CONCEPTS OF ELEMENTARY PARTICLE PHYSICS

OXFORD MASTER SERIES IN PHYSICS

The Oxford Master Series is designed for final year undergraduate and beginning graduate students in physics and
related disciplines. It has been driven by a perceived gap in the literature today. While basic undergraduate physics texts
often show little or no connection with the huge explosion of research over the last two decades, more advanced and
specialized texts tend to be rather daunting for students. In this series, all topics and their consequences are treated at a
simple level, while pointers to recent developments are provided at various stages. The emphasis is on clear physical
principles like symmetry, quantum mechanics, and electromagnetism which underlie the whole of physics. At the
same time, the subjects are related to real measurements and to the experimental techniques and devices currently
used by physicists in academe and industry. Books in this series are written as course books, and include ample
tutorial material, examples, illustrations, revision points, and problem sets. They can likewise be used as preparation
for students starting a doctorate in physics and related fields, or for recent graduates starting research in one of these
fields in industry.

CONDENSED MATTER PHYSICS

1. M.T. Dove: Structure and dynamics: an atomic view of materials
2. J. Singleton: Band theory and electronic properties of solids
3. A.M. Fox: Optical properties of solids, second edition
4. S.J. Blundell: Magnetism in condensed matter
5. J.F. Annett: Superconductivity, superfluids, and condensates
6. R.A.L. Jones: Soft condensed matter
17. S. Tautz: Surfaces of condensed matter
18. H. Bruus: Theoretical microfluidics
19. C.L. Dennis, J.F. Gregg: The art of spintronics: an introduction
21. T.T. Heikkilä: The physics of nanoelectronics: transport and fluctuation phenomena at low temperatures
22. M. Geoghegan, G. Hadziioannou: Polymer electronics

ATOMIC, OPTICAL, AND LASER PHYSICS

7. C.J. Foot: Atomic physics
8. G.A. Brooker: Modern classical optics
9. S.M. Hooker, C.E. Webb: Laser physics
15. A.M. Fox: Quantum optics: an introduction
16. S.M. Barnett: Quantum information
23. P. Blood: Quantum confined laser devices

PARTICLE PHYSICS, ASTROPHYSICS, AND COSMOLOGY

10. D.H. Perkins: Particle astrophysics, second edition
11. Ta-Pei Cheng: Relativity, gravitation and cosmology, second edition
24. G. Barr, R. Devenish, R. Walczak, T. Weidberg: Particle physics in the LHC era
26. M. E. Peskin: Concepts of elementary particle physics

STATISTICAL, COMPUTATIONAL, AND THEORETICAL PHYSICS

12. M. Maggiore: A modern introduction to quantum field theory
13. W. Krauth: Statistical mechanics: algorithms and computations
14. J.P. Sethna: Statistical mechanics: entropy, order parameters, and complexity, second edition
20. S.N. Dorogovtsev: Lectures on complex networks
25. R. Soto: Kinetic theory and transport phenomena
27. M. Maggiore: A modern introduction to classical electrodynamics
Concepts of Elementary Particle
Physics

Michael E. Peskin
SLAC National Accelerator Laboratory, Stanford University

1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Michael E. Peskin 2019
The moral rights of the author have been asserted
First Edition published in 2019
Reprinted with corrections in 2023

All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2019930484
Data available
ISBN 978–0–19–881218–0 (hbk.)
ISBN 978–0–19–881219–7 (pbk.)
DOI: 10.1093/oso/9780198812180.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Preface

This is a textbook of elementary particle physics, intended for students

who have a secure knowledge of special relativity and have completed
an undergraduate course in quantum mechanics.
Particle physics has now reached the end of a major stage in its de-
velopment. The primary forces that act within the atomic nucleus, the
strong and weak interactions, now have a fundamental description, with
equations that are similar in form to Maxwell’s equations. These forces
are summarized in a compact mathematical description, called the Stan-
dard Model of particle physics. The purpose of this book is to explain
what the Standard Model is and how its various ingredients are required
by the results of elementary particle experiments.
Increasingly, there is a gap between the study of elementary particles
and other areas of physical science. While other areas of physics seem to
apply directly to materials science, modern electronics, and even biology,
particle physics describes an increasingly remote regime of very small
distances. Physicists in other areas are put o↵by the sheer size and
expense of elementary particle experiments, and by the esoteric terms by
which particle physicists explain themselves. Particle physics is bound
up with relativistic quantum field theory, a highly technical subject, and
this adds to the difficulty of understanding it.
Still, there is much to appreciate in particle physics if it can be made
accessible. Particle physics contains ideas of great beauty. It reveals
some of the most deep and surprising ideas in physics through direct
connections between theory and experimental results. In this textbook,
I attempt to present particle physics and the Standard Model in a way
that brings the key ideas forward. I hope that it will give students an
entryway into this subject, and will help others gain a better under-
standing of the intellectual value of our recent discoveries.
The presentation of elementary particle physics in this book has been
shaped by many years of discussion with experimental and theoretical
physicists. Particle physicists form a global community that brings to-
gether many di↵erent points of view and di↵erent national styles. This
diversity has been a key source of new ideas that have driven the field
forward. It has also been a source of intuitive pictures that make it pos-
sible to visualize physical processes in the distant and abstract domain of
the subnuclear forces. I have tried to bring as many of these pictures as
possible into my discussion here. My own way of thinking about particle
physics has been shaped by my connection with the great laboratories
at Cornell University and SLAC. I am indebted to many colleagues at
vi Preface

these laboratories for central parts of the development given here.

I have been reminded often during the writing of this book that many
of the great figures responsible for the formulation of the Standard Model
have passed on to that symposium in the beyond. In only the past few
years, we have lost Sidney Drell, Martin Perl, Richard Taylor, Kenneth
Wilson, and, most recently, Burton Richter. All of these people influ-
enced me personally and profoundly a↵ected my thinking about particle
physics. It is a challenge for us who follow them not only to finish their
work but also to open new chapters in the development of fundamental
physics. I hope that this book will provide useful background for those
who wish to do so.
The core of this presentation was developed as a set of lectures for
CERN summer students in 1997; I thank Luis Alvarez-Gaumé for the
invitation to present these lectures. I have presented parts of this mate-
rial at a number of summer schools and courses, in particular, the course
on elementary physics at the Perimeter Scholars International program
at the Perimeter Institute. Most recently, I have polished this material
by my teaching of the course Physics 152/252 at Stanford University. I
am grateful to Patricia Burchat for giving me this opportunity, and for
much advice on teaching a course at this level. I thank the students in all
of these courses for their patience with preliminary versions of this book
and their attention to errors they contained. I thank Sonke Adlung, Har-
riet Konishi, Sal Moore, and their team at Oxford University Press for
their interest in this project. I thank Roy Brener, Tim Cohen, Serge Den-
das, Caterina Doglioni, Christopher Hill, Andre Hoang, Sunghoon Jung,
Andrew Larkoski, Aaron Pierce, Daniel Schroeder, Bruce Schumm, and
André David Tinoco for valuable comments on the presentation, and
Jongmin Yoon for an especially careful reading of the manuscript. Most
of all, I thank my colleagues in the SLAC Theory Group for their ad-
vice and criticism that has benefited my understanding of elementary
particle physics.

Michael E. Peskin
Sunnyvale, CA
August, 2018
Figure permissions

We are grateful to the following colleagues, organizations, and publish-

ers, who retain the original copyright, for permission to include figures
in this book. The figures are labeled below by our figure numbers. The
precise citation for each figure is given in the figure caption.
American Physical Society. Figures 5.3, 5.4 (left and right), 9.2, 9.3,
15.2, 15.3, 19.2, 19.4, 19.6, 20.4, 20.5, and 20.6 are reproduced from
Physical Review Letters. Figures 5.6, 10.6, 10.7, 13.7, 13.8, 20.2, 21.8,
and the figure in Exercise 7.2(m) are reproduced from Physical Review
D. Figures 5.8 and 5.9 are reproduced from Reviews of Modern Physics.
Annual Reviews. Figure 4.1 is reproduced from the Annual Review of
Nuclear and Particle Science.
John Campbell. Figure 13.5 is used by permission of the author.
CDF Collaboration. Figures 13.2, 13.3, and 13.4, which are screen shots
from the CDF trigger, are used by permission of the collaboration.
CERN. The event displays on the cover and Figures 6.5, 6.6, 13.9, 17.1,
17.2, 21.1, 21.3, 21.5, 21.7, 21.9, and 21.10 are reproduced from materials
and reports published on the CERN web site and are used by permission
of the CERN Press Office.
CKMfitter Collaboration. Figure 19.5 is reproduced from the web site
of the CKMfitter Collaboration and is used by permission of the collab-
oration.
Elsevier. Figure 15.1 is reproduced from Nuclear Physics A. Figure 12.1
is reproduced from Nuclear Physics B. Figures 10.2, 10.5, 19.1, 19.3,
21.4, 21.6, and the figure in Exercise 7.2(l) are reproduced from Physics
Letters B. Figures 17.5 and 17.8 are reproduced from Physics Reports.
Macmillan. Figure 5.1 is reproduced from Nature, c 1947, and is used
by permission of Macmillan Publisher Ltd.
Lawrence Berkeley Laboratory. Fig. 5.2, c 2014, is reproduced by
permission of the regents of the University of California through the
Lawrence Berkeley National Laboratory
Kevin McFarland. Figures 15.4, from the web site of Kevin McFarland,
is used by his permission and that of the NuTeV Collaboration.
Particle Data Group. Figures 6.1, 6.2, 6.3, 6.4, 8.1, 9.5, 10.1, 11.3, 13.1,
and 21.2 are taken from editions of the Review of Particle Physics and
are used by permission of the Particle Data Group.
viii Figure permissions

Physics Institute of the University of Bonn. Figures 5.5 and 5.7 are re-
produced from the proceedings of the 1981 International Symposium on
Lepton and Photon Internations at High Energy and used by permission
of the institute.
Toshinori Mori. Figure 17.4 is used by permission of the author.
Aldo Serenelli. Figure 20.3 is used by permission of the author.
SLAC. Figures 6.7, 8.2, 9.1, 10.3, and 10.4 are reproduced from the
web site of the SLAC National Accelerator Laboratory and are used by
permission of the laboratory achivist.
Springer. Figures 10.9, 12.2, 12.3, 17.3, 17.6, and 20.1 are reproduced
from the European Journal of Physics C. Figures 8.3 and 15.5 are re-
produced from Zeitschrift für Physik C.
Taylor & Francis. Figures 9.4 and 11.2 are reproduced from An In-
troduction to Quantum Field Theory, by Michael E. Peskin and Daniel
V. Schroeder, c 1995, by permission of Westview Press, an imprint of
Perseus Books, LLC. This book was recently acquired by the Taylor &
Francis Group, a subsidiary of Informa UK Ltd. and is now published
under the imprint of CRC Press.
ZEUS Collaboration. Figure 9.6, from the ZEUS Collaboration web site,
is used by permission of the collaboration.
Contents

I Preliminaries and Tools 1

1 Introduction 3

2 Symmetries of Space-Time 9
2.1 Relativistic particle kinematics 10
2.2 Natural units 12
2.3 A little theory of discrete groups 14
2.4 A little theory of continuous groups 18
2.5 Discrete space-time symmetries 22
Exercises 23

3 Relativistic Wave Equations 25

3.1 The Klein-Gordon equation 25
3.2 Fields and particles 28
3.3 Maxwell’s equations 29
3.4 The Dirac equation 32
3.5 Relativistic normalization of states 37
3.6 Spin and statistics 39
Exercises 40

4 The Hydrogen Atom and Positronium 43

4.1 The ideal hydrogen atom 43
4.2 Fine structure and hyperfine structure 44
4.3 Positronium 48
Exercises 51

5 The Quark Model 55

5.1 The discovery of the hadrons 55
5.2 Charmonium 58
5.3 The light mesons 65
5.4 The heavy mesons 68
5.5 The baryons 69
Exercises 74

6 Detectors of Elementary Particles 77

6.1 Energy loss by ionization 78
6.2 Electromagnetic showers 81
6.3 Further e↵ects of nuclear scattering 85
6.4 Energy loss through macroscopic properties of the medium 87
x Contents

6.5 Detector systems for collider physics 87

Exercises 93

7 Tools for Calculation 95

7.1 Observables in particle experiments 95
7.2 Master formulae for partial width and cross sections 97
7.3 Phase space 99
7.4 Example: ⇡ + ⇡ scattering at the ⇢ resonance 101
Exercises 106

II The Strong Interaction 111

8 Electron-Positron Annihilation 113
8.1 The reaction e+ e ! µ+ µ 113
8.2 Properties of massless spin- 12 fermions 115
8.3 Evaluation of the matrix elements for e+ e ! µ+ µ 118
8.4 Evaluation of the cross section for e+ e ! µ+ µ 121
8.5 e+ e annihilation to hadrons 122
Exercises 127

9 Deep Inelastic Electron Scattering 129

9.1 The SLAC-MIT experiment 129
9.2 The parton model 133
9.3 Crossing symmetry 134
9.4 Cross section for electron-quark scattering 139
9.5 The cross section for deep inelastic scattering 141
9.6 Bjorken scaling 144
Exercises 146

10 The Gluon 149

10.1 Measurement of parton distribution functions 149
10.2 Photon emission in e+ e ! qq 155
10.3 Three-jet events in e+ e annihilation 160
10.4 E↵ects of gluon emission on pdfs 164
Exercises 168

11 Quantum Chromodynamics 169

11.1 Lagrangian dynamics and gauge invariance 169
11.2 More about Lie groups 171
11.3 Non-Abelian gauge symmetry 173
11.4 Formulation of QCD 176
11.5 Gluon emission in QCD 176
11.6 Vacuum polarization 177
11.7 Asymptotic freedom 179
Exercises 184

12 Partons and Jets 187

12.1 Altarelli-Parisi evolution of parton distribution functions 187
 Contents xi

12.2 The structure of jets 191

Exercises 197

13 QCD at Hadron Colliders 199

13.1 Hadron scattering at low momentum transfer 199
13.2 Hadron scattering at large momentum transfer 204
13.3 Jet structure observables for hadron collisions 208
13.4 The width of a jet in hadron-hadron collisions 209
13.5 Production of the top quark 213
Exercises 215

14 Chiral Symmetry 217

14.1 Symmetries of QCD with zero quark masses 217
14.2 Spontaneous symmetry breaking 219
14.3 Goldstone bosons 223
14.4 Properties of ⇡ mesons as Goldstone bosons 224
Exercises 228

III The Weak Interaction 231

15 The Current-Current Model of the Weak Interaction 233
15.1 Development of the V A theory of the weak interaction 234
15.2 Predictions of the V A theory for leptons 235
15.3 Predictions of the V A theory for pion decay 243
15.4 Predictions of the V A theory for neutrino scattering 245
Exercises 248

16 Gauge Theories with Spontaneous Symmetry Breaking 251

16.1 Field equations for a massive photon 251
16.2 Model field equations with a non-Abelian gauge symmetry 253
16.3 The Glashow-Salam-Weinberg electroweak model 255
16.4 The neutral current weak interaction 260
Exercises 263

17 The W and Z Bosons 265

17.1 Properties of the W boson 265
17.2 W production in pp collisions 269
17.3 Properties of the Z boson 270
17.4 Precision tests of the electroweak model 271
Exercises 280

18 Quark Mixing Angles and Weak Decays 283

18.1 The Cabibbo mixing angle 283
18.2 Quark and lepton mass terms in the SU (2) ⇥ U (1) model 285
18.3 Discrete space-time symmetries and generation mixing 287
18.4 The Standard Model of particle physics 289
18.5 Quark mixing including heavy quarks 290
Exercises 293
xii Contents

19 CP Violation 295
19.1 CP violation in the K 0 –K 0 system 295
19.2 Electric dipole moments 302
19.3 CP violation in the B 0 –B 0 system 303
Exercises 310

20 Neutrino Masses and Mixings 313

20.1 Neutrino mass and decay 313
20.2 Adding neutrino mass to the Standard Model 315
20.3 Measurements of neutrino flavor mixing 319
Exercises 324

21 The Higgs Boson 327

21.1 Constraints on the Higgs field from the weak interaction 327
21.2 Expected properties of the Higgs boson 329
21.3 Measurements of Higgs boson properties at the LHC 332
Exercises 340

IV Epilogue 343
22 Epilogue 345

A Notation 355

B Conversion factors and physical constants 359

C Formulae for the creation and destruction of elementary

particles 361

D Master formulae for the computation of cross sections

and partial widths 363

E QCD formulae for hadron collisions 365

References 367

Index 377
 Part I

Preliminaries and Tools

Introduction
1
The aim of this book is to describe the interactions of nature that act
on elementary particles at distances of the size of an atomic nucleus.
At this time, physicists know about four distinct fundamental inter-
actions. Two of these are macroscopic—gravity and electromagnetism.
Gravity has been known since the beginning of history and has been un-
derstood quantitatively since the time of Newton. Electrical and mag-
netic phenomena have also been known since ancient times. The unified
theory of electromagnetism was given its definitive form by Maxwell in
1865. Through all of these developments, there was no sign that there
could be additional fundamental forces. These would appear only when
physicists could probe matter at very small distances.
The first evidence for additional interactions of nature was Becquerel’s
discovery of radioactivity in 1896. In 1911, Rutherford discovered that
the atom consists of electrons surrounding a very tiny, positively charged
nucleus. As physicists learned more about atomic structure, it became
increasingly clear that the known macroscopic forces of nature could not
give the full explanation. By the middle of the 20th century, experiments
had revealed a series of questions that could not be resolved without new
particles and interactions. These included: These simple questions give the starting
point for the exploration of subnuclear
• What is radioactivity? Why do some atomic nuclei emit high- physics.
energy particles? What specific reactions are responsible? What
are the particles that are emitted in radioactive decay?
• What holds the atomic nucleus together? The nucleus is made of
positively charged protons and neutral neutrons. Electromagnetic
forces destabilize the nucleus—as we see from the fact that heavy
nuclei are unstable with respect to fission. What is the counter-
balancing attractive force?
• What are protons and neutrons made of? These particles have
properties that indicate that they are not elementary pointlike
particles. What gives them structure? What kinds of particles are
inside?

Experiments designed to study these issues produced more confusion

before they produced more understanding. The proton and the neutron
turned out to be the first of hundreds of particles interacting through

Concepts of Elementary Particle Physics. Michael E. Peskin.

c Michael E. Peskin 2019. Published in 2019 by Oxford University Press.
DOI: 10.1093/oso/9780198812180.001.0001
4 Introduction

the nuclear force. The electron turned out to be only one of three appar-
ently pointlike particles with electric charge but no strong interactions.
All of these particles were observed to interact with one another through
a web of new, short-ranged interactions. Finally, as the 1960’s turned to
the 1970’s, the new interactions were sorted into two basic forces—called
the strong and the weak interaction—and simple mathematical expres-
sions for these forces were constructed. Today, physicists refer to these
expressions collectively as “the Standard Model of particle physics”.
Sometimes, authors or lecturers present the table of elementary par-
ticles of the Standard Model and imply that this is all there is to the
story. It is not. The way that the forces of nature act on the elementary
particles is beautiful and intricate. Often, the telling details of these
interactions show up through remarkable aspects of the data when we
examine elementary particle behavior experimentally.
These ideas elicit a related question: Of all the ways that nature could
be built, how do we know that the Standard Model is the correct one?
It seems hardly possible that we could pin down the exact nature of new
fundamental interactions beyond gravity and electromagnetism. All of
the phenomena associated with the new forces occur at distances smaller
than an atomic nucleus, and in a regime where both special relativity
It is important to remember that the and quantum mechanics play an essential role.
theory of particle physics must be stud- In this book, I will explain the answers to these questions. It turns
ied together with the understanding
of how experiments are done and how
out that the new forces have common properties and can be built up
their results are interpreted. from simple ingredients. The presence of these ingredients is revealed
by well-chosen experiments. The dynamics of the new interactions be-
comes more clear at higher energies. With the benefit of hindsight, we
can begin our study today by studying these dynamical ingredients in
their simplest form, working out the consequences of these laws, and
comparing the resulting formulae to data from high energy accelerator
experiments that illustrate the correctness of these formulae in a very
direct way.
Our quest for a fundamental theory of nature is far from complete.
In the final chapter of the book, I will discuss a number of issues about
fundamental forces for which we still have no understanding. It is also
possible, as we probe more deeply into the structure of nature, that we
will uncover new interactions that work at even smaller distances than
those currently explored. But, at least, one chapter of the story, open
since 1896, is now finished. I hope that, working through this book,
you will not only understand how to work with the underlying theories
describing the strong and weak interactions, but also that you will be
amazed at the wealth of evidence that supports the connection of these
theories to the real world.
Outline of the book. The book is organized into three Parts. Part I introduces the basic
materials that we will use to probe the nature of new forces at short
distances. Parts II and III use this as a foundation to build up the
Standard Model theories of the strong and weak interactions.
Part I Part I begins with basic theory that underlies the subject of particle
physics. Even before we attempt to write theories of the subnuclear
 5

forces, we expect that those theories will obey the laws of quantum
mechanics and special relativity. I will provide some methods for using
these important principles to make predictions about the outcome of
elementary particle collisions.
In addition, I will describe the types of matter in the theories of
strong and weak interactions, the basic elementary particles that in-
teract through these forces. It turns out that there are two types of
matter particles that are elementary at the level of our current under-
standing. Of these, one type, the leptons, are seen in our experiments as particles of the Standard Model: the
individual particles. There are six known leptons. Three have electric leptons
charge: the electron (e), the muon (µ), and the tau lepton (⌧ ). The
other three are the neutrinos, particles that are electrically neutral and
extremely weakly interacting. Despite this, the evidence for neutrinos
as ordinary relativistic particles is very persuasive; I will discuss this in
Part III.
Matter particles of the other type, the quarks, are hidden from view.
Quarks appear as constituents of particles such as protons and neutrons particles of the Standard Model: the
that interact through the strong interaction. There are many known quarks
strongly interacting particles, collectively called hadrons. I will explain
the properties of the most prominent ones, and show that they are nat-
urally considered in families. On the other hand, no experiment has
ever seen an isolated quark. It is actually a prediction of the Standard
Model that quarks can never appear singly. This makes it especially
challenging to learn their properties. One piece of evidence that the
description of quarks in the Standard Model is correct is found from the
fact it gives a simple explanation for the quantum numbers of observed
hadrons and their assortment into families. I will discuss this also in
Part I. In the process, I will give names to the hadrons that appear
most often in experiments, so that we can discuss experimental methods
more concretely.
In a relativistic quantum theory, forces are also associated with parti-
cles that can be thought to transmit them. The Standard Model contains
four types of such particles. These are the photon, the carrier of the elec- particles of the Standard Model: the
tromagnetic interaction, the gluon, the carriers of the strong interaction, bosons
the W and Z bosons, the carriers of the weak interaction, and the Higgs
boson, which plays a more subtle role. You will have already encoun-
tered the photon in your study of quantum mechanics. I will introduce
the gluon in Part II and the W , Z, and Higgs bosons in Part III.
To understand experimental findings about elementary particles, we
will need to know at least the basics of how experiments on elementary
particles are done, and what sorts of quantities describing their proper-
ties are measureable. I will discuss this material also in Part I.
Part II begins with a discussion of the most important experiments Part II
that give insight into the underlying character of the strong interaction.
One might guess intuitively that the most convincing data on the strong
interaction comes from the study of collisions of hadrons with other
hadrons. That is incorrect. The experiments that were most crucial in
understanding the nature of strong interaction involved electron scatter-
6 Introduction

ing from protons and the annihilation of electrons and positrons at high
energy. This latter process has an initial state with no hadrons at all.
I will begin Part II with a discussion of the features of these processes
at high energy. Our analysis will introduce the concept of the current-
current interaction, which is an essential part of the physics of both the
strong and weak interactions. Starting from the surprisingly simple fea-
tures of these electromagnetic probes of strongly interacting particles,
we will develop a series of arguments that pass back and forth between
theory and experiment and eventually lead to a unique proposal for the
underlying theory of the strong interaction. We will then describe tests
of this theory of increasing sophistication, finally encompassing recent
results from experiments at the Large Hadron Collider.
The final chapter of Part II presents our current understanding of the
masses of quarks. At first sight, it might seems that it is straightforward
to measure the mass of a quark, although the fact that quarks cannot
be observed individually makes this more difficult. In fact, it turns
out that the question of the quark masses brings in a number of new,
subtle concepts. In particular, we will need to understand the idea of
spontaneous symmetry breaking, a concept that will also prove to be an
essential part of the theory of the weak interaction.
Part III Part III presents the description of the weak interaction. Here I
will begin from a proposal for the nature of the weak interaction that
again is based on a current-current interaction. I will present some
quite counterintuitive, and even startling, predictions of that proposal
and show that they are actually reproduced by experiment. From this
starting point, again in dialogue between theory and experiment, we
will build up the full theory. My discussion will include the precision
study of the carriers of the weak interaction, the W and Z bosons, and
the newest ingredients in this theory, the masses of neutrinos and the
properties of the Higgs boson.
In comparing theory and experiment, I will generally present deriva-
tions of the formulae that give the most important theoretical predic-
tions. To fully understand this book, it is important that you work
through these derivations rather than just skimming the mathematics.
For each one, take a large sheet of blank paper and carry out the cal-
culation yourself. The details of the calculation will reveal insights that
cannot by obtained from a purely qualitative discussion. You will find
that these insights accumulate as you go through the book, as successive
derivations cover similar ground from new perspectives.
Still, these derivations will be simplified with respect to a complete
treatment of the Standard Model of particle physics. Most of the pro-
cesses that I will consider will be studied in the limit of very high ener-
gies, where the mathematical analysis can be reduced as much as possible
and made more transparent. This will be sufficient to discover the equa-
tions of the Standard Model and understand their basic tests. But a full
discussion of the implications of the Standard Model would cover a more
complete list of reactions, including some whose theoretical analysis is
quite complex. A treatment of particle physics at that level is beyond
 7

the scope of this book.

In particular, many aspects of the theory of elementary particles can-
not be understood without a deep understanding of quantum field the-
ory. This book will explain those aspects of quantum field theory that
are absolutely necessary for the presentation, but will omit any sophis-
ticated discussion of this subject. A full description of the properties of
elementary particles needs more.
For students who would like to study further in particle physics, there
are many excellent references written from di↵erent and complementary
points of view. I have put a list of the most useful texts at the beginning
of the References.
A particularly useful reference work is the Review of Particle Physics
assembled by the Particle Data Group (Patrignani et al. 2016). This
volume compiles the basic properties of all known elementary particles
and provides up-to-date reviews of the major topics in this subject. All
elementary particle masses and other physical quantities quoted in this
book but not explicitly referenced are taken from the summary tables
given in that source.
Symmetries of Space-Time
2
We do not have complete freedom in postulating new laws of nature. Any
laws that we postulate should be consistent with well-established symme-
tries and invariance principles. On distance scales smaller than an atom,
space-time is invariant with respect to translations of space and time.
Space-time is also invariant with respect to rotations and boosts, the
symmetry transformations of special relativity. Many aspects of exper-
iments on elementary particles test the principles of energy-momentum
conservation, rotational invariance, the constancy of the speed of light,
and the special-relativity relation of mass, momentum, and energy. So
far, no discrepancy has been seen. So it makes sense to apply these pow-
erful constraints to any proposal for elementary particle interactions.
Perhaps you consider this statement too strong. As we explore new
realms in physics, we might well discover that the basic principles applied
in more familiar settings are no longer valid. In the early 20th century,
real crises brought on by the understanding of atoms and light forced
physicists to abandon Newtonian space-time in favor of that of Einstein
and Minkowski, and to abandon the principles of classical mechanics in
favor of the very di↵erent tools of quantum mechanics. By setting rela-
tivity and quantum mechanics as absolute principles to be respected in
the subnuclear world, we are making a conservative choice of orienta-
tion. There have been many suggestions of more radical approaches to
formulating laws of elementary particles. Some of these have even led to
new insights: The bootstrap of Geo↵rey Chew, in which there is no fun-
damental Hamiltonian, is still finding new applications in quantum field
theory (Simmons-Duffin 2017); string theory, which radically modifies
space-time structure, is a candidate for the overall unification of parti-
cle interactions with quantum gravity (Zwiebach 2004, Polchinski 2005).
However, the most successful routes to the theory of subnuclear interac-
tions have taken translation invariance, special relativity, and standard
quantum mechanics as absolutes. In this book, I will make the assump-
tion that special relativity and quantum mechanics are correct in the
realm of elementary particle interactions, and I will use their principles
in a strong way to organize my exploration of elementary particle forces.
This being so, it will be useful to formulate the constraints from space-
time symmetries in such a way that we can apply them easily. We
would like to use the actual transformation laws associated with these

Concepts of Elementary Particle Physics. Michael E. Peskin.

c Michael E. Peskin 2019. Published in 2019 by Oxford University Press.
DOI: 10.1093/oso/9780198812180.001.0001
10 Symmetries of Space-Time

symmetries as little as possible. Instead, we should formulate questions

in such a way that the answers are expressions invariant under space-
time symmetries. Generally, there will be a small and well-constrained
set of possible invariants. If we are lucky, only one of these will be
consistent with experiment.

2.1 Relativistic particle kinematics

As a first step in simplifying the use of constraints from special relativity,
I will discuss the kinematics of particle interactions. Any isolated parti-
cle is characterized by an energy and a vector momentum. In special rel-
ativity, these are unified into a 4-vector. I will write energy-momentum
Representation of the energy and mo- 4-vectors in energy units and notate them with an index µ = 0, 1, 2, 3,
mentum of a particle in 4-vector nota-
tion. pµ = (E,~pc)µ . (2.1)

I will now review aspects of the formalism of special relativity. Prob-

ably you have seen these formulae before in terms of rulers, clocks, and
moving trains. Now we will need to use them in earnest, because elemen-
tary particle collisions generally occur at energies at which it is essential
to use relativistic formulae.
Under a boost by v along the 3̂ direction, the energy-momentum
4-vector transforms as p ! p0 , with
1 v 1 v
E0 = p (E + p3 c) ,p30 c = p (p3 c + E) ,
1 v 2 /c2 c 1 v 2 /c2 c
1,20 1,2
p c=p c. (2.2)

It is convenient to write this as a matrix transformation

0 1
0 0
B 0 1 0 0 C
p0 = ⇤p with⇤= @ A , (2.3)
0 0 1 0
0 0
where
v 1
= =p . (2.4)
c 1 2

In multiplying matrices and vectors in this book, I will use the con-
vention that repeated indices are summed over. Then, for example, I
will write (2.3) as
p0µ = ⇤µ ⌫ p⌫ , (2.5)
In this book, unless it is explicitly in- omitting the explicit summation sign for the index ⌫. Lorentz trans-
dicated otherwise, repeated indices are formations leave invariant the Minkowski space vector product
summed over. This convention is one
of Einstein’s lesser, but still much ap-
preciated, innovations.
p · q = Ep E q p~ · ~q . (2.6)

To keep track of the minus sign in this product, I will make use of
raised and lowered Lorentz indices. Lorentz transformations preserve
 2.1 Relativistic particle kinematics 11

the metric tensor

0 1 0 1
1 0 0 0 1 0 0 0
B0 1 0 0 C B0 1 0 0 C
⌘µ⌫ = @ A , ⌘ µ⌫ =@ A . (2.7)
0 0 1 0 0 0 1 0
0 0 0 1 0 0 0 1
Using this matrix, and the summation convention, we can write (2.6) as

p · q = pµ ⌘µ⌫ q ⌫ . (2.8)

Alternatively, let q with a lowered index be defined by I will use raised and lowered Lorentz
indices to keep track of the minus
qµ = ⌘µ⌫ q ⌫ = (Eq , ~q )µ . (2.9) sign in the Minkowski vector prod-
uct. Please pay attention to the po-
sition of indices—raised or lowered—
The invariant product of p and q is written throughout this book.

p · q = pµ qµ . (2.10)

To form an invariant, we always combine a raised index with a lowered

index. As the equations in this book become more complex, we will
find this trick very useful in keeping track of the Minkowski space minus
signs.
A particularly important Lorentz invariant is the square of a momen-
tum 4-vector,
p · p ⌘ p2 = E 2 |~ p| 2 c 2 . (2.11)
Being an invariant, this quantity is independent of the state of motion
of the particle. In the rest frame

pµ = (E0 , ~0)µ . (2.12)

I will define the mass of a particle as its rest-frame energy The mass of a particle is a Lorentz-
invariant quantity that characterizes
(mc2 ) ⌘ E0 . (2.13) that particle in any reference frame.

Since p2 is an invariant, the expression

(mc2 )2 = p2 = E 2 p| 2 c 2
|~ (2.14)

is true in any frame of reference.

In this book, I will write particle momenta in two standard ways

pµ = (Ep , ~ pc
)µ or pµ = mc2 (1, ~ )µ , (2.15)

where Definitions of the quantities Ep , ,

associated with relativistic particle mo-
|~
p|c tion.
p|2 + (mc)2 )1/2 ,
Ep = c(|~ = , = (1 2
) 1/2
. (2.16)
Ep
Especially, the symbol Ep will always be used in this book to represent
this standard function of momentum and mass. I will refer to a 4-vector
with E = Ep as being “on the mass shell”.
12 Symmetries of Space-Time

To illustrate these conventions, I will now work out some simple but
important exercises in relativistic kinematics. Imagine that a particle of
mass M , at rest, decays to two lighter particles, of masses m1 and m2 . In
the simplest case, both particles have zero mass: m1 = m2 = 0. Then,
energy-momentum conservation dictates that the two particle energies
are equal, with the value M c2 /2. Then, if the final particles move in the
3̂ direction, we can write their 4-vectors as

pµ1 = (M c2 /2, 0, 0, M c2 /2)µ pµ2 = (M c2 /2, 0, 0, M c2 /2)µ . (2.17)

The next case, which will appear often in the experiments we will
consider, is that with m1 nonzero but m2 = 0. In the rest frame of the
original particle, the momenta of the two final particles will be equal
and opposite. With a little algebra, one can determine

pµ1 = (Ep , 0, 0, pc)µ , pµ2 = (pc, 0, 0, pc)µ (2.18)

These kinematic formulae will be used (for motion in the 3̂ direction), where
very often in this book.
M 2 + m21 2 M 2 m21 2
Ep = c , pc = c . (2.19)
2M 2M
It is easy to check that these formulae satisfy the constraints of to-
tal energy-momentum conservation and that pµ1 satisfies the mass-shell
constraint (2.14).
Finally, we might consider the general case of nonzero m1 and m2 .
Here, it takes a little more algebra to arrive at the final formulae

pµ1 = (E1 , 0, 0, pc)µ , pµ2 = (E2 , 0, 0, pc)µ (2.20)

with
M 2 + m21 m22 M2 m21 + m22 2
E1 = c2 , E2 = c , (2.21)
2M 2M
and
c
p= ( (M, m1 , m2 ))1/2 , (2.22)
2M
where the kinematic function is defined by

(M, m1 , m2 ) = M 4 2M 2 (m21 + m22 ) + (m21 m22 )2 . (2.23)

These three sets of formulae apply equally well to reactions with two
particles in the initial state and two particles in the final state. It is
only necessary to replace M c2 with the center of mass energy ECM of
the reaction.

2.2 Natural units

In the discussion of the previous chapter, I needed to introduce many
factors of c in order to make the treatment of energy, momentum, and
mass more uniform. This is a fact of life in the description of high
 2.2 Natural units 13

energy particles. Ideally, we should take advantage of the worldview of

relativity to pass seamlessly among these concepts. Equally well, our
discussions of particle dynamics will take place in a regime in which
quantum mechanics plays an essential role. To make the best use of
quantum concepts, we should be able to pass easily between the concepts
of momentum and wavenumber, or energy and frequency.
To make these transitions most easily, I will, in this book, adopt nat-
ural units,
h̄ = c = 1 . (2.24)
That is, I will measure momentum and mass in energy units, and I will
measure distances and times in inverse units of energy. For convenience The conventions that define natural
in discussing elementary particle physics, I will typically use the en- units.
ergy units MeV or GeV. This will eliminate a great deal of unnecessary
baggage that we would otherwise need to carry around in our formulae.
For example, to write the mass of the electron, I will write
27
not me = 0.91 ⇥ 10 g but rather me = 0.51 MeV . (2.25)

An electron with a momentum of the order of its rest energy has, ac-
cording to the Heisenberg uncertainty principle, a position uncertainty
h̄ 11
= 3.9 ⇥ 10 cm , (2.26)
me c
which I will equally well write as
1 1
= (0.51 MeV) . (2.27)
me
Natural units make it very intuitive to estimate energies, lengths, and
times in the regime of elementary particle physics. For example, the Natural units are useful for estimation.
lightest strongly interacting particle, the ⇡ meson, has a mass

m⇡ c2 = 140 MeV . (2.28)

This corresponds to a distance

h̄ 13
= 1.4 ⇥ 10 cm (2.29)
m⇡ c
and a time
h̄
= 0.47 ⇥ 10 22 sec . (2.30)
m⇡ c2
These give—within a factor 2 or so—the size of the proton and the The material in this book will be easier
lifetimes of typical unstable hadrons. So, the use of m⇡ gives a good to grasp if you make yourself comfort-
able with the use of natural units. This
first estimate of all dimensionful strong interaction quantities. To obtain will both simplify formulae and simplify
an estimate in the desired units—MeV, cm, sec—we would decorate many estimates of energies, distances,
the simple expression m⇡ with appropriate factors of h̄ and c and then and times.
evaluate as above.
It may make you uncomfortable at first to discard factors of h̄ and c.
Get used to it. That will make it much easier for you to perform calcu-
lations of the sort that we will do in this book. Some useful conversion
14 Symmetries of Space-Time

factors for moving between distance, time, and energy units are given in
Appendix B.
One interesting quantity to put into natural units is the strength of
the electric charge of the electron or proton. In this book, the constant
e will be a positive quantity equal to the electric charge of the proton.
The electric charge of the electron will be ( e). More generally, I will
write the charge of a particle as Qe, with Q = +1 for a proton and
Q = 1 for an electron.
The Coulomb potential between charges 1 and 2 is given in standard
notation by
Q1 Q2 e2
V (r) = . (2.31)
4⇡✏0 r
I will use units for electromagnetism in which also

✏ 0 = µ0 = 1 . (2.32)

Then the Coulomb potential reads

Q 1 Q2 e2 1
V (r) = . (2.33)
4⇡ r
Since r, in natural units, has the dimensions of (energy) 1 , the value of
the electric charge must have a form in which it is dimensionless. Indeed,

e2
↵⌘ (2.34)
4⇡✏0 h̄c
is a dimensionless number, called the fine structure constant, with the
value
↵ = 1 / 137.036 . (2.35)
The intrinsic strengths of the basic el- There are two remarkable things about this equation. First, it is sur-
ementary particle interactions are not prising that there is a dimensionless number ↵ that characterizes the
apparent from the size of their e↵ects—
or from their names. Here is a preview.
strength of the electromagnetic interaction. Second, that number is
small, signalling that the electromagnetic interaction is a weak inter-
Group theory plays an important role action. One of the goals of this book will be to determine whether the
in quantum mechanics, and this im- strong and weak subnuclear interactions can be characterized in the same
portance extends to the study of el-
ementary particle physics. You have
way, and whether these interactions—looking beyond their names—are
encountered group theory concepts in intrinsically strong or weak. I will discuss estimates of the strong and
your quantum mechanics course, but it weak interaction coupling strengths at appropriate points in the course.
is likely that those arguments did not It will turn out that the strong interaction is weak, at least when mea-
make explicit reference to group the-
ory. In particle physics, we lean much
sured under the correct conditions. It will also turn out that the weak
more heavily on group theory, and so interaction is also weak in dimensionless terms. It is weaker than the
it is best to discuss these concepts strong interactions, but not as weak as electromagnetism.
formally and give them their proper
names. Please, then, study Sections 2.3
and 2.4 carefully, especially if you are
uncomfortable with mathematical ab- 2.3 A little theory of discrete groups
straction. With careful reading, you
will see that the concepts I describe Group theory is a very important tool for elementary particle physics.
generalize physical arguments that are In this section and the next, I will review how group theory is used in
already familiar to you.
quantum mechanics, and I will discuss some properties of groups that we
 2.3 A little theory of discrete groups 15

will meet in this book. To do this, I will mainly discuss systems that you
have already studied in your quantum mechanics course, giving a new
description of these systems using a more formal and abstract mathe-
matical language,. Please do not be put o↵by this. The mathematical
terms and concepts that I will introduce will generalize to and, I hope,
illuminate, many new systems that we will study in this book.
In quantum mechanics, we deal with groups on two levels. First,
there are abstract groups. In mathematics, a group is a set of elements
G = {a, b, . . .} with a multiplication law defined, so that ab is defined and
is an element of G. The multiplication law satisfies the three properties Here are the axioms that define a
group.
(1) Multiplication is associative: a(bc) = (ab)c.
(2) G contains an identity element 1 such that, for any element of G,
1a = a1 = a.
(3) For each a in G, there is an inverse element a 1 in G such that
aa 1 = a 1 a = 1.
Every symmetry of nature normally encountered in physics satisfies these
axioms and is described by an abstract group.
Next, we need to relate the abstract group to operators that act on
the space of states in a quantum mechanics problem. A group trans-
formation is a symmetry of a quantum-mechanical system if it leaves
the dynamics of that system invariant. This will be true if the trans-
formations commute with the Hamiltonian H. It is useful to state this
relationship more precisely.
In quantum mechanics, the quantum states are vectors in a Hilbert
space. Symmetries act to convert one of these states into another. A
symmetry transformation carries each state into another one in such
as way that the whole Hilbert space is mapped into itself, preserving
norms, that is, preserving quantum mechanical probabilities. Thus it
must act as a unitary transformation of the Hilbert space. A group G
is then described in a quantum mechanics problem as a set of unitary
transformations {U (a)}, one for each element of G, that obey the mul-
tiplication law of G. That is, if a, b, c are elements of G and ab = c, the
transformations in the set should satisfy U(a) U(b) = U (c). Such a set
of unitary transformations is called a unitary representation of G. The action of a group on the Hilbert
The group G will be a symmetry of the quantum mechanics problem space of states in quantum mechanics is
described through unitary representa-
if, for all elements a of G, the corresponding unitary transformations tions of the group, that is, through uni-
commute with the Hamiltonian, tary matrices with the same multipli-
cation law as the corresponding group
[ U (a), H ] = 0 . (2.36) elements. Thus, unitary group repre-
sentations will be used in many aspects
A direct consequence of (2.36) is that, if | i is an eigenstate of H with of the physics discussed in this book.
energy E, U (a) | i will also be an eigenstate of H with the same energy.
The inverse operation to a will be represented by the inverse transfor-
mation: U (a 1 ) = U(a) 1 = U (a)† .
Without reference to quantum mechanics, but just thinking about the
pure mathematics of groups, we can ask the following question: Given
a group G with elements {a}, can we find a set of finite-dimensional
16 Symmetries of Space-Time

unitary matrices Ua that obey the multiplication law of the group, that
is, such that Ua Ub = Uc ? This is a classical mathematics problem, and
mathematicians have classified the possible answers for all of the groups
commonly encountered in physics, and for more complex examples, in-
cluding one called the “Monster” group. These finite-dimensional uni-
tary representions of groups will appear in quantum mechanics problems
as tranformations among a finite number of eigenstates of the Hamilto-
nian that implement a symmetry of the problem.
The theory of group representations is a deep subject that can become
quite technically complex. It is certainly not necessary to master this
subject in order to study elementary particle physics. In this book, we
will make use of only the simplest examples. However, it is often good to
recognize that a physics problem of interest involves the representations
of a relevant symmetry group. This gives a starting point to analyze
the problem and connects the solution to those of other problems with
which we might be more familiar.
Here is a simple example: Consider the abstract group called Z2 that
contains two elements {1, 1} satisfying the multiplication law

1 · 1 = ( 1)( 1) = 1 , 1 · ( 1) = ( 1) · 1 = ( 1) . (2.37)

We can represent Z2 in a quantum mechanics problem by 2 ⇥ 2 unitary

matrices. To make this concrete, consider a system with two particles
⇡ + and ⇡ . Define the operator C to transform
↵ ↵ ↵ ↵
C ⇡+ = ⇡ , C ⇡ = ⇡+ . (2.38)

The action of C on this 2-dimensional space is given by the matrix

✓ ◆ ✓ + ◆
0 1 |⇡ i
acting on . (2.39)
1 0 |⇡ i

If H is the Hamiltonian for this quantum-mechanical system and [C, H] =

0, that would imply that the masses and decay rates of ⇡ + and ⇡ must
be equal. On the same Hilbert space, we can define the trivial operation
↵ ↵ ↵ ↵
1 ⇡+ = ⇡+ , 1 ⇡ = ⇡ . (2.40)

This is given by
✓ ◆ ✓ ◆
1 0 |⇡ + i
acting on . (2.41)
0 1 |⇡ i

The unitary matrices {1, C} form a 2-dimensional unitary representation

of the group Z2 . If these matrices commute with H, we say that H has
Z2 symmetry.
We could have discussed the relation of C to H and its eigenstates
without making explicit reference to the fact that the unitary matrix C
is part of a group representation. However, using the language of group
theory connects this example to others that we might have studied. Not
 2.3 A little theory of discrete groups 17

all groups are as simple to understand as Z2 . The more complicated the

group, the more useful this connection is.
A group G is called Abelian if, for all a, b in G, ab = ba. A unitarity
representation of an Abelian group G consists of unitary matrices that
commute with one another. This means that they can be simultaneously
diagonalized. The operation of the group is then reduced to simple
An Abelian group is described by its
numbers. In the example above, the matrices (2.41) and (2.39) are eigenstates and their eigenvalues. The
diagonalized in a common basis. It is conventional to use C also as a eigenvalues are precisely what physi-
symbol for the eigenvalue of C on one of its eigenstates. In this case, cists call the quantum numbers of a
the eigenstates are state.
↵ ↵ p
C = +1 : [ ⇡ + + ⇡ ]/ 2
↵ ↵ p
C = 1 : [ ⇡+ ⇡ ]/ 2 . (2.42)
Because C 2 = 1, operating twice with the matrix C must give back the
original state: C · C | i = | i. This must, in particular, be true for an
eigenstate. Then the eigenvalues of C can only be ±1. We say that the
first state in (2.42) has C = +1 and the second has C = 1.
Symmetries of the Hamiltonian may involve transformations of space-
time coordinates, such as the special relativity transformations discussed
in Section 2.1. These are called space-time symmetries. In examples
like the one above, the symmetries relate di↵erent particles or quantum Space-time symmetries vs. internal
states without reference to space-time. These are called internal sym- symmetries.
metries. A given abstract group such as Z2 may describe a space-time
or an internal symmetry.
If G contains two elements a, b that do not commute, ab 6= ba, it is
called a non-Abelian group. If G is non-Abelian, and {Ua } is a finite-
dimensional unitary representation of G, it is generally not possible to
simultaneously diagonalize all of the unitary matrices in {Ua }. However,
by a change of basis, we can reduce these matrices to a common block-
diagonal form 0 1
U1 0 0
U R ! @ 0 U2 0 A , (2.43)
0 0 U3
where the blocks U1 , U2 , U3 , · · · are as small as possible. These minimal-
size unitary transformations representing G are called irreducible unitary
representations of G. For an irreducible representation {Ui }, the size of
the matrices is called the dimension di of the representation. The notion
of irreducible representations is probably more familiar to you in the
context of continuous groups. I will put your knowledge of the rotation The concept of an irreducible group
group into this context in the next section. representation. Many physics problems
in quantum mechanics are solved by
Mathematicians have shown that, for each discrete group G with a breaking up a larger Hilbert space into
finite number of elements, there is only a limited number of inequivalent irreducible representations of an appro-
finite-dimensional irreducible representations. Any other matrix repre- priate symmetry group.
sentation of the group is reducible, in the sense of (2.43), into a sum of
these basic representations. It can be proved that, for a discrete group
G with n elements, the inequivalent unitary transformations satisfy
X
d2i = n . (2.44)
i
18 Symmetries of Space-Time

Once we have found irreducible representations that add up to (2.44), we

have completely determined the structure of the possible representations
of G.
An example is given by the group of⇧ 3 of permutations on three
elements. We can represent such a permutation as the result of trans-
forming the set of labels [123] to a set of labels in another order. With
this representation, the group has 6 elements that can be written
{ [123] , [231] , [312] , [132] , [321] , [213] }. (2.45)
Permutations multiply a · b = c by composition, for example,
[231] · [231] = [312]
[132] · [312] = [321] . (2.46)
That is, applying the two permutations in order (right to left) gives the
resulting permutation as shown.
The 6 permutations in (2.45) can be associated with 6 states in a
Hilbert space. In this representation, the representation matrices are
6 ⇥ 6 matrices with entries 0 and 1. It can be shown that this is a
reducible representation. It contains two 1-dimensional irreducible rep-
resentations. One of these is the trivial representation that multiplies
each element by 1. Another is the representation that multiplies a state
by +1 for an even or cyclic permutation—the first three elements of
(2.45)—and multiplies a state by 1 for an odd permutation—the last
three elements of (2.45). There is also one 2-dimension representation,
presented in Exercise 2.3. These three irreducible representations to-
gether satisfy (2.44).

2.4 A little theory of continuous groups

The concepts reviewed in the previous section extend to the situation
of groups with a continous set of elements. Important examples are
the basic space-time symmetries: the group of spatial translations, the
group of spatial rotations, and the group of Lorentz transformations,
which includes rotations and boosts.
The group of space translations has the simplest structure. All trans-
lations commute with one another. You learned in quantum mechan-
ics that translations are implemented by unitary transformations. For
translations by a in one dimension
U (a) = exp[ iaP ] (2.47)
where P is the operator measuring the total momentum of the system.
The action of a space translation in
This is made most clear by considering the wavefunction of a plane wave
quantum mechanics gives a simple ex- of momentum p,
ample of a unitary representation of an hx|pi = eipx . (2.48)
Abelian group.
Acting on the state |pi with (2.47), we find
hx| U (a) |pi = eip(x a)
, (2.49)
 2.4 A little theory of continuous groups 19

which is the same wavefunction displaced by a. Using the language in-

troduced in the previous section, we say that the set of unitary operators
{U (a)} is a unitary representation of the group of space translations.
The expression of each U (a) as an exponential implies a relation be-
tween the group of translations and the Hermitian operator P . We
describe this relationship by saying that P is the generator of {U (a)} or
the generator of the group of translations.
The statement that P is Hermitian is equivalent to the statement that
the U (a) are unitary,

U (a)† = exp[+iaP † ] = exp[+iaP ] = U (a) 1

. (2.50)

Then, continuous unitary transformations are generated by Hermitian

operators. In quantum mechanics, Hermitian operators correspond to
observables.
Observables have time-independent values if the corresponding opera-
tors commute with the Hamiltonian of the quantum mechanics problem.
In this example, momentum is conserved if [P, H] = 0. Through the cor-
respondence (2.47), this statement is exactly equivalent to the statement
that [U (a), H] = 0, that is, that the equations of motion of the system
are invariant under translations. This relation is completely general. If In quantum mechanics, every symme-
Q is a Hermitian operator on the Hilbert space, the statement that Q is try that leaves the Hamiltonian invari-
ant is associated with a conserved quan-
a conserved quantity, tity. This follows from the connection
[Q, H] = 0 , (2.51) between Hermitan operators and uni-
tary symmetry transformations.
is equivalent to the statement that Q generates a symmetry of the equa-
tions of motion,

[UQ (a), H] = 0 for UQ (a) = exp[ iaQ] . (2.52)

This is the quantum-mechanical version of Noether’s theorem in classical

mechanics: Every symmetry of the equations of motion is associated
with a conservation law, and vice versa.
The group of translations is an Abelian group, since all translations
commute with one another. This implies that all of the matrices U (a)
can be simultaneously diagonalized. Actually, for every U (a), the eigen-
states of U (a) are the eigenstates of P , that is, states of definite momen-
tum. Each eigenstate of P gives a one-dimensional unitary representa-
tion of the translation group.
A non-Abelian continuous group that should be familiar to you is the
rotation group in 3 dimensions. In quantum mechanics, rotations are The action of rotations in quantum me-
implemented on the Hilbert space by the unitary operators chanics gives an example of the unitary
representation of a non-Abelian group.

U (~
↵) = exp[ i~ ~
↵ · J] (2.53)

where ↵~ gives the axis and angle of the rotation and J~ are the opera-
As in the previous example, the con-
tors of angular momentum. These operators satisfy the commutation servation law of angular momentum is
relation associated with the symmetry of invari-
[J i , J j ] = i✏ijk J k . (2.54) ance under rotations.
20 Symmetries of Space-Time

It can be shown that, if Hermitian operators J i satisfy (2.54), the unitary

operators constructed from them satisfy the composition rules of 3d
rotations. That is, if
U ( ~ )U (~
↵) = U (~ ) , (2.55)
then the rotation ~ is the one that results from rotating first through
~ and then through ~ . The operators J i are thus the generators of
↵
rotations. In fact the complete structure of the group of rotations is
specified by the commutation relation (2.54).
In quantum mechanics, finite-dimensional matrix representations of
the rotation group play an important role. The quantum states of atoms
are organized into multiplets of definite angular momentum, for example,
the 2P or 3D states of the hydrogen atom. States of definite angular mo-
mentum give the finite-dimensional irreducible matrix representations of
the rotation group.
Through the correspondence (2.53), a finite-dimensional representa-
tion of the rotation group is generated by a set of finite-dimensional
matrices that satisfy (2.54). The simplest such representations are the
trivial, 1-dimensional representation

J i = 0, (2.56)

the 2-dimensional representation

1 2 3
J1 = , J2 = , J1 = , (2.57)
2 2 2
i
where are the Pauli sigma matrices
✓ ◆ ✓ ◆ ✓ ◆
1 0 1 2 0 i 3 1 0
= , = , = , (2.58)
1 0 i 0 0 1
and the 3-dimensional representation
0 1 0 1 0 1
0 0 0 0 0 i 0 i 0
J1 = @ 0 0 i A , J2 = @ 0 0 0 A , J3 = @ i 0 0A .
0 i 0 i 0 0 0 0 0
(2.59)
It is instructive to check explicitly that (2.57) and (2.59) satisfy (2.54).
The three representations given here are those of spin 0, spin 12 , and
spin 1. We will meet these representations again and again in the ap-
plications I will discuss in this book. Similarly, for every integer or
half-integer value j, there is a set of three (2j + 1) ⇥ (2j + 1) matrices
satisfying these commutation relations. This is the spin j representation
The reduction of a set of states of an of the rotation group.
atom with orbital and spin angular mo- One of the standard problems in atomic physics is to decompose a
menta (`, s) into states of total angu- set of quantum states into irreducible representations of the rotation
lar momentum j is an example of the
reduction of a reducible representation
group. For example, states of an atom may be labelled by orbital
of a continuous group—in this case, angular momentum ` and spin angular momentum s. This gives a set
the rotation group—into a sum of ir- of states with (2` + 1)(2s + 1) elements. The total angular momentum
reducible representations. j takes values
|` s|  j  (` + s) . (2.60)
 2.4 A little theory of continuous groups 21

~ H] = 0, each value of j gives a set of (2j + 1) states with the

Since [J,
same energy. In Section 4.1, we will translate this group theory exercise
into a statement about the energy levels of the hydrogen atom.
We can consider the group of rotations in 3 dimensions as an abstract
group whose multiplication law is defined by the composition of rota-
tions. This group is called SO(3). Similarly, there is an abstract group
of rotations in d dimensions, called SO(d). The case d = 2 is simple; it is
the group of rotations of a circle, an Abelian group of translations of an
angle , with identified with ( + 2⇡). This abstract group is the same
one that we meet when we consider the group of phase transformations
i i↵ i
e !e e . (2.61)

This is a transformation by a 1 ⇥ 1 unitary matrix, so we also call this

group U (1).
General n ⇥ n unitary matrices form a representation of an abstract
group called U (n). Any n ⇥ n unitary matrix can be written in the form
of (2.47) as generated by a set of n ⇥ n Hermitian matrices

U = exp[ i↵a ta ] . (2.62)

The sum over a runs over a basis of n ⇥ n Hermitian matrices, which

contains n2 elements. One of these elements is the unit matrix,

t0 = 1 (2.63)

This matrix commutes with all of the other ta . If we omit this element Definition of the group SU (n).
from the set of Hermitian matrices, we obtain a non-Abelian group of
matrices with n2 1 generators, the n ⇥ n Hermitian matrices with zero
trace. The group generated by these n2 1 matrices is called SU (n). It
is the group of n ⇥ n unitary matrices with determinant 1.
For n = 2, the Pauli sigma matrices (2.58) form a basis for the 2 ⇥ 2
traceless Hermitian matrices. Thus, SO(3) and SU (2) are names for the
same abstract group. (Mathematicians make a distinction between these
groups, but the di↵erence will not be relevant to the calculations done
in this book.) This abstract group describes rotations in three dimen-
sions, but it will also describe some internal symmetries of elemementary
particles that we will meet in the course of our discussion.
A continuous group of transformations generated by Hermitian matri-
ces, in the form (2.62), is called a Lie group. The commutation algebra
of the generators ta , This equation, which expresses the non-
[ta , tb ] = if abc tc (2.64) commuting nature of the generators of
a Lie group, contains the full informa-
is called the Lie algebra of the group. The constants f abc are called tion about the representations and the
the structure constants of the Lie algebra. It can be shown that we can geometry of the group.
always choose a basis for the ta such that the structure constants f abc are
completely antisymmetric in [abc]. These definitions straightforwardly
generalize the presentation that I have given of the rotation group in 3
dimension. In the case of the rotation group,

f abc = ✏abc . (2.65)

22 Symmetries of Space-Time

In the same way as for the rotation group, the Lie algebra of the gen-
erators determines the multiplication law of any two elements of the
group.
In this book, we will meet only special cases of Lie groups. The
particular groups U (1) = SO(2), SU (2) = SO(3), and SU (3) will
have important roles in our story. Still, the abstract properties of Lie
groups will be useful to us in understanding how to apply these groups
to particle physics. I will introduce some further formalism of Lie groups
when we will need it in Chapter 11.

2.5 Discrete space-time symmetries

The symmetries of special relativity include the continuous symmetries
of rotations and Lorentz transformations. But they also include two
distinct space-time transformations that leave the metric tensor (2.7)
invariant but cannot be constructed as a product of continuous rota-
tions and boosts. This will turn out to be an important issue for ele-
mentary particle physics. According to Noether’s theorem, conservation
of energy-momentum is equivalent to the invariance of the equations
of motion with respect to space-time translations, and the conserva-
tion of angular momentum is equivalent to the invariance of the equa-
tions of motion with respect to rotations and boosts. However, there is
no fundamental principle that requires that extra, discrete space-time
transformations must be symmetries of the Hamiltonian or that quanti-
ties associated with these extra discrete symmetries must be conserved.
This is a separate question that can only be answered by experiment.
We will see in Part III that the answer given to this question is quite
surprising.
The two space-time transformations that are not part of the contin-
Minkowski space has two extra space- uous Lorentz group are parity (P ) and time reversal (T ). These
time symmetries: parity P and time- space-time operations satisfy
reversal T .
P2 = 1 T2 = 1 (2.66)
In quantum mechanics, these transformations are implemented by oper-
ators with eigenvalues ±1. I will also refer to the eigenvalue of a quantum
state as the value P or T for that state. Continuous Lorentz invariance
does not imply that these values P and T are conserved. However, P
and T are observed to be conserved in electromagnetism and atomic
physics. The study of energy levels of nuclei confirms that P and T are
also conserved by the strong nuclear interaction.
Parity is defined as the operation on 4-vectors
xµ = (x0 , ~ )xµ ! (x0 , ~x)µ . (2.67)
A rotation matrix, for example,
0 1
1 0 0 0
B 0 cos ✓ sin ✓ 0C
⇤=@ A , (2.68)
0 sin ✓ cos ✓ 0
0 0 0 1
 Exercises 23

or, indeed, any matrix that implements a continuous Lorentz transfor-

mation, has
det⇤= +1 . (2.69)
But (2.67) is implemented by a matrix with det⇤= 1. Thus, this
matrix cannot be generated as a product of continuous rotations. Time
reversal is defined similarly as the operation

xµ = (x0 , ~ )xµ ! ( x0 , ~ )xµ . (2.70)

By the same logic, time reversal cannot be continuously generated.

In quantum mechanics, an isolated particle can also have an intrinsic
parity. That is, under parity, its quantum state of momentum ~k can
transform as
E E E
P A(~k) = + A( ~k) or A( ~k) . (2.71)
A quantum particle can have intrinsic
We refer to these two cases as intrinsic parity (+1) or ( 1). A particle parity +1 or 1.
can also have an intrinsic quantum number under time reversal.
In quantum mechanics, time reversal is implemented by an anti-unitary
operator. In this book, I will avoid detailed analysis of time-reversal
properties as much as possible.
There is one more discrete transformation that is closely related to
these space-time operations. As we will see in the next chapter, quantum
field theory implies that, for each particle in nature, there must exist an
antiparticle with the same mass and opposite values of all conserved
charges. We can then define an operation called charge conjugation (C)
that converts each particle to its antiparticle and vice versa. C then It is useful to consider charge conjuga-
also naturally satisfies tion C as a discrete space-time trans-
formation on the same level as P and
C2 = 1 . (2.72) T.
Quantum states can have intrinsic values of C equal to +1 or 1. C
is observed to be conserved in electromagnetic and strong nuclear reac-
tions.
I have already explained that it is a question for experiment whether
P , C, and T are conserved by all interactions in nature. However, it is
a theorem in quantum field theory that the combination CP T must be
a symmetry of all particle interactions. This statement can be tested
experimentally and, so far, it holds up. We will take up the issue of
the separate conservation of P , C, and T in our discussion of the weak
interaction in Part III.

Exercises
(2.1) Consider the decay of a particle of mass M , at both nonzero. With an appropriate choice of axes,
rest, into two particles with masses m1 and m2 , the momentum vectors of the final particles can be
24 Exercises

written (c) The 2-dimensional representation that assigns

p1 = (E1 , 0, 0, p) p2 = (E2 , 0, 0, p) (2.73) ✓ ◆ ✓ ◆
1 0 0 1
[123] ! [231] !
with E12 = p2 + m21 , E22 = p2 + m22 . 0 1 1 1
✓ ◆ ✓ ◆
(a) Show that 1 1 0 1
[312] ! [213] !
 1/2 1 0 1 0
✓ ◆ ✓ ◆
p = M 4 2M 2 (m21 +m22 )+(m21 m22 )2 /2M 1 1 1 0
[321] ! [132] !
(2.74) 0 1 1 1
(b) Take the limit m2 ! 0 and show that this (2.75)
formula reproduces the result (2.19) for the
decay into one massive and one massless par- (2.4) This problem explores the non-Abelian nature of
ticle. the Lorentz group.
(c) Find formulae for E1 and E2 in terms of M ,
m1 , m 2 . (a) The 4⇥4 matrix⇤ 3 ( ) that represents a boost
by in the 3̂ direction is given by (2.3). Write
(2.2) Using natural units, estimate the following quanti- the corresponding 4⇥4 matrix⇤ 1 ( ) that rep-
ties: resents a boost by in the 1̂ direction.
(a) If the photon has a mass, the electric fields (b) Compute the composite Lorentz transforma-
generated by charges will fall o↵exponentially tion⇤ C = ⇤1 ( ) ⇤3 ( ). The component
at distances larger than the photon Compton ⇤C 0 0 of this matrix should be the compos-
wavelength. It is possible to obtain limits on ite boost C . From this, compute the new
the photon mass by looking for this e↵ect in velocity C .
the solar system. For example, the magnetic
field of Jupiter is found to be a conventional (c) By acting⇤ C on the 4-vector (1, 0, 0, 0), show
dipole field out to many times the radius of that the elements⇤ C i 0 give the direction of
the planet. Estimate the corresponding upper the boost. Show from this that the new veloc-
limit on the photon mass in MeV. ity is ~C = ( , 0, / ). Show that the mag-
nitude of this vector agrees with the result of
(b) The range of the weak interaction is given by part (b).
Compton wavelength of the W boson, which
has a mass of 80.4 GeV. Estimate this length (d) The matrix⇤ C is not symmetric, so it cannot
in cm. be a pure boost. It is, in fact, a combination
of a boost and a rotation. To understand this
(c) If the electron is a composite particle with
better, expand the elements of⇤ C in powers
a nonzero size, that will a↵ect the ob-
of for small , keeping terms up to order
served rate for electron-electron and electron- 2
.
positron scattering. Given that these rates
are in good agreement with the predictions (e) Write the⇤matrix for a pure boost to the ve-
for pointlike electrons up to a center of mass locity ~C . This matrix should be symmetric.
energy of 200 GeV, estimate the upper limit The space-space part should be
on the size of the electron, in cm.
i j
ij C C
(2.3) Show that the following are unitary representations +( C 1) 2
. (2.76)
of the permutation group⇧ 3 by verifying that they C

satisfy the multiplication law of⇧ 3 :

(f) Expand the matrix found in part (e) to order
(a) The 1-dimension representation in which all 2
. Show that it explains the symmetric part
six permutations in (2.45) are represented by of the result found in part (d). Identify the
1. remaining antisymmetric part as an infinites-
(b) The 1-dimension representation in which imal rotation in the 3̂-1̂ plane. The rotation
[123], [231], and [312] are represented by 1 that results from the non-commuting nature
and [213], [321], and [132] are represented of boosts in di↵erent directions is called a
by 1. Wigner rotation.
Relativistic Wave
Equations
3
In the previous chapter, I developed some simple rules for the treatment
of special relativity that will aid us in our search for the laws of ele-
mentary particle interaction. In this chapter, I will discuss some of the
concepts that we will need to use quantum mechanics e↵ectively.
The standard treatment of 1-particle quantum mechanics will not be
adequate for our purposes. First of all, the Schrödinger equation is not
Lorentz-invariant. In that equation, time and space appear asymmet-
rically. In a relativistic theory, the wavefunctions of quantum particles
must obey wave equations in which time and space appear symmetri-
cally in accord with special relativity. In this chapter, I will discuss three
of the most important of these equations.
Standard quantum mechanics is inadequate in another way. In el- The Schrödinger equation is not ade-
ementary particle reactions, the number of particles can change as in- quate to describe elementary particles.
We need a theoretical framework that
dividual particles are created and destroyed. We have already noted in is relativistic, and that allows particles
the previous chapter that every particle must have an antiparticle with to be created and destroyed.
the same mass. Typically, elementary particle interactions allow the
creation of a particle together with its antiparticle, or the annihilation
of a particle with its antiparticle. Then, quantum mechanics must be
generalized to a multiparticle theory.
Both generalizations are accomplished in relativistic quantum field
theory. However, there is no space in this small book for a complete
description of quantum field theory, or even for a derivation of its major
implications. Instead, I will use this chapter to explain some essential
points of quantum field theory that will be needed for our analysis. In
Chapter 7, I will explain how we use quantum field theory to make pre-
dictions for elementary particle reactions, and I will give some shortcuts
and heuristics that will allow us to apply these ideas easily.

3.1 The Klein-Gordon equation

A wave equation is said to be invariant under a group of symmetries
if, for any solution, the symmetry transform of that solution is another
solution of the wave equation. For a scalar field, the Lorentz transform

Concepts of Elementary Particle Physics. Michael E. Peskin.

c Michael E. Peskin 2019. Published in 2019 by Oxford University Press.
DOI: 10.1093/oso/9780198812180.001.0001
26 Relativistic Wave Equations

of a waveform is the same waveform evaluated at Lorentz-transformed

Definition of a relativistically invariant points. In an equation,
wave equation.
0 1
(x) ! (x) = (⇤ x) . (3.1)

Canonically,⇤ 1 appears in this formula so that, if (x) has a maximum

at x = a, 0 (x) will have a maximum at the Lorentz-transformed point
Illustration of the transformation of a x = ⇤a. A Lorentz-invariant theory of waves should have the property
scalar field as in (3.1): that, if (x) solves the wave equation, then 0 (x) in (3.1) does also.
The simplest equation satisfying this property is the Klein-Gordon
equation ✓ 2 ◆
@ 2 2
r +m (t,~x) = 0 . (3.2)
@t2
.
In this equation, t and ~x appear in a symmetric way. The 4-gradient
@ @
@µ = ( , )µ (3.3)
@t @~x
transforms under Lorentz transformations as a 4-vector with a lowered
index. That is, the quantities
@ ~ @2
p · @ = (E + p~ · r) and @2 = ( r2 ) (3.4)
@t @t2
are Lorentz-invariant operators. Using (3.4), we can write the Klein-
The Klein-Gordon equation. Gordon equation (3.2) in a more manifestly Lorentz-invariant form,

(@ 2 + m2 ) (x) = 0 . (3.5)

We can also see the invariance of (3.5) by examining the solutions of this
equation explicitly. These are
·~
iEt+i~px ip·x
(x) = e =e , (3.6)

where pµ = (E,~p)µ is a 4-vector satisfying

E2 p|2 = p2 = m2 .
|~ (3.7)

This criterion is Lorentz-invariant. The Lorentz-invariance of the 4-

vector product is the statement that
1
p · x = ( ⇤p) · (⇤x) or (⇤p) · x = p · (⇤ x) . (3.8)

Then the boost of the solution (3.6) is

1
0 ip·⇤ x i(⇤p)·x
(x) = e =e , (3.9)

which is also a solution of the equation.

The Klein-Gordon equation has the odd feature, from the point of
view of a quantum-mechanical interpretation, that it has solutions cor-
responding both to positive and negative energy. Solving (3.7) for E,
we find that both solutions

E = ±Ep (3.10)
 3.1 The Klein-Gordon equation 27

are acceptable. This is a common property of all relativistic wave equa-

tions. Quantum field theory gives an attractive way to understand the
negative energy solutions, which I will explain below.
Another way to derive the relativistic invariance of the Klein-Gordon
equation is to write a variational principle that gives rise to this equation.
You might be used to the variational principles of Lagrangian mechanics.
In that formalism, we write an action functional S
Z
S[x(t), ẋ(t)] = dtL(x, ẋ) . (3.11)

The principle that S is stationary with respect to all variations of the so-
lution x(t) yields the equation of motion of the system. Mathematically,
if x(t) ! x(t) + x(t), then we can write S in the form
Z 
S[x(t), ẋ(t)] = dt x (t) E[x(t), ẋ(t), ẍ(t)] . (3.12)

Then the equation of motion is E = 0.

To obtain a relativistic equation of motion, we start with a relativis-
tically invariant expression for the action S. The action S should be a
function of the waveform (x). Instead of an integral over t only, I will
integrate symmetrically over all of Minkowski space. Then the action
principle takes the form By choosing an action S in this form,
Z we guarantee that the action is rela-
tivistically invariant. Then the equa-
S[ (x)] = d4 x L( ,@ µ ) (3.13) tion of motion following from the vari-
ational principle must be a relativistic
The function L is called the Lagrange density. I will choose the Lagrange field equation.
density to be relativistically invariant. Then S is the invariant integral
of a invariant function and thus is guaranteed to be Lorentz-invariant.
To illustrate how we apply this formalism, I will propose a simple form
for L. Consider, then, Lagrangian formulation of the Klein-
Gordon equation.
1 µ
L= @ @µ m2 2
. (3.14)
2
There are no uncontracted 4-vector indices. Thus, this expression, and,
by extension its integral S over all space-time, is Lorentz-invariant. The
variation of L with respect to (x) is

L = @µ @µ m2 (3.15)

Putting this under the integral d4 x and integrating by parts in the first
term, we find
Z 
S = d x (x) ( @ 2 m2 ) (x) .
4
(3.16)

The variational principle states that the Lagrangian equation of motion

is the condition that S vanishes for an arbitrary variation of (x). In
(3.16), this condition implies that the quantity in brackets must vanish.
This gives exactly the Klein-Gordon equation (3.5).