourier Transforms,

World Scientific, 2003.

[45] Takhtajan, L., Quantum Mechanics for Mathematicians, AMS, 2008.

[46] Taylor, M., Noncommutative Harmonic Analysis, AMS, 1986. Benjamin,

1969.

[47] Teleman, C., Representation Theory Course Notes,

math.berkeley.edu/~teleman/math/RepThry.pdf

[48] Tung, W-K., Group Theory in Physics, World Scientific Publishing 1985.

[49] Warner, F., Foundations of Differentiable Manifolds and Lie Groups

Springer-Verlag, 1983.

[50] Weil, A., Sur certain groupes doperateurs unitaires, Acta. Math. 111(1964)

143-211.

[51] Weinberg, S., The Quantum Theory of Fields I, Cambridge University

Press, 1995.

[52] Weyl, H., The Theory of Groups and Quantum Mechanics, Dover, 1950.

(First edition in German, 1929)

[53] Wigner, E., The Unreasonable Effectiveness of Mathematics in the Natural

Sciences, Communications in Pure and Applied Mathematics, vol. 13, No.

I (February 1960).

351

[54] Witten, E., Supersymmetry and Morse Theory, J. Differential Geometry,

17(1982) 661-692.

[55] Woodhouse, N. M. J., Special Relativity, Springer-Verlag, 2002.

[56] Zee, A., Quantum Field Theory in a Nutshell, Princeton University Press,

2003.

[57] Zinn-Justin, J., Path Integrals in Quantum Mechanics, Oxford University

Press, 2005.

[58] Zurek, W., Decoherence and the Transition from Quantum to Classical

Revisited, http://arxiv.org/abs/quant-ph/0306072

352

- Dhaka University EMBA Hand BookDiunggah olehnaimenim
- Abstract Algebra Cheat SheetDiunggah olehlpauling
- Basic Theory of Affine Group SchemesDiunggah olehVincent Lin
- Mnl Dec2013 Crc 18thjan14Diunggah olehMukta Deokar
- Linear Algebra (Math1014) Course OutlineDiunggah olehncisisthebest
- F5 Maths Annual Scheme of Work_2010Diunggah olehFikrah Imayu
- QuTiP-2.0.0-DOCDiunggah olehstellasll
- ContextualDiunggah olehjosemariaxx
- fisika komputasi matlabDiunggah olehRizki Kurniawan Hadi
- GroupDiunggah olehlatec
- BSc Computer Science Www.alagappauniversity.ac.in Alagappa University Model Question Paper Dist (1)Diunggah olehradhasubbu
- Magic SquaresDiunggah olehbabugaru
- Operations on Matrices - Linear AlgebraDiunggah olehdocsdownforfree
- Aimsun Macroscopic Modelling Manual v6_1Diunggah olehjuan0001
- Trident 360-Degree Matirx System, Jan 2012 DentistryDiunggah olehCarolyn Wilkinson
- JEE Mathematics SyllabusDiunggah olehVimarsh Srivastava
- 3P2Lb.pdfDiunggah olehSauravSachidanand
- Science (1)Diunggah olehNistoroschi Giorgiana
- 0702842.pdfDiunggah olehShubham Girdhar
- MIT2_019S11_SK4Diunggah olehRuly Irawan