Motionmountain Volume4

Christoph Schiller
MOTION MOUNTAIN
the adventure of physics – vol.iv
quantum theory: the smallest change
www.motionmountain.net
Christoph Schiller
Motion Mountain
The Adventure of Physics

Volume IV
Quantum Theory:
The Smallest Change
Edition 24.1, available as free pdf at

www.motionmountain.net
Editio vicesima tertia.
Proprietas scriptoris © Christophori Schiller

secundo anno Olympiadis vicesimae nonae.
Omnia proprietatis iura reservantur et vindicantur.

Imitatio prohibita sine auctoris permissione.
Non licet pecuniam expetere pro aliquo, quod
partem horum verborum continet; liber
pro omnibus semper gratuitus erat et manet.
Twenty-third edition, ISBN 978-300-021946-7.
Copyright © 2009 by Christoph Schiller,

the second year of the 29th Olympiad.
This pdf file is licensed under the Creative Commons

Attribution-Noncommercial-No Derivative Works 3.0 Germany
Licence, whose full text can be found on the website
creativecommons.org/licenses/by-nc-nd/3.0/de,
with the additional restriction that reproduction, distribution and use,
in whole or in part, in any product or service, be it
commercial or not, is not allowed without the written consent of
the copyright owner. The pdf file was and remains free for everybody
to read, store and print for personal use, and to distribute
electronically, but only in unmodified form and at no charge.
To Britta, Esther and Justus Aaron
τῷ ἐμοὶ δαὶμονι
Die Menschen stärken, die Sachen klären.
PR EFAC E
“ ”
Primum movere, deinde docere.*
Antiquity
This book is written for anybody who is curious about nature and motion. Have you
Motion Mountain – The Adventure of Physics pdf file available free of charge at www.motionmountain.net
ever asked: Why do people, animals, things, images and space move? The answer leads
to many adventures; this volume presents those due the discovery that there is a smallest
change in nature. This smallest change leads to what is called quantum theory. In the
structure of modern physics, shown in Figure 1, quantum physics covers three points;
this volume covers the introduction to the point in the lower right: the foundations of
quantum theory.
The present introduction to quantum physics arose from a threefold aim I have pur-
sued since 1990: to present the basics of motion in a way that is simple, up to date and
captivating.
In order to be simple, the text focuses on concepts, while keeping mathematics to the
necessary minimum. Understanding the concepts of physics is given precedence over
using formulae in calculations. The whole text is within the reach of an undergraduate.
In order to be up to date, the text is enriched by the many gems – both theoretical and
empirical – that are scattered throughout the scientific literature.
In order to be captivating, the text tries to startle the reader as much as possible. Read-
ing a book on general physics should be like going to a magic show. We watch, we are
astonished, we do not believe our eyes, we think, and finally we understand the trick.
When we look at nature, we often have the same experience. Indeed, every page presents
at least one surprise or provocation for the reader to think about. Numerous interesting
challenges are proposed.
The motto of the text, die Menschen stärken, die Sachen klären, a famous statement by
Copyright © Christoph Schiller November 1997–July 2010
Hartmut von Hentig on pedagogy, translates as: ‘To fortify people, to clarify things.’ Clar-
ifying things requires courage, as changing habits of thought produces fear, often hidden
by anger. But by overcoming our fears we grow in strength. And we experience intense
and beautiful emotions. All great adventures in life allow this, and exploring motion is
one of them.
Munich, 27 July 2010.
* ‘First move, then teach.’ In modern languages, the mentioned type of moving (the heart) is called motivat-
ing; both terms go back to the same Latin root.
8 preface
PHYSICS: Unified description of motion Why does motion

Describing motion with action. Adventures: understanding occur? What are
motion, intense joy with space, time and
thinking, catching a quantum particles?
glimpse of bliss,
calculating
masses
Quantum
General relativity theory with gravity Quantum field theory
Adventures: the Adventures: bouncing Adventures: building
night sky, measu- neutrons, under- accelerators, under-
ring curved space, standing tree standing quarks, stars,
exploring black growth bombs and the basis of
life, matter, radiation
holes and the
universe, space How do small
and time things move?
What are things?
How do
Classical gravity everyday, Special relativity Quantum theory
Adventures: fast and large Adventures: light, Adventures: death,
climbing, skiing, things move? magnetism, length sexuality, biology,
space travel, contraction, time enjoying art and
the wonders of dilation and colours, all high-tech
astronomy and E0 = mc2 business, medicine,
geology chemistry, evolution
G c h, e, k
Galilean physics, heat and electricity

Adventures: sport, music, sailing, cooking,
describing beauty and understanding its origin,
using electricity and computers,
understanding the brain and people
F I G U R E 1 A complete map of physics: the connections are defined by the speed of light c, the
gravitational constant G, the Planck constant h, the Boltzmann constant k and the elementary charge e.
Advice for learners
In my experience as a teacher, there was one learning method that never failed to trans-
form unsuccessful pupils into successful ones: if you read a book for study, summarize
every section you read, in your own words, aloud. If you are unable to do so, read the
section again. Repeat this until you can clearly summarize what you read in your own
words, aloud. You can do this alone in a room, or with friends, or while walking. If you
do this with everything you read, you will reduce your learning and reading time signif-
icantly. In addition, you will enjoy learning from good texts much more and hate bad
texts much less. Masters of the method can use it even while listening to a lecture, in a
low voice, thus avoiding to ever take notes.
preface 9
Using this book
Text in green, as found in many marginal notes, marks a link that can be clicked in a pdf
reader. Such green links are either bibliographic references, footnotes, cross references
to other pages, challenge solutions, or pointers to websites.
Solutions and hints for challenges are given in the appendix. Challenges are classified
as research level (r), difficult (d), standard student level (s) and easy (e). Challenges of
type r, d or s for which no solution has yet been included in the book are marked (ny).
A request
The text is and will remain free to download from the internet. In exchange, I would
be delighted to receive an email from you at [email protected], especially on the
following issues:
Challenge 1 s — What was unclear and should be improved?
— What story, topic, riddle, picture or movie did you miss?
— What should be corrected?
Alternatively, you can provide feedback online, on www.motionmountain.net/wiki. The
feedback will be used to improve the next edition. On behalf of all readers, thank you in
advance for your input. For a particularly useful contribution you will be mentioned – if
you want – in the acknowledgements, receive a reward, or both. But above all, enjoy the
reading!

Contents
14 1 Minimum action – quantum theory for poets
The effects of the quantum of action on rest 17 • The consequences of the quan-
tum of action for objects 18 • Why ‘quantum’? 19 • The effect of the quantum of
action on motion 22 • Quantum limits and surprises 23 • Transformation, life
and Democritus 24 • Randomness – a consequence of the quantum of action 28 •
Waves – a consequence of the quantum of action 29 • Particles – a consequence of
the quantum of action 30 • Quantum information 30 • Curiosities and fun chal-
lenges about the quantum of action 31 • The dangers of buying a can of beans 32
• A summary: quantum physics, the law and indoctrination 34
35 2 L ight – the strange consequences of the quantum of action
What is the faintest lamp? 35 • Photons 39 • What is light? 41 • The size of pho-
tons 42 • Are photons countable? – Squeezed light 42 • The positions of pho-
tons 44 • Are photons necessary? 47 • How can a wave be made up of parti-
cles? 49 • Can light move faster than light? – Virtual photons 54 • Indetermi-
nacy of electric fields 55 • Curiosities and fun challenges about photons 56 • A
summary on light: particle and waves 57
60 3 Motion of mat ter – beyond cl assical physics
Wine glasses, pencils and atoms – no rest 60 • No infinite precision 61 • Cool
gas 61 • Flows and the quantization of matter 62 • Fluid flows and quantons 62 •
Knocking tables and quantized conductivity 62 • Matter quantons and their mo-
tion – matter waves 63 • Mass and acceleration of quantons 66 • Why are atoms
not flat? Why do shapes exist? 67 • Rotation, quantization of angular momentum,
and the lack of north poles 68 • Rotation of quantons 70 • Silver, Stern and Ger-
lach – polarization of quantons 70 • Curiosities and fun challenges about quantum
matter 72 • First summary on the motion of quantum particles 72
74 4 The quantum description of mat ter and its motion
Visualizing the wave function: rotating arrows and probability clouds 75 • The
state evolution – the Schrödinger equation 77 • Self-interference of quantons 79 •
The speed of quantons 79 • Dispersion of quantons 79 • Tunnelling and limits on
memory – damping of quantons 80 • The quantum phase 83 • The least action
principle in quantum physics 86 • The motion of quantons with spin 87 • Rela-
tivistic wave equations 89 • Composite vs. elementary quantons 90 • Curiosities
and fun challenges about quantum motion of matter 92 • A summary on motion
of quantons 93
94 5 Permu tation of particles – are particles like gloves?
Why does indistinguishability appear in nature? 96 • Can particles be counted? 97
• What is permutation symmetry? 98 • Indistinguishability and symmetry 98 •
The behaviour of photons 99 • Bunching and antibunching 100 • The energy de-
pendence of permutation symmetry 100 • Indistinguishability in quantum field
theory 101 • How accurately is permutation symmetry verified? 102 • Copies,
clones and gloves 103
105 6 Rotations and statistics – visualizing spin
Quantum particles and symmetry 105 • Types of quantum particles 106 • The belt
trick and its extension 111 • Angels, Pauli’s exclusion principle and the hardness
12 contents
of matter 113 • Spin, statistics and composition 115 • Is spin a rotation about an
axis? 115 • Why is fencing with laser beams impossible? 116 • Rotation requires
antiparticles 117 • A summary on spin and indistinguishability 118 • Limits and
open questions of quantum statistics 118
120 7 Superpositions and probabilities – quantum theory withou t ide-
olo gy
Why are people either dead or alive? 120 • Summary on decoherence, life and
death 126 • What is a system? What is an object? 127 • Is quantum theory non-
local? – A bit about the Einstein–Podolsky–Rosen paradox 128 • Curiosities and
fun challenges about superpositions 130 • What is all the fuss about measurements
in quantum theory? 132 • Hidden variables 137 • Summary on probabilities and
determinism 139 • What is the difference between space and time? 141 • Are we
good observers? 141 • What relates information theory, cryptology and quantum
theory? 142 • Is the universe a computer? 143 • Does the universe have a wave
function? And initial conditions? 143
145 8 Colours and other interactions bet ween light and mat ter
The causes of colour 145 • Using the rainbow to determine what stars are made
of 154 • What determines the colours of atoms? 155 • Relativistic hydrogen 159
• Relativistic wave equations – again 160 • Getting a feeling for the Dirac equa-
tion 162 • Antimatter 163 • Virtual particles 164 • Curiosities and fun chal-
lenges about colour 165 • Material properties 166 • The strength of electromag-
netism 166 • A summary on colours and materials 167
169 9 Q uantum physics in a nu tshell
Physical results of quantum theory 169 • Motion of quantum particles 170 •
Achievements in precision 172 • Is quantum theory magic? 173 • Quantum theory
can do more 174
175 a Units, measurements and constants
SI units 175 • Planck’s natural units 178 • Other unit systems 179 • Curiosities
and fun challenges about units 181 • Precision and accuracy of measurements 182
• Limits to precision 183 • Physical constants 183 • Useful numbers 188
190 b Numbers and vector spaces
Numbers as mathematical structures 190 • Complex numbers 192 • Quater-
nions 193 • Octonions 198 • Other types of numbers 200 • Vector spaces 201 •
Mathematical curiosities and fun challenges 203
204 Challenge hints and solu tions
209 Biblio graphy

224 Credits
Film credits 225 • Image credits 225
Quantum Theory:
The Smallest Change
In our quest to understand how things move,

we discover that there is a minimal change in nature,
implying that motion is fuzzy,
that boxes are never tight,
that matter is composed of elementary units,
and that light and interactions are streams of particles.
The minimal change explains why antimatter exists,
why particles are unlike gloves,
why copying machines do not exist,
that probabilities are reasonable,
and how how all colours in nature are formed.
Chapter 1
MINIMUM ACTION – QUANTUM

THEORY FOR POET S
“ ”
Natura [in operationibus suis] non facit saltus.*
15th century
C
limbing Motion Mountain up to this point, we completed three legs. We
ame across Galileo’s mechanics (the description of motion for kids), then
ontinued with Einstein’s relativity (the description of motion for science-fiction
enthusiasts), and finally explored Maxwell’s electrodynamics (the description of motion
for business people). These three classical descriptions of motion are impressive, beau-
tiful and useful. However, they have a small problem: they are wrong. The reason is
simple: none of them describes life.
Whenever we observe a flower or a butterfly, such as those of Figure 2, we enjoy the
bright colours, the motion, the wild smell, the soft and delicate shape or the fine details
of their symmetries. None of the three classical descriptions of nature can explain any
of these properties; neither do they explain the impression that the flower makes on our
senses. Classical physics can describe certain aspects of the impression, but it cannot ex-
plain their origins. For such an explanation, we need quantum theory. In fact, we will
discover that in life, every type of pleasure is an example of quantum motion. Take any
example of a pleasant situation:** for example, a beautiful evening sky, a waterfall, a ca-
Challenge 2 s ress, or a happy child. Classical physics is not able to explain it: the colours, shapes and
sizes involved remain mysterious.
In the early days of physics, this limitation was not seen as a shortcoming, because nei-
ther senses nor material properties were thought to be related to motion – and pleasure
was not considered a serious subject of investigation for a respectable researcher. How-
Page 314 ever, we have since learned that our senses of touch, smell and sight are primarily detec-
tors of motion. Without motion, there would be no senses. Furthermore, all detectors are
made of matter. During the exploration on electromagnetism we began to understand
that all properties of matter are due to motions of charged constituents. Density, stiff-
ness, colour, and all other material properties result from the electromagnetic behaviour
of the Lego bricks of matter: namely, the molecules, the atoms and the electrons. Thus,
the properties of matter are also consequences of motion. Moreover, we saw that these
tiny constituents are not correctly described by classical electrodynamics. We even found
Ref. 1 * ‘Nature [in its workings] makes no jumps.’

** The photograph on page 13 shows a female glow worm, Lampyris noctiluca, as found in the United King-
dom (© John Tyler, www.johntyler.co.uk/gwfacts.htm).
minimum action – quantum theory for poets 15
F I G U R E 2 Some examples of quantum machines (© Linda de Volder)
Page 118 that light itself does not behaves classically. Therefore the inability of classical physics to
describe matter, light and the senses is indeed due to its intrinsic limitations.
In fact, every failure of classical physics can be traced back to a single, fundamental
Ref. 2 discovery made in 1899 by Max Planck:*
⊳ In nature, actions smaller than ħ = 1.06 ⋅ 10−34 Js are not observed.
All attempts to observe actions smaller than this fail.** In other words, in nature – as in a
good film – there is always some action. The existence of a minimal action – the so-called
* Max Planck (1858–1947), professor of physics in Berlin, was a central figure in thermostatics. He discov-
ered and named Boltzmann’s constant k and the quantum of action h, often called Planck’s constant. His
introduction of the quantum hypothesis gave birth to quantum theory. He also made the works of Einstein
known in the physical community, and later organized a job for him in Berlin. He received the Nobel Prize
for physics in 1918. He was an important figure in the German scientific establishment; he also was one of
the very few who had the courage to tell Adolf Hitler face to face that it was a bad idea to fire Jewish pro-
fessors. (He got an outburst of anger as answer.) Famously modest, with many tragedies in his personal life,
he was esteemed by everybody who knew him.
** In fact, the quantum principle cited here is a slight simplification: the constant originally introduced
by Planck was the (unreduced) h = 2πħ. The factor 2π leading to the final quantum principle was found
somewhat later, by other researchers.
This somewhat unconventional, but didactically useful, approach to quantum theory is due to Niels Bohr.
Ref. 3, Ref. 4 Nowadays, it is hardly ever encountered in the literature, despite its simplicity.
Niels Bohr (b. 1885 Copenhagen, d. 1962 Copenhagen) made Copenhagen University into the new centre
16 1 minimum action – quantum theory for poets
F I G U R E 3 Max Planck (1858–1947)
F I G U R E 4 Niels Bohr (1885–1962)
Challenge 3 s quantum principle – is in complete contrast with classical physics. (Why?) However, it has
passed an enormous number of experimental tests, many of which we will encounter in
this part of our mountain ascent. Therefore, ħ, which is pronounced ‘aitch-bar’, is called
the quantum of action, or alternatively Planck’s constant. Planck discovered the principle
Page 118 when studying the properties of incandescent light, i.e., light emanating from hot bodies.
But the quantum principle also applies to motion of matter, and even, as we will see later,
to motion of space-time.
The quantum principle states that no experiment can measure an action smaller than
ħ. For a long time, Einstein tried to devise experiments to overcome this limit. But he
failed in all his attempts: nature does not allow it, as Bohr showed again and again. We re-
call that physical action is a measure for the change happening in a system. The quantum
principle can thus rephrased as

⊳ In nature, a change smaller than ħ = 1.06 ⋅ 10−34 Js cannot be observed.
In physics – as in the theatre – action is a measure of the change occurring in a system.
Therefore, a minimum action implies that there is a minimum change in nature. Thus the
quantum of action would perhaps be better named the quantum of change. If we compare
two observations, there will always be change between them.
of development of quantum theory, overshadowing Göttingen. He developed the description of the atom
in terms of quantum theory, for which he received the 1922 Nobel Prize in Physics. He had to flee Den-
mark in 1943 after the German invasion, because of his Jewish background, but returned there after the war,
continuing to attract the best physicists across the world.
TA B L E 1 How to convince yourself and others that there is a minimum

action, or minimum change ħ in nature. Compare this table with the two
tables in volume II, that about maximum speed on page 24, and that
about maximum force on page 97.
Issue Method
The action value ħ is check all observations

observer-invariant
Local change or action values < ħ check all observations
are not observed
Change or action values < ħ are check all observations
either non-local or not due to
energy transport
Local change or action values < ħ check all attempts
cannot be produced
Local change or action values < ħ solve all paradoxes
cannot be imagined
A smallest local change or action 1 – show that all
value ħ is consistent consequences, however
weird, are confirmed by
observation
2 – deduce quantum theory
from it and check it
Can a minimum change really exist in nature? Table 1 shows that we need to explore
three points to accept the idea. We need to show that no smaller change is observed in
nature, that no smaller change can ever be observed, and show that all consequences of
this smallest change, however weird they may be, apply to nature. In fact, this is all of
quantum physics. Therefore, these checks are all we do in the remaining of this part of
our adventure. But before we explore some of the experiments that confirm the existence
of a smallest change, we present some of its more surprising consequences.
The effects of the quantum of action on rest

Since action is a measure of change, a minimum observable action means that two suc-
cessive observations of the same system always differ by at least ħ. In every system, there
is always something happening. As a consequence, in nature there is no rest. Everything
Page 14 moves, all the time, at least a little bit. Natura facit saltus.* True, these jumps are tiny, as ħ
is too small to be observable by any of our senses. But for example, the quantum of action
implies that in a mountain – an archetypal ‘system at rest’ – all the atoms and electrons
are continually buzzing around. Rest can be observed only macroscopically, and only as
a long-time or many-particle average.
Since there is a minimum action for all observers, and since there is no rest, in nature
there is no perfectly straight or perfectly uniform motion. Forget all you have learnt so far.
* ‘Nature makes jumps.’

Inertial motion is an approximation! An object can move in straight, uniform motion

only approximately, and only when observed over long distances or long times. We will
see later that the more massive the object is, the better the approximation is. (Can you
Challenge 4 s confirm this?) So macroscopic observers can still talk about space-time symmetries; and
special relativity can thus be reconciled with quantum theory.
Also free fall, or motion along a geodesic, exists only as a long-time average. So gen-
eral relativity, which is based on the existence of freely-falling observers, cannot be cor-
rect when actions of the order of ħ are involved. Indeed, the reconciliation of the quan-
tum principle with general relativity – and thus with curved space – is a big challenge.
(The solution is simple only for weak, everyday fields.) The issues involved are so mind-
shattering that they form a separate, final, part of this mountain ascent. We thus explore
situations without gravity first.
The consequences of the quantum of action for objects
Have you ever wondered why leaves are green? You probably know that they are green
because they absorb blue (short-wavelength) and red (long-wavelength) light, while al-
lowing green (medium-wavelength) light to be reflected. How can a system filter out the
small and the large, and let the middle pass through? To do so, leaves must somehow
measure the frequency. But we have seen that classical physics does not allow measure-
ment of time (or length) intervals, as any measurement requires a measurement unit, and
Page 335 classical physics does not allow such units to be defined. On the other hand, it takes only
a few lines to confirm that with the help of the quantum of action ħ (and the Boltzmann
constant k, which Planck’s discovered at the same time), fundamental units for all mea-
surable quantities can be defined, including time and therefore frequency. (Can you find
Challenge 5 s a combination of c, G and ħ that gives a time? It will only take a few minutes.) When
Planck had found all such combinations, he was as happy as a child; he knew straight
away that he had made a fundamental discovery, even though (in 1899) quantum theory
did not yet exist. He even told his seven-year-old son Erwin about it, while walking with
Ref. 5 him through the woods around Berlin. Planck explained to his son that he had made
a discovery as important as universal gravity. Indeed, Planck knew that he had found
the key to understanding many of the effects that were then unexplained. In particular,
without the quantum of action, colours could not exist: every colour is a quantum effect.*
Planck also realized that the quantum of action allows us to understand the size of
Challenge 7 e all things. (Can you find the combination of c, G and ħ that yields a length?) With the
quantum of action, it was finally possible to determine the maximum size of mountains,
Page 266 of trees and of humans. Planck knew that the quantum of action confirmed what Galileo
had already deduced long before him: that sizes are due to fundamental, minimal scales
in nature.
The size of objects is related to the size of atoms. In turn, the size of atoms is a direct
consequence of the quantum of action. Can you derive an approximation for the size
of atoms, knowing that it is given by the motion of electrons of mass me and charge e,
Challenge 8 s constrained by the quantum of action? This connection, a simple formula, was discovered
* In fact, it is also possible to define all measurement units in terms of c, G and e, the electron charge. Why
Challenge 6 s is this not satisfactory?
F I G U R E 5 An artist’s impression of a water molecule
in 1910 by Arthur Erich Haas, 15 years before quantum theory was formulated. At the
time, he was widely ridiculed. Nowadays, the formula is found in all textbooks.*
In determining the size of atoms, the quantum of action has another important con-
sequence: Gulliver’s travels are impossible. There are no tiny people and no giant ones.
Classically, nothing speaks against the idea; but the quantum of action prevents it. Can
Challenge 9 s you supply the detailed argument?
But if rest does not exist, how can shapes exist? Any shape of everyday life, includ-
ing that of a flower, is the result of body parts remaining at rest with respect to each
other. Now, all shapes result from interactions between the constituents of matter, as
shown most clearly in the shapes of molecules. But how can a molecule, such as the wa-
ter molecule H2 O, shown in Figure 5, have a shape? In fact, a molecule does not have a
fixed shape, but its shape fluctuates, as would be expected from the quantum of action.
Despite the fluctuations, every molecule does have an average shape, because different
angles and distances correspond to different energies. Again, these average length and
angle values only exist because the quantum of action yields fundamental length scales
in nature. Without the quantum of action, there would be no shapes in nature.
The mass of an object is also a consequence of the quantum of action, as we will see
later on. Since all material properties – such as density, colour, stiffness or polarizability
– are defined as combinations of length, time and mass units, we find that all material
properties arise from the quantum of action.
In short, the quantum of action determines the size, shape, colour, mass, and all other
properties of objects, from stones to whipped cream. Measurements are only possible at
all because of the existence of the quantum of action.
Why ‘quantum’?
Quantum effects surround us on all sides. However, since the minimum action is so small,
its effects on motion appear mostly, but not exclusively, in microscopic systems. The study
of such systems was called quantum mechanics by Max Born, one of the major contribu-
* Before the discovery of ħ, the only simple length scale for the electron was the combination
e 2 /(4πε0 me c 2 ) ≈ 3 fm; this is ten thousand times smaller than an atom. In addition, any length scale con-
taining e is a quantum effect as well, and not a classical length scale.
TA B L E 2 Some small systems in motion and the observed action values for their changes
System and change Action Motion
Light
Smallest amount of light absorbed by a coloured surface 1ħ quantum
Smallest impact when light reflects from mirror 2ħ quantum
Smallest consciously visible amount of light c. 5 ħ quantum
Smallest amount of light absorbed in flower petal 1ħ quantum
Blackening of photographic film c. 3 ħ quantum
Photographic flash c. 1017 ħ classical
Electricity
Electron ejected from atom or molecule c. 1–2 ħ quantum
Electron extracted from metal c. 1–2 ħ quantum
Electron motion inside microprocessor c. 2–6 ħ quantum
Signal transport in nerves, from one molecule to the next c. 5 ħ quantum
Current flow in lightning bolt c. 1038 ħ classical
Materials
Tearing apart two neighbouring iron atoms c. 1–2 ħ quantum
Breaking a steel bar c. 1035 ħ classical
Basic process in superconductivity 1ħ quantum
Basic process in transistors 1ħ quantum
Basic magnetization process 1ħ quantum
Chemistry
Atom collision in liquid at room temperature 1ħ quantum
Shape oscillation of water molecule c. 1 − 5 ħ quantum
Shape change of molecule, e.g. in chemical reaction c. 1 − 5 ħ quantum
Single chemical reaction curling a hair c. 2 − 6 ħ quantum
Tearing apart two mozzarella molecules c. 300 ħ quantum
Smelling one molecule c. 10 ħ quantum
Burning fuel in a cylinder in an average car engine explosion c. 1037 ħ classical
Life
Air molecule hitting eardrum c. 2 ħ quantum
Smallest sound signal detectable by the ear Challenge 10 ny
Single DNA duplication step during cell division c. 100 ħ quantum

Ovule fertilization c. 1014 ħ classical
Smallest step in molecular motor c. 5 ħ quantum
Sperm motion by one cell length c. 1015 ħ classical
Cell division c. 1019 ħ classical
Fruit fly’s wing beat c. 1024 ħ classical
Person walking one body length c. 2 ⋅ 1036 ħ classical
Nuclei and stars
Nuclear fusion reaction in star c. 1 − 5 ħ quantum
Explosion of gamma-ray burster c. 1080 ħ classical
F I G U R E 6 Max Born (1882–1970)
tors to the field.* Later, the term quantum theory became more popular.
Quantum theory arises from the existence of minimum measurable values in nature,
generalizing the idea Galileo had in the seventeenth century. As discussed in detail earl-
Vol. I, page 264 ier on, it was Galileo’s insistence on ‘piccolissimi quanti’ that got him into trouble. We
will soon discover that the idea of a smallest change is necessary for a precise and ac-
curate description of nature. Born adopted Galileo’s term for the new branch of physics.
The English language adopted the Latin singular ‘quantum’ instead of the Italian plural
‘quanti’ or the German plural ‘Quanten’.
Note that the term ‘quantum’ does not imply that all measurement values are multiples
of a smallest one: this is so only in a few rare cases.
Quantum theory is the description of microscopic motion. But when is quantum the-
ory necessary? Table 2 shows that all processes on atomic and molecular scales, including
biological and chemical processes, involve actions with values that are near the quantum
of action. So do processes of light emission and absorption. These phenomena can only
be described with quantum theory.
Table 2 also shows that the term ‘microscopic’ has a different meaning for a physicist
and for a biologist. For a biologist, a system is microscopic if it requires a microscope for
its observation. For a physicist, a system is microscopic if its characteristic action is of the
order of the quantum of action. In other words, for a physicist a system is microscopic if it
is not visible in a (light) microscope. To increase the confusion, some quantum physicists
nowadays call their own class of microscopic systems ‘mesoscopic’, while others call their
systems ‘nanoscopic’. Both terms were introduced only to attract attention and funding:
* Max Born (b. 1882 Breslau, d. 1970 Göttingen) first studied mathematics, then turned to physics. A profes-
sor at Göttingen University, he made the city one of the world centres of physics. He developed quantum
mechanics with his assistants Werner Heisenberg and Pascual Jordan, and then applied it to scattering,
solid-state physics, optics and liquids. He was the first to understood that the state function describes a
Ref. 6 probability amplitude. He and Wolf together wrote what is still the main textbook on optics.
Born attracted to Göttingen the most brilliant talents of the time, receiving as visitors Hund, Pauli, Nord-
heim, Oppenheimer, Goeppert-Mayer, Condon, Pauling, Fock, Frenkel, Tamm, Dirac, Mott, Klein, Heitler,
London, von Neumann, Teller, Wigner, and dozens of others. Being Jewish, Born lost his job in 1933; he
emigrated, and became professor in Edinburgh, where he stayed for 20 years. Physics at Göttingen never
recovered from this loss. For his elucidation of the meaning of the wave function he received the 1954 Nobel
Prize in Physics.
F I G U R E 7 Werner Heisenberg (1901–1976)
they are useless.
The effect of the quantum of action on motion
There is another way to characterize the difference between a microscopic, or quantum,
system and a macroscopic, or classical, one. A minimum action implies that the differ-
ence between the action values S of two successive observations of the same system, a
time Δt apart, is limited. One has
ħ
S(t + Δt) − S(t) = (E + ΔE)(t + Δt) − Et = EΔt + tΔE + ΔEΔt ⩾ . (1)
2
The factor 1/2 arises from averaging. Now the values of the energy E and time t – but not
of ΔE or Δt – can be set to zero if we choose a suitable observer. Thus, the existence of a
quantum of action implies that in any system the evolution is constrained by
ħ
ΔEΔt ⩾ , (2)
2
where E is the energy of the system and t is its age, so that ΔE is the change of energy
and Δt is the time between two successive observations.
Challenge 11 e By a similar reasoning , we find that for any system the position and momentum are
constrained by
ħ
ΔxΔp ⩾ , (3)
2
where Δx is the indeterminacy in position and Δp is the indeterminacy in momen-
tum. These two famous relations were called indeterminacy relations by their discoverer,
Werner Heisenberg.* In English they are often called ‘uncertainty relations’; however,
* It is often said that the indeterminacy relation for energy and time has a different weight from that for
momentum and position. This is a wrong idea, propagated by the older generation of physicists, which has
survived through many textbooks for over 70 years. Just forget it. It is essential to remember that all four
quantities appearing in the inequalities describe the internal properties of the system. In particular, t is a
time variable deduced from changes observed inside the system, and not the time coordinate measured by
this term is incorrect. The quantities are not uncertain, but undetermined: because of the
quantum of action, system observables have no definite value. There is no way to ascribe
a precise value to momentum, position, or any other observable of a quantum system.
Any system whose indeterminacy is of the order of ħ is a quantum system; if the
indeterminacy product is much larger, the system is classical, and classical physics is
sufficient for its description. So even though classical physics assumes that there are no
measurement indeterminacies in nature, a system is classical only if its indeterminacies
are large compared to the minimum possible ones!
In short, quantum theory is necessary whenever we try to measure some quantity as
precisely as possible. In fact, every measurement is itself a quantum process. Therefore,
measurement precision is limited. The quantum of action, through the indeterminacy re-
lations, shows that motion cannot be observed to infinite precision. In other words, the mi-
croscopic world is fuzzy. This fact has many important consequences and many strange
ones. For example, if motion cannot be observed with infinite precision, the very con-
cept of motion needs to be handled with great care, as it cannot be applied in certain
situations. In a sense, the rest of our quest is just an exploration of the implications of
this result.
In fact, as long as space-time is flat, it turns out that we can retain the concept of
motion to describe observations, provided we remain aware of the limitations implied
by the quantum principle.
Quantum limits and surprises

The quantum of action, with its implied fuzziness, implies the existence of short-time
deviations from energy, momentum and angular-momentum conservation in micro-
Page 191 scopic systems. In the first part of our mountain ascent, we realized that any type of non-
conservation implies the existence of surprises in nature. Well, here are some of them.
Since precisely uniform motion does not exist, a system moving in one dimension
only – such as the hand of a clock – always has the possibility of moving a bit in the
opposite direction, thus leading to incorrect readings. Indeed, quantum theory predicts
that clocks have essential limitations, and that perfect clocks do not exist.
It is also impossible to avoid that an object makes small displacement sideways. In
fact, quantum theory implies that, strictly speaking, neither uniform nor one-dimensional
motion exists.
Quantum limitations apply also to metre rules. It is impossible to ensure that the rule
is completely at rest with respect to the object being measured. Thus the quantum of
an outside clock; similarly, the position x is not the external space coordinate, but the position characteriz-
Ref. 7 ing the system.
Werner Heisenberg (1901–1976) was an important German theoretical physicist and an excellent table-
tennis and tennis player. In 1925, as a young man, he developed, with some help from Max Born and Pas-
cual Jordan, the first version of quantum theory; from it he deduced the indeterminacy relations. For these
achievements he received the Nobel Prize for physics in 1932. He also worked on nuclear physics and on
turbulence. During the Second World War, he worked on the German nuclear-fission programme. After the
war, he published several successful books on philosophical questions in physics, slowly turned into a crank,
and tried unsuccessfully – with some half-hearted help from Wolfgang Pauli – to find a unified description
of nature based on quantum theory, the ‘world formula’.
action implies again, on one hand, that measurements are possible, and on the other,
that their accuracy is limited.
It also follows from the quantum of action that any inertial or freely-falling observer
must be large, as only large systems approximate inertial motion. An observer cannot
be microscopic. If humans were not macroscopic, they could neither observe nor study
motion.
Because of the finite accuracy with which microscopic motion can be observed, faster-
than-light motion is possible in the microscopic domain! Quantum theory thus predicts
tachyons, at least over short time intervals. For the same reason, motion backwards in
time is possible over microscopic times and distances. In short, a quantum of action im-
plies the existence of microscopic time travel. However, this remains impossible in the
macroscopic domain, such as everyday life.
But there is more: the quantum of action implies that there is no permanence in na-
ture. Imagine a moving car suddenly disappearing for good. In such a situation, neither
momentum nor energy would be conserved. The action change for such a disappearance
is large compared to ħ, so that its observation would contradict even classical physics –
Challenge 12 s as you may wish to check. However, the quantum of action allows a microscopic particle,
such as an electron, to disappear for a short time, provided it reappears afterwards.
The quantum of action also implies that the vacuum is not empty. If one looks at empty
space twice, the two observations being separated by a tiny time interval, some energy
will be observed the second time. If the time interval is short enough, then because of
the quantum of action, matter particles will be observed. Indeed, particles can appear
anywhere from nowhere, and disappear just afterwards: the action limit requires it. In
summary, nature exhibits short-term appearance and disappearance of matter. In other
words, the classical idea of an empty vacuum is correct only when the vacuum is observed
over a long time.
The quantum of action implies that compass needles cannot work. If we look twice in
quick succession at a compass needle, or even at a house, we usually observe that it stays
oriented in the same direction. But since physical action has the same dimensions as
Challenge 13 e angular momentum, a minimum value for action implies a minimum value for angular
momentum. Even a macroscopic object has a minimum value for its rotation. In other
words, quantum theory predicts that everything rotates. Something can be non-rotating
only approximately, when observations are separated by long time intervals.
For microscopic systems, the quantum limits on rotation have specific effects. If the
rotation angle can be observed – as for molecules – the system behaves like macroscopic
object: its position and orientation are fuzzy. But for a system whose rotation angle can-
not be observed, the quantum of action limits the angular momentum to multiples of
ħ/2. In particular, all microscopic bound systems – such as molecules, atoms, or nuclei
– contain rotational motion and rotating components.
Transformation, life and Democritus

At the beginning of our adventure, we mentioned that the Greeks distinguished three
Vol. I, page 21 types of changes: transport, growth, and transformation. We also mentioned that Dem-
ocritus had deduced that all these types of changes – including life and death – were in
fact the same, and due to the motion of atoms. The quantum of action makes exactly this
E
m p
F I G U R E 8 Hills are never high

0 Δx
enough
point.
First of all, a minimum action implies that cages in zoos are dangerous and banks are
not safe. A cage is a feature that it needs a lot of energy to overcome. Mathematically, the
wall of a cage is an energy hill, like the one shown in Figure 8. Imagine that a particle
with momentum p approaches one side of the hill, which is assumed to have width Δx.
In everyday life – and thus in classical physics – the particle will never be observed
on the other side of the hill if its kinetic energy p2 /2m is less than the height E of the
hill. But imagine that the missing momentum to overcome the hill, Δp = 󵀂2mE − p,
satisfies ΔxΔp ⩽ ħ/2. The particle will have the possibility to overcome the hill, despite
its insufficient energy. The quantum of action thus implies that a hill of width
ħ/2
Δx ⩽ (4)
󵀂2mE − p
is not an obstacle to a particle of mass m. But this is not all. Since the value of the par-
ticle momentum p is itself uncertain, a particle can overcome the hill even if the hill is
wider than the value (4) – although the broader it is, the lower the probability will be. So
any particle can overcome any obstacle. This is called the tunnelling effect, for obvious
reasons.
In short, the minimum-action principle implies that there are no tight boxes in nature.
Thanks to the tunnelling effect, matter is not impenetrable, in contrast to everyday, clas-
Challenge 14 s sical observation. Can you explain why lion cages work despite the quantum of action?
By the way, the quantum of action also implies that a particle with a kinetic energy
greater than the energy height of a hill can be reflected by the hill. Classically, this is
Challenge 15 ny impossible. Can you explain it?

The minimum-action principle also implies that bookshelves are dangerous. Why?
Shelves are obstacles to motion. A book on a shelf is in the same situation as the mass in
Figure 9: the mass is surrounded by energy hills hindering its escape to the outer, lower-
energy world. But thanks to the tunnelling effect, escape is always possible. The same
picture applies to a branch of a tree, a nail in a wall, or anything attached to anything
else. Things can never be permanently fixed together. In particular, we will discover that
every example of light emission – even radioactivity – results from this effect. The quan-
tum of action thus implies that decay is part of nature. Note that decay often appears in
everyday life, under a different name: breaking. In fact, all breakages require the quantum
Ref. 8 of action for their description. Obviously, the cause of breaking is often classical, but the
E1
m E2
F I G U R E 9 Leaving enclosures
m
F I G U R E 10 Identical objects with
crossing paths
mechanism of breaking is always quantum. Only objects that obey quantum theory can
break. In short, there are no stable excited systems in nature. For the same reason, by the
Challenge 16 ny way, no memory can be perfect. (Can you confirm this?)
Taking a more general view, ageing and death also result from the quantum of action.
Death, like ageing, is a composition of breaking processes. Breaking is a form of decay,
and is due to tunnelling. Death is thus a quantum process. Classically, death does not
Challenge 17 s exist. Might this be the reason why so many people believe in immortality or eternal
youth?
We will also discover that the quantum of action is the reason for the importance of
the action observable in classical physics. In fact, the existence of a minimal action is the
reason for the least-action principle of classical physics.
A minimum action also implies that matter cannot be continuous, but must be com-
posed of smallest entities. Indeed, any flow of a truly continuous material would contra-
Challenge 18 s dict the quantum principle. Can you give the precise argument? Of course, at this point
in our adventure, the non-continuity of matter is no longer a surprise. But the quantum
of action implies that even radiation cannot be continuous. As Albert Einstein was the
first to state clearly, light is made of quantum particles.
More generally, the quantum of action implies that in nature all flows and all waves are
made of microscopic particles. The term ‘microscopic’ (or ‘quantum’) is essential, as such
particles do not behave like little stones. We have already encountered several differences,
and we will encounter others shortly. For these reasons, there should be a special name
for microscopic particles; but so far all proposals, of which quanton is the most popular,
M
m1
m2
m
m3
F I G U R E 11 Transformation through
reaction
have failed to catch on.
The quantum of action has several strange consequences for microscopic particles.
Take two such particles with the same mass and composition. Imagine that their paths
cross, and that at the crossing they approach each other very closely, as shown in
Figure 10. A minimum action implies that in such a situation, if the distance becomes
small enough, the two particles can switch roles, without anybody being able to avoid, or
notice, it. Thus, in a volume of gas it is impossible – thanks to the quantum of action – to
follow particles moving around and to say which particle is which. Can you confirm this
Challenge 19 s deduction, and specify the conditions, using the indeterminacy relations? In summary,
in nature it is impossible to distinguish between identical particles. Can you guess what
Challenge 20 s happens in the case of light?
But matter deserves still more attention. Imagine again two particles – even two dif-
ferent ones – approaching each other very closely, as shown in Figure 11. We know that
if the approach distance gets small, things get fuzzy. Now, the minimum-action princi-
ple makes it possible for something to happen in that small domain as long as resulting
outgoing products have the same total linear momentum, angular momentum and en-
ergy as the incoming ones. Indeed, ruling out such processes would imply that arbitrarily
small actions could be observed, thus eliminating nature’s fuzziness, as you may wish to
Challenge 21 e check for yourself. In short, a minimum action allows transformation of matter. One also
says that the quantum of action allows particle reactions. In fact, we will discover that all
kinds of reactions in nature, including breathing, digestion, and all other chemical and
nuclear reactions, are due just to the existence of the quantum of action.
One type of process that is especially dear to us is growth. The quantum of action
implies that all growth happens in small steps. Indeed, all growth processes in nature are
quantum processes.
Above all, as mentioned already, the quantum of action explains life. Only the quan-
tum of action makes reproduction and heredity possible. In short, birth, sexuality and
death are consequences of the quantum of action.
So Democritus was both right and wrong. He was right in deducing fundamental
constituents for matter and radiation. He was right in unifying all change in nature –
from transport to transformation and growth – as motion of particles. But he was wrong
in assuming that the small particles behave like stones.
F I G U R E 12 A famous quantum effect: how do train windows manage to show two superimposed
images? (photo © Greta Mansour)
Randomness – a consequence of the quantum of action
What happens if we try to measure changes smaller than the quantum of action? Na-
ture has a simple answer: we get random results. If we build an experiment that tries to
produce a change of the size of a quarter of the quantum of action, the experiment will
produce a change of one quantum of action in a quarter of the cases, and no change in
three quarters of the cases, thus giving an average of one quarter of ħ.
The quantum of action leads to randomness. This can be seen also in another way.
Because of the indeterminacy relations, it is impossible to obtain definite values for both
the momentum and the position of a particle. Obviously, this is also impossible for the
individual components of an experimental set-up or an observer. Therefore, initial con-
ditions – both for a system and for an experimental set-up – cannot be exactly duplicated.
A minimum action thus implies that whenever an experiment on a microscopic system
is performed twice, the outcomes will be different. The outcomes could only be the same
if both the system and the observer were in exactly the same configuration each time.
This turns out to be impossible, because of the second principle of thermodynamics and
because of the quantum of action. Therefore, microscopic systems behave randomly. Obvi-
ously, there will be some average outcome; but in all cases, microscopic observations are
probabilistic. Many find this conclusion of quantum theory the most difficult to swallow.
The quantum of action implies that the behaviour of quantum systems is strikingly dif-
ferent from that of classical systems. But the conclusion is unavoidable: nature behaves
randomly.
Can we observe randomness in everyday life? Yes. Every window proves that nature
behaves randomly on a microscopic scale. Everybody knows that one can use a train win-
dow either to look at the outside landscape or, by concentrating on the reflected image, to
observe some interesting person inside the carriage. In other words, glass reflects some
of the light particles and lets some others pass through. More precisely, glass reflects a
random selection of light particles; yet the average proportion is constant. Partial reflec-
tion is thus similar to the tunnelling effect. Indeed, the partial reflection of photons in
glass is a result of the quantum of action. Again, the situation can be described by classi-
cal physics, but the precise amount of reflection cannot be explained without quantum
theory. Without the quantum of action, train journeys would be much more boring.
F I G U R E 13 A particle and a screen with two nearby slits
Waves – a consequence of the quantum of action
The quantum of action implies an important result about the paths of particles. If a par-
ticle travels from one point to another, there is no way to say which path it has taken
in between. Indeed, in order to distinguish between two possible, but slightly different,
paths, actions smaller than ħ would have to be measured reliably. In particular, if a par-
ticle is sent through a screen with two sufficiently close slits, as illustrated in Figure 13,
it is impossible to say which slit the particle passed through. This impossibility is funda-
mental.
We already know phenomena of motion for which it is not possible to say with preci-
sion how something moves or which path is taken behind two slits: waves behave in this
Page 247 way. All waves are subject to the indeterminacy relations
1 1
ΔωΔt ⩾ and ΔkΔx ⩾ . (5)
2 2
A wave is a type of motion described by a phase that changes over space and time. This
turns out to hold for all motion. In particular, this holds for matter.
We saw above that quantum systems are subject to
ħ ħ
ΔEΔt ⩾ and ΔpΔx ⩾ . (6)
2 2
We are thus led to ascribe a frequency and a wavelength to a quantum system:
2π
E = ħω and p = ħk = ħ . (7)
λ
The energy–frequency relation was deduced by Albert Einstein in 1905; it is found to be
valid in particular for all examples of emission and absorption of light. In 1923 and 1924,
Louis de Broglie* predicted that the relation should hold also for all quantum matter
Page 63 particles. The experimental confirmation came a few years later. (This is thus another
* Louis de Broglie (b. 1892 Dieppe, d. 1987 Paris), French physicist and professor at the Sorbonne. The
energy–frequency relation for light had earned Albert Einstein his Nobel Prize already in 1921. De Broglie
example of a discovery that was made about 20 years too late.) In short, the quantum
of action implies that matter particles behave like waves. In particular, the quantum of
action implies the existence of interference for matter.
Particles – a consequence of the quantum of action

The quantum of action, the smallest change, implies that flows cannot be arbitrary weak.
Page 289 This applies to all flows: in particular, it applies to rivers, solid matter flows, gas flows,
light beams, energy flows, entropy flows, momentum flows, angular momentum flows,
probability flows, signals of all kind, electrical charge flows, colour charge flows and weak
charge flows.
Water flows in rivers, like any other matter flow, cannot be arbitrary small: the quan-
tum of action implies that there is a smallest matter flow in nature. Depending on the
situation, the smallest matter flow is a molecule, an atom or a smaller particle. Indeed,
the quantum of action is also at the origin of the observation of a smallest charge in elec-
tric current. Since all matter can flow, the quantum of action implies that all matter has
particle aspects.
In the same way, the quantum of action, the smallest change, implies that light cannot
be arbitrarily faint. There is a smallest illumination in nature; it is called a photon or a
light quantum. Now, light is a wave, and the argument can be made for any other wave
as well. In short, the quantum of action thus implies that all waves have particle aspects.
This has been proved for light waves, water waves, X rays, sound waves, plasma waves,
fluid whirls and any other wave type that has ever been observed. (Gravitational waves
have not yet been observed; it is expected that their particle-like aspects, the gravitons,
exist also in this case.)
In summary, the quantum of action states: if it moves, it is made of quantum parti-
cles, or quantons. Later on we will explore and specify the exact differences between a
quantum particle and a small stone or a grain of sand. We will discover that matter quan-
tons move differently, behave differently under rotation, and behave differently under
exchange.
Quantum information
In computer science, the smallest unit of change is called a ‘bit change’. The existence
of a smallest change in nature implies that computer science – or information science
– can be used to describe nature, and in particular quantum theory. This analogy has
attracted much research in the past decades, and explored many interesting questions: Is
there unlimited information storage possible? Can information be read out and copied
completely? Can information be transmitted while keeping it secret? Can information
transmission and storage be performed independently of noise? Can quantum physics
be used to make new types of computers? So far, the answer to all these questions is
negative; but the hope to change the situation is not dead yet.
expanded it to predict the wave nature of the electron (and of all other quantum matter particles): this was
the essence of his doctoral thesis. The prediction was confirmed experimentally a few years later, in 1927. For
the prediction of the wave nature of matter, de Broglie received the Nobel Prize for physics in 1929. Being
an aristocrat, he did no more research after that. For example, it was Schrödinger who then wrote down the
wave equation, even though de Broglie could equally have done so.
The analogy between quantum theory and information science has limitations: infor-
mation science can describe only the ‘software’ side of devices. For a physicist, the ‘hard-
ware’ side of nature is central. The hardware of nature enters the description whenever
the actual value ħ of the quantum of action must be introduced.
As we explore the similarities and differences between nature and information science,
we will discover that the quantum of action implies that macroscopic physical systems
cannot be copied – or ‘cloned’, as quantum theorists like to say. Nature does not allow
copies of macroscopic objects. In other words, perfect copying machines do not exist.
The quantum of action makes it impossible to gather and use all information in a way
that allows production of a perfect copy.
The exploration of copying machines will remind us again that the precise order in
which measurements are performed in an experiment matters. When the order of mea-
surements can be reversed without affecting the net result, physicists speak of ‘commu-
tation’. The quantum of action implies that physical observables do not commute.
We will also find that the quantum of action implies that systems are not always in-
Page 127 dependent, but can be entangled. This term, introduced by Erwin Schrödinger, describes
one of the most absurd consequences of quantum theory. Entanglement makes every-
thing in nature connected to everything else. Entanglement produces effects that seem
(but are not) faster than light. Entanglement produces a (fake) form of non-locality. En-
Ref. 9 tanglement also implies that trustworthy communication cannot exist.
We will also discover that decoherence is an ubiquitous process in nature that influ-
ences all quantum systems; it allows measurements on one hand and makes quantum
Page 132 computers impossible on the other.
Curiosities and fun challenges about the quantum of action

Even if we accept that no experiment performed so far contradicts the minimum action,
we still have to check that the minimum action does not contradict reason. In particular,
the minimum action must also be consistent with all imagined experiments. This is not
self-evident.
∗∗
Angular momentum has the same dimensions as action. A smallest action implies that
there is a smallest angular momentum in nature. How can this be, given that some par-
Challenge 22 s ticles have spin zero, i.e., have no angular momentum?
∗∗
Could we have started the whole discussion of quantum theory by stating that there is a
Challenge 23 s minimum angular momentum instead of a minimum action?
∗∗
Niels Bohr, besides propagating the idea of a minimum action, was also an enthusiast of
the so-called complementarity principle. This is the idea that certain pairs of observables
of a system – such as position and momentum – have linked precision: if one of the pair
is known to high precision, the other is necessarily known with low precision. Can you
Challenge 24 s deduce this principle from the minimum action?
∗∗
When electromagnetic fields come into play, the value of the action (usually) depends on
the choice of the vector potential, and thus on the choice of gauge. We saw in the section
on electrodynamics that a suitable choice of gauge can change the value of the action
by adding or subtracting any desired amount. Nevertheless, there is a smallest action in
nature. This is possible, because in quantum theory, physical gauge changes cannot add
or subtract any amount, but only multiples of twice the minimum value. Thus they do
not allow us to go below the minimum action.
∗∗
Adult plants stop growing in the dark. Without light, the reactions necessary for growth
Challenge 25 s cease. Can you show that this is a quantum effect, not explainable by classical physics?
∗∗
Most quantum processes in everyday life are electromagnetic. Can you show that the
quantum of action must also hold for nuclear processes, i.e., for processes that are not
Challenge 26 s electromagnetic?
∗∗
Challenge 27 s Is the quantum of action independent of the observer, even near the speed of light? This
question was the reason why Planck contacted the young Einstein, inviting him to Berlin,
thus introducing him to the international physics community.
∗∗
The quantum of action implies that tiny people, such as Tom Thumb, cannot exist. The
quantum of action implies that fractals cannot exist in nature. The quantum of action
implies that ‘Moore’s law’ of semiconductor electronics, which states that the number of
Challenge 28 ny transistors on a chip doubles every two years, cannot be correct. Why not?
∗∗
Take a horseshoe. The distance between the two ends is not fixed, since otherwise their
position and velocity would be known at the same time, contradicting the indeterminacy
relation. Of course, this reasoning is also valid for any other solid object. In short, both
quantum mechanics and special relativity show that rigid bodies do not exist, albeit for
different reasons.
The dangers of buying a can of beans

Another way to show the absurd consequences of quantum theory is given by the ul-
timate product warning, which according to certain well-informed lawyers should be
Ref. 10 printed on every can of beans and on every product package. It shows in detail how
deeply our human condition fools us.
Warning: care should be taken when looking at this product:
It emits heat radiation.
Bright light has the effect to compress this product.
Warning: care should be taken when touching this product:

Part of it could heat up while another part cools down, causing severe burns.
Warning: care should be taken when handling this product:
This product consists of at least 99.999 999 999 999 % empty space.
This product contains particles moving with speeds higher than one million kilometres per
hour.
Every kilogram of this product contains the same amount of energy as liberated by about
one hundred nuclear bombs.*
In case this product is brought in contact with antimatter, a catastrophic explosion will
occur.
In case this product is rotated, it will emit gravitational radiation.
Warning: care should be taken when transporting this product:
The force needed depends on its velocity, as does its weight.
This product will emit additional radiation when accelerated.
This product attracts, with a force that increases with decreasing distance, every other ob-
ject around, including its purchaser’s kids.
Warning: care should be taken when storing this product:
It is impossible to keep this product in a specific place and at rest at the same time.
Except when stored underground at a depth of several kilometres, over time cosmic radia-
tion will render this product radioactive.
This product may disintegrate in the next 1035 years.
It could cool down and lift itself into the air.
This product warps space and time in its vicinity, including the storage container.
Even if stored in a closed container, this product is influenced and influences all other
objects in the universe, including your parents in law.
This product can disappear from its present location and reappear at any random place in
the universe, including your neighbour’s garage.
Warning: care should be taken when travelling away from this product:
It will arrive at the expiration date before the purchaser does so.
Warning: care should be taken when using this product:

Any use whatsoever will increase the entropy of the universe.
The constituents of this product are exactly the same as those of any other object in the
universe, including those of rotten fish.
The impression of a certain paranoid side to quantum physics is purely coincidental.
* A standard nuclear warhead has an explosive yield of about 0.2 megatons (implied is the standard explosive
Ref. 11 trinitrotoluene or TNT), about thirteen times the yield of the Hiroshima bomb, which was 15 kilotonne. A
megatonne is defined as 1 Pcal=4.2 PJ, even though TNT delivers about 5 % slightly less energy than this
value. In other words, a megaton is the energy content of about 47 g of matter. That is less than a handful
for most solids or liquids.
A summary: quantum physics, the law and indoctrination

Don’t all the deductions from the quantum of action presented so far look wrong, or
at least crazy? In fact, if you or your lawyer made some of the statements on quantum
physics in court, maybe even under oath, you might end up in prison! However, all the
above statements are correct: they are all confirmed by experiment. And there are many
more surprises to come. You may have noticed that, in the preceding examples, we have
made no explicit reference to electricity, to the nuclear interactions, or to gravity. In these
domains the surprises are even more astonishing. Observation of antimatter, electric cur-
rent without resistance, the motion inside muscles, vacuum energy, nuclear reactions in
stars, and maybe soon the boiling of empty space, will fascinate you as much as they have
fascinated, and still fascinate, thousands of researchers.
In particular, the consequences of the quantum of action for the early universe are
Challenge 29 d mind-boggling. Just try to explore for yourself its consequences for the big bang. To-
gether, all these topics will lead us a long way towards the top of Motion Mountain. The
consequences of the quantum of action are so strange, so incredible, and so numerous,
that quantum physics can rightly be called the description of motion for crazy scientists.
In a sense, this generalizes our previous definition of quantum physics as the description
of motion related to pleasure.
Unfortunately, it is sometimes said that ‘nobody understands quantum theory’. This
is wrong. In fact, it is worse than wrong: it is indoctrination and disinformation. Dicta-
torships use indoctrination and disinformation to prevent people from making up their
own mind and from enjoying life. But the consequences of the quantum of action can
be understood and enjoyed by everybody. In order to do so, our first task on our way
towards the top of Motion Mountain will be to use the quantum of action to study of our
classical standard of motion: the motion of light.
“
Nie und nirgends hat es Materie ohne
”
Bewegung gegeben, oder kann es sie geben.
Friedrich Engels, Anti-Dühring.*
Ref. 12 * ‘Never and nowhere has matter existed, nor can it exist, without motion.’ Friedrich Engels (1820–1895)
was one of the theoreticians of Marxism, often also called Communism.
Chapter 2
LIGHT – THE STR ANGE

CONSEQUENC ES OF THE QUANTUM
OF AC TION
“
Alle Wesen leben vom Lichte,
”
jedes glückliche Geschöpfe.
Friedrich Schiller, Wilhelm Tell.*
What is the faintest lamp?
S
ince all the colours of materials are quantum effects, it becomes mandatory to
tudy the properties of light itself. If there is a smallest change, then there
hould also be a smallest illumination in nature. This was already predicted in an-
Ref. 13 cient Greece, for example by Epicurus (341–271 bce ). He stated that light is a stream of
little particles, so that the smallest possible illumination would be that due to a single
light particle. Today, the particles are called light quanta or photons. Incredibly, Epicurus
himself could have checked his prediction with an experiment.
Ref. 14 In the 1930s Brumberg and Vavilov found a beautiful way to check the existence of
photons using the naked eye and a lamp. Our eyes do not allow us to consciously detect
single photons, but Brumberg and Vavilov found a way to circumvent this limitation.
In fact, the experiment is so simple that it could have been performed many centuries
earlier; but nobody had had a sufficiently daring imagination to try it.
Brumberg and Vavilov constructed a small shutter that could be opened for time inter-
vals of 0.1 s. From the other side, in a completely dark room, they illuminated the opening
with extremely weak green light: about 200 aW at 505 nm, as shown in Figure 14. At that
intensity, whenever the shutter opens, on average about 50 photons can pass. This is just
the sensitivity threshold of the eye. To perform the experiment, they repeatedly looked
into the open shutter. The result was simple but surprising. Sometimes they observed
light, and sometimes they did not. Whether they did or did not was completely random.
Brumberg and Vavilov gave the simple explanation that at low lamp powers, because of
fluctuations, the number of photons is above the eye threshold half the time, and below it
the other half. The fluctuations are random, and so the conscious detection of light is as
well. This would not happen if light were a continuous stream: in that case, the eye would
detect light at every opening of the shutter. (At higher light intensities, the percentage of
non-observations quickly decreases, in accordance with the explanation given.) In short,
light is made of photons. Nobody knows how the theory of light would have developed
if this simple experiment had been performed 100 or even 2000 years earlier.
* ‘From light all beings live, each fair-created thing.’ Friedrich Schiller (b. 1759 Marbach, d. 1805 Weimar),
German poet, playwright and historian.
36 2 light
lamp strong shutter head, after

filter 45 minutes
in complete
F I G U R E 14 How to experience single
darkness
photon effects (see text)
photographic
glass film
F I G U R E 15 How does a
white-light spectrum appear at
white red extremely long screen distances?
green (The short-screen-distance
violet spectrum shown, © Andrew
Young, is optimized for CRT
display, not for colour printing,
as explained on mintaka.sdsu.
edu/GF/explain/optics/
rendering.html.)
The detection of photons becomes clearer if we use devices to help us. A simple way is
to start with a screen behind a prism illuminated with white light, as shown in Figure 15.
The light is split into colours. As the screen is placed further and further away, the illumi-
nation intensity cannot become arbitrarily small, as that would contradict the quantum
of action. To check this prediction, we only need some black-and-white photographic
film. Film is blackened by daylight of any colour; it becomes dark grey at medium inten-
sities and light grey at lower intensities. Looking at an extremely light grey film under
the microscope, we discover that, even under uniform illumination, the grey shade is
actually composed of black spots, arranged more or less densely. All these spots have the
same size, as shown in Figure 16. This regular size suggests that a photographic film reacts
to single photons. Detailed research confirms this conjecture; in the twentieth century,
the producers of photographic films have elucidated the underlying mechanism in all its
details.
Single photons can be detected most elegantly with electronic devices. Such devices
can be photomultipliers, photodiodes, multichannel plates or rod cells in the eye; a se-
lection is shown in Figure 17. Also these detectors show that low-intensity light does not
produce a homogeneous colour: on the contrary, low-intensity produces a random pat-
light – the strange consequences of the quantum of action 37
F I G U R E 16 Exposed photographic film at increasing magnification (© Rich Evans)
F I G U R E 17 Detectors that allow photon counting: photomultiplier tubes (left), an avalanche
photodiode (top right, c. 1 cm) and a multichannel plate (bottom right, c. 10 cm) (© Hamamatsu
Photonics)
tern of equal spots, even when observing typical wave phenomena such as interference
patterns, as shown in Figure 18.
The observation is general: whenever sensitive light detectors are constructed with the
aim of ‘seeing’ as accurately as possible (and thus in environments as dark as possible),
one finds that light manifests as a stream of light quanta. Nowadays they are usually called
photons, a term that appeared in 1926. Light of low or high intensity corresponds to a
stream with a small or large number of photons.
A particularly interesting example of a low-intensity source of light is a single atom.
Atoms are tiny spheres. When atoms radiate light or X-rays, the radiation should be emit-
ted as a spherical wave. But in all experiments, the light emitted by an atom is never found
to form a spherical wave, in contrast to what we might expect from everyday physics.
38 2 light
F I G U R E 18 Light waves are made of particles: observation of photons – black spots in these negatives
– in a low intensity double slit experiment, with exposure times of 1, 2 and 5 s, using an image
intensifier (© Delft University of Technology)
light detectors
radiating
atom
F I G U R E 19 An atom radiating one

photon triggers only one detector and
recoils in only one direction
Whenever a radiating atom is surrounded by many detectors, only a single detector is

triggered. Only the average over many emissions and detections yields a spherical shape.
The experiments shows clearly that partial photons cannot be detected.
All experiments in dim light show that the continuum description of light is incorrect.
All these experiments thus prove directly that light is a stream of particles, as Epicurus
had proposed in ancient Greece. More precise measurements confirm the role of the
quantum of action: every photon leads to the same amount of change. All photons of
the same frequency blacken a film or trigger a scintillation screen in the same way. The
amount of change induced by a single photon is indeed the minimal amount of change
that light can produce.
If there were no minimum action, light could be packaged into arbitrarily small
amounts. But this is not possible. In other words, the classical description of light by
Page 73 a continuous vector potential A(t, x), or electromagnetic field F(t, x), whose evolution is
described by a principle of least action, is wrong. Continuous functions do not describe
the observed particle effects. A modified description is required. The modification has to
be significant only at low light intensities, since at high intensities the classical Lagrangian
accurately describes all experimental observations.*
At which intensities does light cease to behave as a continuous wave? Human eyesight
does not allow us to consciously distinguish single photons, although experiments show
Ref. 15 that the hardware of the eye is in principle able to do so. The faintest stars that can be
seen at night produce a light intensity of about 0.6 nW/m2 . Since the pupil of the eye is
small, and we are not able to see individual photons, photons must have energies smaller
than 100 aJ. Brumberg and Vavilov’s experiment yields an upper limit of around 20 aJ.
An exact value for the quantum of action found in light must be deduced from labo-
ratory experiment. Today, recording and counting individual photons is a standard pro-
cedure. Photon counters are part of many spectroscopy set-ups, such as those used to
measure tiny concentrations of materials. For example, they are used to detect drugs in
human hair.
Photons
In general, all experiments show that a beam of light of frequency f or angular frequency
ω, which determines its colour, is accurately described as a stream of photons, each with
the same energy E given by
E = ħ 2π f = ħ ω . (8)
This relation was first deduced by Max Planck in 1899. He showed that for light, the
smallest measurable action is given by the quantum of action ħ. In summary, colour is a
property of photons. A coloured light beam is a hailstorm of corresponding photons.
The value of Planck’s constant can be determined from measurements of black bodies
Page 118 or other light sources. All such measurements coincide and yield
ħ = 1.1 ⋅ 10−34 Js , (9)
a value so small that we can understand why photons go unnoticed by humans. For ex-
Challenge 30 e ample, a green photon with a wavelength of 555 nm has an energy of 0.37 aJ. Indeed, in
normal light conditions the photons are so numerous that the continuum approximation
for the electromagnetic field is highly accurate. In the dark, the insensitivity of the signal
processing of the human eye – in particular the slowness of the light receptors – makes
Ref. 15 photon counting impossible. However, the eye is not far from the maximum possible
sensitivity. From the numbers given above about dim stars, we can estimate that humans
Challenge 31 ny are able to see consciously, under ideal conditions, flashes of about half a dozen photons.
Let us explore the other properties of photons. Above all, photons have no measur-
Challenge 32 ny able (rest) mass and no measurable electric charge. Can you confirm this? In fact, exper-
iments can only give an upper limit for both quantities. The present experimental upper
* The transition from the classical case to the quantum case used to be called quantization. This concept,
and the ideas behind it, are only of historical interest today.
40 2 light
X-ray
detector
deflected
photon with photon after
wavelength λ deflection
the collision,
angle
with wave-
length λ+Δλ
X-ray collision
source in
X-ray sample X-ray electron
sample
source detector after the
collision
F I G U R E 20 A modern version of Compton’s experiment fits on a table. The experiment shows that
photons have momentum: X-rays – and thus the photons they consist of – change frequency when
they hit the electrons in matter in exactly the same way as predicted from colliding particles. (© Helene
Hoffmann)
Ref. 16 limit for the (rest) mass of a photon is 10−52 kg, and for the charge is 5 ⋅ 10−30 times the
electron charge. These limits are so small that we can safely say that both the mass and
the charge of the photon vanish.
We know that light can hit objects. Since the energy and the speed of photons is
Challenge 33 ny known, we guess that the photon momentum is given by
E 2π
p= =ħ or p = ħk . (10)
c λ
In other words, if light is made of particles, we should be able to play billiard with them.
Ref. 17 This is indeed possible, as Arthur Compton showed in a famous experiment in 1923.
He directed X-rays, which are high-energy photons, onto graphite, a material in which
electrons move almost freely. He found that whenever the electrons in the material are
hit by the X-ray photons, the deflected X-rays change colour. As expected, the strength
of the hit is related to the deflection angle of the photon, as deduced from Figure 20.
From the colour change and the reflection angle, Compton confirmed that the photon
momentum indeed satisfies the expression p = ħ k.
All other experiments agree that photons have momentum. For example, when an
atom emits light, the atom feels a recoil. The momentum again turns out to be given by
the expression p = ħ k. In short, the quantum of action determines the momentum of

the photon.
The value of a photon’s momentum respects the indeterminacy principle. Just as it is
impossible to measure exactly both the wavelength of a wave and the position of its crest,
so it is impossible to measure both the momentum and the position of a photon. Can you
Challenge 34 ny confirm this? In other words, the value of the photon momentum is a direct consequence
of the quantum of action.
From our study of classical physics, we know that light has a property beyond its
colour: light can be polarized. That is only a complicated way to say that light can turn
Page 102 the objects that it shines on. In other words, light has an angular momentum oriented
along the axis of propagation. What about photons? Measurements consistently find that
each light quantum carries an angular momentum given by L = ħ. It is called its helicity.
The quantity is similar to one found for massive particles: one therefore also speaks of
the spin of a photon. In short, photons somehow ‘turn’ – in a direction either parallel
or antiparallel to the direction of motion. Again, the magnitude of the photon helicity,
Challenge 35 ny or spin, is no surprise; it confirms the classical relation L = E/ω between energy and
Vol. III, page 102 angular momentum that we found in the section on classical electrodynamics. Note that,
counterintuitively, the angular momentum of a single photon is fixed, and thus indepen-
dent of its energy. Even the most energetic photons have L = ħ. Of course, the value of the
helicity also respects the limit given by the quantum of action. The many consequences
of the helicity value ħ will become clear shortly.
What is light?
“
La lumière est un mouvement luminaire de
”
corps lumineux.
Blaise Pascal*
In the seventeenth century, Blaise Pascal used the above statement about light to make
fun of certain physicists, ridiculing the blatant use of a circular definition. Of course, he
was right: in his time, the definition was indeed circular, as no meaning could be given to
any of the terms. But whenever physicists study an observation with care, philosophers
lose out. All those originally undefined terms now have a definite meaning and the cir-
cular definition is resolved. Light is indeed a type of motion; this motion can rightly be
called ‘luminary’ because, in contrast to the motion of material bodies, it has the unique
property 󰑣 = c; the luminous bodies, today called photons, are characterized, and differ-
entiated from all other particles, by their dispersion relation E = cp, their energy E = ħω,
their spin L = ħ, the vanishing of all other quantum numbers, and the property of being
the quanta of the electromagnetic field.
In short, light is a stream of photons. It is indeed a ‘luminary movement of luminous
bodies’. Photons provide our first example of a general property of the world on small
scales: all waves and all flows in nature are made of quantum particles. Large numbers
of (coherent) quantum particles – or quantons – behave and form as waves. We will see
shortly that this is the case even for matter. Quantons are the fundamental constituents of
all waves and all flows, without exception. Thus, the everyday continuum description of
light is similar in many respects to the description of water as a continuous fluid: photons
are the atoms of light, and continuity is an approximation valid for large numbers of
particles. Single quantons often behave like classical particles.

Physics books used to discuss at length a so-called wave–particle duality. Let us be
clear from the start: quantons, or quantum particles, are neither classical waves nor clas-
sical particles. In the microscopic world, quantons are the fundamental objects.
However, there is much that is still unclear. Where, inside matter, do these monochro-
matic photons come from? Even more interestingly, if light is made of quantons, all elec-
tromagnetic fields, even static ones, must be made of photons as well. However, in static
* ‘Light is the luminary movement of luminous bodies.’ Blaise Pascal (b. 1623 Clermont, d. 1662 Paris), im-
portant French mathematician and physicist up to the age of 26, after which he became a theologian and
philosopher.
42 2 light
fields nothing is flowing. How is this apparent contradiction resolved? And what implica-
tions does the particle aspect have for these static fields? What is the difference between
quantons and classical particles? The properties of photons require more careful study.
Let us go on.
The size of photons

First of all, we might ask: what are these photons made of? All experiments so far, per-
formed down to the present limit of about 10−20 m, give the same answer: ‘we can’t find
anything’. This is consistent with both a vanishing mass and a vanishing size of photons.
Indeed, we would intuitively expect a body with a finite size to have a finite mass. Thus,
although experiments can give only an upper limit, it is consistent to claim that a photon
has zero size.
A particle with zero size cannot have any constituents. Thus a photon cannot be di-
vided into smaller entities: photons are not composite. For this reason, they are called ele-
mentary particles. We will soon give some further strong arguments for this result. (Can
Challenge 36 s you find one?) Nevertheless, the conclusion is strange. How can a photon have vanishing
size, have no constituents, and still be something? This is a hard question; the answer will
appear only in the last volume of our adventure. At the moment we simply have to accept
the situation as it is. We therefore turn to an easier question.
Are photons countable? – Squeezed light
“ ”
Also gibt es sie doch.
Max Planck*
We saw above that the simplest way to count photons is to distribute them across a large
screen and then to absorb them. But this method is not entirely satisfactory, as it destroys
the photons. How can we count photons without destroying them?
One way is to reflect photons in a mirror and measure the recoil of the mirror. It
seems almost unbelievable, but nowadays this effect is becoming measurable even for
small numbers of photons. For example, it has to be taken into account in relation to the
Vol. II, page 158 laser mirrors used in gravitational wave detectors, whose position has to be measured
with high precision.
Another way of counting photons without destroying them involves the use of special
high-quality laser cavities. It is possible to count photons by the effect they have on atoms
cleverly placed inside such a cavity.

In other words, light intensity can indeed be measured without absorption. These
measurement show an important issue: even the best light beams, from the most sophis-
ticated lasers, fluctuate in intensity. There are no steady beams. This comes as no surprise:
if a light beam did not fluctuate, observing it twice would yield a vanishing value for the
action. However, there is a minimum action in nature, namely ħ. Thus any beam and any
flow in nature must fluctuate. But there is more.
* ‘Thus they do exist after all.’ Max Planck, in his later years, said this after standing silently, for a long time,
in front of an apparatus that counted single photons by producing a click for each photon it detected. For
a large part of his life, Planck was sceptical of the photon concept, even though his own experiments and
conclusions were the starting point for its introduction.
A light beam is described by its intensity and its phase. The change – or action – that
occurs while a beam moves is given by the variation in the product of intensity and phase.
Experiments confirm the obvious deduction: the intensity and phase of a beam behave
like the momentum and position of a particle, in that they obey an indeterminacy rela-
tion. You can deduce it yourself, in the same way as we deduced Heisenberg’s relations.
Using as characteristic intensity I = E/ω, the energy divided by the angular frequency,
and calling the phase φ, we get*
ħ
ΔI Δφ ⩾ . (12)
2
For light emitted from an ordinary lamp, the product on the left-hand side of the above
inequality is much larger than the quantum of action. On the other hand, laser beams
can (almost) reach the limit. Laser light in which the two indeterminacies differ greatly
from each other is called non-classical light or squeezed light; it is used in many mod-
ern research applications. Such light beams have to be treated carefully, as the smallest
disturbances transform them back into ordinary laser beams, in which the two indeter-
minacies have the same value. Extreme examples of non-classical light are beams with a
given, fixed, photon number, and thus with an extremely high phase indeterminacy.
The observation of non-classical light highlights a strange fact, valid even for classical
light: the number of photons in a light beam is not a well-defined quantity. In general, it is
undetermined, and it fluctuates. The number of photons at the beginning of a beam is not
necessarily the same as the number at the end of the beam. Photons, unlike stones, cannot
be counted precisely – as long as they are moving and not absorbed. In flight, it is only
possible to determine an approximate number, within the limits set by indeterminacy.
One extreme example, shown in Figure 21, is a light beam with an (almost) fixed phase.
In such a beam, the photon number fluctuates from zero to infinity. In other words, in
order to produce coherent light, such as a laser beam, that approximates a pure sine wave
as perfectly as possible, we must build a source in which the photon number is as unde-
termined as possible.
At the other extreme is a beam with a fixed number of photons: in such a beam of
non-classical light, the phase fluctuates erratically. In contrast, the thermal light that we
encounter in most everyday situations – such as the light from an incandescent lamps
– lies somewhere in between, the phase and intensity indeterminacies being of similar
magnitude.
As an aside, it turns out that even in deep, dark intergalactic space, far from any star,
there are about 400 photons per cubic centimetre. This number, like the number of pho-
Challenge 37 ny tons in a light beam, also has a measurement indeterminacy. Can you estimate it?
In summary, unlike pebbles, photons are countable, but their number is not fixed. And
this it not the only difference between photons and pebbles.
* A large photon number is assumed in the expression. This is obvious, as Δφ cannot grow beyond all
bounds. The exact relations are
ħ
ΔI Δ cos φ ⩾ |⟨sin φ⟩|
2
ħ
ΔI Δ sin φ ⩾ |⟨cos φ⟩| (11)
2
where ⟨x⟩ denotes the expectation value of the observable x.
44 2 light
Thermal light Laser light Non-classical light
Photon sequence: bunching Little or no bunching Anti-bunching
time
Intensity I(t)
time
Probability P(I)
Sub-
Bose-Einstein Poisson Poisson
intensity
Intensity correlation
2 2 2
1 1 1
coherence time
time
Amplitude–
phase
diagram
F I G U R E 21 Three types of light: thermal light, laser light and squeezed light
The positions of photons

Where is a photon when it moves in a beam of light? Quantum theory gives a simple
answer: nowhere in particular. This is proved most spectacularly by experiments with

interferometers, such as the basic interferometer shown in Figure 22. Interferometers
show that even a beam made of a single photon can be split, led along two different paths,
and then recombined. The resulting interference shows that the single photon cannot be
said to have taken either of the two paths. If one of the two paths is blocked, the pattern
on the screen changes. In other words, somehow the photon must have taken both paths
at the same time. Photons cannot be localized: they have no position.*
* One cannot avoid this conclusion by saying that photons are split at the beam splitter: if a detector is
placed in each arm, one finds that they never detect a photon at the same time. Photons cannot be divided.
source detectors
mirrors
beam beam
splitter splitter
possible
two identical light
photons paths
F I G U R E 22 The Mach–Zehnder interferometer and a practical realization, about 0.5 m in size (© Félix
Dieu and Gaël Osowiecki)
This impossibility of localizing photons can be quantified. It is impossible to localize

photons in the direction transverse to the motion. It is less difficult to localize photons
along the direction of motion. In the latter case, the quantum of action implies that the
indeterminacy in the longitudinal position is given by the wavelength of the light. Can
Challenge 38 ny you confirm this?
In particular, this means that photons cannot be simply visualized as short wave trains.
Photons are truly unlocalizable entities, specific to the quantum world.
Now, if photons can almost be localized along their direction of motion, we can ask
how photons are lined up in a light beam. Of course, we have just seen that it does not
make sense to speak of their precise position. But do photons in a perfect beam arrive at
almost-regular intervals?
To the shame of physicists, the study of this question was initiated by two astronomers,
46 2 light
light detector F I G U R E 23
incoming
D1 How to
light
measure
beam
photon
statistics with
an electronic
adjustable coincidence
coincidence
position counter
counter, the
variation being
measured by
varying the
light detector position of a
D2 detector
Ref. 18 Robert Hanbury Brown and Richard Twiss, in 1956. They used a simple method to mea-
sure the probability that the second photon in a light beam arrives at a given time after
the first one. They simply split the beam, put one detector in the first branch, and varied
the position of a second detector in the other branch.
Hanbury Brown and Twiss found that, for coherent light, the clicks in the two counters
– and thus the photons themselves – are correlated. This result is completely contrary to
classical electrodynamics. The result is one of the many that show that photons are indeed
necessary to describe light. To be more precise, their experiment showed that whenever
the first photon hits, the second one is most likely to hit just afterwards. Thus, photons
Page 54 in beams are bunched. (As we will see below, this also implies that photons are bosons.)
Every light beam has an upper time limit for bunching, called the coherence time. For
times longer than the coherence time, the probability for bunching is low, and indepen-
dent of the time interval, as shown in Figure 23. The coherence time characterizes every
light beam, or rather every light source. In fact, it is often easier to think in terms of the
coherence length of a light beam. For thermal lamps, the coherence length is only a few
micrometres: a small multiple of the wavelength. The largest coherence lengths, of over
100 000 km, are obtained with research lasers. Interestingly, coherent light is even found
Ref. 19 in nature: several special stars have been found to emit it.
Although the intensity of a good laser beam is almost constant, the photons do not
arrive at regular intervals. Even the best laser light shows bunching, though with differ-
ent statistics and to a lesser degree than lamp light. Light whose photons arrive regularly,
thus exhibiting so-called (photon) anti-bunching, is obviously non-classical in the sense
defined above; such light can be produced only by special experimental arrangements.
Extreme examples of this phenomenon are being investigated at present by several re-
search groups aiming to construct light sources that emit one photon at a time, at regular
time intervals, as reliably as possible.
In summary, experiments force us to conclude that light is made of photons, but that
photons cannot be localized in light beams. It makes no sense to talk about the position
of a photon in general; the idea makes sense only in some special situations, and then
only approximately and as a statistical average.
Ekin kinetic energy of

emitted electrons
lamp
electrons
Ekin=h (ω−ωt)
threshold
F I G U R E 24 The kinetic
metal plate energy of electrons emitted
frequency of lamp light ω
in vacuum in the photoelectric effect
Are photons necessary?
In light of the results uncovered so far, the answer to the above question is obvious. But
the issue is tricky. In textbooks, the photoelectric effect is usually cited as the first and
most obvious experimental proof of the existence of photons. In 1887, Heinrich Hertz
observed that for certain metals, such as lithium or caesium, incident ultraviolet light
leads to charging of the metal. Later studies of the effect showed that the light causes
emission of electrons, and that the energy of the ejected electrons does not depend on
the intensity of the light, but only on the difference between ħ times its frequency and
a material-dependent threshold energy. Figure 24 summarizes the experiment and the
measurements.
In classical physics, the photoelectric effect is difficult to explain. But in 1905, Albert
Ref. 20 Einstein deduced the measurements from the assumption that light is made of photons
of energy E = ħω. He imagined that this energy is used partly to take the electron over
the threshold, and partly to give it kinetic energy. More photons only lead to more elec-
trons, not to faster ones. In 1921, Einstein received the Nobel Prize for the explanation
of the photoelectric effect. But Einstein was a genius: he deduced the correct result by a
somewhat incorrect reasoning. The (small) mistake was the assumption that a classical,
continuous light beam would produce a different effect. In fact, it is easy to see that a
classical, continuous electromagnetic field interacting with discrete matter, made of dis-
crete atoms containing discrete electrons, would lead to exactly the same result, as long as
the motion of electrons is described by quantum theory. Several researchers confirmed

Ref. 21 this early in the twentieth century. The photoelectric effect by itself does not imply the
existence of photons.
In the past, many researchers were unconvinced that the photoelectric effect shows
the existence of photons. Historically, the most important argument for the necessity of
light quanta was given by Henri Poincaré. In 1911 and 1912, aged 57 and only a few months
before his death, he published two influential papers proving that the radiation law of
black bodies – in which the quantum of action had been discovered by Max Planck –
Ref. 22 requires the existence of photons. He also showed that the amount of radiation emitted
by a hot body is finite only because of the quantum nature of the processes leading to
light emission. A description of these processes in terms of classical electrodynamics
48 2 light
would lead to infinite amounts of radiated energy. Poincaré’s two influential papers con-
vinced most physicists that it was worthwhile to study quantum phenomena in more
detail. Poincaré did not know about the action limit S ⩾ ħ; yet his argument is based on
the observation that light of a given frequency has a minimum intensity, namely a single
photon. Such a one-photon beam may be split into two beams, for example by using a
half-silvered mirror. However, more than one photon is never found in those two beams
together.
Another interesting experiment that requires photons is the observation of ‘molecules
Ref. 23 of photons’. In 1995, Jacobson et al. predicted that the de Broglie wavelength of a packet of
photons could be observed. According to quantum theory, the packet wavelength is given
by the wavelength of a single photon divided by the number of photons in the packet. The
team argued that the packet wavelength could be observable if such a packet could be
split and recombined without destroying the cohesion within it. In 1999, this effect was
indeed observed by de Pádua and his research group in Brazil. They used a careful set-up
with a nonlinear crystal to create what they call a biphoton, and observed its interference
properties, finding a reduction in the effective wavelength by the predicted factor of two.
Since then, packages with three and even four entangled photons have been created and
Ref. 24 observed.
Yet another argument for the necessity of photons is the above-mentioned recoil felt
Page 38 by atoms emitting light. The recoil measured in these cases is best explained by the emis-
sion of a photon in a particular direction. In contrast, classical electrodynamics predicts
the emission of a spherical wave, with no preferred direction.
Page 44 Obviously, the observation of non-classical light, also called squeezed light, also argues
for the existence of photons, as squeezed light proves that photons are indeed an intrinsic
Ref. 25 aspect of light, necessary even when interactions with matter play no role. The same is
true for the Hanbury Brown–Twiss effect.
Finally, the spontaneous decay of excited atomic states also requires the existence of
photons. This cannot be explained by a continuum description of light.
In summary, the concept of a photon is indeed necessary for a precise description
of light; but the details are often subtle, as the properties of photons are unusual and
require a change in our habits of thought. To avoid these issues, most textbooks stop
discussing photons after coming to the photoelectric effect. This is a pity, as it is only then
that things get interesting. Ponder the following. Obviously, all electromagnetic fields are
made of photons. At present, photons can be counted for gamma rays, X-rays, ultraviolet
light, visible light and infrared light. However, for lower frequencies, such as radio waves,
photons have not yet been detected. Can you imagine what would be necessary to count
Challenge 39 ny the photons emitted from a radio station?
This issue leads directly to the most important question of all:
lasers or other F I G U R E 25 Light

pocket lamps coherent light source crossing light
How can a wave be made up of particles?
“
Fünfzig Jahre intensiven Nachdenkens haben
mich der Antwort auf die Frage ‘Was sind
Lichtquanten?’ nicht näher gebracht. Natürlich
bildet sich heute jeder Wicht ein, er wisse die
”
Antwort. Doch da täuscht er sich.
Albert Einstein, 1951 *
If a light wave is made of particles, one must be able to explain each and every wave
property in terms of photons. The experiments mentioned above already hint that this is
possible only because photons are quantum particles. Let us take a more detailed look at
this.
Light can cross other light undisturbed. This observation is not hard to explain with
photons; since photons do not interact with each other, and are point-like, they ‘never’ hit
each other. In fact, there is an extremely small positive probability for their interaction,
as will be found below, but this effect is not observable in everyday life.
But some problems remain. If two light beams of identical frequency and fixed phase
Page 89 relation cross, we observe alternating bright and dark regions: so-called interference
fringes.** How do these interference fringes appear? How can it be that photons are not
detected in the dark regions? We already know the only possible answer: the brightness
at a given place corresponds to the probability that a photon will arrive there. The fringes
imply that photons behave like little arrows. Some further thought leads to the following
description:
— The probability of a photon arriving somewhere is given by the square of an arrow.
— The final arrow is the sum of all the arrows arriving there by all possible paths.
— The arrow is always perpendicular to the direction of motion.
— The arrow’s direction stays fixed in space when the photons move.
* ‘Fifty years of intense reflection have not brought me nearer to the answer to the question ‘What are light
quanta?’ Of course nowadays every little mind thinks he knows the answer. But he is wrong.’ Einstein said
this a few years before his death.
** This experiment is only possible if both beams are derived from a single beam by splitting, or if two
Challenge 40 s expensive high-precision lasers are used. (Why?)
50 2 light
two lasers or screen

point sources
S1
S2
the arrow model:
t1
t2
t3 F I G U R E 26 Interference and the description of

light with arrows (at three instants of time)
— The length of an arrow shrinks with the square of the distance travelled.
— Photons emitted by single-coloured sources are emitted with arrows of constant
length pointing in the direction ωt; in other words, such sources spit out photons
with a rotating mouth.
— Photons emitted by thermal sources, such as pocket lamps, are emitted with arrows
of constant length pointing in random directions.
With this model* we can explain the stripes seen in laser experiments, such as those of
Figure 25 and Figure 26. You can check that in some regions the two arrows travelling
through the two slits add up to zero for all times. No photons are detected there. In other
regions, the arrows always add up to the maximal value. These regions are always bright.
Regions in between have intermediate shades. Obviously, in the case of pocket lamps,
the brightness also behaves as expected: the averages simply add up, as in the common
region shown in the left-hand diagram of Figure 25.

You may wish to calculate the distance between the lines, given the source distance,
Challenge 41 ny the colour and the distance to the screen.
Obviously, the photon model implies that an interference pattern like this is built up
as the sum of a large number of single-photon hits. Using low-intensity beams, we should
therefore be able to see how these little spots slowly build up an interference pattern by
accumulating in the bright regions and never hitting the dark regions. This is indeed the
Page 38 case. All experiments have confirmed this description.
* The model gives a correct description of light except that it neglects polarization. To add it, it is necessary
to combine arrows that rotate in both senses around the direction of motion.
screen
source image
mirror
F I G U R E 27 Light reflected by a
mirror, and the corresponding
arrow sum arrows (at an instant of time)
It is important to note that interference between two light beams is not the result of
two different photons cancelling each other out or being added together. Such cancella-
tion would contradict conservation of energy and momentum. Interference is an effect
applicable to each photon separately, because each photon is spread out over the whole
set-up: each photon takes all possible paths. As Paul Dirac said, each photon interferes
Ref. 26 only with itself. Interference only works because photons are quantons, and not classical
particles.
Dirac’s oft-quoted statement leads to a famous paradox: if a photon can interfere only
with itself, how can two laser beams from two different lasers interfere with each other?
The answer given by quantum physics is simple but strange: in the region where the
beams interfere, it is impossible to say from which source a photon has come. The pho-
tons in the crossing region cannot be said to come from a specific source. Photons in
the interference region are quantons, which indeed interfere only with themselves. In
that region, one cannot truly say that light is a flow of photons. Despite regular claims to
the contrary, Dirac’s statement is correct. It is a strange consequence of the quantum of
action.
Waves also show diffraction. To understand this phenomenon with photons, let us
start with a simple mirror, and study reflection first. Photons (like all quantum particles)
move from source to detector by all possible paths. As Richard Feynman,* who discov-
* Richard (‘Dick’) Phillips Feynman (b. 1918 New York City, d. 1988), US-American physicist. One of the
founders of quantum electrodynamics, he also discovered the ‘sum-over-histories’ reformulation of quan-
tum theory, made important contributions to the theory of the weak interaction and to quantum gravity,
and co-authored a famous textbook, the Feynman Lectures on Physics. He is one of those theoretical physi-
cists who made his career mainly by performing complex calculations – but he backtracked with age, most
successfully in his teachings and physics books, which are all worth reading. He was deeply dedicated to
physics and to enlarging knowledge, and was a collector of surprising physical explanations. He helped
building the nuclear bomb, wrote papers in topless bars, avoided to take any professional responsibility,
and was famously arrogant and disrespectful of authority. He wrote several popular books on the events
52 2 light
source point
arrow sum
at point
usual vanishes
mirror
screen
source image
arrow sum
at image
striped
mirror F I G U R E 28 The light reflected by a
badly-placed mirror and by a grating
ered this explanation, liked to stress, the term ‘all’ has to be taken literally. This is not a
big deal in the explanation of interference. But in order to understand a mirror, we have
to include all possibilities, however crazy they seem, as shown in Figure 27.
As stated above, a light source emits rotating arrows. To determine the probability
that light arrives at a certain location within the image, we have to add up all the ar-
rows arriving at the same time at that location. For each path, the arrow orientation at
the image is shown – for convenience only – below the corresponding segment of the
mirror. The angle and length of the arriving arrow depends on the path. Note that the
sum of all the arrows does not vanish: light does indeed arrive at the image. Moreover,
the largest contribution comes from the paths near to the middle. If we were to perform
the same calculation for another image location, (almost) no light would get there. So
the rule that reflection occurs with the incoming angle equal to the outgoing angle is an
approximation, following from the arrow model of light.

In fact, a detailed calculation, with more arrows, shows that the approximation is quite
precise: the errors are much smaller than the wavelength of the light.
The proof that light does indeed take all these strange paths is given by a more spe-
cialized mirror. As show in Figure 28, we can repeat the experiment with a mirror that
reflects only along certain stripes. In this case, the stripes have been carefully chosen so
that the corresponding path lengths lead to arrows with a bias in one direction, namely
to the left. The arrow addition now shows that such a specialized mirror – usually called
of his life. Though he tried to surpass the genius of Wolfgang Pauli throughout his life, he failed in this
endeavour. He shared the 1965 Nobel Prize in Physics for his work on quantum electrodynamics.
light beam
air
water
F I G U R E 29 If light were made of little stones, they would

move faster in water
a grating – allows light to be reflected in unusual directions. Indeed, this behaviour is
standard for waves: it is called diffraction. In short, the arrow model for photons allows
us to describe this wave property of light, provided that photons follow the ‘crazy’ prob-
ability scheme. Do not get upset! As was said above, quantum theory is the theory for
crazy people.
You may wish to check that the arrow model, with the approximations it generates
by summing over all possible paths, automatically ensures that the quantum of action is
Challenge 42 ny indeed the smallest action that can be observed.
All waves have a signal velocity. The signal velocity also depends on the medium in
which they propagate. As a consequence, waves show refraction when they move from
one medium into another with different signal velocity. Interestingly, the naive particle
picture of photons as little stones would imply that light is faster in materials with high
Challenge 43 ny refractive indices: the so-called dense materials. Can you confirm this? However, exper-
iments show that light in dense materials moves slowly. The wave picture has no diffi-
Challenge 44 ny culty explaining this observation. (Can you confirm this?) Historically, this was one of
the arguments against the particle theory of light. In contrast, the arrow model of light
Challenge 45 ny presented above is able to explain refraction properly. It is not difficult: try it.
Waves also reflect partially from materials such as glass. This is one of the most diffi-
cult wave properties to explain with photons. But it is one of the few effects that is not
explained by a classical wave theory of light. However, it is explained by the arrow model,
as we will find shortly. Partial reflection confirms the first two rules of the arrow model.
Page 49 Partial reflection shows that photons indeed behave randomly: some are reflected and
other are not, without any selection criterion. The distinction is purely statistical. More
about this issue shortly.
In waves, the fields oscillate in time and space. One way to show how waves can be
made of particles is to show how to build up a sine wave using a large number of photons.
Ref. 27 A sine wave is a coherent state of light. The way to build them up was explained by Roy
Glauber. In fact, to build a pure sine wave, we need a superposition of a beam with one
photon, a beam with two photons, a beam with three photons, and so on. Together, they
give a perfect sine wave. As expected, its photon number fluctuates to the highest possible
degree.
54 2 light
If we repeat the calculation for non-ideal beams, we find that the indeterminacy rela-
tion for energy and time is respected: every emitted beam will possess a certain spectral
width. Purely monochromatic light does not exist. Similarly, no system that emits a wave
at random can produce a monochromatic wave. All experiments confirm these results.
Waves can be polarized. So far, we have disregarded this property. In the photon pic-
ture, polarization is the result of carefully superposing beams of photons spinning clock-
wise and anticlockwise. Indeed, we know that linear polarization can be seen as a result
of superposing circularly-polarized light of both signs, using the proper phase. What
seemed a curiosity in classical optics turns out to be a fundamental justification for quan-
tum theory.
Photons are indistinguishable. When two photons of the same colour cross, there is
no way to say afterwards which of the two is which. The quantum of action makes this
impossible. The indistinguishability of photons has an interesting consequence. It is im-
possible to say which emitted photon corresponds to which arriving photon. In other
words, there is no way to follow the path of a photon, as we are used to following the
Page 45 path of a billiard ball. In addition, the experiment by Hanbury Brown and Twiss, implies
Ref. 28 that photons are bosons. Particles with this property are called bosons. We will discover
Page 94 more details about the indistinguishability of bosons later in.
In summary, we find that light waves can indeed be built of particles. However, this is
only possible with the proviso that photons are not precisely countable, that they are not
localizable, that they have no size, no charge and no mass, that they carry an (approxi-
mate) phase, that they carry spin, that they are indistinguishable bosons, that they can
take any path whatsoever, that one cannot pinpoint their origin, and that their probabil-
ity of arriving somewhere is determined by the square of the sum of amplitudes for all
possible paths. In other words, light can be made of particles only if these particles have
very special quantum properties. These quantum properties allow them to behave like
waves when they are present in large numbers.
Can light move faster than light? – Virtual photons

In a vacuum, light can move faster than c, as well as slower than c. The quantum principle
provides the details. As long as this principle is obeyed, the speed of a short light flash
can differ – though only by a tiny amount – from the ‘official’ value. Can you estimate
Challenge 46 ny the allowable difference in arrival time for a light flash coming from the dawn of time?
The arrow explanation gives the same result. If we take into account the crazy possi-
bility that photons can move with any speed, we find that all speeds very different from
c cancel out. The only variation that remains, translated into distances, is the indetermi-
Challenge 47 ny nacy of about one wavelength in the longitudinal direction, which we mentioned above.
More bizarre consequences of the quantum of action appear when we study static elec-
tric fields, such as the field around a charged metal sphere. Obviously, such a field must
also be made of photons. How do they move? It turns out that static electric fields are
made of virtual photons. Virtual photons are photons that do not appear as free parti-
cles: they only appear for an extremely short time before they disappear again. In the
case of a static electric field, they are longitudinally polarized, and do not carry energy
away. Virtual photons, like other virtual particles, are ‘shadows’ of particles that obey
ΔxΔp ⩽ ħ/2 . (13)
Rather than obeying the usual indeterminacy relation, they obey the opposite relation,
which expresses their very brief appearance. Despite their intrinsically short life, and de-
spite the impossibility of detecting them directly, virtual particles have important effects.
Page 164 We will explore virtual particles in detail shortly.
In fact, the vector potential A allows four polarizations, corresponding to the four
coordinates (t, x, y, z). It turns out that for the photons one usually talks about – the
free or real photons – the polarizations in the t and z directions cancel out, so that one
observes only the x and y polarizations in actual experiments.
For bound or virtual photons, the situation is different. All four polarizations are pos-
sible. Indeed, the z and t polarizations are the ones that can be said to be the building
blocks of static electric and magnetic fields.
In other words, static electric and magnetic fields are continuous flows of virtual pho-
tons. In contrast to real photons, virtual photons can have mass, can have spin directions
not pointing along the path of motion, and can have momentum opposite to their direc-
tion of motion. Exchange of virtual photons leads to the attraction of bodies of different
charge. In fact, virtual photons necessarily appear in any description of electromagnetic
Vol. V, page 84 interactions. Later on we will discuss their effects further – including the famous attrac-
tion of neutral bodies.
In summary, light can indeed move faster than light, though only by an amount al-
lowed by the quantum of action. For everyday situations, i.e., for high values of the action,
all quantum effects average out, including light velocities different from c.
There is another point that we should mention here. Not only the position, but also
Ref. 29 the energy of a single photon can be undefined. For example, certain materials split one
photon of energy ħω into two photons, whose two energies add up to the original one.
Quantum mechanics implies that the energy partitioning is known only when the energy
of one of the two photons is measured. Only at that very instant is the energy of the
second photon known. Before the measurement, both photons have undefined energies.
The process of energy fixing takes place instantaneously, even if the second photon is far
Page 128 away. We will explain below the background to this and similar strange effects, which
seem to be faster than light. In fact, despite the appearance, these observations do not
Challenge 48 ny involve faster-than-light transmission of energy or information.
Indeterminacy of electric fields

We have seen that the quantum of action implies an indeterminacy for light intensity.
Since light is an electromagnetic wave, this indeterminacy implies similar, separate limits
for electric and magnetic fields at a given point in space. This conclusion was first drawn
Ref. 30 in 1933 by Bohr and Rosenfeld. They started from the effects of the fields on a test particle
of mass m and charge q, which are described by:
ma = q(E + 󰑣 × b) . (14)
56 2 light
p1
α1 air
water
p2
α2
F I G U R E 30 Refraction and photons
Since it is impossible to measure both the momentum and the position of a particle, they
Challenge 49 ny deduced an indeterminacy for the electrical field, given by
ħ
ΔE = , (15)
q Δx t
where t is the measurement time and Δx is the position indeterminacy. Thus every value
of an electric field, and similarly of a magnetic field, possesses an indeterminacy. The
state of the electromagnetic field behaves like the state of matter in this respect: both
follow an indeterminacy relation.
Curiosities and fun challenges about photons

Can one explain refraction with photons? Newton was not able to do so, but today we can.
In refraction by a horizontal surface, as shown in Figure 30, the situation is translationally
invariant along the horizontal direction. Therefore, the momentum component along
this direction is also conserved: p1 sin α1 = p2 sin α2 . The photon energy E = E1 = E2
is obviously conserved. The index of refraction n is defined in terms of momentum and
energy as
cp
n= . (16)
E
Challenge 50 e The ‘law’ of refraction follows:

sin α1
=n. (17)
sin α2
There is an important issue here. In a material, the velocity of a photon 󰑣 = δE/δp

in a light ray differs from the phase velocity u = E/p that enters into the calculation. In
summary, inside matter, the concept of photon must be used with extreme care.
∗∗
F I G U R E 31 The blue shades of the sky and the
colours of clouds are due to various degrees of
Rayleigh, Mie and Tyndall scattering (© Giorgio di
Iorio)
If en electromagnetic wave has amplitude A, the photon density d is
A2
d= . (18)
ħω
Challenge 51 ny Can you show this?

∗∗
A typical effect of the quantum ‘laws’ is the yellow colour of the lamps used for street illu-
mination in most cities. They emit pure yellow light of a single frequency; that is why no
other colours can be distinguished in their light. According to classical electrodynamics,
harmonics of that light frequency should also be emitted. Experiments show, however,
that this is not the case; so classical electrodynamics is thus wrong. Is this argument cor-
Challenge 52 ny rect?
∗∗
What happens to photons that hit an object but are not absorbed or transmitted? Gen-
erally speaking, they are scattered. The details of the scattering depend on the object, as
Table 3 shows.
All scattering properties are material dependent. Their study explains many colours
Page 145 of transparent materials, as we will see below.
A summary on light: particle and waves

In summary, light is a stream of photons. A single photon is the smallest possible light
intensity of a given colour. Photons, like all quantons, are quite different from everyday
58 2 light
TA B L E 3 Types of light scattering
S c at t e r i n g S c at t e r e r D eta i l s Examples
type
Rayleigh scattering atoms, molecules elastic, angle changes blue sky, red evening
as 1/λ4 , scatterers sky, blue cigarette
smaller than λ/10 smoke
Mie scattering transparent objects, elastic, angle changes blue sky, red
droplets as 1/λ0.5 to 1/λ2 evenings, blue
distant mountains
Tyndall scattering non-transparent objects elastic, angle weakly smog, white clouds,
or not wavelength- fog, white cigarette
dependent smoke
Smekal–Raman excited atoms, molecules inelastic, light gains used in lidar
scattering energy investigations of the
atmosphere
Inverse Raman atoms, molecules inelastic, light loses used in material
scattering energy research
Thomson scattering electrons elastic used for electron
density
determination
Compton scattering electrons inelastic, X-ray lose proves particle
energy nature of light
page 40
Brillouin scattering acoustic phonons, density inelastic, frequency used to study
variations in solids/fluids shift of a few GHz phonons and to
diagnose optical
fibres
X-ray scattering crystalline solids elastic, due to used to determine
interference at crystal structures
crystal planes
life particles. In fact, we can argue that the only (classical) particle aspects of photons are
their quantized energy, momentum and spin. In all other respects, photons are not like
little stones. Photons cannot be localized in light beams. Photons are indistinguishable.
Photons are bosons. Photons have no mass, no charge and no size. It is more accurate to
Ref. 31 say that photons are calculating devices to precisely describe observations about light.
The strange properties of photons are the reason why earlier attempts to describe light
as a stream of (classical) particles, such as the attempt of Newton, failed miserably, and
were rightly ridiculed by other scientists. Indeed, Newton upheld his theory against all
experimental evidence – especially with regard to light’s wave properties – which is some-
thing that a physicist should never do. Only after people had accepted that light is a wave,
and then discovered and understood that quantum particles are different from classical
particles, was the approach successful.
The quantum of action implies that all waves are streams of quantons. In fact, all waves
are correlated streams of quantons. This is true for light, for any other form of radiation,
and for all forms of matter waves.

The indeterminacy relations show that even a single quanton can be regarded as a
wave; however, whenever it interacts with the rest of the world, it behaves as a particle. In
fact, it is essential that all waves be made of quantons: if they were not, then interactions
would be non-local, and objects could not be localized at all, contrary to experience. To
decide between the wave and particle descriptions, we can use the following criterion.
Whenever matter and light interact, it is more appropriate to describe electromagnetic
radiation as a wave if the wavelength λ satisfies
ħc
λ≫ , (19)
kT
where k = 1.4 ⋅ 10−23 J/K is Boltzmann’s constant. If the wavelength is much smaller than
the quantity on the right-hand side, the particle description is most appropriate. If the
two sides are of the same order of magnitude, both descriptions play a role. Can you
Challenge 53 e explain the criterion?

Chapter 3
MOTION OF MAT TER – BEYOND

C L A S SIC A L PH YSIC S
“ ”
All great things begin as blasphemies.
George Bernard Shaw
T
he existence of a smallest action has numerous important consequences for
he motion of matter. We start with a few experimental results that show
hat the quantum of action is indeed the smallest measurable action, also in
the case of matter. Then we show that the quantum of action implies the existence of a
phase and thus of the wave properties of matter, including the same description that we
already found for light: matter particles behave like rotating arrows.
Wine glasses, pencils and atoms – no rest
“ ”
Otium cum dignitate.*
Cicero, De oratore.
If the quantum of action is the smallest observable change in a physical system, then two
observations of the same system must always differ. Thus there cannot be perfect rest in
nature. Is that true? Experiments show that this is indeed the case.
A simple consequence of the quantum of action is the impossibility of completely fill-
ing a glass of wine. If we call a glass at maximum capacity (including surface tension
effects, to make the argument precise) ‘full’, we immediately see that the situation re-
quires the liquid’s surface to be completely at rest. But a completely quiet surface would
admit two successive observations that differ by less than ħ. We could try to reduce all
motions by reducing the temperature of the system. But absolute rest would imply reach-
ing absolute zero temperature. Experiments show that this is impossible. (Indeed, this
impossibility, the so-called third ‘law’ of thermodynamics, is equivalent to the existence

of a minimum action.) There is no rest in nature. In other words, the quantum of action
proves the old truth that a glass of wine is always partially empty and partially full.
The absence of microscopic rest, predicted by the quantum of action, is confirmed
in many experiments. For example, a pencil standing on its tip cannot remain vertical,
as shown in Figure 32, even if it is isolated from all disturbances, such as vibrations, air
molecules and thermal motion. This follows from the indeterminacy relation. In fact, it
Challenge 54 d is even possible to calculate the time after which a pencil must have fallen over.
But the most important consequence of the absence of rest is another. The absence of
* ‘Rest with dignity.’

motion of mat ter – beyond cl assical physics 61
axis F I G U R E 32 A falling pencil
rest for the electrons inside atoms prevents them from falling into the nuclei, despite their
mutual attraction. In short, the existence and the size of atoms, and thus of all matter, is
a direct consequence of the absence of microscopic rest. We will explore this in more
Page 67 detail below. We only exist and live because of the quantum of action.
No infinite precision
The quantum of action prevents the observation of rest in many ways. In order to check
whether an object is at rest, we need to observe its position with high precision. Because
of the wave properties of light, we need a high-energy photon: only a high-energy photon
has a small wavelength and thus allows a precise position measurement. As a result of
this high energy, however, the object is disturbed. Worse, the disturbance itself is not
precisely measurable; so there is no way to determine the original position even by taking
the disturbance into account. In short, perfect rest cannot be observed even in principle.
Indeed, all experiments in which systems have been observed with high precision con-
firm that perfect rest does not exist. The absence of rest has been confirmed for electrons,
neutrons, protons, ions, atoms, molecules, atomic condensates, crystals, and objects with
a mass of up to a tonne, as used in certain gravitational wave detectors. No object is ever
at rest.
The same argument also shows that no measurement, of any observable, can ever be
performed to infinite precision. This is another of the far-reaching consequences of the
quantum of action.
Cool gas
The quantum of action implies that rest is impossible in nature. In fact, even at extremely
low temperatures, all particles inside matter are in motion. This fundamental lack of rest
is said to be due to the so-called zero-point fluctuations. A good example is provided by
the recent measurements of Bose–Einstein condensates. They are trapped gases, with a
small number of atoms (between ten and a few million), cooled to extremely low tem-
peratures (around 1 nK). These cool gases can be observed with high precision. Using
elaborate experimental techniques, Bose–Einstein condensates can be put into states for
which ΔpΔx is almost exactly equal to ħ/2 – though never lower than this value. These
experiments confirm directly that there is no rest, but a fundamental fuzziness in nature.
62 3 motion of mat ter – beyond cl assical physics
This leads to an interesting puzzle. In a normal object, the distance between the atoms
Challenge 55 s is much larger than their de Broglie wavelength. (Can you confirm this?) But today it is
possible to cool objects to extremely low temperatures. At sufficiently low temperatures,
Ref. 33 less than 1 nK, the wavelength of the atoms may be larger than their separation. Can you
Challenge 56 s imagine what happens in such cases?
Flows and the quantization of matter
“ ”
Die Bewegung ist die Daseinsform der Materie.
Friedrich Engels, Anti-Dühring.*
Not only does the quantum of action make rest impossible, but it makes any situation
that does not change in time impossible. The most important examples of (apparently)
stationary situations are flows. The quantum of action implies that no flow can be station-
ary. More precisely, a smallest action implies that no flow can be continuous. All flows
fluctuate and are made of smallest entities: in nature, all flows are made of quantum parti-
cles. We saw above that this is valid for light. Two simple types of flow from our everyday
experience directly confirm this consequence from the quantum of action: flows of fluids
and flows of electricity.
Fluid flows and quantons

The flow of matter also exhibits smallest units. We mentioned early on in our adventure
Vol. I, page 302 that a consequence of the particulate structure of liquids is that oil or any other smooth
liquid produces noise when it flows through even the smoothest of pipes. We mentioned
that the noise we hear in our ears in situations of absolute silence – for example, in a
snowy and windless landscape in the mountains or in an anechoic chamber – is due
to the granularity of blood flow in the veins. Experiments show that all flows of matter
produce vibrations. This is a consequence of the quantum of action, and of the resulting
granularity of matter.
Knocking tables and quantized conductivity

If electrical current were a continuous flow, it would be possible to observe action values
as small as desired. The simplest counter-example was discovered in 1996, by José Costa-
Ref. 34, Ref. 35 Krämer and his colleagues. They put two metal wires on top of each other on a kitchen
table and attached a battery, a current-voltage converter (or simply a resistor) and a stor-
age oscilloscope to them. Then they measured the electrical current while knocking on
the table. That is all.
Knocking the table breaks the contact between the two wires. In the last millisecond
before the wires detach, the conductivity and thus the electrical current diminished in
regular steps of about 7 μA, as can easily be seen on the oscilloscope (see Figure 33). This
simple experiment could have beaten, if it had been performed a few years earlier, a num-
ber of other, enormously expensive experiments which discovered this quantization at
costs of several million euro each, using complex set-ups at extremely low temperatures.
Ref. 12 * ‘Motion is matter’s way of being.’

F I G U R E 33 Steps
in the flow of
electricity in metal
wire crossings: the
set-up, the
nanowires at the
basis of the effect,
and three
measurement
results (© José
Costa-Krämer,
AAPT from Ref. 35)
In fact, the quantization of conductivity appears in any electrical contact with a small
cross-section. In such situations the quantum of action implies that the conductivity can
Challenge 57 ny only be a multiple of 2e 2 /ħ ≈ (12 906 Ω)−1 . Can you confirm this result? Note that elec-
trical conductivity can be as small as required; only the quantized electrical conductivity
has the minimum value of 2e 2 /ħ.
Many more elaborate experiments confirm the observation of conductance steps.
They force us to conclude that there is a smallest electric charge in nature. This small-
est charge has the same value as the charge of an electron. Indeed, electrons turn out to
be part of every atom, in a construction to be explained shortly. In metals, a large num-
ber of electrons can move freely: that is why metals conduct electricity so well and work
as mirrors.
In short, matter and electricity flow in smallest units. Depending on the material, the
smallest flowing units of matter may be ‘molecules’, ‘atoms’, ‘ions’, or ‘electrons’. Electrons,
ions, atoms and molecules are quantum particles, or quantons. The quantum of action
implies that matter is made of quantons. Matter quantons share some properties with or-
dinary stones, but also differ from them in many ways. A stone has position and momen-
tum, mass and acceleration, size, shape, structure, orientation and angular momentum,
and colour. Let us explore each of these properties for quantons, and see how they are
related to the quantum of action.
Matter quantons and their motion – matter waves

Ref. 36 In 1923 and 1924, the French physicist Louis de Broglie pondered the consequences of
the quantum of action for matter particles. In the case of light, the quantum of action
connects wave behaviour to particle behaviour. He reasoned that the same should apply
to matter. It dawned to him that streams of matter particles with the same momentum
should behave as waves, just as light quanta do. He thus predicted that like for light,
F I G U R E 34 Electrons beams diffract and interfere at

multiple slits (© Claus Jönsson)
F I G U R E 35 Formation over time of the interference pattern of electrons, here in a low-intensity

double-slit experiment: (a) 8 electrons, (b) 270 electrons, (c) 2000 electrons, (d) 6000 electrons, after 20
minutes of exposure. The last image corresponds to the situation shown in the previous figure.
(© Tonomura Akira/Hitachi)
coherent matter flows should have a wavelength and angular frequency given by
2π ħ E
λ= and ω= , (20)
p ħ
where p and E are the momentum and the energy, respectively, of the single particles.
Equivalently, we can write the relations as
p = ħk and E = ħω . (21)
These relations state that matter also behaves as a wave.

Soon after de Broglie’s prediction, experiments began to confirm the statement. It is
indeed observed that matter streams can diffract, refract and interfere, with the values
predicted by de Broglie. Because of the smallness of the wavelength, careful experiments
are needed to detect these effects. But one by one, all experimental confirmations of the
wave properties of light were repeated for matter beams. For example, just as light is
diffracted when it passes around an edge or through a slit, matter is also diffracted in
these situations. This is true even for electrons, the simplest particles of matter, as shown
Ref. 37 in Figure 34. The experiment with electrons is quite difficult. It was first performed by
Claus Jönsson in Tübingen in 1961; in the year 2002 it was voted the most beautiful ex-
periment in all of physics. Many years after Jönsson, the experiment was repeated with a
modified electron microscope, as shown in Figure 35.
Inspired by light interferometers, researchers began to build matter interferome-
Ref. 38 ters. Today these work with beams of electrons, nucleons, nuclei, atoms, or even large
Vol. III, page 87 molecules. Just as observations of light interference prove the wave character of light, so
the interference patterns observed with matter beams prove the wave character of matter.
Like light, matter is made of particles; like light, matter behaves as a wave when large
numbers of particles with the same momentum are involved. But although beams of
large molecules behave as waves, everyday objects – such as cars on a motorway – do not.
There are two main reasons for this. First, for cars on a motorway the relevant wavelength
is extremely small. Secondly, the speeds of the cars vary too much: streams of cars with
the same speed cannot be made coherent.
If matter behaves like a wave, we can draw a strange conclusion. For any wave, the
position and the wavelength cannot both be sharply defined simultaneously: the indeter-
minacies of the wave number k = 2π/λ and of the position X obey the relation
1
ΔkΔX ≥ . (22)
2
Similarly, for every wave the angular frequency ω = 2π f and the instant T of its peak
amplitude cannot both be sharply defined. Their indeterminacies are related by
1
ΔωΔT ≥ . (23)
2
Using de Broglie’s wave properties of matter (21), we get
ħ ħ
ΔpΔX ⩾ and ΔEΔT ⩾ . (24)
2 2
These famous relations are called Heisenberg’s indeterminacy relations. They were discov-
ered by the German physicist Werner Heisenberg in 1925. They are valid for all quantum
particles, be they matter or radiation. The indeterminacy relations state that there is no
way to simultaneously ascribe a precise momentum and position to a quantum system,
nor to simultaneously ascribe a precise energy and age. The more accurately one quan-
tity is known, the less accurately the other is.* As a result, matter quantons – rather like
stones, but unlike photons – can be localized, but always only approximately.
Both indeterminacy relations have been checked experimentally in great detail. All
experiments confirm them. In fact, every experiment proving that matter behaves like a
wave is a confirmation of the indeterminacy relation – and vice versa.
When two variables are linked by indeterminacy relations, one says that they are com-
plementary to each other. Niels Bohr systematically explored all possible such pairs. You
Challenge 58 s can can also do that for yourself. Bohr was deeply fascinated by the existence of a com-
plementarity principle, and he later extended it in philosophical directions. In a famous
scene, somebody asked him what was the quantity complementary to precision. He an-
swered: ‘clarity’.
We remark that the usual, real, matter quantons always move more slowly than light.
Due to the inherent fuzziness of quantum motion, it should not come to a surprise that
there are some exceptions. Indeed, in some extremely special cases, the quantum of ac-
tion allows the existence of particles that move faster than light – so-called virtual parti-
Page 164 cles – which we will meet later on.
In short, the quantum of action means that matter quantons do not behave like point-
like stones, but as waves. In particular, like for waves, the values of position and mo-
mentum cannot both be exactly defined for quantons. The values are fuzzy – position
and momentum are undetermined. The more precisely one of the two is known, the less
precisely the other is known.
Mass and acceleration of quantons

Matter quantons, like stones, have mass. Hits by single electrons, atoms or molecules can
be detected. Quantons can also be slowed down or accelerated. We have already explored
Page 28 some of these experiments in the section on electrodynamics. However, quantons differ
Challenge 59 s from pebbles. Using the time–energy indeterminacy relation, you can deduce that
2mc 3
a⩽ . (25)
ħ
Thus there is a maximum acceleration for quantons.** Indeed, no particle has ever been
Ref. 39 observed with a higher acceleration than this value. In fact, no particle has ever been
* A policeman stops the car being driven by Werner Heisenberg. ‘Do you know how fast you were driving?’
‘No, but I know exactly where I was!’
** We note that this acceleration limit is different from the acceleration limit due to general relativity:
c4
a⩽ . (26)
4Gm
In particular, the quantum limit (25) applies to microscopic particles, whereas the general-relativistic limit
applies to macroscopic systems. Can you confirm that in each domain the relevant limit is the smaller of
Challenge 60 e the two?
F I G U R E 36 Probability clouds: a hydrogen atom in its spherical ground state (left) and in a
non-spherical excited state (right) as seen by an observer travelling around it (QuickTime film produced
with Dean Dauger’s software package ‘Atom in a Box’, available at daugerresearch.com)
observed with an acceleration anywhere near this value.
Why are atoms not flat? Why do shapes exist?

The quantum of action determines all sizes in nature. In particular, it determines all
shapes. Let us explore this topic.
Experiments show that all composed quantons have structures of finite sizes and often
with complex shapes. All the size and shape of composed quantons is due to the motion
of their constituents.
In 1901, Jean Perrin, and independently, in 1904, Nagaoka Hantaro, proposed that
Ref. 40 atoms are small ‘solar systems’. In 1913, Niels Bohr used this idea, combining it with the
quantum of action, and found that he could predict a number of properties of the hy-
Ref. 41 drogen atom that had not until then been understood. These calculations, though often
found in textbooks, are only of historical interest today. Even Bohr knew that they were
not correct, because they assumed that hydrogen atoms were flat. This is observed not to
Challenge 61 e be the case; moreover, it contradicts the quantum of action. The quantum of action im-
plies that the motion of quantum constituents is fuzzy. Therefore, all composed quantons
must be made of clouds of constituents.
In short, the quantum of action predicts that atoms are spherical clouds. Experiment
and theory show that the shape of any atom or molecule is due to the cloud, or probability
distribution, of its electrons. The quantum of action thus states that atoms or molecules
are not hard balls, as Democritus or Dalton believed, but that they are clouds. Atomic
electron clouds are not infinitely hard, but can to a certain degree mix and interpenetrate.
Indeed, this mixing leads to molecules, liquids, solids, flowers and people.
Matter is made of clouds. On the other hand, while atoms are spherical, molecules
have more complex shapes. A detailed exploration shows that all shapes, from the sim-
plest molecules to the shape of people, are due to the interactions between electrons and
nuclei of the constituent atoms. Nowadays, many shapes can be calculated to high preci-
sion. Small molecules, like water, have shapes that are fairly rigid, though endowed with
source
F I G U R E 37 The quantization of
angular momentum
a certain degree of elasticity. Large molecules, such as polymers or peptides, have flexi-
ble shapes. These shape changes are essential for their effects inside cells and thus for our
survival. A large body of biophysical and biochemical research is exploring molecular
shape effects.
In summary, the quantum of action thus implies that shapes exist and that they fluc-
tuate. For example, if a long molecule is held fixed at its two ends, it cannot remain at
rest in between. Such experiments are easy to perform nowadays, for example with DNA;
they again confirm that perfect rest does not exist, and that the quantum of action is at
the basis of chemistry and life.
All shapes are due to the quantum of action. Every object with a non-spherical shape
is able to rotate. Let us explore what the quantum of action can say about rotation.
Rotation, quantization of angular momentum, and the lack of

north poles
“
Tristo è quel discepolo che non avanza il suo
”
maestro.
Leonardo da Vinci*
In everyday life, rotation is a frequent type of motion. Wheels are all around us. It turns
out that the quantum of action has important consequences for rotational motion. First
of all, we note that action and angular momentum have the same physical dimension:
both are measured in Js or Nms. It only takes a little thought to show that if matter or
radiation has a momentum and wavelength related by the quantum of action, then an-
gular momentum is fixed in multiples of the quantum of action. This famous argument
Ref. 42 is due to Dicke and Wittke.
* ‘Sad is that disciple who does not surpass his master.’ This statement is sculpted in large letters in the
chemistry aula of the University of Rome La Sapienza.
Imagine a circular fence, made of N steel bars spaced apart at a distance a = 2πR/N,
as shown in Figure 37. At the centre of the fence, imagine a source of matter or radiation
that can emit particles towards the fence in any chosen direction. The linear momentum
of such a particle is p = ħk = 2πħ/λ. At the fence slits, the wave will interfere. Out-
side the fence, the direction of the motion of the particle is determined by the condition
of positive interference. In other words, the angle θ, describing the direction of motion
outside the fence, is given by a sin θ = Mλ, where M is an integer. Through the deflec-
tion due to the interference process, the fence receives a linear momentum p sin θ, or an
angular momentum L = pR sin θ. Using all these expressions, we find that the angular
momentum transferred to the fence is
L = N Mħ . (27)
In other words, the angular momentum of the fence is an integer multiple of ħ. Fences
can only have integer intrinsic angular momenta (in units of ħ). Of course, this is only
a hint, not a proof. Nevertheless, the generalization of argument to all bodies is correct.
The measured intrinsic angular momentum of bodies is always a multiple of ħ. Quantum
theory thus states that every object’s angular momentum increases in steps. Angular mo-
mentum is quantized.
But rotation has more interesting aspects. Thanks to the quantum of action, just as
linear momentum is usually fuzzy, so is angular momentum. There is an indeterminacy
Ref. 43 relation for angular momentum L. The complementary variable is the phase angle φ of
Ref. 44 the rotation. The indeterminacy relation can be expressed in several ways. The simplest
Page 43 – though not the most precise – is
ħ
ΔL Δφ ⩾ . (28)
2
(This is obviously an approximation: the relation is only valid for large angular momenta.
It cannot be valid for small values, as Δφ by definition cannot grow beyond 2π. In par-
ticular, angular-momentum eigenstates have ΔL = 0.*) The indeterminacy of angular
momentum appears for all macroscopic bodies, or alternatively, for all cases when the
angular phase of the system can be measured.
The quantization and indeterminacy of angular momentum have important conse-
quences. Classically speaking, the poles of the Earth are the places that do not move
when observed by a non-rotating observer. Therefore, at those places matter would have
a defined position and a defined momentum. However, the quantum of action forbids
this. There cannot be a North Pole on Earth. More precisely, the idea of a rotational axis
is an approximation, not valid in general.
* An exact formulation of the indeterminacy relation for angular momentum is
ħ
ΔL Δφ ⩾ |1 − 2πP(π)| , (29)
2
where P(π) is the normalized probability that the angular position has the value π. For an angular-
momentum eigenstate, one has Δφ = π/󵀂3 and P(π) = 1/2π. This expression has been tested and con-
Ref. 45 firmed by experiments.
Rotation of quantons
Even more interesting are the effects of the quantum of action on microscopic particles,
such as atoms, molecules or nuclei. We note again that action and angular momentum
have the same units. The precision with which angular momentum can be measured
depends on the precision of the rotation angle. But if a microscopic particle rotates, this
rotation might be unobservable: a situation in fundamental contrast with the case of
macroscopic objects. Experiments indeed confirm that many microscopic particles have
unobservable rotation angles. For example, in many (but not all) cases, an atomic nucleus
rotated by half a turn cannot be distinguished from the unrotated nucleus.
If a microscopic particle has a smallest unobservable rotation angle, the quantum of
action implies that the angular momentum of that particle cannot be zero. It must always
be rotating. Therefore we need to check, for each particle, what its smallest unobservable
angle of rotation is. Physicists have checked all particles in nature in experiments, and
found smallest unobservable angles (depending on the particle type) of 0, 4π, 2π, 4π/3,
π, 4π/5, 2π/3, etc.
Let us take an example. Certain nuclei have a smallest unobservable rotation angle
of half a turn. This is the case for a prolate nucleus (one that looks like a rugby ball)
turning around its short axis. Both the largest observable rotation and the indeterminacy
are thus a quarter turn. Since the change, or action, produced by a rotation is the number
of turns multiplied by the angular momentum, we find that the angular momentum of
this nucleus is 2 ⋅ ħ.
As a general result, we deduce from the minimum angle values that the angular mo-
mentum of a microscopic particle can be 0, ħ/2, ħ, 3ħ/2, 2ħ, 5ħ/2, 3ħ, etc. In other words,
the intrinsic angular momentum of a particle, usually called its spin, is an integer mul-
tiple of ħ/2. Spin describes how a particle behaves under rotations. (It turns out that all
spin-0 particles are composed of other particles, thus respecting the quantum of action
as the limit for rotational motion in nature.)
How can a particle rotate? At this point, we do not yet know how to picture the rota-
tion. But we can feel it – just as we showed that light is made of rotating entities: all matter,
including electrons, can be polarized. This is shown clearly by the famous Stern–Gerlach
experiment.
Silver, Stern and Gerlach – polarization of quantons

After a year of hard work, in 1922, Otto Stern and Walther Gerlach* completed a beautiful
experiment to investigate the polarization of matter quantons. They knew that inhomoge-
neous magnetic fields act as polarizers for rotating charges. Rotating charges are present
in every atom. Therefore they let a beam of silver atoms, extracted from an oven by evap-
oration, pass an inhomogeneous magnetic field. They found that the beam splits into two
Ref. 46 separate beams, as shown in Figure 38. No atoms leave the magnetic field region in in-
termediate directions. This is in full contrast to what would be expected from classical
physics.
* Otto Stern (1888–1969) and Walther Gerlach (1889–1979), both German physicists, worked together at
the University of Frankfurt. For his subsequent measurement of the anomalous magnetic moment of the
proton, Stern received the Nobel Prize for physics in 1943, after he had to flee Germany.
observation
classical
prediction
silver
z beam
N
∂B
∂z
S
aperture
silver F I G U R E 38 The
oven beam Stern–Gerlach
experiment
The splitting into two beams is an intrinsic property of silver atoms; today we know
that it is due to their spin. Silver atoms have spin ħ/2, and depending on their orien-
tation in space, they are deflected either in the direction of the field inhomogeneity or
against it. The splitting of the beam is a pure quantum effect: there are no intermediate
options. Indeed, the Stern–Gerlach experiment provides one of the clearest demonstra-
tions that classical physics does not work well in the microscopic domain. In 1922, the
result seemed so strange that it was studied in great detail all over the world.
When one of the two beams – say the ‘up’ beam – is passed through a second set-up,
all the atoms end up in the ‘up’ beam. The other possible exit, the ‘down’ beam, remains
unused in this case. In other words, the up and down beams, in contrast to the original
beam, cannot be split. This is not surprising.
But if the second set-up is rotated by π/2 with respect to the first, again two beams
– ‘right’ and ‘left’ – are formed, and it does not matter whether the incoming beam is
directly from the oven or from the ‘up’ part of the beam. A partially-rotated set-up yields
a partial, uneven split. The proportions depend on the angle.
We note directly that if we split the beam from the oven first vertically and then hori-
Challenge 62 e zontally, we get a different result from splitting the beam in the opposite order. Splitting
processes do not commute. (When the order of two operations makes a difference to
the net result, physicists call them ‘non-commutative’.) Since all measurements are also
physical processes, we deduce that, in general, measurements and processes in quantum
systems are non-commutative.
Beam splitting is direction-dependent. Matter beams behave almost in the same way
as polarized light beams. Indeed, the inhomogeneous magnetic field acts on matter some-
what like a polarizer acts on light. The up and down beams, taken together, define a po-
larization direction. In fact, the polarization direction can be rotated (with the help of
a homogeneous magnetic field). Indeed, a rotated beam in a unrotated magnet behaves
like an unrotated beam in a rotated magnet.
Curiosities and fun challenges about quantum matter
“
It is possible to walk while reading, but not to
”
read while walking.
Serge Pahaut
The quantum of action implies that there are no fractals in nature. Everything is made of
particles. And particles are clouds. Quantum theory requires that all shapes in nature be
‘fuzzy’ clouds.
∗∗
Can atoms rotate? Can an atom that falls on the floor roll under the table? Can atoms be
put into high-speed rotation? The answer is ‘no’ to all these questions, because angular
Ref. 58 momentum is quantized; moreover, atoms are not solid objects, but clouds. The macro-
scopic case of an object turning more and more slowly until it stops does not exist in the
microscopic world. The quantum of action does not allow it.
∗∗
Light is refracted when it enters dense matter. Do matter waves behave similarly? Yes,
they do. In 1995, David Pritchard showed this for sodium waves entering a gas of helium
Ref. 60 and xenon.
∗∗
Many quantum effects yield curves that show steps. An important example is the molar
heat of hydrogen H2 gas. In creasing the temperature from 20 to 8 000 K, the molar heat
is shows two steps, first from 3R/2 to 5R/2, and then to 7R/2. Can you explain why?
∗∗
Most examples of quantum motion given so far are due to electromagnetic effects. Can
Challenge 63 ny you argue that the quantum of action must also apply to nuclear motion?
First summary on the motion of quantum particles

In summary, the ‘digital’ beam splitting seen in the Stern–Gerlach experiment and the
wave properties of matter force us to rethink our description of motion. In special relativ-
ity, the existence of a maximum speed forced us to introduce the concept of space-time,
and then to refine our description of motion. In general relativity, the maximum force
obliged us to introduce the concepts of horizon and curvature, and then again to refine
our description of motion. At this point, the existence of the quantum of action and the
wave behaviour of matter force us to take the same two similar steps. We first introduce
the concept of a wave function, and then we refine our description of motion.
Motion Mountain – The Adventure of Physics pdf file available free of charge at www.motionmountain.net Copyright © Christoph Schiller November 1997–July 2010
73
motion of mat ter – beyond cl assical physics
Chapter 4
THE QUANTUM DESC R I PTION OF

MAT TER AND I T S MOTION
I
n everyday life and in classical physics, we say that a system has a position, that
t is oriented in a certain direction, that it has an axis of rotation, and that
t is in a state with specific momentum. In classical physics, we can talk in this
way because the state – the situation a system ‘is’ in and the properties a system ‘has’ –
and the results of measurement coincide. They coincide because measurements can be
imagined to have a negligible effect on the system.
Because of the existence of a smallest action, the interaction necessary to perform
a measurement on a system cannot be made arbitrarily small. Therefore, the quantum
of action makes it impossible for us to continue saying that a system has momentum,
position or an axis of rotation. We are forced to use the idea of the rotating arrow and to
introduce the concept of wave function or state function. Let us see why and how.
The Stern–Gerlach experiment shows that the measured values of spin orientation
are not intrinsic, but result from the measurement process itself (in this case, the interac-
tion with the inhomogeneous field). This is in contrast to the spin magnitude, which is
intrinsic and state-independent.
Therefore, the quantum of action forces us to distinguish three entities:
— the state of the system;
— the operation of measurement;
— the result of the measurement.
In contrast to the classical, everyday case, the state of a quantum system (the properties a
system ‘has’) is not described by the outcomes of measurements. The simplest illustration
of this difference is the system made of a single particle in the Stern–Gerlach experiment.
The experiment shows that a spin measurement on a general (oven) particle state some-
times gives ‘up’ (say +1), and sometimes gives ‘down’ (say −1). So a general atom, in an
oven state, has no intrinsic orientation. Only after the measurement, an atom is either in
an ‘up’ state or in a ‘down’ state. It is also found that feeding ‘up’ states into the measure-
ment apparatus gives ‘up’ states: thus certain special states, called eigenstates, do remain
unaffected by measurement. Finally, the experiment shows that states can be rotated by
applied fields: they have a direction in space.
These details can be formulated in a straightforward way. Since measurements are
operations that take a state as input and produce as output an output state and a mea-
surement result, we can say:
— States are described by rotating arrows.
the quantum description of mat ter 75
— Measurements of observables are operations on the state vectors.

— Measurement results are real numbers; and like in classical physics, they usually de-
pend on the observer.
Thus we have distinguished three quantities that are not distinguished in classical physics:
states, measurement of observables and measurement results. Given this distinction,
quantum theory follows quite simply, as we shall see.
Quantum physics describes observables as operators, and thus as transformations in
Hilbert space, because any measurement is an interaction with a system and thus a trans-
formation of its state. The Stern–Gerlach experiment shows this clearly: the interaction
with the field transforms some atoms in one way, and other atoms in another way.
— Mathematically, states are complex vectors in an abstract space. The space of all possi-
Page 202 ble states is a Hilbert space.
— Mathematically, measurements are linear transformations, more precisely, they are
described by self-adjoint, or Hermitean, operators (or matrices).
— Mathematically, changes of viewpoint are described by unitary operators (or matrices)
that act on states and on measurement operators.
Quantum-mechanical experiments also show that a measurement of an observable
can only give a result that is an eigenvalue of this transformation. The resulting states,
those exceptional states that are not influenced when the corresponding variable is mea-
sured, are the eigenvectors. Therefore every expert on motion must know what an eigen-
value and an eigenvector is.
For any linear transformation T, those special vectors ψ that are transformed into
multiples of themselves,
T ψ = λψ (30)
are called eigenvectors (or eigenstates), and the multiplication factor λ is called the asso-
ciated eigenvalue. Experiments show that the state of the system after a measurement
is given by the eigenvector corresponding to the measured eigenvalue. In the Stern–
Gerlach experiment, the eigenstates are the ‘up’ and the ‘down’ states. Eigenstates are
those states that do not change when the corresponding variable is measured. Eigenval-
ues of Hermitean operators are always real, so that consistency is ensured.
In summary, the quantum of action obliges us to distinguish between three concepts
that are mixed together in classical physics: the state of a system, a measurement on the
system, and the measurement result. The quantum of action forces us to change the vo-
cabulary with which we describe nature, and obliges to use more differentiated concepts.
Now follows the main step: the description of motion with these concepts. This is what
is usually called ‘quantum theory’.
Visualizing the wave function: rotating arrows and probability

clouds
The state of a quantum particle is described by a wave function. To be able to visualize the
wave function, we first imagine that a quantum particle is localized as much as possible.
In this case, the wave function for a free quanton can be described simply by a rotating
arrow.
76 4 the quantum description of mat ter
F I G U R E 39 The motion of a wave

function, the quantum state, through a
double slit, showing both the particle
and the wave properties of matter. The
density of the state is displayed by
brightness, and the local phase is
encoded in the colour. (QuickTime film
© Bernd Thaller)
When a localized quanton travels through space, the attached arrow rotates. If the
particle is non-relativistic and if spin can be neglected, the rotation takes place in a plane
perpendicular to the direction of motion. The end of the arrow then traces a helix around
the direction of motion. In this case, the state at a given time is described by the angle
of the arrow. This angle is the quantum phase. The quantum phase is responsible for the
wave properties of matter, as we will see. The wavelength and the frequency of the helix
are determined by the momentum and the kinetic energy of the particle.
If the particle is not localized – but still non-relativistic and with negligible spin effects
– the state, or the wave function, defines a rotating arrow at each point in space. The
rotation still takes place in a plane perpendicular to the direction of motion. But now we
have a distribution of arrows that trace helices parallel to the direction of motion. At a
given point in space and time, the state has a quantum phase and a length of the arrow.
Figure 39 Shows the evolution of a wave function. The direction of the arrow at each
point is shown by the colour at the specific point. The wave function ψ(t, x) for non-
relativistic particles with negligible spin effects is described by a length and a phase: it
is a complex number at each point in space. The phase is essential for interference and
many other wave effects. What measurable property does the amplitude, the length of
the local arrow, describe? The answer was given by the German physicist Max Born:the
amplitude of the wave function is a probability amplitude. The square of the amplitude,
i.e., the quantity |ψ(t, x)|2 gives the probability to find the particle at the place x at time
t.
Note that even though the wave function can be seen as defining an arrow at every
point in space, the wave function as a whole can also be described as one, single vector,
this time in a Hilbert space. For free particles, the Hilbert space is infinite dimensional!
Nevertheless, it is not hard to calculate in such spaces. The scalar product of two wave
functions is the spatial integral of the product of the complex conjugate of the first func-
tion and the (unconjugated) second function. With this definition, all vector concepts
(unit vectors, null vectors, basis vectors, etc.) can be meaningfully applied to wave func-
tions.
In short, we can imagine the state or wave function of non-relativistic quantum parti-
cles as an arrow at every point in space. The rotation frequency of the arrow is its kinetic
energy; the wavelength of the arrow motion – the period of the helical curve that the tip
of the arrow traces during motion – is the momentum of the quantum particle. An arrow
at each point in space is a field; since the field is concentrated in the region where the
particle is located, and since the amplitude of the field is related to the probability to find
the particle, the wave field can best be called a cloud.
The state or wave function of a quantum particle is a rotating cloud. One aspect of this
rotating cloud is unusual: the cloud is made of little arrows. Every point of the cloud is
described by a local density and a local orientation. This last property is not present in
any cloud of everyday life. Therefore, for many decades it was tacitly assumed that no
simple visualization of such a cloud is possible. Only the last years have shown that there
is a simple visualization for such clouds; this visualization is presented in the last volume
Vol. VI, page 160 of this series.
The state evolution – the Schrödinger equation

The description of the state of a non-relativistic quanton with negligible spin effects as
rotating cloud completely determines how the wave function evolves in time. For such
quantum particles the evolution follows from the total energy, the sum of kinetic and
potential energy T + V , and the properties of matter waves.
The local rate of change of the state arrow ψ is produced by the local total energy, or
Hamiltonian, H = T + V :
∂
iħ ψ = Hψ . (31)
∂t
This famous equation is Schrödinger’s equation of motion.* This evolution equation ap-
plies to all quantum systems and is one of the high points of modern physics.
Ref. 47 In fact, Erwin Schrödinger had found his equation in two different ways. In his first
Ref. 48 paper, he deduced it from a variational principle. In his second paper, he deduced the
evolution equation directly, by asking a simple question: how does the state evolve? He
knew that the state of a quanton behaves both like a wave and like a particle. A wave is
described by a field, which he denoted ψ(t, x). If the state ψ behaves like a wave, then
the corresponding wave function’ must be an amplitude W multiplied by a phase factor
* Erwin Schrödinger (b. 1887 Vienna, d. 1961 Vienna) was famous for being a physicien bohémien, always
living in a household with two women. In 1925 he discovered the equation that brought him international
fame, and the Nobel Prize for physics in 1933. He was also the first to show that the radiation discovered
by Victor Hess in Vienna was indeed coming from the cosmos. He left Germany, and then again Austria,
out of dislike for National Socialism, and was a professor in Dublin for many years. There he published his
famous and influential book What is life?. In it, he came close to predicting the then-unknown nucleic acid
DNA from theoretical insight alone.
F I G U R E 40 Erwin Schrödinger (1887–1961)
eikx−ωt . The state can thus be written as
ψ(t, x) = W(t, x)eikx−ωt . (32)
The amplitude W is the length of the local arrow; the phase is the orientation of the local
arrow. Equivalently, the amplitude is the local density of the cloud, and the phase is the
local orientation of the cloud.
We know that the quantum wave must also behave like a particle of mass m. In par-
ticular, the non-relativistic relation between energy and momentum E = p2 /2m + V (x)
– where V (x) is the potential at position x – must be fulfilled for these waves. The two
Page 63 de Broglie relations (21) for matter wavelength and frequency then imply
∂ψ −ħ2 2
iħ = Hψ = ∇ ψ + V (x)ψ . (33)
∂t 2m
As mentioned, this equation states how the arrow wave associated to a particle, the wave
function ψ, evolves over time. In 1926, this wave equation for the complex field ψ became
instantly famous when Schrödinger used it, by inserting the potential felt by an electron
near a proton, to calculate the energy levels of the hydrogen atom. In a hydrogen atom,
light is emitted by the single electron inside that atom; therefore a precise description of
the motion of the electron in a hydrogen atom allows us to describe the light frequen-
cies it can emit. (We will perform the calculation and the comparison with experiment
Page 155 below.) First of all, the Schrödinger equation explained that only discrete colours are emit-
ted by hydrogen; in addition, the frequencies of the emitted light were found to be in
agreement with the prediction of the equation to five decimal places. This was an impor-
tant result, especially if we keep in mind that classical physics cannot even explain the
existence of atoms, let alone their light emission! In contrast, quantum physics explains
all properties of atoms and their colours to high precision. The discovery of the quantum
of action led the description of the motion of matter to a new high point.
In fact, the exact description of matter is found when spin and the relativistic energy–
Page 159 momentum relation are taken into account. We do this below. No deviations between
relativistic calculations and experiments have ever been found. Even today, predictions
and measurements of atomic spectra remain the most precise and accurate in the whole
study of nature: in the cases that experimental precision allows it, the calculated values
agree with experiments to 13 decimal places.
Self-interference of quantons
Waves interfere. We saw above that all experiments confirm that all quantum parti-
cles show interference. Figure 39 shows that interference is a direct consequence of the
Schrödinger equation. The film shows the solution of the Schrödinger equation for a dou-
ble slit. The film visualizes how the wave function behaves when a double slit induces
diffraction and interference. It turns out that the Schrödinger equation indeed repro-
duces and explains the observations shown in Figure 34 and Figure 35: interference is
due to the evolution of rotating clouds.
Like in all interference phenomena, the local intensity of the interference pattern turns
out to be proportional to the square |W|2 of the state amplitude. We also note that even
though the wave function is spread out over the whole detection screen just before it
hits the screen, it nevertheless yields only a localized spot on the screen. This effect, the
Page 128 so-called collapse of the wavefunction, is explored in detail below.
The speed of quantons

Let us delve a little into the details of the description given by the Schrödinger equation
(33). The equation expresses a simple connection: the classical speed of a matter particle
is the group velocity of the wave function ψ. We know from classical physics that the
group velocity is not always well defined: in cases where the group dissolves in several
peaks, the concept of group velocity is not of much use. These are the cases in which
quantum motion is very different from classical motion, as we will soon discover. But for
the other, more well-behaved cases, we find that the wave function moves in the same
way as a classical particle does.
The Schrödinger equation makes another point: velocity and position of matter are
not independent variables, and cannot be chosen at will. Indeed, the initial condition of
a system is given by the initial value of the wave function alone. No derivatives have to
be (or can be) specified. In other words, quantum systems are described by a first-order
evolution equation, in stark contrast to classical systems.
Dispersion of quantons Copyright © Christoph Schiller November 1997–July 2010
For free quantum particles, the evolution equation has a simple consequence, shown in
Figure 41. Imagine a wave function that is localized around a given starting position.
This wave function describes a quantum system at rest. When time passes, this wave
function will spread out in space. Indeed, the evolution equation is similar, mathemati-
cally, to a diffusion equation. In the same way that a drop of ink in water spreads out, also
a localized quantum particle will spread out in space. The most probable position stays
unchanged, but the probability to find the particle at large distances from the starting
position increases over time. For quantum particles, this spreading effect is indeed ob-
served in experiment; it is a consequence of the wave aspect of matter. For macroscopic
objects, the spreading effect is not observed, however: cars rarely move away from park-
ing spaces. Indeed, quantum theory predicts that for macroscopic systems, the effect of
F I G U R E 41 The evolution of a wave

function (lowest curve) with zero
momentum, and the motion of its
parts with positive and negative
momenta. Local phase is encoded in
the colour. (QuickTime film © Bernd
Thaller)
Challenge 64 ny spreading is negligibly small. Can you show why?

In summary, the wave aspect of matter leads to the spreading of wave functions. Wave
functions show dispersion.
Tunnelling and limits on memory – damping of quantons

‘Common sense’ says that a slow ball cannot roll over a high hill. More precisely, classical
physics says that if the kinetic energy T is smaller than the potential energy V that the
ball would have at the top of the hill, then the ball cannot reach the top of the hill. In
contrast, according to quantum theory, there is a positive probability of passing the hill
for any energy of the ball.
In quantum theory, hills and obstacles are described by potential barriers, and objects
by wave functions. Any initial wave function will spread beyond any potential barrier of
finite height and width. The wave function will also be non-vanishing at the location
of the barrier. In short, any object can overcome any hill or barrier. This effect is called
the tunnelling effect. It is in complete contrast to everyday experience – and to classical
mechanics.
The tunnelling effect results from a new aspect contained in the quantum description
of hills: the statement that all obstacles in nature can be overcome with a finite effort. No
obstacle is infinitely difficult to surmount. Indeed, only for a potential of infinite height
would the wave function vanish and fail to spread to the other side. But such potentials
exist only as approximations; in nature potentials are always of finite value.
Challenge 65 ny How large is the tunnelling effect? Calculation shows that the transmission probability
F I G U R E 42 The tunnelling of a wave

function through a potential hill (the
rectangular column): most of the
wave function is reflected, and part of
the wave function passes to the
other side. Local phase is encoded in
the colour. (QuickTime film © Bernd
Thaller)
V
m T
0 Δx F I G U R E 43 Climbing a hill
P is given approximately by
2󰑤
16T(V − T) − 󵀄2m(V − T)
P≈ e ħ (34)
V2
where 󰑤 is the width of the hill, 󰑣 its height, and m and T the mass and the kinetic energy
of the particle. For a system of large number of particles, the probability is the product
of the probabilities for the different particles.
Let us take the case of a car in a garage, and assume that it is made of 1028 atoms at
room temperature. A typical garage wall has a thickness of 0.1 m and a potential height
of V = 1 keV = 160 aJ for the passage of an atom. We get that the probability of finding
the car outside the garage is
(1028 )
−(1012 ) 40
P ≈ 󶀣10 󶀳 ≈ 10−(10 )
. (35)
Farady cage
with high screen with
electric intereference
potential pattern that
depends on
potential
beam splitter
F I G U R E 44 A localized electric potential
in an interferometer leads to a shift of
charged matter beam the interference pattern
Challenge 66 e The smallness of this value (just try to write it down, to be convinced) is the reason why
it is never taken into account by the police when a car is reported missing. (Actually,
the probability is considerably smaller. Can you name at least one effect that has been
Challenge 67 ny forgotten in this simple calculation?)
Obviously, tunnelling can be important only for small systems, made of a few parti-
cles, and for thin barriers, with a thickness of the order of ħ/󵀄2m(V − T) . For example,
tunnelling of single atoms is observed in solids at high temperature, but is not impor-
tant in daily life. For electrons, the effect is larger: the barrier width 󰑤 for an appreciable
tunnelling effect is
0.5 nm 󵀂aJ
󰑤≈ . (36)
󵀂V − T
At room temperature, the kinetic energy T is of the order of 6 zJ; increasing the tempera-
ture obviously increases the tunnelling. As a result, electrons tunnel quite easily through
barriers that are a few atoms in width. Indeed, every TV tube uses tunnelling at high
temperature to generate the electron beam producing the picture. The necessary heating
is the reason why television tubes take some time to switch on.
The tunnelling of electrons also limits the physical size of computer memories. Mem-
ory chips cannot be made arbitrary small. Silicon integrated circuits with one terabyte
Challenge 68 s (TB) of random-access memory (RAM) will probably never exist. Can you imagine why?
In fact, tunnelling limits the working of any type of memory, including that of our brain.
Indeed, if we were much hotter than 37°C, we could not remember anything!
Since light is made of particles, it can also tunnel through potential barriers. The best
– or highest – potential barriers for light are mirrors; mirrors have barrier heights of the
order of one attojoule. Tunnelling implies that light can be detected behind any mirror.
These so-called evanescent waves have indeed been detected; they are used in various
high-precision experiments.
solenoid
with
current screen with
intereference
pattern that
depends on
magnetic field
beam splitter
F I G U R E 45 Magnetic fields change the

neutral matter beam phase of a spinning particle
The quantum phase
The motion of the cloud for a single microscopic particle is described by two quantities:
the amplitude and the phase. Whereas the amplitude is easy to picture – just think of the
(square root of the) density of a real cloud – the phase takes more effort. States or wave
functions are clouds with a local phase: they are clouds of objects that can be rotated. In
case of an everyday water cloud, local rotation of droplets has no effect of the cloud. In
contrast, in quantum theory, the local rotation of the cloud, thus the change of its phase,
does have a measurable effect. Let us explore this point.
Page 49 The phase of matter waves behaves like the phase of photons: it evolves with time, and
thus increases along the path of a moving particle. The phase can be pictured by a small
rotating arrow. The angular velocity with which the phase rotates is given by the famous
relation ω = E/ħ. In short, we can picture the wave function of a free particle as a moving
cloud that rotates and disperses.
Above all, the phase is that aspect of the wave function that leads to the wave effects of
matter. In particular, the phase of the wave function leads to interference effects. When
two partial wave functions are separated and recombined after a relative phase change,
the phase change will determine the interference. This is the origin of the electron beam
interference observations shown in Figure 34. Without quantum phase, there would be
no extinction and no interference.
The phase of a wave function can be influenced in many ways. The simplest way is the
use of electric fields. If the wave function of a charged particle is split, and one part is
led through a region with an electric field, a phase change will result. The arrangement
is shown in Figure 44. A periodic change of the electric potential yields a periodic shift
of the interference pattern. This is indeed observed.
Another simple case of phase manipulation is shown in Figure 45: also a magnetic
field changes the phase of a spinning particle, and thus influences the interference be-
haviour.
A famous experiment shows the importance of the phase in an even more surprising
Ref. 49 way: the Aharonov–Bohm effect. The effect is famous because it is counter-intuitive and
because it was predicted before it was observed. Look at the set-up shown in Figure 46.
magnetic field (even screen

if only inside the solenoid)
current
vector
potential
charged matter beam
F I G U R E 46 The Aharonov–Bohm effect: the influence of the magnetic vector potential on interference
(left) and a measurement confirmation (right), using a microscopic sample that transports electrons in
thin metal wires (© Doru Cuturela)
F I G U R E 47 The motion of a wave

function around a solenoid showing the

Aharonov–Bohm effect. The density of
the state is displayed by brightness, and
the local phase is encoded in the colour.
(QuickTime film © Bernd Thaller)
A matter wave of charged particles is split into two by a cylinder – positioned at a right
angle to the matter’s path – and the matter wave recombines behind it. Inside the cylinder
there is a magnetic field; outside, there is none. (A simple way to realize such a cylinder is
electrically
charged
wire
screen with
intereference
pattern that
depends on
wire charge
F I G U R E 48 The Aharonov–Casher effect:

beam splitter the influence of charge on the phase
leads to interference even for interfering
polarized neutron beam
neutrons
a long solenoid.) Quantum physics predicts that an interference pattern will be observed,
and that the position of the stripes will depend on the value of the magnetic field. This
happens even though the wave never enters the region with the field! The effect has been
observed in countless experiments.
The Aharonov–Bohm effect is surprising. The reason for the effect is simple: for a
charged particle, the phase of a wave function is determined by the vector potential A,
not by the magnetic field B. The vector potential around a solenoid does not vanish, as
Page 69 we know from the section on electrodynamics, but circulates around the solenoid. This
circulation distinguishes the two sides of the solenoid and leads to a phase shift – one
that indeed depends on the magnetic field value – and thus produces interference, even
though the particle never interacts with the magnetic field.
A further example for phase manipulation is the so-called Aharonov–Casher effect,
which even occurs for neutral particles, as long as they have a magnetic moment, such
as neutrons. The phase of a polarized neutron will be influenced by an electric field, so
that the arrangement shown in Figure 48 will show an interference pattern that depends
on the applied electric potential.
Another case of phase manipulation will be presented later on: also gravitational fields
can be used to rotate wave functions. In fact, even the acceleration due to rotational
motion can do so. In fact, it has been possible to measure the rotation of the Earth by
Ref. 50 observing the change of neutron beam interference patterns.
Another important class of experiments that manipulate the phase of wave functions
are possible with macroscopic quantum phenomena. In superconductivity and in super-
fluidity, the phase of the wave function is regularly manipulated with magnetic and elec-
tric fields. This possibility has many important technical applications. For example, the
Josephson effect is used to measure electric potential differences by measuring the fre-
quency of emitted radio waves, and so-called superconducting quantum interference de-
vices, or SQIDs, are used to measure tiny magnetic fields.
We note that all these experiments confirm that the absolute phase of a wave function
cannot be measured. However, relative phases – phase differences or phase changes – can
Challenge 69 e be measured. Can you confirm this?
All the phase shift effects just explained have been observed in numerous experiments.
The phase of a wave function is an essential aspect of it: the phase is the reason for calling
it wave function in the first place. Like in any wave, the phase evolves over time and it can
be influenced by various external influences. Above all, the experiments show that a lo-
calized quantum particle – thus when the spread of the wave function can be neglected –
is best imagined as a rotating arrow; if the spread cannot be neglected, the wave function
is best imagined as a wave of arrows rotating at each point in space.
The least action principle in quantum physics

In nature, motion happens in a way that minimizes change. Indeed, in classical physics,
Vol. I, page 203 the principle of least action states that in nature, the motion of a particle happens along
that particular path – out of all possible paths with the same end points – for which the
action is minimal. This principle of cosmic laziness was stated mathematically by saying
that in nature, the variation δS of the action is zero. Change minimization explains all
classical evolution equations. We now transfer this idea to the quantum domain.
For quantum systems, we need to redefine both the concept of action and the concept
of variation: first of all, we to find a description of action that is based on operators;
secondly, we need to define the action variation without paths, as the concept of ‘path’
does not exist for quantum systems; thirdly, since there is a smallest action in nature, a
vanishing variation is not a clearly defined concept, and we must overcome this hurdle.
There are two main ways to achieve this: to describe the motion of quantum systems as a
superposition of all possible paths, or to describe action with the help of wave functions.
Both approaches are equivalent.
In the first approach, the path integral formulation, the motion of a quantum particle
is described as a democratic superposition of motions along all possible paths. For each
path, the evolution of the arrow is determined, and at the end point, the arrows from all
paths are added. The action for each path is the number of turns that the arrow performs
along the path. The result from this exercise is that the path for which the arrow makes
the smallest number of turns is usually (but not always!) the most probable path. A more
precise investigation shows that classical, macroscopic systems always follows the path
of smallest action, but quantum systems do not.
In the second approach to quantum physics, action is defined with help of wave func-
tions. In classical physics, we defined the action (or change) as the integral of the La-
grangian between the initial and final points in time, and the Lagrangian itself as the
Page 199 difference between kinetic and potential energy. In quantum physics, the simplest defi-
nition is the quantum action defined by Julian Schwinger. Let us call the initial and final
states of the system ψi and ψf . The action S between these two states is defined as
S = ⟨ψi | 󵐐 L dt | ψf ⟩ , (37)
where L is the Lagrangian (operator). The angle brackets represent the ‘multiplication’ of
states and operators as defined in quantum theory. (We skip the details of notation and
mathematics here.) In simple words, also in quantum theory, action – i.e., the change
occurring in a system – is the integral of the Lagrangian. The Lagrangian operator L is
defined in the same way as in classical physics: the Lagrangian L = T −V is the difference
between the kinetic energy T and the potential energy V operators. The only difference
is that, in quantum theory, the momentum and position variables of classical physics are
replaced by the corresponding operators of quantum physics.*
To transfer the concept of action variation δS to the quantum domain, Julian
Schwinger introduced the straightforward expression
δS = ⟨ψi | δ󵐐 L dt| ψf ⟩ . (38)
The concept of path is not needed in this expression, as the variation of the action is
based on varying wave functions instead of varying particle paths.
The last classical requirement to be transferred to the quantum domain is that, be-
cause nature is lazy, the variation of the action must vanish. However, in the quantum
domain, the variation of the action cannot be zero, as the smallest observable action is
the quantum of action. As Julian Schwinger discovered, there is only one possible way to
express the required minimality of action:
δS = ⟨ψi | δ󵐐 L dt| ψf ⟩ = −iħ δ⟨ψi |ψf ⟩ . (39)
This so-called quantum action principle describes all motion in the quantum domain.
Classically, the right-hand side is zero – since ħ is taken to be zero – and we then recover
the minimum-action principle δS = 0 of classical physics. But in quantum theory, when-
ever we try to achieve small variations, we encounter the quantum of action and changes
of (relative) phase. This is expressed by the right-hand side of the expression.
In simple words, all quantum motion – i.e., the quantum evolution of a state ψ or |ψ⟩
– happens in such a way that the action variation is the same as −i times the quantum
of action ħ times the variation of the scalar product between initial and final states. In
other words, in the actual motion, the intermediate states are fixed by the requirement
that they must lead from the initial state to the final state with the smallest number of
effective turns of the state phase. The factor −i expresses the dependence of the action
on the rotation of the wave function.
In summary, the least action principle is also valid in quantum physics.
The motion of quantons with spin

“ ”
Everything turns.
Anonymous
What is the origin of the quantum phase? Classical physics helps to answer the question.
Quantons can rotate around an axis: we speak of particle spin. But if quantum particles
can spin, they should possess angular momentum. And indeed, experiments confirm
this deduction.
In particular, electrons have spin. The full details of electron spin were deduced from
Ref. 51 experiments by two Dutch students, George Uhlenbeck and Samuel Goudsmit, in 1925.
* More precisely, there is also a condition governing the ordering of operators in a mixed product, so that
the non-commutativity of operators is taken into account. We do not explore this issue here.
They had the guts to publish what Ralph Kronig had also suspected: that electrons rotate
around an axis with a projected component of the angular momentum given by ħ/2. In
fact, this value – often called spin 1/2 for short – is valid for all elementary matter particles.
(In contrast, all elementary radiation particles have spin values of ħ, or spin 1 for short.)
If a spinning particle has angular momentum, it must be possible to rearrange the axis
by applying a torque, to observe precession, to transfer the spin in collisions etc. All this is
Page 70 indeed observed; in fact, the Stern–Gerlach experiment allows all these observations. The
only difference between particle spin and classical angular momentum is that particle
Page 68 spin is quantized, as we deduced above.
In other words, the spin describes how a particle behaves under rotations. One result
of spin is that charged quantum particles also act as small dipole magnets. The strength
of the dipole magnet is described by the so-called д-value. Uhlenbeck and Goudsmit
Page 90 deduced a д-value of 2 for the electron in order to explain the optical spectra. This value
Ref. 52 was explained by Llewellyn Thomas as a relativistic effect a few months later; today one
often speaks of Thomas precession.
By 2004, experimental techniques had become so sensitive that the magnetic effect
of a single electron spin attached to an impurity (in an otherwise non-magnetic ma-
terial) could be detected. Researchers now hope to improve these so-called ‘magnetic-
resonance-force microscopes’ until they reach atomic resolution.
In 1927, the Austrian physicist Wolfgang Pauli* discovered how to include spin 1/2 in
a quantum-mechanical description: instead of a state function described by a single com-
plex number, a state function with two complex components is needed. The reason for
this expansion is simple. In general, the little rotating arrow that describes a quantum
state does not rotate around a fixed plane, as is assumed by the Schrödinger equation;
the plane of rotation has also to be specified at each position in space. This implies that
two additional parameters are required at each space point, bringing the total number
of parameters to four real numbers, or, equivalently, two complex numbers. Nowadays,
Pauli’s equation is mainly of conceptual interest, because – like that of Schrödinger – it
does not comply with special relativity. However, the idea of including the local rotation
plane remains valid. The idea was used by Dirac when he introduced the relativistic de-
scription of the electron, and the idea is also used in all other wave equations for particles
with spin.
In summary, the description of a quanton with spin implies the use of wave functions
that specify two complex numbers at each point in space and time. Copyright © Christoph Schiller November 1997–July 2010
* Wolfgang Ernst Pauli (b. 1900 Vienna, d. 1958 Zürich), at the age of 21, wrote one of the best texts on special
and general relativity. He was the first to calculate the energy levels of hydrogen using quantum theory,
discovered the exclusion principle, incorporated spin into quantum theory, elucidated the relation between
spin and statistics, proved the CPT theorem, and predicted the neutrino. He was admired for his intelligence,
and feared for his biting criticisms, which led to his nickname, ‘conscience of physics’. Despite this, he helped
many people in their research, such as Heisenberg with quantum theory, without claiming any credit for
Ref. 53 himself. He was seen by many, including Einstein, as the greatest and sharpest mind of twentieth-century
physics. He was also famous for the ‘Pauli effect’, i.e., his ability to trigger disasters in laboratories, machines
and his surroundings by his mere presence. As we will see shortly, one can argue that Pauli actually received
the Nobel Prize for physics in 1945 (officially ‘for the discovery of the exclusion principle’) for finally settling
the question of how many angels can dance on the tip of a pin.
Relativistic wave equations

In 1899, Max Planck had discovered the quantum of action. In 1905, Albert Einstein pub-
lished the theory of special relativity, which was based on the idea that the speed of light
c is independent of the speed of the observer. The first question Planck asked himself
was whether the value of the quantum of action would be independent of the speed of
the observer. It was his interest in this question that led him to invite Einstein to Berlin.
With this invitation, he made the patent-office clerk famous in the world of physics.
Experiments show that the quantum of action is indeed independent of the speed of
the observer. All observers find the same minimum value. To include special relativity
into quantum theory, we therefore only need to find the correct quantum Hamiltonian
H.
Given that the classical Hamiltonian of a free particle is given by
H = β󵀆 c 4 m 2 + c 2 p 2 with p = γm󰑣 , (40)
one might ask: what is the corresponding Hamilton operator? The simplest answer was
Ref. 54 given, only in 1950, by L.L. Foldy and S.A. Wouthuysen. The operator is almost the same
one:
1 0 0 0
0 1 0 0
H = β󵀆 c m + c p
4 2 2 2
with β = 󶀫 󶀻 . (41)
0 0 −1 0
0 0 0 −1
The signs appearing in the matrix operator β distinguish between particles and antiparti-
cles. The numbers +1 and −1 appear twice, to take care of the two possible spin directions
for each case.
With this relativistic Hamiltonian operator for spin 1/2 particles – and with all others
– the wave function is described by four complex numbers, two for particles and two for
antiparticles. That each type of particles requires two complex components follows from
the requirement to specify, at each point in space, the length of the arrow, its phase, and
Vol. II, page 65 its plane of rotation. Long ago we also found that relativity automatically introduces an-
Page 163 timatter, and we will explore the issue in detail below. Both matter and antimatter are
part of any relativistic description of quantum effects. The wave function for a particle
has vanishing antiparticle components, and vice versa. In total, the wavefunction for rel-
ativistic spin 1/2 systems has thus four complex components.

The Hamilton operator yields the velocity operator 󰑣 through the same relation that
is valid in classical physics:
d p
󰑣= x=β . (42)
dt 󵀆 c 4 m2 + c 2 p2
This velocity operator shows a continuum of eigenvalues, from minus to plus the speed
of light. The velocity 󰑣 is a constant of motion, as are the momentum p and the energy
E = 󵀆 c 4 m2 + c 2 p2 . (43)
Also the orbital angular momentum L is defined as in classical physics, through
L=x×p. (44)
Ref. 55 The orbital angular momentum L and the spin σ are separate constants of motion. A
particle (or antiparticle) with positive (or negative) angular momentum component has
a wave function with only one non-vanishing component; the other three components
vanish.
But alas, the representation of relativistic motion given by Foldy and Wouthuysen is
not the simplest when it comes to take electromagnetic interactions into account. The
simple identity between the classical and quantum-mechanical descriptions is lost when
electromagnetism is included. We will solve this problem below, when we explore Dirac’s
Page 160 evolution equation for relativistic wave functions.
Composite vs. elementary quantons

When is an object composite, and not elementary? Quantum theory gives several prag-
matic answers. The first one is somewhat strange: an object is composite when its gyro-
Ref. 56 magnetic ratio is different from the one predicted by QED. The gyromagnetic ratio γ is
defined as the ratio between the magnetic moment M and the angular momentum L:
M = γL . (45)
The gyromagnetic ratio γ is measured in units of s−1 T−1 , i.e., C/kg, and determines the
energy levels of magnetic spinning particles in magnetic fields; it will reappear later in
Page 121 the context of magnetic resonance imaging. All candidates for elementary particles have
spin 1/2. The gyromagnetic ratio for spin-1/2 particles of magnetic moment M and mass
m can be written as
M e
γ= =д . (46)
ħ/2 2m
The criterion for being elementary can thus be reduced to a condition on the value of the
dimensionless number д, the so-called д-factor. (The expression eħ/2m is often called
the magneton of the particle. Confusingly, the dimensionless factor д/2 is often called the
gyromagnetic ratio as well.) If the д-factor differs from the value predicted by QED for
point particles – about 2.0 – the object is composite. For example, a 4 He+ helium ion has
Challenge 70 ny spin 1/2 and a д value of 14.7 ⋅ 103 . Indeed, the radius of the helium ion is 3 ⋅ 10−11 m,
obviously a finite value, and the ion is a composite entity. For the proton, one measures
a д-factor of about 5.6. Indeed, experiments yield a finite proton radius of about 0.9 fm
and show that it contains several constituents.
The neutron, which has a magnetic moment despite being electrically neutral, must
therefore be composite. Indeed, its radius is approximately the same as that of the proton.
Similarly, molecules, mountains, stars and people must be composite. According to this
first criterion, the only elementary particles are leptons (i.e., electrons, muons, tauons
and neutrinos), quarks, and intermediate bosons (i.e., photons, W-bosons, Z-bosons and
Page 121 gluons). More details on these particles will be revealed in the chapters on the nucleus.
Another simple criterion for compositeness has just been mentioned: any object with
a measurable size is composite. This criterion yields the same list of elementary particles as
the first. Indeed, the two criteria are related. The simplest model for composite structures
Ref. 57 predicts that the д-factor obeys
R
д−2= (47)
λC
where R is the radius and λC = h/mc is the Compton wavelength of the system. This
expression is surprisingly precise for helium-4 ions, helium-3, tritium ions and protons,
Challenge 71 e as you may wish to check. The tables in Appendix B make the same point.
A third criterion for compositeness is more general: any object larger than its Compton
length is composite. The argument is simple. An object is composite if one can detect
internal motion, i.e., motion of some components. Now the action of any part with mass
mpart moving inside a composed system of size r obeys
Spart < 2π r mpart c < π r m c (48)
where m is the mass of the composite object. On the other hand, following the principle
of quantum theory, this action, to be observable, must be larger than ħ/2. Inserting this
condition, we find that for any composite object*
ħ
r> . (49)
2π m c
The right-hand side differs only by a factor 4π2 from the so-called Compton (wave)length
h
λ= (50)
mc
of an object. Thus any object larger than its own Compton wavelength is composite; and
any object smaller than the right-hand side of expression (49) is elementary. Again, only
leptons, quarks and intermediate bosons pass the test. All other objects are composite.
Challenge 73 ny This third criterion produces the same list as the previous ones. Can you explain why?
A fourth criterion is regularly cited by Steven Weinberg: a particle is elementary if it
appears in the Lagrangian of the standard model of particle physics. Can you show that
Challenge 74 ny this criterion follows from the previous ones?
Interestingly, we are not yet finished with this topic. Even stranger statements about
Vol. VI, page 267 compositeness will appear when gravity is taken into account. Just be patient: it is worth
it.
Challenge 72 ny * Can you find the missing factor of 2? And is the assumption that the components must always be lighter
than the composite a valid one?
F I G U R E 49 A special potential well

that does not disturb a wave function.
Colour indicates phase. (QuickTime
film © Bernd Thaller)
Curiosities and fun challenges about quantum motion of matter
Take the sharpest knife edge or needle tip you can think of: the quantum of action implies
that they are all fuzzy clouds.
∗∗
Do hydrogen atoms exist? Most types of atom have been imaged with microscopes, pho-
tographed under illumination, levitated one by one, and even moved with needles, one
by one, as the picture on page 271 shows. Researchers have even moved single atoms
Ref. 59 by using laser beams to push them. However, not a single one of these experiments has
measured or imaged hydrogen atoms. Is that a reason to doubt the existence of hydrogen
Challenge 75 s atoms? Taking this not-so-serious discussion seriously can be a lot of fun.
∗∗
Two observables can commute for two different reasons: either they are very similar –
such as the coordinates x and x 2 – or they are very different – such as the coordinate x
Challenge 76 ny and the momentum p y . Can you give an explanation for this?
∗∗
Space and time translations commute. Why then do the momentum operator and the
Challenge 77 ny Hamiltonian not commute in general?
∗∗
There are some potentials that have no influence on a wave function. Figure 49 shows
an example. This potential has reflection coefficient zero for all energies; the scattered
wave has no reflected part. The well has the shape of a soliton of the Korteweg–de Vries
equation, which is related to the Schrödinger equation.
∗∗
For a bound system in a non-relativistic state with no angular momentum, one has the
Ref. 61 relation
9ħ2
⟨r 2 ⟩ ⟨T⟩ ⩾ , (51)
8m
where m is the reduced mass and T the kinetic energy of the components, and r is the
Challenge 78 s size of the system. Can you deduce this result, and check it for hydrogen?
∗∗
One often reads that the universe might have been born from a quantum fluctuation.
Challenge 79 ny Can you explain why this statement make no sense?
A summary on motion of quantons

In summary, the motion of quantons can be described in two ways:
— Seen from far away, at low magnification, moving quantum particles and their wave
functions behave as advancing, rotating and precessing arrows. The details of the rota-
tion and precession of the arrow depend on the energy and momentum of the particle.
The squared length of the arrow is the probability to observe a particle. If a particle
can get from a starting point to a final point in several ways, arrows add up.
— At large magnification, quantum particles move like advancing, rotating and precess-
ing clouds. The cloud shape is described by the wave function. The local cloud ori-
entation, the local phase, follows a wobbling motion. The square of the probability
amplitude, the density of the cloud, is the probability for finding the particle at a
given spot.
Rotating arrows resulting from cloud averages combine particle and wave properties.
A full rotation of the arrow corresponds to the quantum of action. These central feature
imply that a non-relativistic particle whose spin can be neglected follows the Schrödinger
equation, and that a relativistic electron follows the Dirac equation. They describe all of
chemistry and biology.
To continue with the greatest efficiency on our path across quantum physics, we ex-
plore three important topics: the issue of indistinguishability of several particles, the topic
of spin, and the issue of the interpretation of the probabilities. Copyright © Christoph Schiller November 1997–July 2010
Chapter 5
PERMUTATION OF PARTICLES – ARE

PARTIC LES LI K E GLOV ES?
W
hy are we able to distinguish twins from each other? Why can we distinguish
hat looks alike, such as a copy from an original? Most of us are convinced that
henever we compare an original with a copy, we can find a difference. This con-
viction turns out to be correct. A short exploration shows that this conviction is in con-
trast with classical physics. The possibility to distinguish originals from copies is a pure
quantum effect.
Indeed, quantum theory has a lot to say about copies and their differences. Think
about any method that allows to distinguish objects: you will find that it runs into trouble
Challenge 80 s for point-like particles. Therefore in the quantum domain something must change about
our ability to distinguish particles and objects. Let us explore the issue.
Some usually forgotten properties of objects are highlighted by studying a pretty com-
binatorial puzzle: the glove problem. It asks:
How many surgical gloves (for the right hand) are necessary if m doctors
need to operate 󰑤 patients in a hygienic way, so that nobody gets in contact
with the body fluids of anybody else?
The same problem also appears in other settings. For example, it also applies to condoms,
men and women – and is then called the condom problem – or to computers, interfaces
Ref. 62 and computer viruses. To be clear, the optimal number of gloves is not the product m󰑤.
In fact, the problem has three subcases.
Challenge 81 s — The simple case m = 󰑤 = 2 already provides the most important ideas needed. Are
you able to find the optimal solution and procedure?
— In the case 󰑤 = 1 and m odd or the case m = 1 and 󰑤 odd, the solution is (m + 1)/2
Challenge 82 e gloves. This is the optimal solution, as you can easily check yourself.
Ref. 63 — A solution with a simple procedure for all other cases is given by ⌈2󰑤/3+m/2⌉ gloves,
where ⌈x⌉ means the smallest integer greater than or equal to x. For example, for two
doctors and three patients this gives only three gloves. (However, this formula does
Challenge 83 e not always give the optimal solution; better values exist in certain subcases.)
Two basic properties of gloves determine the solution to the puzzle. First, gloves have
two sides, an interior and an exterior one. Secondly, gloves can be distinguished from
each other. Do these two properties also apply to particles? We will discuss the issue of
Page 99 double-sidedness in the last part of the mountain ascent. In fact, the question whether
particles can be turned inside out will be of importance for their description and their
motion. (In fact, particles do behave like gloves in the sense that one can distinguish right-
Page 180 handed from left-handed ones.) In the present chapter we concentrate on the second
issue, namely whether objects and particles can always be distinguished. We will find that
elementary particles do not behave like gloves but in an even more surprising manner.
In everyday life, distinction of objects can be achieved in two ways. We are able to
distinguish objects – or people – from each other because they differ in their intrinsic
properties, such as their mass, colour, size or shape. In addition, we are also able to dis-
tinguish objects if they have the same intrinsic properties. Any game of billiard suggests
that by following the path of each ball, we can distinguish it from the others. In short,
objects with identical properties can also be distinguished using their state.
The state of a billiard ball is given by its position and momentum. In the case of billiard
balls, the state allows distinction because the measurement error for the position of the
ball is much smaller than the size of the ball itself. However, in the microscopic domain
this is not the case. Let us take single atoms. Atoms of the same type have the same
intrinsic properties. To distinguish them in collisions, we would need to keep track of
their motion. But we have no chance to achieve this. Already in the nineteenth century it
was shown experimentally that even nature itself is not able to do it! This profound result
was discovered studying systems which incorporate a large number of colliding atoms of
the same type: gases.
Page 297 The calculation of the entropy S of a simple gas, made of N simple particles of mass
m moving in a volume V , gives
S V 3 ln α
= ln 󶁤 3 󶁴 + + (52)
kN Λ 2 N
where k is the Boltzmann constant, ln the natural logarithm, T the temperature, and
Λ = 󵀆2πħ2 /mkT is the thermal wavelength (approximately the de Broglie wavelength
of the particles making up the gas). In this formula, the pure number α is equal to 1
if the particles are distinguishable like billiard balls, and equal to 1/N ! if they are not
distinguishable at all. Measuring the entropy thus allows us to determine α and therefore
whether particles are distinguishable. It turns out that only the second case describes
Challenge 84 e nature. This can easily be checked without even performing the measurement: only in
the second case, α = 1/N ! does the entropy of two volumes of identical gas add up.*
The result, often called Gibbs’ paradox,** thus proves that the microscopic components
Ref. 64 of matter are indistinguishable: in a system of microscopic particles, there is no way to say
which particle is which. Indistinguishability is an experimental property of nature.***

The properties of matter would be completely different without indistinguishability.
Challenge 85 d * Indeed, the entropy values observed by experiment, for a monoatomic gas, are given by the so-called
Sackur–Tetrode formula
S V 5
= ln 󶁤 󶁴+ (53)
kN NΛ 3 2
which follows when α = 1/N ! is inserted above. It was deduced independently by the German physicist Otto
Sackur (1880–1914) and the Dutch physicist Hugo Tetrode (1895–1931). Note that the essential parameter is
the ratio between V /N, the volume per particle, and Λ3 , the de Broglie volume of a particle.
** Josiah Willard Gibbs (1839–1903), US-American physicist who was, with Maxwell and Planck, one of the
three founders of statistical mechanics and thermodynamics; he introduced the concept of ensemble and
F I G U R E 50 Willard Gibbs (1839–1903)
F I G U R E 51 Identical objects with crossing
m paths
For example, we will discover that without it, knifes and swords would not cut. In addi-
tion, the soil would not carry us; we would fall right through it. To illuminate the issue
in more detail, we explore the following question.
Why does indistinguishability appear in nature?

Take two microscopic particles with the same mass, the same composition and the same
shape, such as two atoms. Imagine that their paths cross, and that they approach each
other to small distances at the crossing, as shown in Figure 51. In a gas, both a collision
of atoms or a near miss are examples. Now, experiments show that at small distances
it is impossible to say whether the two particles have switched roles or not. This is the
main reason that makes it impossible in a gas to follow particles moving around and
to determine which particle is which. This impossibility is a direct consequence of the
quantum of action.
For a path that brings two approaching particles very close to each other, a role switch
requires only a small amount of change, i.e., only a small (physical) action. However, we
know that there is a smallest observable action in nature. Keeping track of each particle
at small distances would require action values smaller than the minimal action observed
in nature. The existence of a smallest action thus makes it impossible to keep track of
microscopic particles when they come too near to each other. Any description of several
the term thermodynamic phase.

*** When radioactivity was discovered, people thought that it contradicted the indistinguishability of atoms,
as decay seems to single out certain atoms compared to others. But quantum theory then showed that this
is not the case and that atoms do remain indistinguishable.
particles must thus take into account that after a close encounter, it is impossible to say
which particle is which.
If we remember that quantum theory describes particles as clouds, the indistinguisha-
bility appears more natural. Whenever two clouds meet and depart again, it is impossible
to say which one is which.
If two particles are kept distant enough, one does have an effective distinguishability;
indistinguishability thus appears only when the particles come close. In short, indistin-
guishability is a consequence of the existence of a minimal action in nature. This result
leads straight away to the next question:
Can particles be counted?

In everyday life, objects can be counted because they can be distinguished. Since quan-
tum particles cannot always be distinguished, we need some care in determining how
to count them. The first step in counting particles is the definition of what is meant by
a situation without any particle at all. This seems an easy thing to do, but later on we
will encounter situations where already this step runs into difficulties. In any case, the
first step is thus the specification of the vacuum. Any counting method requires that the
situation without particles is clearly separated from situations with particles.
The second step is the specification of an observable useful for determining particle
number. The easiest way is to chose one of those quantum numbers which add up under
composition, such as electric charge.* Counting is then performed by measuring the total
charge and dividing by the unit charge.
This method has several advantages. First of all, it is not important whether parti-
cles are distinguishable or not; counting always works. Secondly, virtual particles are not
Page 89 counted. This is a welcome state of affairs, as we will see, because for virtual particles, i.e.,
particles for which E 2 ̸= p2 c 2 + m2 c 4 , there is no way to define a particle number anyway.
The side effect of the counting method is that antiparticles count negatively! Also this
consequence is a result of the quantum of action. We saw above that the quantum of
action implies that even in vacuum, particle–antiparticle pairs are observed at sufficiently
high energies. As a result, an antiparticle must count as minus one particle. In other
words, any way of counting particles can produce an error due to this effect. In everyday
life this limitation plays no role, as there is no antimatter around us. The issue does play
a role at higher energies, however. It turns out that there is no general way to count the
exact number of particles and antiparticles separately; only the sum can be defined. In
short, quantum theory shows that particle counting is never perfect.

In summary, nature does provide a way to count particles even if they cannot be dis-
tinguished, though only for everyday, low energy conditions; due to the quantum of ac-
tion, antiparticles count negatively. Antiparticles thus provide a limit to the counting of
particles at high energies, when the mass–energy relation becomes important.
* In everyday life, the weight or mass is commonly used as observable. However, it cannot be used in the
quantum domain, except for simple cases. Can you give at least two reasons, one from special relativity and
Challenge 86 s one from general relativity?
What is permutation symmetry?

Since particles are countable but indistinguishable, there exists a symmetry of nature
for systems composed of several identical particles. Permutation symmetry, also called
exchange symmetry, is the property of nature that observations are unchanged under ex-
change of identical particles. Together with space-time symmetry, gauge symmetry and
the not yet encountered renormalization symmetry, permutation symmetry forms one
of the four pillars of quantum theory. Permutation symmetry is a property of composed
systems, i.e., of systems made of many (identical) subsystems. Only for such systems does
indistinguishability play a role.
In other words, ‘indistinguishable’ is not the same as ‘identical’. Two particles are not
the same; they are more like copies of each other. On the other hand, everyday life ex-
perience shows us that two copies can always be distinguished under close inspection,
so that the term is not fully appropriate either. In the microscopic domain, particles are
countable and completely indistinguishable.* Particles are perfect copies of each other.
We will discover shortly that permutation is partial rotation. Permutation symmetry
Challenge 87 e thus is a symmetry under partial rotations. Can you find out why?
Indistinguishability and symmetry

The indistinguishability of particles leads to important conclusions about the description
of their state of motion. This happens because it is impossible to formulate a description
of motion that includes indistinguishability right from the start. (Are you able to con-
Challenge 88 s firm this?) We need to describe a n-particle state with a state Ψ1...i... j...n which assumes
that distinction is possible, as expressed by the ordered indices in the notation, and we
introduce the indistinguishability afterwards.
Indistinguishability means that the exchange of any two particles results in the same
physical system.** Now, two quantum states have the same physical properties if they
differ at most by a phase factor; indistinguishability thus requires
Ψ1...i... j...n = eiα Ψ1... j...i...n (54)
for some unknown angle α. Applying this expression twice, by exchanging the same cou-
ple of indices again, allows us to conclude that e2iα = 1. This implies that
Ψ1...i... j...n = ± Ψ1... j...i...n , (55)

in other words, a wave function is either symmetric or antisymmetric under exchange of

indices. (One can also say that the eigenvalue for the exchange operator is either +1 or
−1.) Quantum theory thus predicts that particles can be indistinguishable in one of two
distinct ways.*** Particles corresponding to symmetric wave functions – those which
* The word ‘indistinguishable’ is so long that many physicists sloppily speak of ‘identical’ particles neverthe-
less. Take care.
** We therefore have the same situation that we encountered already several times: an overspecification of
the mathematical description, here the explicit ordering of the indices, implies a symmetry of this description,
which in our case is a symmetry under exchange of indices, i.e., under exchange of particles.
*** This conclusion applies to three-dimensional space only. In two dimensions there are more possibilities.
detectors
mirrors
beam
source splitter
possible
two identical light F I G U R E 52 Two-photon emission
photons paths and interference: both photons are
always found arriving together, at
the same detector.
transform with a ‘+’ in equation (55) – are called bosons, those corresponding to anti-
symmetric wave functions – those which transform with a ‘−’ in equation (55) – are
called fermions.*
Experiments show that the exchange behaviour depends on the type of particle. Pho-
tons are found to be bosons. On the other hand, electrons, protons and neutrons are
found to be fermions. Also about half of the atoms are found to behave as bosons (at
moderate energies), the other half are fermions. In fact, a composite of an even number
of fermions (at moderate energies) – or of any number of bosons (at any energy) – turns
out to be a boson; a composite of an odd number of fermions is (always) a fermion.
For example, almost all of the known molecules are bosons (electronically speaking).
Fermionic molecules are rather special and even have a special name in chemistry; they
are called radicals and are known for their eagerness to react and to form normal bosonic
molecules. Inside the human body, too many radicals can have adverse effects on health;
it is well known that vitamin C is important because it is effective in reducing the number
of radicals.
To which class of particles do mountains, trees, people and all other macroscopic ob-
Challenge 89 s jects belong?
The behaviour of photons

A simple experiment allows to determine the behaviour of photons. Take a source that
emits two photons of identical frequency and polarization at the same time, as shown
in Figure 52. In the laboratory, such a source can be realized with a down-converter, a
material that converts a photon of frequency 2ω into two photons of frequency ω. Both
photons, after having travelled exactly the same distance, are made to enter the two sides
of a beam splitter (for example, a half-silvered mirror). At the two exits of the beam split-
Ref. 66 ter are two detectors. Experiments show that both photons are always detected together
* The term ‘fermion’ is derived from the name of the Italian physicist and Nobel Prize winner Enrico Fermi
(b. 1901 Roma, d. 1954 Chicago) famous for his all-encompassing genius in theoretical and experimental
physics. He mainly worked on nuclear and elementary particle physics, on spin and on statistics. For his
experimental work he was called ‘quantum engineer’. He is also famous for his lectures, which are still pub-
lished in his own hand-writing, and his brilliant approach to physical problems. Nevertheless, his highly
deserved Nobel Prize was one of the few cases in which the prize was given for a discovery which turned
out to be incorrect.
‘Bosons’ are named after the Indian physicist Satyenra Nath Bose (b. 1894 Calcutta, d. 1974 Calcutta)
Ref. 65 who first described the statistical properties of photons. The work was later expanded by Albert Einstein.
4He shows bunching
classical
prediction
classical
prediction
3He shows anti-bunching
F I G U R E 53 Bunching and antibunching of 3 He and 4 He helium atoms: the measurement result, the
detector and the experiment (from atomoptic.iota.u-psud.fr/research/helium/helium.html, photo
© Denis Boiron, Jerome Chatin)
on the same side, and never separately on opposite sides. This result shows that photons
are bosons. Fermions behave in exactly the opposite way; two fermions are always de-
tected separately on opposite sides, never together on the same side.
Bunching and antibunching

A beautiful way to test the exchange character of a particle is the Hanbury Brown–Twiss
Page 45 experiment described earlier on. First of all, the experiment shows that quantum parti-
cles behave differently than classical particles. In addition, compared to classical particles,
fermions show antibunching – because of Pauli’s exclusion principle – and bosons show
bunching. Hanbury Brown and Twiss performed the experiment with photons. In 2005,
Ref. 67 a French–Dutch research collaboration performed the experiment with atoms. By using
an extremely cold helium gas at 500 nK and a clever detector principle, they were able
to measure the correlation curves typical for the effect. The result, shown in Figure 53
shows that, as predicted by quantum theory, 3 He is a fermion and 4 He is a boson.
The energy dependence of permutation symmetry

If experiments force us to conclude that nobody, not even nature, can distinguish any
two particles of the same type, we deduce that they do not form two separate entities, but
some sort of unity. Our naive, classical sense of particle as a separate entity from the rest
of the world is thus an incorrect description of the phenomenon of ‘particle’. Indeed, no
experiment can track particles with identical intrinsic properties in such a way that they
can be distinguished with certainty. This impossibility has been checked experimentally
F I G U R E 54 Picturing particles as localized excitations (left) or clouds (right)
with all elementary particles, with nuclei, with atoms and with numerous molecules.
How does this fit with everyday life, i.e., with classical physics? Photons do not worry
us much here. Let us focus the discussion on matter particles. We know to be able to
distinguish electrons by pointing to the wire in which they flow, and we can distinguish
our fridge from that of our neighbour. While the quantum of action makes distinction
impossible, everyday life allows it. The simplest explanation is to imagine a microscopic
particle, especially an elementary one, as a bulge, i.e., as a localized excitation of the
vacuum, or as a tiny cloud. Figure 54 shows two such bulges and two clouds representing
particles. It is evident that if particles are too near to each other, it makes no sense to
distinguish them; we cannot say any more which is which.
The bulge image shows that either for large distances or for high potential walls sep-
arating them, distinction of identical particles does become possible. In such situations,
measurements allowing to track them independently do exist. In other words, we can
specify a limit energy at which permutation symmetry of objects or particles separated
by a distance d becomes important. It is given by
cħ
E= . (56)
d
Challenge 90 ny Are you able to confirm the expression? For example, at everyday temperatures we can
distinguish atoms inside a solid from each other, since the energy so calculated is much
higher than the thermal energy of atoms. To have fun, you might want to determine at
Challenge 91 e what energy two truly identical human twins become indistinguishable. Estimating at
what energies the statistical character of trees or fridges will become apparent is then
straightforward.
The bulge image of particles thus purveys the idea that distinguishability exists for
objects in everyday life but not for particles in the microscopic domain. To sum up, in
daily life we are able to distinguish objects and thus people for two reasons: because they
are made of many parts, and because we live in a low energy environment.
The energy issue immediately adds a new aspect to the discussion. How can we de-
scribe fermions and bosons in the presence of virtual particles and of antiparticles?
Indistinguishability in quantum field theory

Quantum field theory, as we will see shortly, simply puts the cloudy bulge idea of
Figure 54 into mathematical language. A situation with no bulge is called vacuum state.
Quantum field theory describes all particles of a given type as excitations of a single fun-
damental field. Particles are indistinguishable because each particle is an excitation of the
same basic substrate and each excitation has the same properties. A situation with one
particle is then described by a vacuum state acted upon by a creation operator. Adding
a second particle is described by adding a second creation operator, and subtracting a
particle by adding a annihilation operator; the latter turns out to be the adjoint of the
former.
Quantum field theory studies how creation and annihilation operators must behave
to describe observations.* It arrives at the following conclusions:
— Fields with half-integer spin are fermions and imply (local) anticommutation.
— Fields with integer spin are bosons and imply (local) commutation.
— For all fields at space-like separations, the commutator, respectively anticommutator,
vanishes.
— Antiparticles of fermions are fermions, and antiparticles of bosons are bosons.
— Virtual particles behave under exchange like their real counterparts.
These connections are at the basis of quantum field theory. They describe how particles
are identical. But why are they? Why are all electrons identical? Quantum field theory
describes electrons as identical excitations of the vacuum, and as such as identical by
construction. Of course, this answer is only partially satisfying. We will find a better one
only in the final part of our mountain ascent.
How accurately is permutation symmetry verified?

A simple but effective experiment testing the fermion behaviour of electrons was carried
Ref. 68 out by Ramberg and Snow. They sent an electric current of 30 A through a copper wire
for one month and looked for X-ray emission. They did not find any. They concluded
that electrons are always in an antisymmetric state, with a symmetric component of less
than
2 ⋅ 10−26 (59)
of the total state. In short, electrons are always in an antisymmetric state: they are
fermions.
The reasoning behind this elegant experiment is the following. If electrons would not
always be fermions, every now and then an electron could fall into the lowest energy
level of a copper atom, leading to X-ray emission. The lack of such X-rays implies that
electrons are fermions to a very high accuracy. X-rays could be emitted only if they were
bosons, at least part of the time. Indeed, two electrons, being fermions, cannot be in the
* Whenever the relation

[b, b† ] = bb† − b† b = 1 (57)
†
holds between the creation operator b and the annihilation operator b, the operators describe a boson. The
dagger can thus be seen as describing the operation of adjoining; a double dagger is equivalent to no dagger.
If the operators for particle creation and annihilation anticommute
{d, d † } = d d † + d † d = 1 (58)
they describe a fermion. The so defined bracket is called the anticommutator bracket.
same quantum state: this restriction is called the Pauli exclusion principle. It applies to
all fermions and is our next topic.
Copies, clones and gloves

Can classical systems be indistinguishable? They can: large molecules are examples –
provided they are made of exactly the same isotopes. Can large classical systems, made
of a mole or more particles be indistinguishable? This simple question effectively asks
whether a perfect copy, or (physical) clone of a system is possible.
It could be argued that any factory for mass-produced goods, such as one producing
shirt buttons or paper clips, shows that copies are possible. But the appearance is deceiv-
ing. On a microscope there is usually some difference. Is this always the case? In 1982, the
Dutch physicist Dennis Dieks and independently, the US-American physicists Wootters
Ref. 69 and Zurek, published simple proofs that quantum systems cannot be copied. This is the
famous no-cloning theorem.
A copying machine is a machine that takes an original, reads out its properties and
produces a copy, leaving the original unchanged. This seems definition seems straight-
forward. However, we know that if we extract information from an original, we have to
interact with it. As a result, the system will change at least by the quantum of action. We
thus expect that due to quantum theory, copies and originals can never be identical.*
Quantum theory proves this in detail. A copying machine is described by an operator
that maps the state of an original system to the state of the copy. In other words, a copying
machine is linear. This linearity leads to a problem. Simply stated, if a copying machine
were able to copy originals either in state |A⟩ or in state |B⟩, it could not decide what to do
if the state of the original were |A⟩+|B⟩. On one hand, the copy should be |A⟩+|B⟩; on the
other hand, the linearity of the copier forbids this. Indeed, a copier is a device described
by an operator U that changes the starting state |s⟩c of the copy in the following way:
— If the original is in state |A⟩, a copier acts as
U|A⟩|s⟩c = |A⟩|A⟩c . (60)
— If the original is in state |B⟩, a copier acts as
U|B⟩|s⟩c = |B⟩|B⟩c . (61)

As a result of these two requirements, an original in the state |A + B⟩ is treated by the

copier as
U|A + B⟩|s⟩c = |A⟩|A⟩c + |B⟩|B⟩c . (62)
This is in contrast to what we want, which would be
Uwanted |A + B⟩|s⟩c = (|A⟩ + |B⟩)(|A⟩c + |B⟩c ) . (63)
* This seems to provide a solution against banknote forgeries. In fact, Steve Wiesner proposed to use quan-
tum theory already in 1970; he imagined to use polarizations of stored single photons as bits of serial num-
Challenge 92 ny bers. Can you explain why this cannot work?
In other words, a copy machine cannot copy a state completely.* This is the no-cloning
theorem.
The impossibility of copying is implicit in quantum theory. If we were able to clone
systems, we could to measure a variable of a system and a second variable on its copy.
We would be thus able to beat the indeterminacy relation. This is impossible. Copies are
and always must be imperfect.
Other researchers then explored how near to perfection a copy can be, especially in the
Ref. 70 case of classical systems. To make a long story short, these investigations show that also
the copying or cloning of macroscopic systems is impossible. In simple words, copying
machines do not exist. Copies can always be distinguished from originals if observations
are made with sufficient care. In particular, this is the case for biological clones; biological
clones are identical twins born following separate pregnancies. They differ in their finger
prints, iris scans, physical and emotional memories, brain structures, and in many other
Challenge 93 s aspects. (Can you specify a few more?) In short, biological clones, like identical twins,
are not copies of each other.
The lack of quantum mechanical copying machines is disappointing. Such machines,
or teleportation devices, could be fed with two different inputs, such as a lion and a goat,
and produce a superposition: a chimaera. Quantum theory shows that all these imagi-
nary beings cannot be realized.
In summary, everyday life objects such as photocopies, billiard balls or twins are al-
ways distinguishable. There are two reasons: first, quantum effects play no role in every-
day life, so that there is no danger of unobservable exchange; secondly, perfect clones of
classical systems do not exist anyway, so that there always are tiny differences between
any two objects, even if they look identical at first sight. Gloves, being classical systems,
can thus always be distinguished.
* The no-cloning theorem puts severe limitations on quantum computers, as computations often need
copies of intermediate results. It also shows that faster-than-light communication is impossible in EPR exper-
iments. In compensation, quantum cryptography becomes possible – at least in the laboratory. Indeed, the
no-cloning theorem shows that nobody can copy a quantum message without being noticed. The specific
ways to use this result in cryptography are the 1984 Bennett–Brassard protocol and the 1991 Ekert protocol.
Chapter 6
ROTATIONS AND STATISTICS

– V I SUALI Z I NG SPI N
S
pin is the observation that matter beams can be polarized: rays can be rotated.
pin thus describes how particles behave under rotations. Particles are thus not
imple spheres shrunk to points. We also saw that spin describes a fundamental
difference between quantum particles and gloves: spin specifies the indistinguishability
of quantum particles and quantum systems in general. We now explore this connection
in more detail.
Quantum particles and symmetry

Ref. 71 The general background for the appearance of spin was clarified by Eugene Wigner in
1939.* He started by recapitulating that any quantum particle, if elementary, must be-
have like an irreducible representation of the set of all viewpoint changes. This set of view-
point changes forms the symmetry group of flat space-time, the so-called inhomogeneous
Lorentz group. Why?
Vol. I, page 214 We have seen in the chapter on symmetry that the symmetry of any composite system
leads to certain requirements for its components. If the components do not follow these
requirements, they cannot build a symmetric composite. We know from everyday life
and precision experiments that all physical systems are symmetric under translation in
time and space, under rotation in space, under boosts, and – in many cases – under
mirror reflection, matter–antimatter exchange and motion reversal.
We know these symmetries known from everyday life; for example, the usefulness
of what we call ‘experience’ in everyday life is simply a consequence of time translation
symmetry. The set of all these common symmetries, more precisely, of all these symmetry
transformations, is called the inhomogeneous Lorentz group.
These symmetries, i.e., these changes of viewpoints, lead to certain requirements for
the components of physical systems, i.e., for the elementary quantum particles. In math-
ematical language, the requirement is expressed by saying that elementary particles must
be irreducible representations of the symmetry group.
Following Wigner, every textbook on quantum theory carries out this reasoning sys-
tematically. One obtains a list of all possible irreducible representations, in other words, a
list of all possible ways that elementary particles can behave. ** Cataloguing the possibil-
* Eugene Wigner (b. 1902 Budapest, d. 1995 Princeton), Hungarian–US-American theoretical physicist, re-
ceived the Nobel Prize for physics in 1963. He wrote over 500 papers, many about various aspects of sym-
metry in nature. He was also famous for being the most polite physicist in the world.
** To be of physical relevance for quantum theory, representations have to be unitary. The full list of irre-
106 rotations and statistics – visualizing spin
ities, one finds first of all that every elementary particle is described by four-momentum
– no news so far – and by an internal angular momentum, the spin. Four-momentum re-
sults from the translation symmetry of nature, and spin from its rotation symmetry. The
momentum value describes how a particle behaves under translation, i.e., under pos-
ition and time shift of viewpoints. The spin value describes how an object behaves under
rotations in three dimensions, i.e., under orientation change of viewpoints.* As is well
known, the magnitude of four-momentum is an invariant property, given by the mass,
whereas its orientation in space-time is free. Similarly, the magnitude of spin is an invari-
ant property, and its orientation has various possibilities with respect to the direction
of motion. In particular, the spin of massive quantum particles behaves differently from
that of massless quantum particles.
For massive quantum particles, the inhomogeneous Lorentz group implies that the in-
variant magnitude of spin is 󵀄 J(J + 1) ħ, often written, by oversimplification, as J. Since
the value J specifies the magnitude of the angular momentum, it gives the representation
under rotations of a given particle type. The exploration shows that the spin magnitude
J can be any multiple of 1/2, i.e., it can take the values 0, 1/2, 1, 3/2, 2, 5/2, etc. Experi-
ments show that electrons, protons and neutrons have spin 1/2, the W and Z particles
spin 1 and helium atoms spin 0. In addition, the representation of spin J is 2J + 1 di-
mensional, meaning that the spatial orientation of the spin has 2J + 1 possible values.
For electrons, with J = 1/2, there are thus two possibilities; they are usually called ‘up’
and ‘down’. Spin thus only takes discrete values. This is in contrast with linear momen-
tum, whose representations are infinite dimensional and whose possible values form a
continuous range.
Also massless quantum particles are characterized by the value of their spin. It can
take the same values as in the massive case. For example, photons and gluons have spin
1. For massless particles, the representations are one-dimensional, so that massless parti-
cles are completely described by their helicity, defined as the projection of the spin onto
the direction of motion. Massless particles can have positive or negative helicity, often
also called right-handed and left-handed polarization. There is no other freedom for the
orientation of spin in the massless case.
To complete the list of particle properties, the remaining symmetries must be in-
cluded. Since motion inversion, spatial parity and charge inversion are parities, each el-
ementary particle has to be described by three additional numbers, called T, P and C,
each of which can only take the values +1 or −1. Being parities, these numbers must be
multiplied to yield the value for a composed system.
In short, symmetry investigations lead to the classification of all particles by their

mass, their momentum, their spin and their P, C and T parities.
Types of quantum particles

The spin values observed for all quantum particles in nature is given in Table 4. The
parities and all known intrinsic properties of the elementary particles are given in Table 5.
ducible unitary representations of viewpoint changes thus provides the range of possibilities for any particle
that wants to be elementary.
* The group of physical rotations is also called SO(3), since mathematically it is described by the group of
Special Orthogonal 3 by 3 matrices.
rotations and statistics – visualizing spin 107
F I G U R E 55 An argument showing why

rotations by 4π are equivalent to no
rotation at all
TA B L E 4 Particle spin as representation of the rotation group
Spin System Massive examples Massless examples
[ħ] unchanged after elementary composite elementary
rotation by
0 any angle none a,b mesons, nuclei, none b
atoms
1/2 2 turns e, μ, τ, q, nuclei, atoms, none, as neutrinos have a tiny mass
󰜈e , 󰜈 μ , 󰜈τ molecules
1 1 turn W, Z mesons, nuclei, д, γ
atoms, molecules,
toasters
3/2 2/3 turn none b baryons, nuclei, none b
atoms
c
2 1/2 turn none nuclei ‘graviton’
5/2 2/5 turn none nuclei none
d
3 1/3 turn none nuclei none
d d d d
etc. etc. etc. etc. etc. d
a. Whether the Higgs boson exists, and whether it is is elementary is still unknown.
b. Supersymmetry predicts particles in these and other boxes.

c. The graviton has not yet been observed.
d. Nuclei exist with spins values up to at least 101/2 and 51 (in units of ħ). Ref. 72
Spin and parities together are called quantum numbers. The other intrinsic properties are
related to interactions, such as electric charge or isospin. We will explore them in the next
volume. But let us return to spin.
TA B L E 5 Elementary particle properties
Particle Mass m a L ifetime τ Isospin I, Charge, L epton

or energy spin J, c isospin, &
width, b parit y P, strange- baryon e
main decay charge ness, c num-
modes parit y C charm, bers
beau t y, LB
topness:
QISCBT
Elementary radiation (bosons)

photon γ 0 (<10−53 kg) stable I(J PC ) = 000000 0, 0
0, 1(1−− )
W± 80.398(25) GeV/c 2 2.124(41) GeV J = 1 ±100000
0, 0
67.60(27)% hadrons,
32.12(36)% l + 󰜈
Z 91.1876(21) GeV/c 2 2.65(2) ⋅ 10−25 s J = 1 000000 0, 0
or 2.4952(23) GeV/c 2
69.91(6)% hadrons,
10.0974(69)% l + l −
gluon 0 stable I(J P ) = 0(1− ) 000000 0, 0
Elementary matter (fermions): leptons
electron e 9.109 382 15(45) ⋅ > 13 ⋅ 1030 s J = 12 −100 000 1, 0
−31
10 kg = 81.871 0438(41) pJ/c 2
= 0.510 998 910(13) MeV/c 2 = 0.000 548 579 909 43(23) u

gyromagnetic ratio μe /μB = −1.001 159 652 1811(7)
electric dipole moment d = (0.7 ± 0.7) ⋅ 10−29 e m f
muon μ 0.188 353 130(11) yg 2.197 03(4) μs J = 12 −100000 1, 0
99 % e − 󰜈̄e 󰜈 μ
= 105.658 3668(38) MeV/c 2 = 0.113 428 9256(29) u
gyromagnetic ratio μ μ /(eħ/2m μ ) = −1.001 165 9208(6)
electric dipole moment d = (3.7 ± 3.4) ⋅ 10−21 e m
1
tau τ 1.776 84(17) GeV/c 2 290.6(1.0) fs J= 2
−100000 1, 0
1
el. neutrino < 2 eV/c 2 J= 2
1, 0
󰜈e
1
muon < 2 eV/c 2 J= 2
1, 0
neutrino 󰜈 μ
1
tau neutrino < 2 eV/c 2 J= 2
1, 0
󰜈τ
Elementary matter (fermions): quarks д
+
up u 1.5 to 3.3 MeV/c 2 see proton I(J P ) = 12 ( 12 ) + 23 + 12 0000 0, 1
3
TA B L E 5 (Continued) Elementary particle properties
Particle Mass m a L ifetime τ Isospin I, Charge, L epton

or energy spin J, c isospin, &
width, b parit y P, strange- baryon e
main decay charge ness, c num-
modes parit y C charm, bers
beau t y, LB
topness:
QISCBT
+
down d 3.5 to 6 MeV/c 2 see proton I(J P ) = 12 ( 12 ) − 13 − 12 0000 0, 1
3
+
strange s 70 to 130 MeV/c 2 I(J P ) = 0( 12 ) − 13 0−1000 0, 1
3
+
charm c 1.27(11) GeV/c 2 I(J P ) = 0( 12 ) + 23 00+100 0, 1
3
+
bottom b 4.20(17) GeV/c 2 τ = 1.33(11) ps I(J P ) = 0( 12 ) − 13 000−10 0, 1
3
+
top t 171.2(2.1) GeV/c 2 I(J P ) = 0( 12 ) + 23 0000+1 0, 1
3
Hypothetical elementary matter (boson)

Higgs h H > 114 GeV/c 2 J=0
Notes:
a. See also the table of SI prefixes on page 176. About the eV/c 2 mass unit, see page 180.
b. The energy width Γ of a particle is related to its lifetime τ by the indeterminacy relation Γτ = ħ. There
is a difference between the half-life t1/2 and the lifetime τ of a particle: they are related by t1/2 = τ ln 2,
where ln 2 ≈ 0.693 147 18; the half-life is thus shorter than the lifetime. The unified atomic mass unit u is
defined as 1/12 of the mass of a carbon 12 atom at rest and in its ground state. One has 1 u = 121 m(12 C) =
1.660 5402(10) yg.
c. To keep the table short, the header does not explicitly mention colour, the charge of the strong interactions.
This has to be added to the list of basic object properties. Quantum numbers containing the word ‘parity’
are multiplicative; all others are additive. Time parity T (not to be confused with topness T), better called
motion inversion parity, is equal to CP. The isospin I (or IZ ) is defined only for up and down quarks and their
composites, such as the proton and the neutron. In the literature one also sees references to the so-called
G-parity, defined as G = (−1)IC .
The header also does not mention the weak charge of the particles. The details on weak charge д, or, more
precisely, on the weak isospin, a quantum number assigned to all left-handed fermions (and right-handed
anti-fermions), but to no right-handed fermion (and no left-handed antifermion), are given in the section
Page 180 on the weak interactions.
d. ‘Beauty’ is now commonly called bottomness; similarly, ‘truth’ is now commonly called topness. The signs
of the quantum numbers S, I, C, B, T can be defined in different ways. In the standard assignment shown
here, the sign of each of the non-vanishing quantum numbers is given by the sign of the charge of the
corresponding quark.
e. If supersymmetry exists, R-parity must be added to this column. R-parity is a multiplicative quantum
number related to the lepton number L, the baryon number B and the spin J through the definition
R = (−1)3B+L+2J . All particles from the standard model are R-even in this case, whereas their superpart-
ners would be R-odd.
Ref. 74, Ref. 73 f . The electron radius is less than 10−22 m. It is possible to store single electrons in traps for many months.
д. See page 172 for the precise definition and meaning of the quark masses.
h. Currently a hypothetical particle.
A central result of quantum theory is that spin 1/2 is a possibility in nature, even
though this value does not appear in everyday life. For a system to have spin 1/2 means
that for such a system only a rotation of 720 degrees is equivalent to one of 0 degrees,
F I G U R E 56 Two belt buckles connected by a belt,

visualizing two spin 1/2 particles
F I G U R E 57 The belt trick: a double
rotation of the belt buckle is equivalent to
no rotation (QuickTime film © Greg Egan)
while one of 360 degrees is not. No such systems exist in everyday life, but they do exist
in microscopic systems: electrons, silver atoms and molecular radicals have spin 1/2. A
full list of spins of particles is given in Table 4.
The mathematician Hermann Weyl used a simple image explaining the connection
Page 45 between spin 1/2 and invariance under rotation by 4π. Take two cones, touching each
other at their tips as well as along a line. Hold one cone and roll the other around it, as
shown in Figure 55. When the rolling cone, after a full turn around the other cone, has
come back to the original position, it has rotated by some angle. If the cones are wide,
the rotation angle is small. If the cones are very thin, like needles, the moving cone has
rotated by (almost) 720 degrees. A rotation of 720 degrees is thus similar to one by 0
degrees. If we imagine the cone angle to vary continuously, this visualization also shows
that a 720 degree rotation can be continuously deformed into a 0 degree rotation, whereas
a 360 degree rotation cannot.
To sum up, the list of possible representations thus shows that rotations require the
existence of spin. But why then do experiments show that all fermions have half-integer
spin and that all bosons have integer spin? Why do electrons obey the Pauli exclusion
principle? At first, it is not clear what the spin has to do with the statistical properties of
a particle.
In fact, there are several ways to show that rotations and statistics are connected. The
first proof, due to Wolfgang Pauli, used the details of quantum field theory and was so
Ref. 75 complicated that its essential ingredients were hidden. It took several decades to convince
everybody that a simple observation about belts was the central part of the proof.
α=0 α = 2π α = 4π
F I G U R E 58 The human arm as spin 1/2 model
The belt trick and its extension

The well-known belt trick – also called scissor trick or plate trick – was often used by
Ref. 76 Dirac to explain the features of spin 1/2. Starting from Figure 54, which models particles
as indistinguishable excitations, it is not difficult to imagine a sort of sheet connecting
them, similar to a belt connecting two belt buckles, as shown in Figure 56. The buckles
represent the particles. If one belt buckle is rotated by 2π along any axis, a twist is inserted
into the belt. Now rotate the same buckle by another 2π, bringing the total to 4π. It turns
out that the ensuing double twist can easily be undone without moving or rotating the
buckles. The animation of Figure 57 shows the details. You may want to do this yourself,
Challenge 94 e using a real belt or a strip of paper, in order to believe it. In short, belt buckles return to
their original state only after rotations by 4π, and not after rotations by 2π.
Now look again at Figure 56. If you take the two buckles and simply swap positions, a
twist is introduced into the belt. Now swap them again: this will undo the twist. In short,
two belt buckles return to their original state only after a double exchange, and not after
a single exchange.
In other words, if we take each buckle to represent a particle and a twist to mean
a factor −1, the belt exactly describes the phase behaviour of spin 1/2 wave functions,
both under rotation and under exchange. In particular, we see that rotation and exchange
behaviour are related.
The human body has such a belt built in: the arm. Just take your hand, put an object on
it for clarity such as a cup, and turn the hand and object by 2π by twisting the arm. After
a second rotation the whole system will be untangled again. This is sometimes called the
plate trick. The trick is even more impressive when many arms are used. You can put your
Challenge 95 e two hands (if you chose the correct starting position) under the cup or you can take a
friend or two who each keep a hand attached to the cup. The feat can still be performed:
Challenge 96 e the whole system untangles after two full turns.
This leads us to the most complete way to show the connection between rotation and
exchange. Just glue any number of threads, belts or tubes, say half a metre long, to an
asymmetric object. (With many such tails, is not appropriate any more to call it a belt
buckle.) Like the arm of a human being, each band is supposed to go to infinity and be
attached there. If the object, which represents the particle, is rotated by 2π, twists ap-
pear in its tails. If the object is rotated by an additional turn, to a total of 4π, as shown
in Figure 59, all twists and tangles can be made to disappear, without moving or turn-
ing the object. You really have to experience this in order to believe it. And the trick
really works with any number of bands glued to the object. The website www.evl.uic.
edu/hypercomplex/html/dirac.html provides a beautiful animation showing this. Again
we find that an object attached to belts behaves under rotations like a spin 1/2 particle.
rotating the buckle

either by 4π
F I G U R E 59 The extended belt trick,
modelling the rotation behaviour of a spin
1/2 particle: independently of the number of
or simply rearranging bands or tubes or strings attached, the two
the bands gives the situations can be transformed into each other,
other situation either by rotating the central object by 4π or
by keeping the central object fixed and
moving the bands around it
F I G U R E 60 Extended belt
models for two spin 1/2
particles
Similarly, the belt trick can be extended to many bands also for the topic of exchange.
Take two buckles that are connected with many bands or threads, like in Figure 60. An
exchange of the two buckles produces quite a tangle, even if one takes paths that go ‘be-
tween’ the bands; nevertheless, in both cases a second exchange leads back to the original
situation.
But this is not all. Take two particles with any number of tails, as shown on the right
side of Figure 60. You can also add belts going from one to the other particle. If you ex-
change the positions of two such spin 1/2 particles, always keeping the ends at infinity
fixed, a tangled mess is created. But incredibly, if you exchange the two objects a second
Challenge 97 e time, everything untangles neatly, independently of the number of attached strings. You
might want to test yourself that the behaviour is also valid if additional particles are in-
volved, as long as you always exchange the same two particles twice. Unfortunately, no
animation or video showing this is yet available on the internet. In any case, we conclude
that objects attached to belts behave like a spin 1/2 particle also under exchange.
All these observations together form the spin–statistics theorem for spin 1/2 particles:
spin and exchange behaviour are related. Indeed, these almost ‘experimental’ arguments
Ref. 77 can be put into exact mathematical language by studying the behaviour of the configura-
tion space of particles. These investigations result in the following statements:
⊳ Objects of spin 1/2 are fermions.*

⊳ Exchange and rotation of spin 1/2 particles are similar processes.
In short, objects that behave like spin 1/2 particles under rotations also behave like spin
1/2 particles under exchange. And vice versa. The exchange behaviour of particles deter-
mines their statistical properties; the rotation behaviour determines their spin. By ex-
tending the belt trick to several buckles, each with several belts, we thus visualized the
spin–statistics theorem for fermions.
Note that all these arguments require three dimensions, because there are no tangles
(or knots) in fewer or more dimensions.** And indeed, spin exists only in three spatial
dimensions.
Here is a challenge. A spin 1/2 object can be modelled with one belt attached to it. If
you want to model the spin behaviour with attached one-dimensional strings instead of
Challenge 99 s bands, what is the minimum number of strings required?
Angels, Pauli’s exclusion principle and the hardness of matter

Why are we able to knock on a door? Why can stones not fly through tree trunks? How
does the mountain we are walking on carry us? Why can’t we walk across walls? In classi-
cal physics, we avoided this issue, by taking solidity as a defining property of matter. But
doing so, we cheated: we have seen that matter consists mainly of low density electron
clouds. Thus we have to understand the issue without any sneaky way out. The answer
is famous: interpenetration of bodies is made impossible by Pauli’s exclusion principle
among the electrons inside atoms. The exclusion principle states:
⊳ Two fermions cannot occupy the same quantum state.
All experiments known confirm the statement.

Why do electrons and other fermions obey the Pauli exclusion principle? The answer
Ref. 78 can be given with a beautifully simple argument. We know that exchanging two fermions
produces a minus sign in the total wave function. Imagine these two fermions being, as
a classical physicist would say, located at the same spot, or as a quantum physicist would
say, in the same state. If that could be possible, an exchange would change nothing in the
system. But an exchange of fermions must produce a minus sign for the total state. Both
possibilities – no change at all as well as a minus sign – cannot be realized at the same
time. There is only one way out: two fermions must avoid to ever be in the same state.
This is Pauli’s exclusion principle.
* A mathematical observable behaving like a spin 1/2 particle is neither a vector nor a tensor, as you may
Challenge 98 e want to check. An additional concept is necessary; such an observable is called a spinor. We will introduce
Page 160 it in detail later on.
** Of course, knots and tangles do exist in higher dimensions. Instead of considering knotted one-
dimensional lines, one can consider knotted planes or knotted higher-dimensional hyperplanes. For ex-
ample, deformable planes can be knotted in four dimensions and deformable 3-spaces in five dimensions.
However, the effective dimensions that produce the knot are always three.
The exclusion principle is the reason that two pieces of matter in everyday life cannot
penetrate each other, but have to repel each other. For example, bells only work because
of the exclusion principle. A bell would not work if the colliding pieces that produce the
sound would interpenetrate. But in any example of two interpenetrating pieces electrons
from two atoms would have to be at the same spot: they would have to be in the same
states. This is forbidden. Pauli’s exclusion principle forbids interpenetration of matter.
Why don’t we fall through the floor, even though gravity pulls us down, but remain
standing on its surface? Again, the reason is Pauli’s exclusion principle. Why does the
floor not fall? It does not fall, because the matter of the Earth cannot be compressed
further. Why? Pauli’s exclusion principle does not allow atoms to be compressed. In other
words, the exclusion principle implies that matter cannot be compressed indefinitely, as
at a certain stage an effective Pauli pressure appears, so that a compression limit ensues.
For this reason for example, planets or neutron stars do not collapse under their own
gravity.
The exclusion principle is the reason that atoms are extended electron clouds. In fact,
the exclusion principle forces the electrons to form shells, and when one shell is filled, a
next one is started. The size of the atom is the size of the last shell. Without the exclusion
principle, atoms would be point-like. The same argument applies to the protons (and the
neutrons) in nuclei.
The exclusion principle also answers the question about how many angels can dance
on the top of a pin. (Note that angels, if at all, must be made of fermions, as you might
Challenge 100 s want to deduce from the information known about them.) Both theory and experiment
Ref. 79 confirm the answer already given by Thomas Aquinas in the Middle Ages: Only one! The
fermion exclusion principle could also be called ‘angel exclusion principle’. To stay in the
topic, the principle also shows that ghosts cannot be objects, as ghosts are supposed to
be able to traverse walls.
Whatever the interpretation, the exclusion principle keeps things in shape; without
it, there would be no three-dimensional objects. Only the exclusion principle keeps the
cloudy atoms of nature from merging, holding them apart. This repulsion keeps the size
of planets to a finite value, and that of neutron stars. All shapes of solids and fluids are a
direct consequence of the exclusion principle. In other words, when we knock on a table
or on a door, we prove experimentally that both objects are made of fermions.
Since permutation properties and spin properties of fermions are so well described
by the belt model, we could be led to the conclusion that these properties might really
be consequence of such belt-like connections between particles and the outside world.
Maybe for some reason we only observe the belt buckles, not the belts themselves. In the
final part of this walk we will discover whether this idea is correct.
So far, we have only considered spin 1/2 particles. We will not talk much about sys-
tems with odd spin of higher value, such as 3/2 or 5/2. Such systems can all be seen as
Challenge 101 ny being composed of spin 1/2 entities. Can you confirm this?
We did not talk about lower spins than 1/2 either. A famous theorem states that spin
value between 0 and 1/2 is impossible, because the largest angle that can be measured in
three dimensions is 4π.
F I G U R E 61 Some visualizations of spin

J=0 J = 1/2 J=1 representations
There is no way to measure a larger angle;* the quantum of action makes this impos-
sible. Thus there cannot be any spin value between 0 and 1/2 in nature.
Spin, statistics and composition
Under rotations, integer spin particles behave differently from half-integer particles. Inte-
ger spin particles do not show the strange sign changes under rotations by 2π. In the belt
imagery, integer spin particles need no attached strings. In particular, a spin 0 particle
obviously corresponds to a sphere. Models for other spin values are shown in Figure 61.
Exploring their properties in the same way as above, we arrive at the full spin–statistics
theorem:
⊳ Exchange and rotation of objects are similar processes.

⊳ Objects of half-integer spin are fermions. They obey the Pauli exclusion prin-
ciple.
⊳ Objects of integer spin are bosons.
Challenge 102 ny You might prove by yourself that this suffices to show the following:
⊳ Composites of bosons, as well as composites of an even number of fermions

(at low energy), are bosons; composites of an uneven number of fermions are
fermions.**
These connections express basic characteristics of the three-dimensional world in which

we live.
Is spin a rotation about an axis?

The spin of a particle behaves experimentally like an intrinsic angular momentum, adds
up like angular momentum, is conserved as part of angular momentum, is described
* This statement, like all statements about spin 1/2, is tied to the three-dimensionality of space. In two
dimensions, other largest angles and other ‘spin’ values are possible.
** This sentence implies that spin 1 and higher can also be achieved with tails; can you find such a represen-
Challenge 103 ny tation?
Note that composite fermions can be bosons only up to that energy at which the composition breaks
down. Otherwise, by packing fermions into bosons, we could have fermions in the same state.
like angular momentum and has a name synonymous with angular momentum. Despite
all this, for many decades a strange myth was spread in many physics courses and text-
books around the world, namely that spin 1/2 is not a rotation about an axis. The myth
maintains that any rotating object must have integer spin. Since half integer spin is not
possible in classical physics, it is argued that such spin is not due to rotation. It is time to
finish with this example of incorrect thinking.
Electrons do have spin 1/2 and are charged. Electrons and all other charged particles
with spin 1/2 do have a magnetic moment.* A magnetic moment is expected for any
rotating charge. In other words, spin 1/2 does behave like rotation. However, assuming
that a particle consists of a continuous charge distribution in rotational motion gives the
wrong value for the magnetic moment. In the early days of the twentieth century, when
physicists were still thinking in classical terms, they concluded that spin 1/2 particles
thus cannot be rotating. This myth has survived through many textbooks. The correct
deduction, however, is that the assumption of continuous charge distribution is wrong.
Indeed, charge is quantized; nobody today expects that elementary charge is continu-
ously spread over space, as that would contradict its quantization.
Let us recall what rotation is. Both the belt trick for spin 1/2 as well as the integer
spin case remind us: a rotation of one body around another is a fraction or a multiple
of an exchange. What we call a rotating body in everyday life is a body continuously
exchanging the positions of its parts. Rotation and exchange are the same.
Now, we just found that spin is exchange behaviour. Since rotation is exchange and
spin is exchange, it follows that spin is rotation. Since we deduced, like Wigner, spin from
rotation invariance, this consequence is not a surprise.
The belt model of a spin 1/2 particle tells us that such a particle can rotate continuously
Page 110 without any hindrance. In short, we are allowed to maintain that spin is rotation about an
axis, without any contradiction to observations, even for spin 1/2. The belt model helps
us to keep two things in mind: we must assume that in the belt model only the buckle
Ref. 80 can be observed and does interact, not the belt(s), and we must assume that elementary
charge is not continuously distributed in space.**
Why is fencing with laser beams impossible?

When a sword is approaching dangerously, we can stop it with a second sword. Many old
films use such scenes. When a laser beam is approaching, it is impossible to fend it off
with a second beam, despite all science fiction films showing so. Banging two laser beams
against each other is impossible. The above explanation of the spin–statistics theorem
shows why.
The electrons in the swords are fermions and obey the Pauli exclusion principle.
Fermions make matter impenetrable. On the other hand, the photons in laser beams
are bosons. Two bosons can be in the same state; bosons allow interpenetration. Matter
is impenetrable because at the fundamental level it is composed of fermions. Radiation is
Challenge 104 ny * This can easily be measured in a an experiment; however, not one of the Stern–Gerlach type. Why not?
** Obviously, the detailed structure of the electron still remains unclear at this point. Any angular momen-
tum S is given classically by S = Θω; however, neither the moment of inertia Θ, connected to the rotation
radius and electron mass, nor the angular velocity ω are known at this point. We have to wait quite a while,
until the final part of our adventure, to find out more.
F I G U R E 62 Equivalence of exchange and rotation in

space-time
composed of bosons; light beams can cross each other. The distinction between fermions
and bosons thus explains why objects can be touched while images cannot. In the first
Page 86 part of our mountain ascent we started by noting this difference; now we know its origin.
Rotation requires antiparticles

The connection between rotation and antiparticles may be the most astonishing conclu-
sion from the experiments showing the existence of spin. So far, we have seen that rota-
tion requires the existence of spin, that spin appears when relativity is introduced into
Vol. II, page 65 quantum theory, and that relativity requires antimatter. Taking these three statements to-
gether, the conclusion of the title is not surprising any more: rotation requires antiparti-
cles. Interestingly, there is a simple argument making the same point with the belt model,
if it is extended from space alone to full space-time.
To learn how to think in space-time, let us take a particle spin 1, i.e., a particle look-
ing like a detached belt buckle in three dimensions. When moving in a 2+1 dimensional
Challenge 105 ny space-time, it is described by a ribbon. Playing around with ribbons in space-time, in-
stead of belts in space, provides many interesting conclusions. For example, Figure 62
shows that wrapping a rubber ribbon around the fingers can show that a rotation of a
body by 2π in presence of a second one is the same as exchanging the positions of the
two bodies.* Both sides of the hand transform the same initial condition, at one edge
of the hand, to the same final condition at the other edge. We have thus successfully ex-
tended a known result from space to space-time: rotation and exchange are equivalent.
If you think that Figure 62 is not a satisfying explanation, you are right. A more sat-
isfying explanation must include a smooth sequence of steps realizing the equivalence
between rotation and exchange. This is shown in Figure 63. We assume that each parti-
cle is described by a segment; in the figure, the two segments lie horizontally. The leftmost
diagram shows two particles: one at rest and one being rotated by 2π. The deformation
of the ribbons shows that this process is equivalent to the exchange in position of two
particles, which is shown in the rightmost diagram.
But the main point is made by the intermediate diagrams. One notes that the sequence
that shows the equivalence between rotation and exchange requires the use of a loop.
* Obviously, the full argument would need to check the full spin 1/2 model of Figure 59 in four-dimensional
Challenge 106 ny space-time. But this is not an easy task; there is no good visualization yet.
t t t t t
x x x x x
F I G U R E 63 Belts in space-time: rotation and antiparticles
But such a loop is the appearance of a particle–antiparticle pair. In other words, without
antiparticles, the equivalence of rotation and exchange would not hold. Rotation in space-
time requires the existence of antiparticles.
A summary on spin and indistinguishability

The quantum of action implies that physical systems are made of two types of indistin-
guishable particles: bosons and fermions. The two possible exchange behaviours are re-
lated to their spin value, because exchange is related to rotation.
Experiments show that radiation is made of elementary particles that behave as
bosons. They have integer spin. Two or more bosons, such as two photons, can share
the same state. This sharing makes laser light possible.
Experiments show that matter is made of elementary particles that behave as fermions.
They have half-integer spin. They obey Pauli’s exclusion principle: two fermions cannot
be in the same state. The exclusion principle between electrons explains the structure and
(partly) the size of atoms, as well as the chemical behaviour of atoms, as we will find out
later on. Together with the electrostatic repulsion of electrons, the exclusion principle
explains the incompressibility of matter and its lack of impenetrability.
Despite the indistinguishability of particles, the quantum of action also implies that
exact copies of macroscopic objects cannot be produced, because quantum states for
macroscopic objects cannot be copied.
Finally, the connection between spin and rotation implies that antiparticles exist. It
also implies that spin is intrinsically a three-dimensional phenomenon.
Challenge 107 e Can you summarize why matter is hard, but radiation is not?
Limits and open questions of quantum statistics

The topic of statistics is an important research field in theoretical and experimental
physics. In particular, researchers have searched and still are searching for generaliza-
tions of the possible exchange behaviours of particles.
In two spatial dimensions, the result of an exchange of the wave function is not de-
scribed by a sign, but by a continuous phase. Two-dimensional objects behaving in this
way, called anyons because they can have ‘any’ spin, have experimental importance, since
in many experiments in solid state physics the set-up is effectively two-dimensional. The
fractional quantum Hall effect, perhaps the most interesting discovery of modern exper-
Page 73 imental physics, has pushed anyons onto the stage of modern research.
Other theorists generalized the concept of fermions in other ways, introducing

Ref. 81 parafermions, parabosons, plektons and other hypothetical concepts. Oscar Greenberg
has spent most of his professional life on this issue. His conclusion is that in 3 + 1 space-
time dimensions, only fermions and bosons exist. (Can you show that this implies that
Challenge 108 s the ghosts appearing in Scottish tales do not exist?)
From a different viewpoint, the belt model of spin 1/2 invites to study the behaviour
of braids, open links and knots. (In mathematics, braids and open links are made of
strands extending to infinity.) This fascinating part of mathematical physics has become
important with in modern unified theories, which all state that particles, especially at
high energies, are not point-like, but extended entities.
Still another generalization of statistical behaviour at high energies is the concept of
quantum group. In all of these cases, the quest is to understand what happens to per-
mutation symmetry in a unified theory of nature. A glimpse of the difficulties appears
already above: how can Figures 54, 59 and 63 be reconciled and combined? We will settle
Vol. VI, page 160 this issue in the final part of our mountain ascent.

Chapter 7
SUPERPOSITIONS AND
PROBABILITIES – QUANTUM THEORY
W I THOUT I DEOLO GY
“
The fact that an adequate philosophical
presentation has been so long delayed is no
doubt caused by the fact that Niels Bohr
brainwashed a whole generation of theorists
Ref. 82 into thinking that the job was done fifty years
”
ago.
Murray Gell-Mann
W hy is this famous physical issue arousing such strong emotions? In particular,

ho is brainwashed, Gell-Mann, the discoverer of the quarks, or most of the
orld’s physicists working on quantum theory who follow Niels Bohr’s opinion?
In the twentieth century, quantum mechanics has thrown many in disarray. Indeed, it
radically changed the two most basic concepts of classical physics: state and system. The
state is not described any more by the specific values taken by position and momentum,
but by the specific wave function ‘taken’ by the position and momentum operators.* In
addition, in classical physics a system was described as a set of permanent aspects of na-
ture; permanence was defined as negligible interaction with the environment. Quantum
mechanics shows that this definition has to be modified as well.
The description of nature with quantum theory is unfamiliar for two reasons: it allows
superpositions and it leads to probabilities. Let us clarify these issues. A clarification is
essential if we want to avoid getting lost on our way to the top of Motion Mountain, as
happened to quite a number of people since quantum theory appeared, including Gell-
Mann.
Why are people either dead or alive?

The evolution equation of quantum mechanics is linear in the wave function; linearity
implies the existence of superpositions. Therefore we can imagine and try to construct
systems where the state ψ is a superposition of two radically distinct situations, such as
those of a dead and of a living cat. This famous fictional animal is called Schrödinger’s
cat after the originator of the example. Is it possible to produce it? And how would it
evolve in time? We can ask the same two questions in other situations. For example, can
we produce a superposition of a state where a car is inside a closed garage with a state
* It is equivalent, but maybe conceptually clearer, to say that the state is described by a complete set of
commuting operators. In fact, the discussion is somewhat simplified in the Heisenberg picture. However,
here we study the issue in the Schrödinger picture, using wave functions.
quantum theory withou t ideolo gy 121
Every such àrtistic impression’ is wrong.

F I G U R E 64 An artist’s
(Why?) impression of a
macroscopic superposition
is impossible
where the car is outside? What happens then?

Such strange situations are not usually observed in everyday life. The reason for this
rareness is an important aspect of what is often called the ‘interpretation’ of quantum
mechanics. In fact, such strange situations are possible, and the superposition of macro-
scopically distinct states has actually been observed in a few cases, though not for cats,
people or cars. To get an idea of the constraints, let us specify the situation in more de-
tail.*
The object of discussion are linear superpositions of the type ψ = aψa + bψb , where
ψa and ψb are macroscopically distinct states of the system under discussion, and where
a and b are some complex coefficients. States are called macroscopically distinct when
each state corresponds to a different macroscopic situation, i.e., when the two states can
be distinguished using the concepts or measurement methods of classical physics. In par-
ticular, this means that the physical action necessary to transform one state into the other
must be much larger than ħ. For example, two different positions of any body composed
of a large number of molecules are macroscopically distinct.
A ‘strange’ situation is thus a superposition of macroscopically distinct states. Let us
work out the essence of macroscopic superpositions more clearly. Given two macroscop-
ically distinct states ψa and ψb , a superposition of the type ψ = aψa + bψb is called a pure
state. Since the states ψa and ψb can interfere, one also talks about a (phase) coherent su-
perposition. In the case of a superposition of macroscopically distinct states, the scalar
product ψa† ψb is obviously vanishing. In case of a coherent superposition, the coefficient
product a∗ b is different from zero. This fact can also be expressed with the help of the
density matrix ρ of the system, defined as ρ = ψ ⊗ ψ † . In the present case it is given by
ρpure = ψ ⊗ ψ † = |a|2 ψa ⊗ ψa† + |b|2 ψb ⊗ ψb† + a b∗ ψa ⊗ ψb† + a∗ b ψb ⊗ ψa†

|a|2 a b∗ ψa†
= (ψa , ψb ) 󶀥 󶀵 󶀥 †󶀵 . (64)
a∗ b |b|2 ψb
We can then say that whenever the system is in a pure state, its density matrix, or density
functional, contains off-diagonal terms of the same order of magnitude as the diagonal
* Most what can be said about this topic has been said by two people: John von Neumann, who in the
Ref. 83 nineteen-thirties stressed the differences between evolution and decoherence, and by Hans Dieter Zeh, who
Ref. 84 in the nineteen-seventies stressed the importance of baths and the environment in the decoherence process.
122 superpositions and probabilities
ones.* Such a density matrix corresponds to the above-mentioned strange situations that
we do not observe in daily life.
We now have a look at the opposite situation, a density matrix for macroscopic distinct
states with vanishing off-diagonal elements. For two state, the example
ρ = |a|2 ψa ⊗ ψa† + |b|2 ψb ⊗ ψb†

|a|2 0 ψa†
= (ψa , ψb ) 󶀥 2󶀵 󶀥 †󶀵 (66)
0 |b| ψb
describes a system which possesses no phase coherence at all. (Here, ⊗ denotes the non-
commutative dyadic product or tensor product which produces a tensor or matrix start-
ing from two vectors.) Such a diagonal density matrix cannot be that of a pure state; it
describes a system which is in the state ψa with probability |a|2 and which is in the state
ψb with probability |b|2 . Such a system is said to be in a mixed state, because its state is
not known, or equivalently, to be in a (phase) incoherent superposition, because interfer-
ence effects cannot be observed in such a situation. A system described by a mixed state
is always either in the state ψa or in the state ψb . In other words, a diagonal density ma-
trix for macroscopically distinct states is not in contrast, but in agreement with everyday
experience. In the picture of density matrices, the non-diagonal elements contain the
difference between normal, i.e., incoherent, and unusual, i.e., coherent, superpositions.
The experimental situation is clear: for macroscopically distinct states, only diagonal
density matrices are observed in everyday life. Any system in a coherent macroscopic
superposition somehow loses its off-diagonal matrix elements. How does this process of
decoherence** take place? The density matrix itself shows the way.
Ref. 85 In thermodynamics, the density matrix for a large system is used for the definition of
Challenge 109 ny its entropy and of all its other thermodynamic quantities. These studies show that
S = −k tr(ρ ln ρ) (67)
where tr denotes the trace, i.e., the sum of all diagonal elements. We also remind ourselves
that a system with a large and constant entropy is called a bath. In simple physical terms,
a bath is a system to which we can ascribe a temperature. More precisely, a (physical) bath,
or (thermodynamic) reservoir, is any large system for which the concept of equilibrium
can be defined. Experiments show that in practice, this is equivalent to the condition
that a bath consists of many interacting subsystems. For this reason, all macroscopic
quantities describing the state of a bath show small, irregular fluctuations, a fact that will
be of central importance shortly.
* Using the density matrix, we can rewrite the evolution equation of a quantum system:
dρ i
ψ̇ = −iHψ becomes = − [H , ρ] . (65)
dt ħ
Both are completely equivalent. (The new expression is sometimes also called the von Neumann equation.)
We won’t actually do any calculations here. The expressions are given so that you recognize them when you
encounter them elsewhere.
** In certain settings, decoherence is called disentanglement, as we will see below.
An everyday bath is also a thermodynamic bath in the physical sense: indeed, a ther-
modynamic bath is similar to an extremely large warm water bath, one for which the
temperature does not change even if one adds some cold or warm water to it. Examples
of physical baths are an intense magnetic field, a large amount of gas, or a large solid.
(The meanings of ‘intense’ and ‘large’ of course depend on the system under study.) The
physical concept of bath (or reservoir) is thus an abstraction and a generalization of the
everyday concept of bath.
It is easy to see from the definition (67) of entropy that the loss of off-diagonal ele-
Challenge 110 s ments corresponds to an increase in entropy. And it is known that any increase in en-
tropy of a reversible system, such as the quantum mechanical system in question, is due
to an interaction with a bath.
Where is the bath interacting with the system? It obviously must be outside the system
one is talking about, i.e., in its environment. Indeed, we know experimentally that any
environment is large and characterized by a temperature. Some examples are listed in
Table 6. Any environment therefore contains a bath. We can even go further: for every
experimental situation, there is a bath interacting with the system. Indeed, every system
which can be observed is not isolated, as it obviously interacts at least with the observer;
and every observer by definition contains a bath, as we will show in more detail shortly.
Usually however, the most important baths we have to take into consideration are the
atmosphere around a system, the radiation or electromagnetic fields interacting with the
system, or, if the system itself is large enough to have a temperature, those degrees of
freedom of the system which are not involved in the superposition under investigation.
Since every system is in contact with baths, every density matrix of a macroscopic
superposition will lose its diagonal elements eventually. At first sight, this direction of
thought is not convincing. The interactions of a system with its environment can be made
extremely small by using clever experimental set-ups; that would imply that the time
for decoherence can be made extremely large. Thus we need to check how much time a
superposition of states needs to decohere. It turns out that there are two standard ways to
estimate the decoherence time: either by modelling the bath as large number of colliding
particles, or by modelling it as a continuous field.
If the bath is described as a set of particles randomly hitting the microscopic system,
it is best characterized by the effective wavelength λeff of the particles and by the average
Challenge 111 ny interval thit between two hits. A straightforward calculation shows that the decoherence
time t d is in any case smaller than this time interval, so that Copyright © Christoph Schiller November 1997–July 2010
1
t d ⩽ thit = , (68)
φσ
where φ is the flux of particles and σ the cross-section for the hit.* Typical values are given
in Table 6. We easily note that for macroscopic objects, decoherence times are extremely
* The decoherence time is derived by studying the evolution of the density matrix ρ(x, x 󳰀 ) of objects local-
󳰀 2
ized at two points x and x 󳰀 . One finds that the off-diagonal elements follow ρ(x, x 󳰀 , t) = ρ(x, x 󳰀 , 0)e−Λt(x−x ) ,
where the localization rate Λ is given by
Λ = k 2 φσeff (69)
where k is the wave number, φ the flux and σeff the cross-section of the collisions, i.e., usually the size of the
Ref. 86 macroscopic object.
TA B L E 6 Common and less common baths with their main properties
B at h t y p e T e m p e r - Wa v e - P a r - Cross Hit time

at u r e length ticle s e c t i o n 1/σφ f o r
flux ( at o m )
T λeff φ σ a t o ma b a l la
matter baths
solid, liquid 300 K 10 pm 1031 /m2 s 10−19 m2 10−12 s 10−25 s
air 300 K 10 pm 1028 /m2 s 10−19 m2 10−9 s 10−22 s
laboratory vacuum 50 mK 10 μm 1018 /m2 s 10−19 m2 10 s 10−12 s
photon baths
sunlight 5800 K 900 nm 1023 /m2 s 10−4 s 10−17 s
‘darkness’ 300 K 20 μm 1021 /m2 s 10−2 s 10−15 s
cosmic microwaves 2.7 K 2 mm 1017 /m2 s 102 s 10−11 s
terrestrial radio waves
Casimir effect very large
Unruh radiation of Earth 40 zK very large
nuclear radiation baths
radioactivity 10 fm 1 /m2 s 10−25 m2 1025 s 1012 s
cosmic radiation >1000 K 10 fm 10−2 /m2 s 10−25 m2 1027 s 1014 s
solar neutrinos ≈ 10 MK 10 fm 1011 /m2 s 10−47 m2 1036 s 1015 s
cosmic neutrinos 2.0 K 3 mm 1017 /m2 s 10−62 m2 1045 s 1024 s
gravitational baths
gravitational radiation 5 ⋅ 1031 K 10−35 m very large
a. Values are rough estimates. The macroscopic ball is assumed to have a 1 mm size.
short. (We also note that nuclear and gravitational effects lead to large decoherence times
and can thus all be neglected.) Scattering leads to fast decoherence. However, for atoms
or smaller systems, the situation is different, as expected.
A second method to estimate the decoherence time is also common. Any interaction
of a system with a bath is described by a relaxation time tr . The term relaxation designates
any process which leads to the return to the equilibrium state. The terms damping and
friction are also used. In the present case, the relaxation time describes the return to equi-
librium of the combination bath and system. Relaxation is an example of an irreversible
evolution. A process is called irreversible if the reversed process, in which every compo-
nent moves in opposite direction, is of very low probability.* For example, it is usual that
One also finds the surprising result that a system hit by a particle of energy Ehit collapses the density
Ref. 87 matrix roughly down to the de Broglie (or thermal de Broglie) wavelength of the hitting particle. Both
results together give the formula above.
* Beware of other definitions which try to make something deeper out of the concept of irreversibility, such
as claims that ‘irreversible’ means that the reversed process is not at all possible. Many so-called ‘contra-
a glass of wine poured into a bowl of water colours the whole water; it is very rarely ob-
served that the wine and the water separate again, since the probability of all water and
wine molecules to change directions together at the same time is rather low, a state of
affairs making the happiness of wine producers and the despair of wine consumers.
Now let us simplify the description of the bath. We approximate it by a single, un-
specified, scalar field which interacts with the quantum system. Due to the continuity
of space, such a field has an infinity of degrees of freedom. They are taken to model the
many degrees of freedom of the bath. The field is assumed to be in an initial state where
its degrees of freedom are excited in a way described by a temperature T. The interac-
tion of the system with the bath, which is at the origin of the relaxation process, can be
described by the repeated transfer of small amounts of energy Ehit until the relaxation
process is completed.
The objects of interest in this discussion, like the mentioned cat, person or car, are
described by a mass m. Their main characteristic is the maximum energy Er which can
be transferred from the system to the environment. This energy describes the interac-
tions between system and environment. The superpositions of macroscopic states we are
interested in are solutions of the Hamiltonian evolution of these systems.
The initial coherence of the superposition, so disturbingly in contrast with our every-
Ref. 88 day experience, disappears exponentially within a decoherence time t d given by*
Ehit eEhit /kT − 1

t d = tr (72)
Er eEhit /kT + 1
where k is the Boltzmann constant and like above, Er is the maximum energy which can
be transferred from the system to the environment. Note that one always has t d ⩽ tr .
After the decoherence time t d is elapsed, the system has evolved from the coherent to
the incoherent superposition of states, or, in other words, the density matrix has lost its
off-diagonal terms. One also says that the phase coherence of this system has been de-
stroyed. Thus, after a time t d , the system is found either in the state ψa or in the state ψb ,
respectively with the probability |a|2 or |b|2 , and not any more in a coherent superposi-
tion which is so much in contradiction with our daily experience. Which final state is
selected depends on the precise state of the bath, whose details were eliminated from the
calculation by taking an average over the states of its microscopic constituents.
dictions’ between the irreversibility of processes and the reversibility of evolution equations are due to this
mistaken interpretation of the term ‘irreversible’.

* This result is derived as in the above case. A system interacting with a bath always has an evolution given
Ref. 89 by the general form
dρ i 1
= − [H , ρ] − 󵠈[V ρ, V j† ] + [V j , ρV j† ] , (70)
dt ħ 2to j j
where ρ is the density matrix, H the Hamiltonian, V the interaction, and to the characteristic time of the
Challenge 112 ny interaction. Are you able to see why? Solving this equation, one finds for the elements far from the diagonal
ρ(t) = ρ0 e−t/t0 . In other words, they disappear with a characteristic time to . In most situations one has a
relation of the form
E
t0 = tr hit = thit (71)
Er
or some variations of it, as in the example above.
The important result is that for all macroscopic objects, the decoherence time t d is
extremely small. In order to see this more clearly, we can study a special simplified case.
A macroscopic object of mass m, like the mentioned cat or car, is assumed to be at the
same time in two locations separated by a distance l, i.e., in a superposition of the two
corresponding states. We further assume that the superposition is due to the object mov-
ing as a quantum mechanical oscillator with frequency ω between the two locations; this
is the simplest possible system that shows superpositions of an object located in two dif-
ferent positions. The energy of the object is then given by Er = mω2 l 2 , and the smallest
transfer energy Ehit = ħω is the difference between the oscillator levels. In a macroscopic
situation, this last energy is much smaller than kT, so that from the preceding expression
Ref. 90 we get
2
Ehit ħ2 λT2
t d = tr = tr = t r 2 (73)
2Er kT 2mkT l 2 l
in which the frequency ω has disappeared. The quantity λT = ħ/󵀂2mkT is called the
thermal de Broglie wavelength of a particle.
It is straightforward to see that for practically all macroscopic objects the typical deco-
herence time t d is extremely short. For example, setting m = 1 g, l = 1 mm and T = 300 K
we get t d /tr = 1.3⋅10−39 . Even if the interaction between the system and the environment
would be so weak that the system would have as relaxation time the age of the universe,
which is about 4 ⋅ 1017 s, the time t d would still be shorter than 5 ⋅ 10−22 s, which is over
a million times faster than the oscillation time of a beam of light (about 2 fs for green
light). For Schrödinger’s cat, the decoherence time would be even shorter. These times
are so short that we cannot even hope to prepare the initial coherent superposition, let
alone to observe its decay or to measure its lifetime.
For microscopic systems however, the situation is different. For example, for an elec-
tron in a solid cooled to liquid helium temperature we have m = 9.1 ⋅ 10−31 kg, and typ-
ically l = 1 nm and T = 4 K; we then get t d ≈ tr and therefore the system can stay in
a coherent superposition until it is relaxed, which confirms that for this case coherent
effects can indeed be observed if the system is kept isolated. A typical example is the be-
Ref. 91 haviour of electrons in superconducting materials. We will mention a few more below.
In 1996 the first actual measurement of decoherence times was published by the Paris
Ref. 92 team around Serge Haroche. It confirmed the relation between the decoherence time
and the relaxation time, thus showing that the two processes have to be distinguished
at microscopic scale. In the meantime, other experiments confirmed the decoherence
Ref. 93 process with its evolution equation, both for small and large values of t d /tr . A particularly
Ref. 94 beautiful experiment has been performed in 2004, where the disappearance of two-slit
interference for C70 molecules was observed when a bath interacts with them.
Summary on decoherence, life and death

Our exploration showed that decoherence results from coupling to a bath in the environ-
ment. Decoherence is a statistical, thermodynamic effect.
The estimates of decoherence times in everyday life told us that both the preparation
and the survival of superpositions of macroscopically different states is made impossible
by the interaction with any bath found in the environment. This is the case even if the
usual measure of this interaction, given by the friction of the motion of the system, is very
small. Even if a macroscopic system is subject to an extremely low friction, leading to a
very long relaxation time, its decoherence time is still vanishingly short. Only carefully
designed and expensive laboratory systems can reach substantial decoherence times.
Our everyday environment is full of baths. Therefore, coherent superpositions of macro-
scopically distinct states never appear in everyday life. Cars cannot be in and out of a garage
at the same time. In short, we cannot be dead and alive at the same time.
What is a system? What is an object?

In classical physics, a system is a part of nature which can be isolated from its environ-
ment. However, quantum mechanics tells us that isolated systems do not exist, since in-
teractions cannot be made vanishingly small. The results above allow us to define the
concept of system with more accuracy. A system is any part of nature which interacts in-
coherently with its environment. In other words, an object is a part of nature interacting
with its environment only through baths.
In particular, a system is called microscopic or quantum mechanical and can described
by a wave function ψ whenever
— it is almost isolated, with tevol = ħ/ΔE < tr , and
Ref. 95 — it is in incoherent interaction with its environment.
In short, a microscopic or quantum mechanical system that is described by a wave func-
tion interacts incoherently and weakly with its environment. (For such a system, the en-
ergy indeterminacy ΔE is larger than the relaxation energy.) In contrast, a bath is never
isolated in the sense just given, because the evolution time of a bath is always much larger
than its relaxation time. Since all macroscopic bodies are in contact with baths – or even
contain one – they cannot be described by a wave function. In particular, it is impossible
to describe any measuring apparatus with the help of a wave function.
We thus conclude that a macroscopic system is a system with a decoherence time much
shorter than any other evolution time of its constituents. Obviously, macroscopic systems
also interact incoherently with their environment. Thus cats, cars and television news
speakers are all macroscopic systems.
One possibility is left over by the two definitions: what happens in the situation in
which the interactions with the environment are coherent? We will encounter some ex-
amples shortly. Following the definition, they are neither microscopic and macroscopic
systems; they are not described by a wave function, and strictly speaking, they are not
systems. In these situations, when the interaction is coherent, one speaks of entanglement
or of entangle ‘systems’; such a particle or set of particles is said to be entangled with its
environment.
Entangled, coherently interacting systems can be divided, but must be disentangled
when doing so. The act of division leads to detached entities; detached entities interact in-
coherently. Quantum theory shows that nature is not made of detached entities, but that
it is made of detachable entities. In quantum theory, the criterion of detachment is the
incoherence of interaction. Coherent superpositions imply the surprising consequence
that there are systems which, even though they look being made of detached parts, are
not. Entanglement poses a limit to detachment. All surprising properties of quantum
space
collapse
t1 t2 t3 t4
F I G U R E 65 Quantum mechanical motion: an electron
wave function (actually its module squared) from the
slit screen
space moment it passes a slit until it hits a screen
mechanics, such as Schrödinger’s cat, are consequences of the classical prejudice that a
system made of two or more parts can obviously be detached into two subsystems with-
out disturbance. But coherent superpositions, or entangled systems, do not allow detach-
ment without disturbance. Whenever we assume to be able to detach entangled systems,
we get strange or incorrect conclusions, such as apparent faster-than-light propagation,
or, as one says today, non-local behaviour. Let us have a look at a few typical examples.
Is quantum theory non-local? – A bit about the

Einstein–Podolsky–Rosen paradox
“
[Mr. Duffy] lived a little distance away from his
”
body ...
James Joyce, A Painful Case
We now explore non-locality in quantum mechanics.* To do this, we explore wave func-

tion collapse in some detail.
We imagine an electron hitting a screen after passing a slit. Following the description
just deduced, the process proceeds schematically as depicted in Figure 65. A film of the
same process can be seen in the lower left corners on these pages, starting at page 74.
The process has a surprising aspect: due to the short decoherence time, during this (and
any other) wave function collapse the maximum of the wave function changes position
faster than light. Is this reasonable?
A situation is called acausal or non-local if energy is transported faster than light. Us-
Challenge 113 s ing Figure 65 you can determine the energy velocity involved, using the results on signal
Page 107 propagation. The result is a value smaller than c. A wave function maximum moving
faster than light does not imply energy moving faster than light.
* This continues a topic that we know already: we have explored a different type of non-locality, in general
Vol. II, page 255 relativity, earlier on.
space
detector 2
detector 1
collapse
time F I G U R E 66 Bohm’s Gedanken experiment
In other words, quantum theory has speeds greater than light, but no energy speeds
Ref. 96 greater than light. In classical electrodynamics, the same happens with the scalar and the
vector potentials if the Coulomb gauge is used. We have also encountered speeds faster
Page 53 than that of light in the motion of shadows and in many other observations. Any physicist
now has two choices: he can be straight, and say that there is no non-locality in nature;
or he can be less straight, and claim there is. In the latter case, he has to claim that even
classical physics is non-local. However, this never happens. On the other hand, there is a
danger in this more provoking usage: a small percentage of those who say that the world
is non-local after a while start to believe that there really are faster-than-light effects in
nature. These people become prisoners of their muddled thinking; on the other hands,
muddled thinking helps to get more easily into newspapers. In short, even though the
definition of non-locality is not unanimous, here we stick to the stricter one, and define
non-locality as energy transport faster than light.
An often cited Gedanken experiment that shows the pitfalls of non-locality was
proposed by Bohm* in the discussion around the so-called Einstein–Podolsky–Rosen
paradox. In the famous EPR paper the three authors try to find a contradiction between
Ref. 97, Ref. 98

quantum mechanics and common sense. Bohm translated their rather confused paper
into a clear Gedanken experiment. When two particles in a spin 0 state move apart, mea-
suring one particle’s spin orientation implies an immediate collapse also of the other par-
ticle’s spin, namely in the exactly opposite direction. This happens instantaneously over
the whole separation distance; no speed limit is obeyed. In other words, entanglement
seems to lead to faster-than-light communication.
However, in Bohm’s experiment, no energy is transported faster than light. No non-
locality is present, despite numerous claims of the contrary by certain authors. The two
* David Joseph Bohm (1917–1992) American–British physicist. He codiscovered the Aharonov–Bohm effect;
he spent a large part of his later life investigating the connections between quantum physics and philosophy.
entangled electrons belong to one system: assuming that they are separate only because
the wave function has two distant maxima is a conceptual mistake. In fact, no signal can
be transmitted with this method; the decoherence is a case of prediction which looks
like a signal without being one. Bohm’s experiment, like any other EPR-like experiment,
does not allow communication faster than light. We already discussed such cases in the
Page 111 section on electrodynamics.
Bohm’s experiment has actually been performed. The first and most famous realiza-
Ref. 99 tion was realized in 1982 by Alain Aspect; he used photons instead of electrons. Like all
latter tests, it has fully confirmed quantum mechanics.
In fact, experiments such as the one by Aspect confirm that it is impossible to treat
either of the two particles as a system by itself; it is impossible to ascribe any physical
property, such as a spin orientation, to either of them alone. (The Heisenberg picture
would express this restriction even more clearly.)
The mentioned two examples of apparent non-locality can be dismissed with the re-
mark that since obviously no energy flux faster than light is involved, no problems with
causality appear. Therefore the following example is more interesting. Take two identical
atoms, one in an excited state, one in the ground state, and call l the distance that sepa-
rates them. Common sense tells that if the first atom returns to its ground state emitting
a photon, the second atom can be excited only after a time t = l/c has been elapsed, i.e.,
after the photon has travelled to the second atom.
Surprisingly, this conclusion is wrong. The atom in its ground state has a non-zero
probability to be excited at the same moment in which the first is de-excited. This has
Ref. 100 been shown most simply by Gerhard Hegerfeldt. The result has even been confirmed
experimentally.
More careful studies show that the result depends on the type of superposition of the
two atoms at the beginning: coherent or incoherent. For incoherent superpositions, the
intuitive result is correct; the counter-intuitive result appears only for coherent superpo-
sitions. Again, a careful discussion shows that no real non-locality of energy is involved.
In summary, faster-than-light speeds in wave function collapse do not contradict the
limit on energy speed of special relativity. Collapse speeds are phase velocities. In nature,
phase velocities are unlimited; unlimited phase velocities never imply energy transport
faster than light.
Curiosities and fun challenges about superpositions Copyright © Christoph Schiller November 1997–July 2010
Challenge 114 s Can a photograph show an object at two different places at the same time?
∗∗
In a few cases, the superposition of different macroscopic states can actually be observed
by lowering the temperature to sufficiently small values and by carefully choosing suit-
ably small masses or distances. Two well-known examples of coherent superpositions
are those observed in gravitational wave detectors and in Josephson junctions. In the
Ref. 90 first case, one observes a mass as heavy as 1000 kg in a superposition of states located
at different points in space: the distance between them is of the order of 10−17 m. In
the second case, in superconducting rings, superpositions of a state in which a macro-
scopic current of the order of 1 pA flows in clockwise direction with one where it flows
Ref. 105 in counter-clockwise direction have been produced.

∗∗
Ref. 101 Superpositions of magnetization in up and down direction at the same time have also be
observed for several materials.
∗∗
Some people wrongly state that an atom that is in a superposition of states centred at
different positions has been photographed. (This lie is even used by some sects to attract
Challenge 115 s believers.) Why is this not true?
∗∗
Since the 1990s, the sport of finding and playing with new systems in coherent macro-
Ref. 102 scopic superpositions has taken off across the world. The challenges lie in the clean exper-
iments necessary. Experiments with single atoms in superpositions of states are among
Ref. 103 the most popular ones.
∗∗
Ref. 104 In 1997, coherent atom waves were extracted from a cloud of sodium atoms.
∗∗
Macroscopic objects usually are in incoherent states. This is the same situation as for
light. The world is full of ‘macroscopic’, i.e., incoherent light: daylight, and all light from
lamps, from fire and from glow-worms is incoherent. Only very special and carefully
constructed sources, such as lasers or small point sources, emit coherent light. Only these
sources allow to study interference effects. In fact, the terms ‘coherent’ and ‘incoherent’
originated in optics, since for light the difference between the two, namely the capacity
to interfere, had been observed centuries before the case of matter.
Coherence and incoherence of light and of matter manifest themselves differently, be-
cause matter can stay at rest but light cannot and because matter is made of fermions,
Page 116 but light is made of bosons. Coherence can be observed easily in systems composed of
bosons, such as light, sound in solids, or electron pairs in superconductors. Coherence
is less easily observed in systems of fermions, such as systems of atoms with their elec-
tron clouds. However, in both cases a decoherence time can be defined. In both cases
coherence in many particle systems is best observed if all particles are in the same state
(superconductivity, laser light) and in both cases the transition from coherent to incoher-
ent is due to the interaction with a bath. A beam is thus incoherent if its particles arrive
randomly in time and in frequency. In everyday life, the rarity of observation of coherent
matter superpositions has the same origin as the rarity of observation of coherent light.
∗∗
We will discuss the relation between the environment and the decay of unstable systems
Vol. V, page 34 later on. The phenomenon is completely described by the concepts given here.
∗∗
Challenge 116 ny Can you find a method to measure the degree of entanglement? Can you do so for a
system made of many particles?

∗∗
The study of entanglement leads to a simple conclusion: teleportation contradicts correla-
Challenge 117 ny tion. Can you confirm the statement?
∗∗
Challenge 118 s Are ghost images in TV sets, often due to spurious reflections, examples of interference?
∗∗
Challenge 119 d What happens when two monochromatic electrons overlap?
∗∗
Some people say that quantum theory could be used for quantum computing, by using
Ref. 106 coherent superpositions of wave functions. Can you give a general reason that makes
this aim very difficult – even though not impossible – even without knowing how such
Challenge 120 s a quantum computer might work, or what the so-called qubits might be?
What is all the fuss about measurements in quantum theory?

Measurements in quantum mechanics are disturbing. They lead to statements in which
probabilities appear. For example, we speak about the probability of finding an electron at
a certain distance from the nucleus of an atom. Statements like this belong to the general
type ‘when the observable A is measured, the probability to find the outcome a is p.’ In
the following we will show that the probabilities in such statements are inevitable for
any measurement, because, as we will show, any measurement and any observation is
a special case of decoherence or disentanglement process. (Historically, the process of
measurement was studied before the more general process of decoherence. That explains
in part why the topic is so confused in many peoples’ minds.)
Page 185 What is a measurement? As already mentioned earlier on, a measurement is any in-
teraction which produces a record or a memory. (Any effect of everyday life is a record;
but this is not true in general. Can you give some examples of effects that are records and
Challenge 121 ny some effects which are not?) Measurements can be performed by machines; when they
are performed by people, they are called observations. In quantum theory, the process of
measurement is not as straightforward as in classical physics. This is seen most strikingly
when a quantum system, such as a single electron, is first made to pass a diffraction slit,
or better – in order to make its wave aspect become apparent – a double slit and then
is made to hit a photographic plate, in order to make also its particle aspect appear. Ex-
periment shows that the blackened dot, the spot where the electron has hit the screen,
cannot be determined in advance. (The same is true for photons or any other particle.)
However, for large numbers of electrons, the spatial distribution of the black dots, the
so-called diffraction pattern, can be calculated in advance with high precision.
The outcome of experiments on microscopic systems thus forces us to use probabil-
ities for the description of microsystems. We find that the probability distribution p(x)
of the spots on the photographic plate can be calculated from the wave function ψ of
the electron at the screen surface and is given by p(x) = |ψ † (x)ψ(x)|2 . This is in fact a
ball
gravity
pegs
F I G U R E 67 A system showing probabilistic behaviour
special case of the general first property of quantum measurements: the measurement of
an observable A for a system in a state ψ gives as result one of the eigenvalues an , and
the probability Pn to get the result an is given by
Pn = |φ†n ψ|2 , (74)
Page 282 where φn is the eigenfunction of the operator A corresponding to the eigenvalue an . This
experimental result requires an explanation.
Experiments also show a second property of quantum measurements: after the mea-
surement, the observed quantum system is in the state φn corresponding to the mea-
sured eigenvalue an . One also says that during the measurement, the wave function has
Ref. 107 collapsed from ψ to φn . By the way, these experimental results can also be generalized to
the more general cases with degenerate and continuous eigenvalues.
At first sight, the sort of probabilities encountered in quantum theory are different
from the probabilities we encounter in everyday life. Take roulette, dice, pachinko ma-
chines or the direction in which a pencil on its tip falls: all have been measured exper-
imentally to be random (assuming no cheating by the designer or operators) to a high
degree of accuracy. These everyday systems do not puzzle us. We unconsciously assume
that the random outcome is due to the small, but uncontrollable variations of the starting
conditions or the environment every time the experiment is repeated.*
But microscopic systems seem to be different. The two properties of quantum mea-
surements just mentioned express what physicists observe in every experiment, even if
the initial conditions are taken to be exactly the same every time. But why then is the
position for a single electron, or most other observables of quantum systems, not pre-
dictable? In other words, what happens during the collapse of the wave function? How
long does it take? In the beginning of quantum theory, there was the perception that the
observed unpredictability is due to the lack of information about the state of the particle.
This lead many to search for so-called ‘hidden variables’. All these attempts were doomed
to fail, however. It took some time for the scientific community to realize that the unpre-
dictability is not due to the lack of information about the state of the particle, which is
* To get a feeling for the limitations of these unconscious assumptions, you may want to read the already
mentioned story of those physicists who built a machine that could predict the outcome of a roulette ball
Vol. I, page 108 from the initial velocity imparted by the croupier.
indeed described completely by the state vector ψ.

In order to uncover the origin of probabilities, let us recall the nature of a measure-
ment, or better, of a general observation. Any observation is the production of a record.
The record can be a visual or auditive memory in our brain, or a written record on paper,
Vol. III, page 184 or a tape recording, or any such type of object. As explained in the previous volume, an
object is a record if it cannot have arisen or disappeared by chance. To avoid the influ-
ence of chance, all records have to be protected as much as possible from the external
world; e.g. one typically puts archives in earthquake safe buildings with fire protection,
keeps documents in a safe, avoids brain injury as much as possible, etc.
On top of this, records have to be protected from their internal fluctuations. These
internal fluctuations are due to the many components any recording device is made of.
If the fluctuations were too large, they would make it impossible to distinguish between
the possible contents of a memory. Now, fluctuations decrease with increasing size of a
system, typically with the square root of the size. For example, if a hand writing is too
small, it is difficult to read if the paper gets brittle; if the magnetic tracks on tapes are
too small, they demagnetize and lose the stored information. In other words, a record is
rendered stable against internal fluctuations by making it of sufficient size. Every record
thus consists of many components and shows small fluctuations.
The importance of size can be expressed in another way: every system with memory,
i.e., every system capable of producing a record, contains a bath. In summary, the state-
ment that any observation is the production of a record can be expressed more precisely
as: Any observation of a system is the result of an interaction between that system and a
bath in the recording apparatus.*
In addition, any observation measuring a physical quantity uses an interaction de-
pending on that same quantity. With these seemingly trivial remarks, we can describe in
more detail the process of observation, or, as it is usually called in the quantum theory,
the measurement process.
Any measurement apparatus, or detector, is characterized by two main aspects, shown
in Figure 68: the interaction it has with the microscopic system, and the bath it contains
Ref. 108 to produce the record. Any description of the measurement process thus is the descrip-
tion of the evolution of the microscopic system and the detector; therefore one needs
the Hamiltonian for the particle, the interaction Hamiltonian, and the bath properties
(such as the relaxation time tr ). The interaction specifies what is measured and the bath
realizes the memory.
We know that only classical thermodynamic systems can be irreversible; quantum
systems are not. We therefore conclude: a measurement system must be described clas-
sically: otherwise it would have no memory and would not be a measurement system: it
would not produce a record! Memory is a classical effect. (More precisely, memory is an
effect that only appears in the classical limit.) Nevertheless, let us see what happens if we
describe the measurement system quantum mechanically.
Let us call A the observable which is measured in the experiment and its eigenfunc-
tions φn . We describe the quantum mechanical system under observation – often a par-
* Since baths imply friction, we can also say: memory needs friction.
the quantum apparatus, e.g. eye, ear,

mechanical or machine, with memory,
system i.e. coupled to a bath
H H int tr
describes is determined describes its

its possible by the type friction, e.g. F I G U R E 68 The concepts used in the
states of measurement due to heat flow description of measurements
ticle – by a state ψ. The full state of the system can always be written as
ψ = ψ p ψother = 󵠈 cn φn ψother . (75)

n
Here, ψ p is the aspect of the (particle or system) state that we want to measure, and ψother
represents all other degrees of freedom, i.e., those not described – spanned, in mathemat-
ical language – by the operator A corresponding to the observable we want to measure.
The numbers cn = |φ†n ψ p | give the expansion of the state ψ p , which is taken to be nor-
malized, in terms of the basis φn . For example, in a typical position measurement, the
functions φn would be the position eigenfunctions and ψother would contain the informa-
tion about the momentum, the spin and all other properties of the particle.
How does the system–detector interaction look like? Let us call the state of the appa-
ratus before the measurement χstart . The measurement apparatus itself, by definition, is
a device which, when it is hit by a particle in the state φn ψother , changes from the state
χstart to the state χn . One then says that the apparatus has measured the eigenvalue an
corresponding to the eigenfunction φn of the operator A. The index n is thus the record
of the measurement; it is called the pointer index or variable. This index tells us in which
state the microscopic system was before the interaction. The important point, taken from
our previous discussion, is that the states χn , being records, are macroscopically distinct,
precisely in the sense of the previous section. Otherwise they would not be records, and
the interaction with the detector would not be a measurement.

Of course, during measurement, the apparatus sensitive to φn changes the part ψother
of the particle state to some other situation ψother,n , which depends on the measurement
and on the apparatus; we do not need to specify it in the following discussion.* But let
us have an intermediate check of our reasoning. Do apparatuses as described here exist?
* How does the interaction look like mathematically? From the description we just gave, we specified the
final state for every initial state. Since the two density matrices are related by
ρf = T ρi T † (76)
Challenge 122 ny we can deduce the Hamiltonian from the matrix T. Are you able to see how?
By the way, one can say in general that an apparatus measuring an observable A has a system interaction
Yes, they do. For example, any photographic plate is a detector for the position of ion-
izing particles. A plate, and in general any apparatus measuring position, does this by
changing its momentum in a way depending on the measured position: the electron on
a photographic plate is stopped. In this case, χstart is a white plate, φn would be a particle
localized at spot n, χn is the function describing a plate blackened at spot n and ψother,n
describes the momentum and spin of the particle after it has hit the photographic plate
at the spot n.
Now we are ready to look at the measurement process itself. For the moment, let us
disregard the bath in the detector, and let us just describe it with a state as well, which
we call χstart . In the time before the interaction between the particle and the detector, the
combined system (including the detector) was in the initial state ψi given simply by
ψi = ψ p χstart = 󵠈 cn φn ψother χstart , (78)
n
where ψ p is the (particle or system) state. After the interaction, using the just mentioned,
experimentally known characteristics of the apparatus, the combined state ψa is
ψa = 󵠈 cn φn ψother,n χn . (79)
n
This evolution from ψi to ψa follows from the evolution equation applied to the particle–
detector combination. Now, the combined state ψa is a superposition of macroscopically
distinct states: it is a superposition of distinct macroscopic states of the detector. In our
example ψa could correspond to a superposition of one state where a spot on the left
upper corner is blackened on an otherwise white plate with another state where a spot
on the right lower corner of the otherwise white plate is blackened. Such a situation is
never observed. Let us see why.
The density matrix ρa of the combined state ψa after the measurement, given by
∗
ρa = ψa ⊗ ψa† = 󵠈 cn cm (φn ψother,n χn ) ⊗ (φm ψother,m χm )† , (80)
n,m
contains large non-diagonal terms, i.e., terms for n ̸= m, whose numerical coefficients
are different from zero. Now let us take the bath back in. From the previous section we
know the effect of a bath on such a macroscopic superposition. We found that a density
matrix such as ρa decoheres extremely rapidly. We assume here that the decoherence
time is negligibly small, in practice thus instantaneous,* so that the off-diagonal terms
Hamiltonian depending on the pointer variable A, and for which one has
[H + Hint , A] = 0 . (77)
* Note however, that an exactly vanishing decoherence time, which would mean a strictly infinite number
of degrees of freedom of the environment, is in contradiction with the evolution equation, and in particu-
lar with unitarity, locality and causality. It is essential in the whole argument not to confuse the logical
consequences of a extremely small decoherence time with those of an exactly vanishing decoherence time.
vanish, and only the final, diagonal density matrix ρf , given by
ρf = 󵠈|cn |2 (φn ψother,n χn ) ⊗ (φn ψother,n χn )† (81)

n
has experimental relevance. As explained above, such a density matrix describes a mixed
state and the numbers Pn = |cn |2 = |φ†n ψ p |2 give the probability of measuring the value
an and of finding the particle in the state φn ψother,n as well as the detector in the state χn .
But this is precisely what the two properties of quantum measurements state.
We therefore find that describing a measurement as an evolution of a quantum system
interacting with a macroscopic detector, itself containing a bath, we can deduce the two
properties of quantum measurements, and thus the collapse of the wave function, from
the quantum mechanical evolution equation. The decoherence time td of the previous
section becomes the time of collapse in the case of a measurement; in addition we find
tcollapse = td < tr . (82)
In other words, the collapse time is always smaller than the relaxation time of the bath.
We thus have a formula for the time the wave function takes to collapse. The first exper-
Ref. 109 imental measurements of the time of collapse have appeared and confirmed this result.
Hidden variables
A large number of people are not satisfied with the arguments just presented. They long
for more mystery in quantum theory. They do not like the idea that probabilities are due
to baths. The most famous prejudice they cultivate is the idea that the probabilities are
due to some hidden aspect of nature which is still unknown to humans. These imagined,
unknown aspects are called hidden variables.
The beautiful thing about quantum mechanics is that it allows both conceptual and
experimental tests on whether such hidden variables exist without the need of knowing
them. Hidden variables controlling the evolution of microscopic system would contra-
dict the result that action values below ħ/2 cannot be detected. This minimum observable
action is the reason for the random behaviour of microscopic systems. A minimal action
thus excludes hidden variables.
Historically, the first argument against hidden variables was given by John von Neu-
mann.* An additional no-go theorem for hidden variables was published by Kochen
Ref. 110 and Specker in 1967, (and independently by Bell in 1969). The theorem states that non-
contextual hidden variables are impossible, if the Hilbert space has a dimension equal or
larger than three. The theorem is about non-contextual variables, i.e., about hidden vari-
ables inside the quantum mechanical system. The Kochen–Specker theorem thus states
that there is no non-contextual hidden variables model, because mathematics forbids
* János von Neumann (b. 1903 Budapest, d. 1957 Washington DC) Hungarian mathematician. One of the
greatest and clearest minds of the twentieth century, he settled already many questions, especially in applied
mathematics and quantum theory, that others still struggle with today. He worked on the atomic and the
hydrogen bomb, on ballistic missiles, and on general defence problems. In another famous project, he build
the first US-American computer, building on his extension of the ideas of Konrad Zuse.
it. This result essentially eliminates all possibilities for hidden variables, because usual
quantum mechanical systems have Hilbert space dimensions much larger than three.
Of course, one cannot avoid noting that about contextual hidden variables, i.e., vari-
ables in the environment, there are no restricting theorems; indeed, their necessity was
shown earlier in this section.
But also common sense eliminates hidden variables, without any recourse to mathe-
matics, with an argument often overlooked. If a quantum mechanical system had inter-
nal hidden variables, the measurement apparatus would have zillions of them.* And that
would mean that it could not work as a measurement system.
Despite all these results, people have also looked for experimental tests on hidden
variables. Most tests are based on the famed Bell’s equation, a beautifully simple relation
published by John Bell** in the 1960s.
The starting idea is to distinguish quantum theory and locally realistic theories using
hidden variables by measuring the polarizations of two correlated photons. Quantum
theory says that the polarization of the photons is fixed only at the time it is measured,
whereas local realistic theories say that it is fixed already in advance. Experiment can be
used to decide which approach is correct.
Imagine the polarization is measured at two distant points A and B, each observer
can measure 1 or −1 in each of his favourite direction. Let each observer choose two
directions, 1 and 2, and call their results a1 , a2 , b1 and b2 . Since the measurement results
all are either 1 or −1, the value of the specific expression (a1 + a2 )b1 + (a2 − a1 )b2 has
always the value ±2.
Ref. 111 Imagine you repeat the experiment many times, assuming that the hidden variables
appear statistically. You then can deduce (a special case of) Bell’s inequality for two hid-
Challenge 123 e den variables; it predicts that
|(a1 b1 ) + (a2 b1 ) + (a2 b2 ) − (a1 b2 )| ⩽ 2 . (83)
Here, the expressions in brackets are the averages of the measurement products over a
large number of samples. This result holds independently of the directions of the involved
polarizers.
On the other hand, for the case that the polarizers 1 and 2 at position A and the
corresponding ones at position B are chosen with angles of π/4, quantum theory predicts
that
|(a1 b1 ) + (a2 b1 ) + (a2 b2 ) − (a1 b2 )| = 2󵀂2 > 2 . (84)
This prediction is in complete contradiction with the hidden variable result.

All experimental checks of Bell’s equation have confirmed standard quantum mechan-
ics. No evidence for hidden variables has ever been found. This is not really surprising,
since the search for such variables is based on a misunderstanding of quantum mechan-
ics or on personal desires on how the world should be, instead of relying on experimental
evidence.
* Which leads to the definition: one zillion is 1023 .

** John Stewart Bell (1928–1990), theoretical physicist who worked mainly on the foundations of quantum
theory.
Another measurable contradiction between quantum theory and locally realistic the-
ories has been predicted by Greenberger, Horn and Zeilinger in systems with three en-
Ref. 112 tangles particles. The various predictions have been confirmed in all experiments.
Summary on probabilities and determinism
“
Geometric demonstramus quia facimus; si
”
physics demonstrare possemus, faceremus.
Giambattista Vico*
From the arguments presented here we draw a number of conclusions which we need
for the rest of our mountain ascent. Note that these conclusions are not yet shared by all
physicists! The whole topic is still touchy.
— Probabilities do not appear in measurements because the state of the quantum system
is unknown or fuzzy, but because the detailed state of the bath in the environment
is unknown. Quantum mechanical probabilities are of statistical origin and are due
to baths in the environment (or in the measurement apparatus). The probabilities are
due to the large number of degrees of freedom contained in any bath. These large
numbers make the outcome of experiments unpredictable. If the state of the bath
were known, the outcome of an experiment could be predicted. The probabilities of
quantum theory are ‘thermodynamic’ in origin.
In other words, there are no fundamental probabilities in nature. All probabilities
in nature are due to decoherence; in particular, all probabilities are due to the statistics
of the many particles – some of which may be virtual – that are part of the baths
in the environment. Modifying well-known words by Albert Einstein, ‘nature really
does not play dice.’ We therefore called ψ the wave function instead of ‘probability
amplitude’, as is often done. An even better name would be state function.
— Any observation in everyday life is a special case of decoherence. What is usually
called the ‘collapse of the wave function’ is a decoherence process due to the interac-
tion with the baths present in the environment or in the measuring apparatus. Be-
cause humans are warm-blooded and have memory, humans themselves are thus
measurement apparatuses. The fact that our body temperature is 37°C is thus the
reason that we see only a single world, and no superpositions. (Actually, there are
Challenge 124 s many additional reasons; can you name a few?)
— A measurement is complete when the microscopic system has interacted with the
bath in the measuring apparatus. Quantum theory as a description of nature does not
require detectors; the evolution equation describes all examples of motion. However,
measurements do require the existence of detectors. Detectors, being machines that
record observations, have to include a bath, i.e., have to be classical, macroscopic
objects. In this context one speaks also of a classical apparatus. This necessity of the
measurement apparatus to be classical had been already stressed in the very early
stages of quantum theory.
* ‘We are able to demonstrate geometrical matters because we make them; if we could prove physical mat-
ters we would be able to make them.’ Giovanni Battista Vico (b. 1668 Napoli, d. 1744 Napoli) important
Italian philosopher and thinker. In this famous statement he points out a fundamental distinction between
mathematics and physics.
— All measurements, being decoherence processes that involve interactions with baths,
are irreversible processes and increase entropy.
— A measurement is a special case of quantum mechanical evolution, namely the evolu-
tion for the combination of a quantum system, a macroscopic detector and the envi-
ronment. Since the evolution equation is relativistically invariant, no causality prob-
lems appear in measurements; neither do locality problems and logical problems ap-
pear.
— Since both the evolution equation and the measurement process does not involve
quantities other than space-time, Hamiltonians, baths and wave-functions, no other
quantity plays a role in measurement. In particular, no human observer nor any con-
sciousness are involved or necessary. Every measurement is complete when the micro-
scopic system has interacted with the bath in the apparatus. The decoherence inherent
in every measurement takes place even if nobody is looking. This trivial consequence
is in agreement with the observations of everyday life, for example with the fact that
the Moon is orbiting the Earth even if nobody looks at it.* Similarly, a tree falling in
the middle of a forest makes noise even if nobody listens. Decoherence is independent
of human observation, of the human mind and of human existence.
— In every measurement the quantum system interacts with the detector. Since there is a
minimum value for the magnitude of action, every observation influences the observed.
Therefore every measurement disturbs the quantum system. Any precise description
of observations must also include the description of this disturbance. In the present
section the disturbance was modelled by the change of the state of the system from
ψother to ψother,n . Without such a change of state, without a disturbance of the quantum
system, a measurement is impossible.
— Since the complete measurement is described by quantum mechanics, unitarity is
and remains the basic property of evolution. There are no non-unitary processes in
quantum mechanics.
— The description of the collapse of the wave function as a decoherence process is an
explanation exactly in the sense in which the term ‘explanation’ was defined earlier
Page 242 on; it describes the relation between an observation and all the other aspects of reality,
in this case the bath in the detector or the environment. The collapse of the wave
function has been both calculated and explained. The collapse is not a question of
‘interpretation’, i.e., of opinion, as unfortunately often is suggested.**
— It is not useful to speculate whether the evolution for a single quantum measurement
could be determined if the state of the environment around the system were known.
Measurements need baths. But a bath, being irreversible, cannot be described by a

wave function, which behaves reversibly.*** Quantum mechanics is deterministic.
Baths are probabilistic.
* The opposite view is sometimes falsely attributed to Niels Bohr. The Moon is obviously in contact with
Challenge 125 s many radiation baths. Can you list a few?
** This implies that the so-called ‘many worlds’ interpretation is wishful thinking. The conclusion is con-
Ref. 113 firmed when studying the details of this religious approach. It is a belief system, not based on facts.
*** This very strong type of determinism will be very much challenged in the last part of this text, in which
it will be shown that time is not a fundamental concept, and therefore that the debate around determinism
looses most of its interest.
In summary, there is no irrationality in quantum theory. Whoever uses quantum the-

ory as argument for superstitions, irrational behaviour, new age beliefs or ideologies is
guilty of disinformation. The statement by Gell-Mann at the beginning of this chapter is
thus such an example. Another is the following well-known, but incorrect statement by
Richard Feynman:
Ref. 114 ... nobody understands quantum mechanics.
Nobel Prizes obviously do not prevent views distorted by ideology.

The process of decoherence allows understanding many other issues. We explore a
few interesting ones.
What is the difference between space and time?
Space and time differ. Objects are localized in space but not in time. Why is this the
case? In nature, most bath–system interactions are mediated by a potential. All poten-
tials are by definition position dependent. Therefore, every potential, being a function
of the position x, commutes with the position observable (and thus with the interaction
Hamiltonian). The decoherence induced by baths – except if special care is taken – thus
first of all destroys the non-diagonal elements for every superposition of states centred
at different locations. In short, objects are localized because they interact with baths via
potentials.
For the same reason, objects also have only one spatial orientation at a time. If the
system–bath interaction is spin-dependent, the bath leads to ‘localization’ in the spin
variable. This occurs for all microscopic systems interacting with magnets. As a result,
macroscopic superpositions of magnetization are almost never observed. Since electrons,
protons and neutrons have a magnetic moment and a spin, this conclusion can even be
extended: everyday objects are never seen in superpositions of different rotation states
because their interactions with baths are spin-dependent.
As a counter-example, most systems are not localized in time, but on the contrary exist
for very long times, because practically all system–bath interactions do not commute
with time. In fact, this is the way a bath is defined to begin with. In short, objects are
permanent because they interact with baths.
Are you able to find an interaction which is momentum-dependent instead of
Challenge 126 s position-dependent? What is the consequence for macroscopic systems?
In other words, in contrast to general relativity, quantum theory produces a distinc-

tion between space and time. In fact, we can define position as the observable that com-
mutes with interaction Hamiltonians. This distinction between space and time is due to
the properties of matter and its interactions. We could not have deduced this distinction
in general relativity.
Are we good observers?

Are humans classical apparatuses? Yes, they are. Even though several prominent physi-
cists claim that free will and probabilities are related, a detailed investigation shows that
Ref. 115 this in not the case. Our senses are classical machines in the sense described above: they
record observations by interaction with a bath. Our brain is also a classical apparatus:
the neurons are embedded in baths. Quantum probabilities do not play a determining
role in the brain.
Any observing entity needs a bath and a memory to record its observations. This
means that observers have to be made of matter; an observer cannot be made of radi-
ation. Our description of nature is thus severely biased: we describe it from the stand-
point of matter. That is a bit like describing the stars by putting the Earth at the centre
of the universe. Can we eliminate this basic anthropomorphism? We will find out as we
Page 71 continue.
What relates information theory, cryptology and quantum

theory?
Physics means talking about observations of nature. Like any observation, also measure-
ments produce information. It is thus possible to translate much (but not all) of quantum
theory into the language of information theory. In particular, the existence of a minimal
change in nature implies that the information about a physical system can never be com-
plete, that information transport has its limits and that information can never be fully
trusted. The details of these studies form a fascinating way to look at the microscopic
world.
The analogy between quantum theory and information theory becomes even more
Ref. 116 interesting when the statements are translated into the language of cryptology. Cryptol-
ogy is the science of transmitting hidden messages that only the intended receiver can
decrypt. In our modern times of constant surveillance, cryptology is an important tool
to protect personal freedom.*
The quantum of action implies that messages can be sent in an (almost) safe way. Lis-
tening to a message is a measurement process. Since there is a smallest action, one can
detect whether somebody has tried to listen to a sent message. A man in the middle at-
tack – somebody who pretends to be the receiver and then sends a copy of the message
to the real, intended receiver – can be avoided by using entangled systems as signals to
transmit the information. Quantum cryptologists therefore usually use communication
systems based on entangled photons.
The major issue of quantum cryptology, a large modern research field, is the key dis-
tribution problem. All secure communication is based on a secret key that is used to
decrypt the message. Even if the communication channel is of the highest security – like
entangled photons – one still has to find a way to send the communication partner the se-
cret key necessary for the decryption of the messages. Finding such methods is the main
aspect of quantum cryptology. However, close investigation shows that all key exchange
methods are limited in their security.
In short, due to the quantum of action, nature provides limits on the possibility of
sending encrypted messages. The statement of these limits is (almost) equivalent to the
statement that change in nature is limited by the quantum of action.
* Cryptology consists of the field of cryptography, the art of coding messages, and the field of cryptoanal-
ysis, the art of deciphering encrypted messages. For a good introduction to cryptology, see the text by
Albrecht Beu telspacher, Jörg S chwenk & Klaus-Dieter Wolfenstätter, Moderne Ver-
fahren der Kryptographie, Vieweg 1995.
Is the universe a computer?

The quantum of action provides a limit to secure information exchange. This connection
also allows to brush aside several incorrect statements often found in the media. Stating
Page 95 that ‘the universe is information’ or that ‘the universe is a computer’ is as unreasonable
as saying that the universe is an observation, a measurement apparatus, a clockwork or a
chewing-gum dispenser. Any expert of motion should beware of these and similarly fishy
statements; people who use them either deceive themselves or try to deceive others.
Does the universe have a wave function? And initial conditions?

The wave function of the universe is frequently invoked in discussions about quantum
theory. Many deduce conclusions from this idea, for example on the irreversibility of
time, on the importance of initial conditions, on changes required to quantum theory
and much more. Are these arguments correct?
Page 202 The first thing to clarify is the meaning of ‘universe’. As explained above, the term can
have two meanings: either the collection of all matter and radiation, or this collection
plus all of space-time. Let us also recall the meaning of ‘wave function’: it describes the
state of a system. The state distinguishes two otherwise identical systems; for example,
position and velocity distinguish two otherwise identical ivory balls on a billiard table.
Alternatively and equivalently, the state describes changes in time.
Does the universe have a state? If we take the wider meaning of universe, it does not.
Page 27 Talking about the state of the universe is a contradiction: by definition, the concept of
state, defined as the non-permanent aspects of an object, is applicable only to parts of
the universe.
We then can take the narrower sense of ‘universe’ – the sum of all matter and radiation
only – and ask the question again. To determine the state of all matter and radiation, we
need a possibility to measure it: we need an environment. But the environment of matter
and radiation is space-time only; initial conditions cannot be determined since we need
measurements to do this, and thus an apparatus. An apparatus is a material system with
a bath attached to it; however, there is no such system outside the universe.
In short, quantum theory does not allow for measurements of the universe; therefore
the universe has no state. Beware of anybody who claims to know something about the
wave function of the universe. Just ask him: If you know the wave function of the uni-
verse, why aren’t you rich?
Despite this conclusion, several famous physicists have proposed evolution equations
Ref. 117 for the wave function of the universe. (The best-known is the Wheeler–DeWitt equation.)
It seems a silly point, but the predictions of these equations have not been compared to
experiments; the arguments just given even make this impossible in principle. Exploring
this direction, so interesting it may seem, must therefore be avoided if we want to reach
the top of Motion Mountain.
There are many additional twists to this story. One twist is that space-time itself, even
without matter, might be a bath. This speculation will be shown to be correct later on;
this result seems to allow speaking of the wave function of all matter. But then again, it
turns out that time is undefined at the scales where space-time is an effective bath; this
implies that the concept of state is not applicable there.
A lack of ‘state’ for the universe is a strong statement. It also implies a lack of initial
conditions! The arguments are precisely the same. This is a tough result. We are so used
to think that the universe has initial conditions that we never question the term. (Even
in this text the mistake might appear every now and then.) But there are no initial con-
ditions of the universe.
We can retain as summary, valid even in the light of the latest research: the universe
has no wave function and no initial conditions, independently of what is meant by ‘uni-
verse’.
Chapter 8
COLOURS AND OTHER

INTER ACTIONS BET W EEN LIGHT
AND MAT TER
“ ”
Rem tene; verba sequentur.*
Cato
S
tones have colours. Why? We know how matter and radiation move. The next
tep is to describe the interaction between them. In other words, what is the
pecific way in which charged quantum particles react to electromagnetic fields,
and vice versa? In this chapter, we first give an overview of the ways that colours in na-
ture result from the quantum of action, i.e., from the interaction between matter quan-
tons and photons. Then we explore the simplest such system: we show how the quantum
of action leads to the colours of hydrogen atoms. After this, we discover that the inter-
action between matter and radiation leads to other surprising effects, especially when
special relativity is taken into account.
The causes of colour

Quantum theory explains all colours in nature. Indeed, all the colours that we observe are
due to charged particles. More precisely, colours are due to to the interactions of charged
particles with photons. All colours are thus quantum effects.
The charged particles at the basis of most colours are electrons and nuclei, including
their composites, from ions, atoms and molecules to fluids and solids. An overview of
Ref. 119 the specific mechanisms that generate colour is given in the following table. The table
Challenge 127 s includes all colours that appear in everyday life. (Can you find one that is missing?) Some
colour issues are still topic of research. For example, until recently it was unclear why
exactly asphalt is black. The exact structure of the chemical compounds, the asphaltenes,
that produce the very dark brown colour was unknown. Only recent research has settled
Ref. 120 this question.
* ‘Know the subject and the words will follow.’ Marcus Porcius Cato, (234–149 bce ) or Cato the elder, Ro-
man politician famous for his speeches and his integrity.
146 8 colours and other interactions
TA B L E 7 Causes of colour
Colour type Example D eta i l s
Class I: Colours due to simple excitations

1. Incandescence and free charge radiation
Carbon arc lamp, hot Colours are due to continuous
steel, lightbulb wire, spectrum emitted by all hot
most stars, magma, lava, matter; colour sequence,
hot melts given by Wien’s rule, is black,
red, orange, yellow, white,
blue-white (molten lead and
silver © Graela)
Wood fire, candle Wood and wax flames are
yellow due to incandescence if
carbon-rich and oxygen-poor
White fireworks, Due to metals burning to
flashlamp, sparklers oxide at high temperature,
such as magnesium, zinc, iron,
aluminium or zirconium
(sparkler © Sarah Domingos)
Nuclear reactors, Due to fast free charges:
synchroton light sources, Vavilov–Čerenkov radiation is
free electron lasers due to speed of particle larger
than the speed of light in
matter, Bremsstrahlung is due
to the deceleration of charged
particles (nuclear reactor core
under water, image from
NASA)
2. Atomic gas excitations
Red neon lamp, blue Colours are due to transitions
argon lamp, UV between atomic energy levels
mercury lamp, yellow (gas discharges © Pslawinski)
sodium street lamps,
most gas lasers, metal
vapour lasers, some
fluorescence
Aurora, In air, blue and red colours are
triboluminescence in due to atomic and molecular
scotch tape, energy levels of nitrogen,
crystalloluminescence whereas green, yellow, orange
in strontium bromate colours are due to oxygen
(aurora © Jan Curtis)
Lightning, arcs, sparks, Colour lines are due to energy
coloured fireworks, most levels of highly excited atoms
coloured flames, some (flames of K, Cu, Cs, B, Ca
electroluminescence © Philip Evans)
colours and other interactions bet ween light and mat ter 147
TA B L E 7 Causes of colour (continued)
3. Vibrations and rotations of molecules

Bluish water, blue ice Colours are due to quantized
when clear, violet iodine, levels of rotation and
red-brown bromine, vibrations in molecules (blue
yellow-green chlorine, iceberg © Marc Shandro)
red flames from CN or
blue-green flames from
CH, some gas lasers
Class II: Colours due to ligand field effects
4. Transition metal compounds
Green malachite Colours are due to electronic
Cu2 CO3 (OH)2 , blue states of the ions; phosphors
cobalt oxide, blue are used in cathodes tubes for
azurite TV/computer displays and on
Cu3 (CO3 )2 (OH)2 , red to fluorescent lamp tubes (green
brown hematite Fe2 O3 , malachite on yellow kasolite, a
green MnO, white uranium mineral, picture
Mn(OH)2 , brown width 5 mm, found in Kolwezi,
manganite, chrome Zaire/Congo, © Stephan
green Cr2 O3 , green Wolfsried, television shadow
praesodymium, pink mask photo © Planemad)
europium and yellow
samarium compounds,
piezochromic and
thermochromic
Cr2 O3 − Al2 O3 UV and
electron phosphors,
scintillation, some
fluorescence, some
lasers
5. Transition metal impurities

Ruby, emerald, Electronic states of transition
alexandrite, perovskites, metal ions are excited by light
corresponding lasers and thus absorb specific
wavelengths (ruby on calcite
from Mogok, Myanmar,
picture width 3 cm, © Rob
Lavinsky)
Class III: Colours due to molecular orbitals

6. Organic compounds
Red haemoglobin in Colours are due to conjugated
blood, green chlorophyll π-bonds, i.e. to alternating
in plants, yellow or single and double bonds in
orange carotenes in molecules; floral pigments are
carrots, flowers and almost all anthocyanins,
yellow autumn leaves, betalains or carotenes; used in
red or purple colourants for foods and
anthocyanins in berries, cosmetics, in textile dyes, in
flowers and red autumn electrochromic displays, in
leaves, blue indigo, red inks for colour printers, in
lycopene in tomatoes, photosensitizers (narcissus
red meat from © Thomas Lüthi, blood on
iron-containing finger © Ian Humes, berries
myoglobin, brown © Nathan Wall, hair courtesy
glucosamine in crust of dusdin)
baked food, brown
tannins, black
eumelanin in human
skin, hair and eye,
iron-rich variation
pheomelanin in
redheads, black melanin
also in cut apples and
bananas as well as in
movable sacks in
chameleons,
brown-black asphalt,
some fluorescence,
chemiluminescence,
phosphorescence,
halochromism,
electrochromism and
thermochromism, dye
lasers
Glow-worms, some Bioluminescence is due to
bacteria and funghi, excited molecules, generally
most deep-sea fish, called luciferines (angler fish,
octopi, jellyfish, and length 4.5 cm, © Steve
other deep-sea animals Haddock)
7. Inorganic charge transfer

Blue sapphire, blue lapis Light induces change of
lazuli, green amazonite, position of an electron from
brown-black magnetite one atom to another; for
Fe3 O4 and most other example, in blue sapphire the
iron minerals (colouring transition is between Ti and
basalt black, beer bottles Fe impurities; many paint
brown, quartz sand pigments use charge transfer
yellow, and many other colours; fluorescent analytical
rocks with brown or red reagents are used in molecular
tones), black graphite, medicine and biology
purple permanganate, (magnetite found in Laach,
orange potassium Germany, picture width
dichromate, yellow 10 mm, © Stephan Wolfsried,
molybdates, red Evelien Willemsen)
hematite Fe2 O3 , some
fluorescence
Class IV: Colours due to energy band effects
8. Metallic bands
Gold (green in Colours in reflection and in
transmission), pyrite, transmission are due to
iron, brass, alloys, silver, transitions of electrons
copper, ruby glass between overlapping bands
(saxophone © Selmer)
9. Pure semiconductor bands
Silicon, GaAs, black Colours are due to electron
galena PbS, red cinnabar transitions between separate
HgS, cadmium yellow bands; colour series is black,
CdS, black CdSe, red red, orange, yellow,
CdSx Se1−x , white ZnO, white/colourless; some used
orange vermillion HgS, as pigments (zinc oxide
colourless diamond, courtesy Walkerma)
black to gold
piezochromic SmS
10. Doped semiconductor bands
Blue, yellow, green and Colours are due to transitions
black diamond; LEDs; between dopants and
semiconductor lasers; semiconductor bands
solar cells; ZnS and (e.g. blue diamond: boron
Znx Cd1−x S based and accepters, black diamond:
other phosphors nitrogen donors) (quantum
dots © Andrey Rogach)
11. Colour centres

Amethyst, smoky quartz, Colours are due to colour
fluorite, green diamonds, centres, i.e. to electrons or to
blue, yellow and brown holes bound at crystal
topaz, brown salt, purple vacancies; colour centres are
colour of irradiated usually are created by
glass containing Mn2+ , radiation (amethyst © Rob
lyoluminescence, some Lavinsky)
fluorescence, F-centre
lasers
Some light-dependent The photochromic colouring
sunglasses is due to colour centres
formed by the UV light of the
Sun
Class V: Colours due to physical and geometrical optics
12. Dispersive refraction and polarization
Cut diamond, cut Spectral decomposition
zirconia, halos and sun (sparkle or ‘fire’ of gemstones)
dogs formed by ice is due to dispersion in crystals
crystals in the air (zirconia photo © Gregory
Phillips)
Rainbow Colours of primary and
secondary bow are due to
dispersion in water droplets
Green flash dispersion in the atmosphere
shifts the sun colours
13. Scattering
Blue sky, blue colouring Blue light is scattered more
of distant mountains, than red light by Rayleigh
red sunset; colour scattering, when scatterers
intensification by (molecules, dust) are smaller
pollution; blue quartz than the wavelength of light

(Tokyo sunset © Altus
Plunkett, blue quartz © David
Lynch)
White colour of hair, The white colour is due to

milk, beer foam, clouds, wavelength-independent Mie
fog, cigarette smoke scattering, i.e. scattering at
coming out of lungs, particles larger than the
snow, whipped cream, wavelength of light (snow
shampoo, stars in man © Andreas Kostner)
gemstones
Blue human skin colour Tyndall blue colours are due
in cold weather, blue to scattering on small particles
and green eyes in in front of a dark background
humans, blue monkey (blue poison frog Dendrobates
skin, blue turkey necks, azureus © Lee Hancock)
most blue fish, blue
reptiles, blue cigarette
smoke
Ruby glass The red colour of Murano
glass is due to scattering by
tiny colloidal gold particles
included in the glass in
combination with the metallic
band structure of gold (ruby
glass © murano-glass-shop.it)
Nonlinearities, Raman Frequency-shifting scattering,
effect, potassium second harmonic generation
dihydrogen phosphate and other nonlinearities of
(KDP) certain materials change the
colour of light impinging with
high intensities (800 nm to
400 nm frequency doubling
ring laser © Jeff Sherman)
14. Interference (without diffraction)
Nacre, oil films, soap Thin film interference
bubbles, coatings on produces a standard colour
camera lenses, eyes of sequence that allows precise

cats in the dark, wings thickness determination
of flies and dragonflies, (abalone shell © Anne Elliot)
fish scales, some snakes,
pearls
Polarization colours of Colours are due to
thin layers of interference, as shown by the
birefringent crystals or dependence on layer
thicker layers of stressed thickness (photoelasticity
polymers courtesy Nevit Dilmen)
Supernumerary Due to interference, as shown

rainbows (see page 88) by the dependence on drop
size
Iridescent beetles, Due to scattering at small
butterflies and bird structures or at nanoparticles,
feathers, iridescent as shown by the angular
colours on banknotes dependence of the colour
and on cars (mallard duck © Simon
Griffith)
15. Diffraction (with interference)
Opal Colours are due to the tiny
spheres included in the water
inside the opal; colours can
change if the opal dries out
(polished Brazilian opal
© Opalsnopals)
Aureole, glory, corona Colours are due to diffraction
at the tiny mist droplets
(aeroplane condensation
cloud iridescence © Franz
Kerschbaum)
Diffraction gratings, Colours are due to diffraction
CDs, vinyl records, and interference at tiny,
some beetles and snakes regular pits (CD illuminated
by flashlamp © Alfons
Reichert)
photonic crystals A modern research topic
Cholesteric liquid Colours are due to diffraction

crystals, certain beetles and interference in internal
material layers (liquid crystal
colours © Ingo Dierking)
Class VI: Colours due to eye limitations

Fechner colours, as on lite.bu. Benham’s top Colours are due to different
edu/vision/applets/Color/ speed response of different
Benham/Benham.html photoreceptors
Internal colour production when Phosphenes Occur through pressure

eyes are stimulated (rubbing, sneeze), or with
electric or magnetic fields
Polarization colours Haidinger’s brush See page 90
Colour illusions, as on www.psy. Appearing and Effects are due to
ritsumei.ac.jp/~akitaoka/color9e. disappearing colours combinations of brain
html processing and eye limitations
False colour output, as described Red light can be seen as Observable with adaptive
on page 141 green optics, if red light is focused
on a green-sensitive cone
Colour-blind or ‘daltonic’ person, Protan, deutan or tritan Each type limits colour
see page 144, with reduced colour perception in a different way
spectrum
Colours fascinate. Fascination always also means business; indeed, a large part of the
chemical industry is dedicated to synthesizing colourants for paints, clothes, food and
cosmetics. Also evolution uses the fascination of colours for its own business: propagat-
ing life. The specialists in this domain are the flowering plants. The chemistry of colour
production in plants is extremely involved and at least as interesting as the production
of colours in factories. Practically all flower colourants, from white, yellow, orange, red
to blue, are from three chemical classes: the carotenoids, the anthocyanins (flavonoids)
and the betalains. These colourants are stored in petals inside dedicated containers, the
Ref. 121 vacuoles. There are many good review articles giving more details.
Even though colours are common in plants and animals, most higher animals do not
produce many colourants themselves. For example, humans produce only one colourant:
melanin. (Hemoglobin, which colours blood red, is not a dedicated colourant, but trans-
ports the oxygen from the lungs through the body. Also the pink myoglobin in the mus-
cles is not a dedicated colourant.) Many higher animals, such as birds, need to eat the
colourants that are so characteristic for their appearance. The yellow colour of legs of pi-
geons is an example. It has been shown that the connection between colour and nutrition
is regularly used by potential mates to judge from the body colours whether a proposing
Ref. 122 partner is sufficiently healthy, and thus sufficiently sexy.
In summary, an exploration of the causes of colours found in nature confirms that all
colours are due to quantum effects. We therefore explore the simplest coloured systems
found in nature: atomic gases.
F I G U R E 69 The spectrum of
daylight: a stacked image of
an extended rainbow,
showing its Fraunhofer lines
(© Nigel Sharp, NOAO, FTS,
NSO, KPNO, AURA, NSF)
Using the rainbow to determine what stars are made of
Near the beginning of the eighteenth century, Bavarian instrument-maker Joseph Fraun-
hofer* and the English physicist William Wollaston noted that the rainbow lacks certain
colours. These colours appear as black lines when the rainbow is spread out in sufficient
breadth. Figure 69 shows the lines in detail; they are called Fraunhofer lines today. In
1860, Gustav Kirchhoff and Robert Bunsen showed that the colours missing in the rain-
bow were exactly those colours that certain elements emit when heated. In this way they
managed to show that sodium, calcium, barium, nickel, magnesium, zinc, copper and
iron are present in the Sun. Looking at the rainbow thus tells us what the Sun is made of.
Of the 476 Fraunhofer lines that Kirchhoff and Bunsen observed, 13 did not corre-
spond to any known element. In 1868, Jules Janssen and Joseph Lockyer independently
predicted that these unknown lines were from a new element. It was eventually found
on Earth, in an uranium mineral called cleveite, in 1895. This new element was called
helium, from the Greek word ἥλιος ‘helios’ – Sun. Today we know that it is the second
most widespread ingredient of the Sun and of the universe, after hydrogen. Despite be-
ing so common, helium is rare on Earth because it is a light noble gas that does not form
chemical compounds. Helium thus tends to rise in the atmosphere and escape into space.
Understanding the colour lines produced by each element had started to become in-
teresting already before the discovery of helium; but afterwards the interest increased
* Joseph Fraunhofer (b. 1787 Straubing, d. 1826 München). Bavarian. Having been orphaned at the age of
11, he learned lens-polishing. He taught himself optics from books. He entered an optical company at the
age of 19, ensuring the success of the business by producing the best available lenses, telescopes, microme-
ters, optical gratings and optical systems of his time. He invented the spectroscope and the heliometer. He
Page 279 discovered and counted 476 lines in the spectrum of the Sun; these lines are now named after him. (Today,
Fraunhofer lines are still used as measurement standards: the second and the metre are defined in terms
of them.) Physicists from all over the world would buy their equipment from him, visit him, and ask for
copies of his publications. Even after his death, his instruments remained unsurpassed for generations. With
his telescopes, in 1837 Bessel was able to make the first measurement of parallax of a star, and in 1846 Jo-
hann Gottfried Galle discovered Neptune. Fraunhofer became a professor in 1819. He died young, from the
consequences of the years spent working with lead and glass powder.
F I G U R E 70 A
low-pressure
hydrogen discharge
in a 20 cm long
glass tube (© Jürgen
Bauer at www.
smart-elements.
com)
further, thanks to the increasing number of applications of colour knowledge in chem-

istry, physics, technology, crystallography, biology and lasers. Colours are big business,
as the fashion industry shows.
Colours are specific light frequencies. Light is an electromagnetic wave and is emitted
by moving charges. For a physicist, colours thus result from the interaction of charged
matter with the electromagnetic field. However, sharp colour lines cannot be explained
by classical electrodynamics. Indeed, only quantum theory can explain them – or any
other colour. In fact, every colour in nature is formed with the help of ħ, the quantum of
action.
What determines the colours of atoms?

The simplest colours to study are the sharp colour lines emitted by single atoms. Single
atoms are mainly found in gases. The simplest atom to study is that of hydrogen. Hot
hydrogen gas, shown in Figure 70, emits light consisting of a handful of sharp spectral
lines, as shown on the left of Figure 71. Already in 1885, the Swiss schoolteacher Johann
Balmer (1828–1898) had discovered that the wavelengths of visible hydrogen lines obey
the formula:
1 1 1
= R󶀤 − 2󶀴 . (85)
λm 4 m
Careful measurements, which included the hydrogen’s spectral lines in the infrared and
in the ultraviolet, allowed Johannes Rydberg (1854–1919) to generalize this formula to:
1 1 1
= R󶀤 − 2󶀴 ,
2
(86)
λmn n m
where n and m > n are positive integers, and the so-called Rydberg constant R has the
value 10.97 μm−1 ; easier to remember, the inverse value is 1/R = 91.16 nm. All the colour
lines emitted by hydrogen satisfy this simple formula. Classical physics cannot explain
this result at all. Thus, quantum theory has a clearly defined challenge here: to explain
the formula and the value of R.
Incidentally, the transition λ21 for hydrogen is called the Lyman-alpha line. Its wave-
length, 121.6 nm, lies in the ultraviolet. It is easily observed with telescopes, since most
of the visible stars consist of excited hydrogen. The Lyman-alpha line is routinely used
to determine the speed of distant stars or galaxies, since the Doppler effect changes the
Hydrogen: spectral lines and

energy levels Continuum of ionized states
Energy
n=
8
954.597 nm
3D5/2 3D5/2
nm n=3 3P3/2 , 3D3/2 3P3/2 , 3D3/2
3S1/2 , 3P1/2 3S1/2 F=2

700 3P1/2 F=1
n=2 2P3/2 2P3/2 F=1
656.2852 nm
2S1/2 , 2P1/2 2S1/2 F=0
656.272 nm 2P1/2
650
F=1
F=0
600
550
500
486.133 nm
n=1
450 1S1/2 F=1

434.047 nm 1S1/2 F=0
410.174 nm
400 nonrelativistic relativistic virtual particle nuclear levels
397.007 nm
(Bohr) levels (Sommerfeld- levels (with at higher scale
Dirac) levels Lamb shift) (hyperfine
(fine structure) structure)
F I G U R E 71 Atomic hydrogen: the visible spectrum of hydrogen (NASA) and its calculated energy levels,
in four approximations of increasing precision. Can you associate the visible lines to the correct level
transitions?
Ref. 123 wavelength when the speed is large. The record so far is a galaxy found in 2004, with a
Lyman-alpha line shifted to 1337 nm. Can you calculate the speed with which it is mov-
Challenge 128 ny ing away from the Earth?
There are many ways to deduce Balmer’s formula from the minimum action. The first
way was found by Niels Bohr. Then, in 1926, Erwin Schrödinger solved his equation of
motion for an electron moving in the electrostatic potential V (r) = e 2 /4πε0 r of a point-
like proton. By doing so, Schrödinger deduced Balmer’s formula and became famous in
the world of physics. However, this important calculation is long and complex. In order
to understand hydrogen colours, it is not necessary to solve an equation of motion; it is
sufficient to compare the energies of the initial and final states. This can be done most
easily by noting that a specific form of the action must be a multiple of ħ/2. This approach
was developed by Einstein, Brillouin and Keller, and is now named after them. It relies
Ref. 124 on the fact that the action S of any quantum system obeys
1 μ
S= 󵐔 dqi pi = 󶀤ni + i 󶀴 ħ (87)
2π 4
for every coordinate qi and its conjugate momentum pi . Here, ni can be zero or any
positive integer, and μi is the so-called Maslov index, an even integer, which in the case
of atoms has the value 2 for the radial and azimuthal coordinates r and θ, and 0 for the
rotation angle φ. The integral is to be taken along a full orbit. In simple words, the action
S is a half-integer multiple of the quantum of action. This result can be used to calculate
the energy levels.
Any rotational motion in a spherical potential V (r) is characterized by a constant
energy E and constant angular momenta L and Lz . Therefore the conjugate momenta for
the coordinates r, θ and φ are
Challenge 129 ny
L2
pr = 󵀌2m(E − V (r)) −
r2
Lz2
pθ = 󵀌 L 2 −
sin2 θ
pφ = L z . (88)
Using these expressions in equation (87) and setting n = nr + nθ + nφ + 1, we get* the

result
1 me 4 R 2.19 aJ 13.6 eV
En = − 2 =− 2 ≈− 2 ≈− . (91)
n 2(4πε0 )2 ħ2 n n n2
These energy levels En are shown in Figure 71.

Using the idea that a hydrogen atom emits a single photon when its electron changes
from state En to Em , we get exactly the formula deduced by Balmer and Rydberg from
Challenge 131 e observations! The match between observation and calculation is about four digits. For
the first time ever, a material property, the colour of hydrogen atoms, had been explained
from a fundamental principle of nature. Key to this explanation was the quantum of
action. (This whole discussion assumes that the electrons in hydrogen atoms that emit
Challenge 132 ny light are in eigenstates. Can you argue why this is the case?)
In short, the quantum of action implies that only certain specific energy values for
* The calculation is straightforward. After insertion of V (r) = e/4πε0 r into equation (88) one needs to
Challenge 130 ny perform the (tricky) integration. Using the general result
1 dz 󵀄 2 B
󵐔 Az + 2Bz − C = −󵀂C + (89)
2π z 󵀂−A
one gets
1 e2 󵀌 m
󶀤nr + 󶀴 ħ + L = nħ = . (90)
2 4πε0 −2E
This leads to the energy formula (91).
an electron are allowed inside an atom. The lowest energy level, for n = 1, is called the
ground state. Its energy value 2.19 aJ is the ionization energy of hydrogen; if that energy is
added to the ground state, the electron is no longer bound to the nucleus. The ionization
energy thus plays the same role for electrons around atoms as does the escape velocity
for satellites around planets.
The calculation also yields the effective radius of the electron orbit in hydrogen. It is
given by
ħ2 4πε0 ħ
r n = n2 2
= = n2 a0 ≈ n2 52.918 937 pm . (92)
me e me cα
In contrast to classical physics, quantum theory allows only certain specific orbits around
Page 159, page 166 the nucleus. (For more details about the fine structure constant α, see below.) The small-
est value, 53 pm for n = 1, is called the Bohr radius, and is denoted by a0 . To be more
precise, these radii are the average sizes of the electron clouds surrounding the nucleus.
Quantum theory thus implies that a hydrogen atom excited to the level n = 500 is about
Ref. 125 12 μm in size: larger than many bacteria! Such blown-up atoms, usually called Rydberg
atoms, have indeed been observed in the laboratory, although they are extremely sensi-
tive to perturbations.
The orbital frequency of electrons in hydrogen is
1 e 4 me 1 me c 2 α 2 6.7 PHz
fn = = ≈ (93)
n3 4ε02 h3 n3 h n3
and the electron speed is
1 e2 αc 2.2 Mm/s 0.007 c

󰑣n = = ≈ ≈ . (94)
n 4πε0 ħ n n n
As expected, the further the electron’s orbit is from the nucleus, the more slowly it moves.
This result can also be checked by experiment: exchanging the electron for a muon allows
us to measure the time dilation of its lifetime. Measurements are in full agreement with
Ref. 126 the calculations.
In 1915, Arnold Sommerfeld understood that the analogy of electron motion with or-
bital motion could be continued in two ways. First of all, electrons can move in ellipses.
The quantization of angular momentum implies that only selected eccentricities are pos-
sible. The higher the angular momentum, the larger the number of possibilities: the first
are shown in Figure 72. The highest eccentricity corresponds to the minimum value l = 0
of the so-called azimuthal quantum number, whereas the case l = n − 1 correspond to
Ref. 127 circular orbits. In addition, the ellipses can have different orientations in space.
The second point Sommerfeld noted was that the speeds of the electron in hydro-
gen are slightly relativistic. The above calculation did not take into account relativistic
effects. However, high-precision measurements show slight differences between the non-
relativistically calculated energy levels and the measured ones.
0 0.2 0.4 0.6 0.8 1 nm

nucleus n=1, l=0
(not to scale)
n=2, l=1 n=2, l=0
n=3, l=2 n=3, l=1
n=3, l=0
F I G U R E 72 The imagined, but not existing and thus false electron orbits of the Bohr–Sommerfeld
model of the hydrogen atom (left) and the correct description, using the probability density of the
electron in the various states (right) (© Wikimedia).
Relativistic hydrogen
In the relativistic case, too, the EBK action has to be a multiple of ħ/2. From the relativistic
Ref. 124 expression for the kinetic energy of the electron
e2
E + mc 2 = 󵀆 p2 c 2 + m2 c 4 − (95)
4πε0 r
Challenge 133 ny we get the expression
E 2me 2 E
p2r = 2mE 󶀤1 + 2
󶀴 + 󶀤1 + 󶀴 . (96)
2mc 4πε0 r mc 2
We now introduce, for convenience, the so-called fine structure constant, as α =

e 2 /(4πε0 ħc) = 󵀄4πħR/mc ≈ 1/137.036. (α is a dimensionless constant; R = 10.97 μm−1
Challenge 134 ny is the Rydberg constant.) The radial EBK action then implies that
mc 2
Enl + mc 2 = . (97)
α2
󵀌1 + 2
󶀤n−l− 21 +󵀄(l+ 21 )2 −α 2 󶀴
This result, first found by Arnold Sommerfeld in 1915, is correct for point-like electrons.
In reality, the electron has spin 1/2; the correct relativistic energy levels thus appear when
F I G U R E 73 Paul Dirac (1902–1984)
we set l = j ± 1/2 in the above formula. The result can be approximated by
R α2 n 3
En j = − 󶀧1 + 󶀧 1
− 󶀷 + ...󶀷 . (98)
n 2 n 2
j+ 2
4
It reproduces the hydrogen spectrum to an extremely high accuracy. If we compare the

result with the non-relativistic one, we note that each non-relativistic level n is split in
n different levels. This is shown in Figure 71. In practice, in precision experiments the
lines of the hydrogen spectrum have a so-called fine structure. The magnitude of the
fine structure depends on α, a fundamental constant of nature. Since the importance of
this fundamental constant was discovered in this context, the name chosen by Arnold
Sommerfeld, the fine structure constant, has been taken over across the world. The fine
structure constant describes the strength of the electromagnetic interaction.
Of course, it is possible to do even better. The introduction of virtual-particle effects
and the coupling of the proton spin gives additional corrections, as shown in Figure 71.
But that is still not all, isotope effects, Doppler shifts and level shifts to environmental
electric or magnetic fields also influence the hydrogen spectrum.
Relativistic wave equations – again
“ ”
The equation was more intelligent than I was.
Paul Dirac, about his equation, repeating
a statement made by Heinrich Hertz.
What is the evolution equation for the wave function in the case that relativity, spin and
interactions with the electromagnetic field are taken into account? We could try to gener-
Page 89 alize the representation of relativistic motion given by Foldy and Wouthuysen to the case
of particles with electromagnetic interactions. Unfortunately, this is not a simple matter.
The simple identity between the classical and quantum-mechanical descriptions is lost if
electromagnetism is included.
Charged quantum particles are best described by another, equivalent representation
of the Hamiltonian, which was discovered much earlier, in 1926, by the British physicist
Paul Dirac.* Dirac found a neat trick to take the square root appearing in the relativistic
* Paul Adrien Maurice Dirac (b. 1902 Bristol, d. 1984 Tallahassee), British physicist, son of a Francophone
Swiss immigrant. He studied electrotechnics in Bristol, then went to Cambridge, where he later became a
F I G U R E 74 The famous
Zitterbewegung: the
superposition of positive and
negative energy states leads
to an oscillation around a
mean vale. Colour indicates
phase; two coloured curves
are shown, as the Dirac
equation in one dimension
has only two components (not
four); the grey curve is the
probability density. (QuickTime
energy operator. In Dirac’s representation, the Hamilton operator is given by
HDirac = βm + α ⋅ p . (99)
The quantities β and the three components (α1 , α2 , α3 ) = α turn out to be complex 4 × 4
matrices.
In Dirac’s representation, the position operator x is not the position of a particle, but
has additional terms; its velocity operator has only the eigenvalues plus or minus the ve-
locity of light; the velocity operator is not simply related to the momentum operator; the
equation of motion contains the famous ‘Zitterbewegung’ term; orbital angular momen-
tum and spin are not separate constants of motion.
So why use this horrible Hamiltonian? Because only the Dirac Hamiltonian can easily
be used for charged particles. Indeed, it is transformed to the Hamiltonian coupled to the
professor, holding the chair that Newton had once held. In the years from 1925 to 1933 he published a stream
of papers, of which several were worth a Nobel Prize; he received this in 1933. He unified special relativity
and quantum theory, predicted antimatter, worked on spin and statistics, predicted magnetic monopoles,
speculated on the law of large numbers, and more besides. His introversion, friendliness and shyness, and
his deep insights into nature, combined with a dedication to beauty in theoretical physics, made him a
legend all over the world during his lifetime. For the latter half of his life he tried, unsuccessfully, to find an
alternative to quantum electrodynamics, of which he was the founder, as he was repelled by the problems
of infinities. He died in Florida, where he lived and worked after his retirement from Cambridge.
Page 71 electromagnetic field by the so-called minimal coupling, i.e., by the substitution
p → p − qA , (100)
that treats electromagnetic momentum like particle momentum. With this prescription,
Dirac’s Hamiltonian describes the motion of charged particles interacting with an elec-
tromagnetic field A. The minimal coupling substitution is not possible in the Foldy–
Wouthuysen Hamiltonian. In the Dirac representation, particles are pure, point-like,
structureless electric charges; in the Foldy–Wouthuysen representation they acquire a
Ref. 128 charge radius and a magnetic-moment interaction. (We will come to the reasons below,
in the section on QED.)
In more detail, the simplest description of an electron (or any other elementary, stable,
electrically-charged particle of spin 1/2) is given by the action S and Lagrangian
S = 󵐐 LQED d 4 x where (101)
1
/ − mc 2 ) ψ −
LQED = ψ (iħcD F F μ󰜈 and where
4μ0 μ󰜈
/ μ = γ μ (∂ μ − ieA μ )
D
The first, matter term in the Lagrangian leads to the Dirac equation: it describes how
elementary, charged, spin 1/2 particles are moved by electromagnetic fields. The sec-
ond, radiation term leads to Maxwell’s equations, and describes how electromagnetic
fields are moved by the charged particle wave function. Together with a few calculating
tricks, these equations describe what is usually called quantum electrodynamics, or QED
for short.
As far as is known today, the relativistic description of the motion of charged mat-
ter and electromagnetic fields given the QED Lagrangian (101) is perfect: no differences
between theory and experiment have ever been found, despite intensive searches and de-
spite a high reward for anybody who would find one. All known predictions completely
correspond with the measurements. In the most spectacular cases, the correspondence
between theory and measurement extends to more than thirteen digits. But even more
interesting than the precision of QED are certain of its features that are missing in classical
electrodynamics. Let’s have a quick tour.
Getting a feeling for the Dirac equation

The QED Lagrangian implies that the wave function of a charged particle in a potential
follows the Dirac equation:
iħcγ μ (∂ μ − ieA μ )ψ = mcψ . (102)
The many indices should not make us forget that this equation simply states that the
eigenvalue of the energy–momentum operator is the rest mass (times the speed of light
c). In other words, the equation states that the wave ψ moves with a phase velocity c.
The wave function ψ has four complex components. Two describe the motion of par-
ticles, and two the motion of antiparticles. Each type of particle needs two complex com-
ponents, because the equation describes spin and particle density. Spin is a rotation, and
a rotation requires three real parameters. Spin and density thus require four real parame-
ters; they can be combined into two complex numbers, both for particles and for antipar-
ticles.
Each of the four components of the wave function of a relativistic spinning particle
Challenge 135 e follows the relativistic Schrödinger–Klein–Gordon equation. This means that the rela-
tivistic energy–momentum relation is followed by each component separately.
The relativistic wave function ψ has the important property that a rotation by 2π
Challenge 136 e changes its sign. Only a rotation by 4π leaves the wave function unchanged. This is the
typical behaviour of spin 1/2 particles. For this reason, the four-component wave func-
tion of a spin 1/2 particle is called a spinor.
Antimatter
‘Antimatter’ is now a household term. Interestingly, the concept appeared before there
was any experimental evidence for it. The relativistic expression for the energy E of an
electron with charge e in the field of a charge Q is
Qe
E+ = 󵀆m2 c 4 + p2 c 2 . (103)
4πε0 r
This expression also allows solutions with negative energy and opposite charge −e. Quan-
tum theory shows that this is a general property, and these solutions correspond to what
is called antimatter.
Indeed, the antimatter companion of the electron was predicted in the 1920s by Paul
Dirac from his equation. Unaware of this prediction, Carl Anderson discovered the anti-
electron in 1932, and called it the positron. (The correct name would have been ‘positon’,
without the ‘r’. This correct form is used in the French language.) Anderson was studying
cosmic rays, and noticed that some ‘electrons’ were turning the wrong way in the mag-
netic field he had applied to his apparatus. He checked his apparatus thoroughly, and
finally deduced that he had found a particle with the same mass as the electron but with
positive electric charge.
The existence of positrons has many strange implications. Already in 1928, before their
discovery, the Swedish theorist Oskar Klein had pointed out that Dirac’s equation for
electrons makes a strange prediction: when an electron hits a sufficiently steep potential
wall, the reflection coefficient is larger than unity. Such a wall will reflect more than is
thrown at it. In addition, a large part of the wave function is transmitted through the
wall. In 1935, after the discovery of the positron, Werner Heisenberg and Hans Euler
Ref. 129 explained the paradox. They found that the Dirac equation predicts that whenever an
electric field exceeds the critical value of
me c 2 me2 c 3
Ec = = = 1.3 EV/m , (104)
eλe eħ
the vacuum will spontaneously generate electron–positron pairs, which are then sepa-
F I G U R E 75 Klein’s paradox:
the motion of a relativistic
wave function that encounters
a very steep potential. Part of
the wave function is
transmitted; this part is
antimatter, as the larger lower
component shows. (QuickTime
rated by the field. As a result, the original field is reduced. This so-called vacuum po-
larization is the reason for the reflection coefficient greater than unity found by Klein.
Indeed, steep potentials correspond to high electric fields.
Vacuum polarization is a weak effect. It has been only observed in collisions of high
Vol. V, page 111 energy, where it the effectively increases the fine structure constant. Later on we will de-
scribe truly gigantic examples of vacuum polarization that are postulated around charged
black holes.
Vacuum polarization shows that, in contrast to everyday life, the number of particles
is not a constant in the microscopic domain. Only the difference between particle number
and antiparticle number turns out to be conserved. Vacuum polarization thus limits our
possibility to count particles in nature!
Of course, the generation of electron–positron pairs is not a creation out of nothing,
but a transformation of energy into matter. Such processes are part of every relativistic
description of nature. Unfortunately, physicists have a habit of calling this transformation
‘pair creation’, thus confusing the issue somewhat.
Virtual particles
Despite what was said so far, actions smaller than the minimal action do have a role to
play. We have already encountered one example: in a collision between two electrons,
Page 54 there is an exchange of virtual photons. We learned that the exchanged virtual photon
cannot be observed. Indeed, the action S for this exchange obeys
S⩽ħ. (105)
In short, virtual particles appear only as mediators in interactions. They cannot be ob-
served. Virtual particles, in contrast to ordinary, real particles, do not obey the relation
E 2 − p2 c 2 = m2 c 4 . For example, the kinetic energy can be negative. Indeed, virtual parti-
cles are the opposite of ‘free’ or real particles. They may be observed in a vacuum if the
measurement time is very short. They are intrinsically short-lived.
Virtual photons are the cause for electrostatic potentials, for magnetic fields, for the
Casimir effect, for spontaneous emission, for the van der Waals force, and for the Lamb
shift in atoms. A more detailed treatment shows that in every situation with virtual pho-
tons there are also, with even lower probability, virtual electrons and virtual positrons.
Massive virtual particles are essential for vacuum polarization, for the limit in the
number of the elements, for black-hole radiation and for Unruh radiation. Massive vir-
tual particles also play a role in the strong interaction, where they hold the nucleons
together in nuclei, and in weak nuclear interaction, where they explain why beta decay
happens and why the Sun shines.
In particular, virtual particle–antiparticles pairs of matter and virtual radiation parti-
cles together form what we call the vacuum. In addition, virtual radiation particles form
what are usually called static fields. Virtual particles are needed for a full description of
all interactions. In particular, virtual particles are responsible for every decay process.
Curiosities and fun challenges about colour

Where is the sea bluest? Sea water is blue because it absorbs red and green light. Sea
water can also be of bright colour if the sea floor reflects light. In addition, sea water can
be green, when it contains small particles that scatter or absorb blue light. Most often,
these particles are soil or plankton. (Satellites can determine plankton content from the
‘greenness’ of the sea.) Thus the sea is especially blue if it is deep, quiet and cold; in that
case, the ground is distant, soil is not mixed into the water, and the plankton content is
low. The Sargasso Sea is deep, quiet and cold for most of the year. It is often called the
bluest of the Earth’s waters.
∗∗
If atoms contain orbiting electrons, the rotation of the Earth, via the Coriolis acceleration,
Ref. 126 should have an effect on their motion. This beautiful prediction is due to Mark Silverman;
the effect is so small, however, that is has not yet been measured.
∗∗
Light is diffracted by material gratings of light. Can matter be diffracted by light gratings?
Surprisingly, it actually can, as predicted by Dirac and Kapitza in 1937. This was accom-
Ref. 133 plished for the first time in 1986, using atoms. For free electrons, the feat is more difficult;
the clearest confirmation came in 2001, when new laser technology was used to perform
a beautiful measurement of the typical diffraction maxima for electrons diffracted by a
light grating.
∗∗
Light is totally reflected when it is directed to a dense material at a large enough angle
so that it cannot enter the material. A group of Russian physicists have shown that if the
Ref. 126 dense material is excited, the intensity of the totally-reflected beam can be amplified. It
is unclear whether this will ever lead to applications.
∗∗
The ways people handle single atoms with electromagnetic fields provide many beautiful
examples of modern applied technologies. Nowadays it is possible to levitate, to trap, to
Vol. I, page 272 excite, to photograph, to deexcite and to move single atoms just by shining light onto
Ref. 130 them. In 1997, the Nobel Prize in Physics has been awarded to the originators of the field,
Steven Chu, Claude Cohen-Tannoudji and William Philips.
∗∗
An example of modern research is the study of hollow atoms, i.e., atoms missing a number
of inner electrons. They have been discovered in 1990 by J.P. Briand and his group. They
appear when a completely ionized atom, i.e., one without any electrons, is brought in
contact with a metal. The acquired electrons then orbit on the outside, leaving the inner
shells empty, in stark contrast with usual atoms. Such hollow atoms can also be formed
Ref. 131 by intense laser irradiation.
∗∗
Ref. 132 Given two mirrors and a few photons, it is possible to capture an atom and keep it floating
between the two mirrors. This feat, one of several ways to isolate single atoms, is now
Challenge 137 ny standard practice in laboratories. Can you imagine how it is done?
Material properties
The quantum of action determines the colours of the hydrogen atoms. In the same way,
it determines the colours of all the other atoms whose Fraunhofer lines are observed
in the infrared, visible and ultraviolet. In fact, also the colour of solids and liquids are
determined by the quantum of action.
The quantum of action also determines all other material properties. The elasticity,
the plasticity, the brittleness, the magnetic and electric properties of materials are equally
fixed by the quantum of action. Many details of this general statement are, however, still
a subject of research. Many material properties are not completely understood, though
none is in contradiction with the quantum of action. Material research is among the most
important fields of modern science, and most advances in the standard of living result
from it. However, we will not explore it much in the following.
In summary, material research has confirmed that quantum physics is the correct de-
scription of materials. And it has confirmed that all material properties of everyday life
are of electromagnetic origin.
The strength of electromagnetism

The great physicist Wolfgang Pauli used to say that after his death, the first thing he would
ask the devil would be to explain Sommerfeld’s fine-structure constant. (Others used to
comment that after the devil had explained it to him, he would think a little, and then
snap: ‘Wrong!’)
Page 159 The fine-structure constant was introduced by Arnold Sommerfeld. It is the dimen-
Ref. 134 sionless constant of nature given by
e2 1
α= ≈ ≈ 0.007 297 352 5376(50) . (106)
4πε0 ħc 137.035 999 679(94)
This number first appeared in explanations of the fine structure of atomic colour spectra;
hence its strange name. Sommerfeld was the first to understand its general importance.
It is central to quantum electrodynamics for several reasons. First of all, it describes the
strength of electromagnetism. Since all charges are multiples of the electron charge, a
higher value for the fine structure constant α would mean a stronger attraction or re-
pulsion between charged bodies. Thus the value of α determines the sizes of atoms, and
indeed of all things, as well as all colours in nature.
Secondly, it is only because the number α is so small that we are able to talk about
particles at all. The argument is somewhat involved; it will be given in detail later on. In
any case, the small value of the fine-structure constant makes it possible to distinguish
particles from each other. If the number were near to or larger than one, particles would
interact so strongly that it would not be possible to observe them or to talk about them
at all.
This leads on to the third reason for the importance of the fine-structure constant.
Since it is a dimensionless number, it implies some yet-unknown mechanism that fixes its
value. Uncovering this mechanism is one of the challenges remaining in our adventure.
As long as the mechanism remains unknown – as was the case in 2007 – we do not
understand the colour and size of a single thing around us!
Small changes in the strength of electromagnetic attraction between electrons and
protons would have numerous important consequences. Can you describe what would
happen to the size of people, to the colour of objects, to the colour of the Sun, or to the
workings of computers, if the strength were to double? And what if it were to gradually
Challenge 138 s drop to half its usual value?
Since the 1920s, explaining the value of α has been seen as one of the toughest chal-
lenges facing modern physics. That is the reason for Pauli’s request to the devil. In 1946,
during his Nobel Prize lecture, he repeated the statement that a theory that does not de-
Ref. 135 termine this number cannot be complete. Since that time, physicists seem to have fallen
into two classes: those who did not dare to take on the challenge, and those who had no
clue. This fascinating story still awaits us.
The problem of the fine-structure constant is so deep that it leads many astray. For
example, it is sometimes said that it is impossible to change physical units in such a way
that ħ, c and e are all equal to 1 at the same time, because to do so would change the
Challenge 139 s number α = 1/137.036.... Can you show that the argument is wrong?
A summary on colours and materials

In summary, the interaction of electromagnetic fields and the electrons inside atoms,
molecules, liquids and solids determines the site, the shape, the colour and the material
Motion Mountain – The Adventure of Physics pdf file available free of charge at www.motionmountain.net Copyright © Christoph Schiller November 1997–July 2010
8 colours and other interactions
properties of all things around us.

168
Chapter 9
QUANTUM PHYSIC S I N A NUT SHELL
C
ompared to classical physics, quantum theory is remarkably more
omplex. The basic idea however, is simple: in nature there is a minimum
hange, or a minimum action, with the value ħ = 1.1 ⋅ 10−34 Js. The minimum
action leads to all the strange observations made in the microscopic domain, such as
wave behaviour of matter, tunnelling, indeterminacy relations, randomness in measure-
ments, quantization of angular momentum, pair creation, decay, indistinguishability
and particle reactions.
The essence of quantum theory is thus the lack of the infinitely small. The mathematics
of quantum theory is often disturbingly involved. Was this part of our walk worth the
effort? It was; the results are profound and the accuracy of the description is excellent.
We first give an overview of these results and then turn to the questions that are still left
open.
Physical results of quantum theory
“ ”
Deorum offensae diis curae.
Voltaire, Traité sur la tolérance.
All of quantum theory can be resumed in two sentences.
⊳ In nature, actions smaller than ħ = 1.1 ⋅ 10−34 Js are not observed.
⊳ All intrinsic properties in nature – with the exception of mass – such as elec-
tric charge, spin, parities, etc., appear as integer numbers; in composed systems
they either add or multiply.
The existence of a smallest action in nature directly leads to the main lesson we learned
about motion in the quantum part of our adventure:
⊳ If it moves, it is made of quantons, or quantum particles.
⊳ In nature, all intrinsic properties – with the exception of mass – such as

electric charge, spin, parities, etc., appear as integer numbers. Since all physical
systems are made of quantons, in composed systems intrinsic properties either
add or multiply.
170 9 quantum physics in a nu tshell
These statements applies to every physical system, thus to all objects and to all images, i.e.,
to all matter and radiation. Moving stuff is made of quantons. Stones, water waves, light,
sound waves, earthquakes, gelatine and everything else we can interact with is made of
quantum particles.
Once we asked: what is matter and what are interactions? Now we know: they are
composites of elementary quantum particles. An elementary quantum particle is a count-
able entity, smaller than its own Compton wavelength, described by energy–momentum,
mass, spin, C, P and T parity. As we will see in the next volume however, this is not yet
the complete list of intrinsic particle properties.
All moving entities are made of quantum particles. To see how deep this result is, you
can apply it to all those moving entities for which it is usually forgotten, such as ghosts,
spirits, angels, nymphs, daemons, devils, gods, goddesses and souls. You can check your-
Challenge 140 e self what happens when their particle nature is taken into account.
Motion of quantum particles
Quantons, or quantum particles, differ from everyday particles: quantum particles inter-
fere: they behave like a mixture of particles and waves. This property follows directly from
the existence of ħ, the smallest action in nature. From the existence of ħ, quantum theory
deduces all its statements about quantum particle motion. We now summarize the main
ones.
There is no rest in nature. All objects obey the indeterminacy principle, which states
that the indeterminacies in position x and momentum p follow
ΔxΔp ⩾ ħ/2 with ħ = 1.1 ⋅ 10−34 Js (107)
and making rest an impossibility. The state of quantum particles is defined by the same
observables as in classical physics, with the difference that observables do not commute.
Classical physics appears in the limit that the Planck constant ħ can effectively be set to
zero.
Quantum theory introduces a probabilistic element into motion. It results from the
minimum action value through the interactions with the baths in the environment of
any system.
Quantum particles behave like waves. The associated de Broglie wavelength λ is given
by the momentum p through
h 2πħ
λ= = (108)
p p
both in the case of matter and of radiation. This relation is the origin of the wave be-
haviour of light and matter. The light particles are called photons; their observation is
now standard practice. Quantum theory states that particle waves, like all waves, inter-
fere, refract, disperse, dampen, can be dampened and can be polarized. This applies to
photons, electrons, atoms and molecules. All waves being made of quantum particles,
all waves can be seen, touched and moved. Light for example, can be ‘seen’ in photon-
photon scattering, can be ‘touched’ using the Compton effect, and can be ‘moved’ by
gravitational bending. Matter particles, such as molecules or atoms, can be seen in elec-
quantum physics in a nu tshell 171
tron microscopes and can be touched and moved with atomic force microscopes. The
interference and diffraction of wave particles is observed daily in the electron micro-
scope.
Matter waves can be imagined as rotating clouds. In the limit of negligible cloud size,
quantum particles can be imagined as rotating little arrows.
Particles cannot be enclosed. Even though matter is impenetrable, quantum theory
shows that tight boxes or insurmountable obstacles do not exist. Waiting long enough
always allows to overcome boundaries, since there is a finite probability to overcome any
obstacle. This process is called tunnelling when seen from the spatial point of view and is
called decay when seen from the temporal point of view. Tunnelling explains the working
of television tubes as well as radioactive decay.
All particles and all particle beams can be rotated. Particles possess an intrinsic an-
gular momentum called spin, specifying their behaviour under rotations. Bosons have
integer spin, fermions have half integer spin. An even number of bound fermions or any
number of bound bosons yield a composite boson; an odd number of bound fermions
or an infinite number of interacting bosons yield a low-energy fermion. Solids are im-
penetrable because of the fermion character of its electrons in the atoms.
Identical particles are indistinguishable. Radiation is made of indistinguishable parti-
cles called bosons, matter of fermions. Under exchange, fermions commute at space-like
separations, whereas bosons anticommute. All other properties of quantum particles are
the same as for classical particles, namely countability, interaction, mass, charge, angu-
lar momentum, energy, momentum, position, as well as impenetrability for matter and
penetrability for radiation. Perfect copying machines do not exist.
In collisions, particles interact locally, through the exchange of other particles. When
matter particles collide, they interact through the exchange of virtual bosons, i.e., off-
shell bosons. Motion change is thus due to particle exchange. Exchange bosons of even
spin mediate only attractive interactions. Exchange bosons of odd spin mediate repulsive
interactions as well.
The properties of collisions imply the existence of antiparticles, as regularly observed
in experiments. Elementary fermions, in contrast to many elementary bosons, differ
from their antiparticles; they can be created and annihilated only in pairs. Apart from
neutrinos, elementary fermions have non-vanishing mass and move slower than light.
Images, made of radiation, are described by the same properties as matter. Images can
only be localized with a precision of the wavelength λ of the radiation producing them.
The appearance of Planck’s constant ħ implies that length scales and time scales exist
in nature. Quantum theory introduces a fundamental jitter in every example of motion.

Thus the infinitely small is eliminated. In this way, lower limits to structural dimensions
and to many other measurable quantities appear. In particular, quantum theory shows
that it is impossible that on the electrons in an atom small creatures live in the same way
that humans live on the Earth circling the Sun. Quantum theory shows the impossibility
of Lilliput.
Clocks and metre bars have finite precision, due to the existence of a smallest action
and due to their interactions with baths. On the other hand, all measurement apparatuses
must contain baths, since otherwise they would not be able to record results.
Ref. 118 Quantum physics leaves no room for cold fusion, astrology, teleportation, telekinesis,
supernatural phenomena, multiple universes, or faster than light phenomena – the EPR
paradox notwithstanding.
Achievements in precision
Apart from the conceptual changes, quantum theory improved the accuracy of predic-
tions from the few – if any – digits common in classical mechanics to the full number of
digits – sometimes thirteen – that can be measured today. The limited precision is usu-
ally not given by the inaccuracy of theory, it is given by the measurement accuracy. In
other words, the agreement is only limited by the amount of money the experimenter is
willing to spend. Table 8 shows this in more detail.
TA B L E 8 Selected comparisons between classical physics, quantum theory and experiment
O b s e r va b l e Classi- Prediction of Measure- Cost

cal q ua n t um ment esti-
predic - theorya m at e
tion
Simple motion of bodies
Indeterminacy 0 ΔxΔp ⩾ ħ/2 (1 ± 10−2 ) ħ/2 10 k€
Matter wavelength none λp = 2πħ (1 ± 10−2 ) ħ 10 k€
Tunnelling rate in alpha 0 1/τ is finite (1 ± 10−2 ) τ 5 k€
decay
Compton wavelength none λc = h/me c (1 ± 10−3 ) λ 20 k€
Pair creation rate 0 σE agrees 100 k€
Radiative decay time in none τ ∼ 1/n3 (1 ± 10−2 ) 5 k€
hydrogen
Smallest angular 0 ħ/2 (1 ± 10−6 ) ħ/2 10 k€
momentum
Casimir effect/pressure 0 p = (π2 ħc)/(240r 4 ) (1 ± 10−3 ) 30 k€
Colours of objects
Spectrum of hot objects diverges λmax = hc/(4.956 kT ) (1 ± 10−4 ) Δλ 10 k€
Lamb shift none Δλ = 1057.86(1) MHz (1 ± 10−6 ) Δλ 50 k€
Rydberg constant none R∞ = me cα 2 /2h (1 ± 10−9 ) R∞ 50 k€
Stefan–Boltzmann none σ = π2 k 4 /60ħ3 c 2 (1 ± 3 ⋅ 10−8 ) σ 20 k€
constant
Wien’s displacement none b = λmax T (1 ± 10−5 ) b 20 k€
constant
Refractive index of water none 1.34 a few % 1 k€
Photon-photon scattering 0 from QED: finite agrees 50 M€
Electron gyromagnetic 1 or 2 2.002 319 304 3(1) 2.002 319 304 30 M€
ratio 3737(82)
Composite matter properties
Atom lifetime ≈ 1 μs ∞ > 1020 a 1€
Molecular size and shape none from QED within 10−3 20 k€
quantum physics in a nu tshell 173
Page 175 a. All these predictions are calculated from the fundamental quantities given in Appendix A.
We notice that the predicted values are not noticeably different from the measured ones.
If we remember that classical physics does not allow to calculate any of the measured
values, we get an idea of the progress quantum physics has brought. But despite this
impressive agreement, there still are unexplained observations: the one we have encoun-
tered so far is the fine structure constant. The measured value is α = 1/137.035 9991(1),
but no reason for this value has appeared. This is the main open problem – some would
even say, the only open problem – of the electromagnetic interaction.
In summary, in the microscopic domain we are left with the impression that quantum
theory is in perfect correspondence with nature; despite prospects of fame and riches,
despite the largest number of researchers ever, no contradiction with observation has
been found yet.
Is quantum theory magic?
Studying nature is like experiencing magic. Nature often looks different from what it is.
During magic we are fooled – but only if we forget our own limitations. Once we start to
see ourselves as part of the game, we start to understand the tricks. That is the fun of it.
The same happens in physics.
∗∗
The world looks irreversible, even though it isn’t. We never remember the future. We are
fooled because we are macroscopic.
∗∗
The world looks decoherent, even though it isn’t. We are fooled again because we are
macroscopic.
∗∗
There are no clocks possible in nature. We are fooled because we are surrounded by a
huge number of particles.
∗∗
Motion seems to disappear, even though it is eternal. We are fooled again, because our
senses cannot experience the microscopic domain.

∗∗
The world seems dependent on the choice of the frame of reference, even though it is not.
We are fooled because we are used to live on the surface of the Earth.
∗∗
Objects seem distinguishable, even though the statistical properties of their components
show that they are not. We are fooled because we live at low energies.
∗∗
Matter looks continuous, even though it isn’t. We are fooled because of the limitations of
our senses.
In short, our human condition permanently fools us. The answer to the title question is
affirmative: quantum theory is magic. That is its main attraction.
Quantum theory can do more

We can summarize this part of our adventure with a simple statement: quantum physics is
the description of matter and radiation without the concept of infinitely small. All change
in nature is described by finite quantities, above all, by the smallest change possible in
nature, the quantum of action ħ.
If we turn back to the start of our exploration of quantum theory, we cannot hide
a certain disappointment. We know that classical physics cannot explain life. Searching
for the details of microscopic motion, we encountered so many interesting aspects that
we have not finished the explanation of life. For example, we know what determines the
speed of electrons in atoms, but we do not know what determines the running speed of
an athlete. In fact, we have not even discussed the properties of any solid or liquid, let
alone those of more complex structures like living beings.
In short, after this introduction into quantum theory, we must still connect it to the
everyday world. The topic of the next volume will be the exploration of the motion of
living things and of the properties of composite materials and stars, using the quantum
of action as a starting point.

Appendix A
UNIT S, MEA SUREMENT S AND

C ON STAN T S
M
easurements are comparisons with standards. Standards are based on a unit.
any different systems of units have been used throughout the world.
ost standards confer power to the organization in charge of them. Such power
can be misused; this is the case today, for example in the computer industry, and was
so in the distant past. The solution is the same in both cases: organize an independent
and global standard. For units, this happened in the eighteenth century: to avoid mis-
use by authoritarian institutions, to eliminate problems with differing, changing and
irreproducible standards, and – this is not a joke – to simplify tax collection, a group
of scientists, politicians and economists agreed on a set of units. It is called the Système
International d’Unités, abbreviated SI, and is defined by an international treaty, the
‘Convention du Mètre’. The units are maintained by an international organization, the
‘Conférence Générale des Poids et Mesures’, and its daughter organizations, the ‘Com-
mission Internationale des Poids et Mesures’ and the ‘Bureau International des Poids et
Ref. 136 Mesures’ (BIPM), which all originated in the times just before the French revolution.
SI units
All SI units are built from seven base units, whose official definitions, translated from
French into English, are given below, together with the dates of their formulation:
‘The second is the duration of 9 192 631 770 periods of the radiation corresponding
to the transition between the two hyperfine levels of the ground state of the caesium 133
atom.’ (1967)*
‘The metre is the length of the path travelled by light in vacuum during a time inter-
val of 1/299 792 458 of a second.’ (1983)
‘The kilogram is the unit of mass; it is equal to the mass of the international prototype
of the kilogram.’ (1901)*
‘The ampere is that constant current which, if maintained in two straight parallel con-
ductors of infinite length, of negligible circular cross-section, and placed 1 metre apart in
vacuum, would produce between these conductors a force equal to 2 ⋅ 10−7 newton per
metre of length.’ (1948)
‘The kelvin, unit of thermodynamic temperature, is the fraction 1/273.16 of the ther-
modynamic temperature of the triple point of water.’ (1967)*
‘The mole is the amount of substance of a system which contains as many elementary
entities as there are atoms in 0.012 kilogram of carbon 12.’ (1971)*
‘The candela is the luminous intensity, in a given direction, of a source that emits
176 a units, measurements and constants
monochromatic radiation of frequency 540 ⋅ 1012 hertz and has a radiant intensity in that
direction of (1/683) watt per steradian.’ (1979)*
Note that both time and length units are defined as certain properties of a standard ex-
ample of motion, namely light. In other words, also the Conférence Générale des Poids
et Mesures makes the point that the observation of motion is a prerequisite for the defini-
tion and construction of time and space. Motion is the fundament each observation and
measurements. By the way, the use of light in the definitions had been proposed already
in 1827 by Jacques Babinet.*
From these basic units, all other units are defined by multiplication and division. Thus,
all SI units have the following properties:
SI units form a system with state-of-the-art precision: all units are defined with a pre-
cision that is higher than the precision of commonly used measurements. Moreover, the
precision of the definitions is regularly being improved. The present relative uncertainty
of the definition of the second is around 10−14 , for the metre about 10−10 , for the kilo-
gram about 10−9 , for the ampere 10−7 , for the mole less than 10−6 , for the kelvin 10−6 and
for the candela 10−3 .
SI units form an absolute system: all units are defined in such a way that they can
be reproduced in every suitably equipped laboratory, independently, and with high pre-
cision. This avoids as much as possible any misuse by the standard-setting organization.
(The kilogram, still defined with the help of an artefact, is the last exception to this re-
quirement; extensive research is under way to eliminate this artefact from the definition
– an international race that will take a few more years. There are two approaches: count-
ing particles, or fixing ħ. The former can be achieved in crystals, the latter using any
formula where ħ appears, such as the formula for the de Broglie wavelength or that of
the Josephson effect.)
SI units form a practical system: the base units are quantities of everyday magnitude.
Frequently used units have standard names and abbreviations. The complete list includes
the seven base units, the supplementary units, the derived units and the admitted units.
The supplementary SI units are two: the unit for (plane) angle, defined as the ratio
of arc length to radius, is the radian (rad). For solid angle, defined as the ratio of the
subtended area to the square of the radius, the unit is the steradian (sr).
The derived units with special names, in their official English spelling, i.e., without
capital letters and accents, are:
* The respective symbols are s, m, kg, A, K, mol and cd. The international prototype of the kilogram is
Vol. I, page 88 a platinum–iridium cylinder kept at the BIPM in Sèvres, in France. For more details on the levels of the
Ref. 137 caesium atom, consult a book on atomic physics. The Celsius scale of temperature θ is defined as: θ/°C =
T/K − 273.15; note the small difference with the number appearing in the definition of the kelvin. SI also
states: ‘When the mole is used, the elementary entities must be specified and may be atoms, molecules, ions,
electrons, other particles, or specified groups of such particles.’ In the definition of the mole, it is understood
that the carbon 12 atoms are unbound, at rest and in their ground state. In the definition of the candela, the
frequency of the light corresponds to 555.5 nm, i.e., green colour, around the wavelength to which the eye
is most sensitive.
* Jacques Babinet (1794–1874), French physicist who published important work in optics.
units, measurements and constants 177
Name A b b r e v i at i o n Name A b b r e v i at i o n
hertz Hz = 1/s newton N = kg m/s2

pascal Pa = N/m2 = kg/m s2 joule J = Nm = kg m2 /s2
watt W = kg m2 /s3 coulomb C = As
volt V = kg m2 /As3 farad F = As/V = A2 s4 /kg m2
ohm Ω = V/A = kg m2 /A2 s3 siemens S = 1/Ω
weber Wb = Vs = kg m2 /As2 tesla T = Wb/m2 = kg/As2 = kg/Cs
henry H = Vs/A = kg m2 /A2 s2 degree Celsius °C (see definition of kelvin)
lumen lm = cd sr lux lx = lm/m2 = cd sr/m2
becquerel Bq = 1/s gray Gy = J/kg = m2 /s2
sievert Sv = J/kg = m2 /s2 katal kat = mol/s
We note that in all definitions of units, the kilogram only appears to the powers of
1, 0 and −1. The final explanation for this fact appeared only recently. Can you try to
Challenge 141 ny formulate the reason?
The admitted non-SI units are minute, hour, day (for time), degree 1∘ = π/180 rad,
minute 1󳰀 = π/10 800 rad, second 1󳰀󳰀 = π/648 000 rad (for angles), litre and tonne. All
other units are to be avoided.
All SI units are made more practical by the introduction of standard names and ab-
breviations for the powers of ten, the so-called prefixes:*
Power Name Power Name Power Name Power Name

101 deca da 10−1 deci d 1018 Exa E 10−18 atto a
102 hecto h 10−2 centi c 1021 Zetta Z 10−21 zepto z
103 kilo k 10−3 milli m 1024 Yotta Y 10−24 yocto y
106 Mega M 10−6 micro μ unofficial: Ref. 138
109 Giga G 10−9 nano n 1027 Xenta X 10−27 xenno x
1012 Tera T 10−12 pico p 1030 Wekta W 10−30 weko w
1015 Peta P 10−15 femto f 1033 Vendekta V 10−33 vendeko v
1036 Udekta U 10−36 udeko u
SI units form a complete system: they cover in a systematic way the complete set of
observables of physics. Moreover, they fix the units of measurement for all other sciences
* Some of these names are invented (yocto to sound similar to Latin octo ‘eight’, zepto to sound similar
to Latin septem, yotta and zetta to resemble them, exa and peta to sound like the Greek words ἑξάκις and
πεντάκις for ‘six times’ and ‘five times’, the unofficial ones to sound similar to the Greek words for nine,
ten, eleven and twelve); some are from Danish/Norwegian (atto from atten ‘eighteen’, femto from femten
‘fifteen’); some are from Latin (from mille ‘thousand’, from centum ‘hundred’, from decem ‘ten’, from nanus
‘dwarf ’); some are from Italian (from piccolo ‘small’); some are Greek (micro is from μικρός ‘small’, deca/deka
from δέκα ‘ten’, hecto from ἑκατόν ‘hundred’, kilo from χίλιοι ‘thousand’, mega from μέγας ‘large’, giga from
γίγας ‘giant’, tera from τέρας ‘monster’).
Translate: I was caught in such a traffic jam that I needed a microcentury for a picoparsec and that my
Challenge 142 e car’s fuel consumption was two tenths of a square millimetre.
as well.
SI units form a universal system: they can be used in trade, in industry, in commerce,
at home, in education and in research. They could even be used by extraterrestrial civi-
lizations, if they existed.
SI units form a coherent system: the product or quotient of two SI units is also an SI
unit. This means that in principle, the same abbreviation, e.g. ‘SI’, could be used for every
unit.
The SI units are not the only possible set that could fulfil all these requirements, but they
are the only existing system that does so.*
Since every measurement is a comparison with a standard, any measurement requires
Challenge 143 e matter to realize the standard (even for a speed standard), and radiation to achieve the
comparison. The concept of measurement thus assumes that matter and radiation exist
and can be clearly separated from each other.
Planck’s natural units
Since the exact form of many equations depends on the system of units used, theoretical
physicists often use unit systems optimized for producing simple equations. The chosen
units and the values of the constants of nature are related. In microscopic physics, the
system of Planck’s natural units is frequently used. They are defined by setting c = 1, ħ =
1, G = 1, k = 1, ε0 = 1/4π and μ0 = 4π. Planck units are thus defined from combinations
of fundamental constants; those corresponding to the fundamental SI units are given in
Table 10.** The table is also useful for converting equations written in natural units back
Challenge 144 e to SI units: just substitute every quantity X by X/XPl .
TA B L E 10 Planck’s (uncorrected) natural units
Name Definition Va l u e
Basic units
the Planck length lPl = 󵀆 ħG/c 3 = 1.616 0(12) ⋅ 10−35 m
the Planck time tPl = 󵀆 ħG/c 5 = 5.390 6(40) ⋅ 10−44 s
the Planck mass mPl = 󵀆ħc/G = 21.767(16) μg
the Planck current IPl = 󵀆4πε0 c /G
6
= 3.479 3(22) ⋅ 1025 A
* Apart from international units, there are also provincial units. Most provincial units still in use are of
Roman origin. The mile comes from milia passum, which used to be one thousand (double) strides of about
1480 mm each; today a nautical mile, once defined as minute of arc on the Earth’s surface, is exactly 1852 m).
The inch comes from uncia/onzia (a twelfth – now of a foot). The pound (from pondere ‘to weigh’) is used
as a translation of libra – balance – which is the origin of its abbreviation lb. Even the habit of counting
in dozens instead of tens is Roman in origin. These and all other similarly funny units – like the system
in which all units start with ‘f ’, and which uses furlong/fortnight as its unit of velocity – are now officially
defined as multiples of SI units.
** The natural units xPl given here are those commonly used today, i.e., those defined using the constant
ħ, and not, as Planck originally did, by using the constant h = 2πħ. The electromagnetic units can also be
defined with other factors than 4πε0 in the expressions: for example, using 4πε0 α, with the fine structure
constant α, gives qPl = e. For the explanation of the numbers between brackets, the standard deviations, see
below.
Name Definition Va l u e
the Planck temperature TPl = 󵀆 ħc 5 /Gk 2 = 1.417 1(91) ⋅ 1032 K
Trivial units
the Planck velocity 󰑣Pl = c = 0.3 Gm/s
the Planck angular momentum LPl = ħ = 1.1 ⋅ 10−34 Js
the Planck action SaPl = ħ = 1.1 ⋅ 10−34 Js
the Planck entropy SePl = k = 13.8 yJ/K
Composed units
the Planck mass density ρPl = c 5 /G 2 ħ = 5.2 ⋅ 1096 kg/m3
the Planck energy EPl = 󵀆 ħc 5 /G = 2.0 GJ = 1.2 ⋅ 1028 eV
the Planck momentum pPl = 󵀆 ħc 3 /G = 6.5 Ns
the Planck power PPl = c /G
5
= 3.6 ⋅ 1052 W
the Planck force FPl = c 4 /G = 1.2 ⋅ 1044 N
the Planck pressure pPl = c 7 /Għ = 4.6 ⋅ 10113 Pa
the Planck acceleration aPl = 󵀆 c 7 /ħG = 5.6 ⋅ 1051 m/s2
the Planck frequency fPl = 󵀆 c 5 /ħG = 1.9 ⋅ 1043 Hz
the Planck electric charge qPl = 󵀆4πε0 cħ = 1.9 aC = 11.7 e
the Planck voltage UPl = 󵀆 c 4 /4πε0 G = 1.0 ⋅ 1027 V
the Planck resistance RPl = 1/4πε0 c = 30.0 Ω
the Planck capacitance CPl = 4πε0 󵀆 ħG/c 3 = 1.8 ⋅ 10−45 F
the Planck inductance LPl = (1/4πε0 )󵀆 ħG/c 7 = 1.6 ⋅ 10−42 H
the Planck electric field EPl = 󵀆 c 7 /4πε0 ħG 2 = 6.5 ⋅ 1061 V/m
the Planck magnetic flux density BPl = 󵀆 c 5 /4πε0 ħG 2 = 2.2 ⋅ 1053 T
The natural units are important for another reason: whenever a quantity is sloppily called
‘infinitely small (or large)’, the correct expression is ‘as small (or as large) as the corre-
sponding corrected Planck unit’. As explained throughout the text, and especially in the
Page 31 final part, this substitution is possible because almost all Planck units provide, within
a correction factor of order 1, the extremal value for the corresponding observable –
some an upper and some a lower limit. Unfortunately, these correction factors are not
yet widely known. The exact extremal value for each observable in nature is obtained
when G is substituted by 4G and 4πε0 by 4πε0 α in all Planck quantities. These extremal
values, or corrected Planck units, are the true natural units. To exceed the extremal values
Challenge 145 s is possible only for some extensive quantities. (Can you find out which ones?)
Other unit systems

A central aim of research in high-energy physics is the calculation of the strengths of
all interactions; therefore it is not practical to set the gravitational constant G to unity,
as in the Planck system of units. For this reason, high-energy physicists often only set
c = ħ = k = 1 and μ0 = 1/ε0 = 4π,* leaving only the gravitational constant G in the

equations.
In this system, only one fundamental unit exists, but its choice is free. Often a stan-
dard length is chosen as the fundamental unit, length being the archetype of a measured
quantity. The most important physical observables are then related by
1/[l 2 ] = [E]2 = [F] = [B] = [Eelectric ] ,

1/[l] = [E] = [m] = [p] = [a] = [ f ] = [I] = [U] = [T] ,
1 = [󰑣] = [q] = [e] = [R] = [Saction ] = [Sentropy ] = ħ = c = k = [α] ,
[l] = 1/[E] = [t] = [C] = [L] and
[l]2 =1/[E]2 = [G] = [P]
(109)
where we write [x] for the unit of quantity x. Using the same unit for time, capacitance
and inductance is not to everybody’s taste, however, and therefore electricians do not use
this system.**
Often, in order to get an impression of the energies needed to observe an effect un-
der study, a standard energy is chosen as fundamental unit. In particle physics the most
common energy unit is the electronvolt (eV), defined as the kinetic energy acquired by
an electron when accelerated by an electrical potential difference of 1 volt (‘protonvolt’
would be a better name). Therefore one has 1 eV = 1.6 ⋅ 10−19 J, or roughly
1
1 eV ≈ 6
aJ (110)
which is easily remembered. The simplification c = ħ = 1 yields G = 6.9 ⋅ 10−57 eV−2 and
allows one to use the unit eV also for mass, momentum, temperature, frequency, time
Challenge 146 e and length, with the respective correspondences 1 eV ≡ 1.8 ⋅ 10−36 kg ≡ 5.4 ⋅ 10−28 Ns
≡ 242 THz ≡ 11.6 kK and 1 eV−1 ≡ 4.1 fs ≡ 1.2 μm.
To get some feeling for the unit eV, the following relations are useful. Room temper-
ature, usually taken as 20°C or 293 K, corresponds to a kinetic energy per particle of
0.025 eV or 4.0 zJ. The highest particle energy measured so far belongs to a cosmic ray
Ref. 141 with an energy of 3 ⋅ 1020 eV or 48 J. Down here on the Earth, an accelerator able to pro-
duce an energy of about 105 GeV or 17 nJ for electrons and antielectrons has been built,
and one able to produce an energy of 14 TeV or 2.2 μJ for protons will be finished soon.
Both are owned by CERN in Geneva and have a circumference of 27 km.
The lowest temperature measured up to now is 280 pK, in a system of rhodium nuclei
* Other definitions for the proportionality constants in electrodynamics lead to the Gaussian unit system
often used in theoretical calculations, the Heaviside–Lorentz unit system, the electrostatic unit system, and
Ref. 139 the electromagnetic unit system, among others.
** In the list, l is length, E energy, F force, Eelectric the electric and B the magnetic field, m mass, p mo-
mentum, a acceleration, f frequency, I electric current, U voltage, T temperature, 󰑣 speed, q charge, R
resistance, P power, G the gravitational constant.
The web page www.chemie.fu-berlin.de/chemistry/general/units_en.html provides a tool to convert var-
ious units into each other.
Researchers in general relativity often use another system, in which the Schwarzschild radius rs =
2Gm/c 2 is used to measure masses, by setting c = G = 1. In this case, mass and length have the same
dimension, and ħ has the dimension of an area.
Ref. 142 held inside a special cooling system. The interior of that cryostat may even be the coolest
point in the whole universe. The kinetic energy per particle corresponding to that temper-
ature is also the smallest ever measured: it corresponds to 24 feV or 3.8 vJ = 3.8 ⋅ 10−33 J.
For isolated particles, the record seems to be for neutrons: kinetic energies as low as
10−7 eV have been achieved, corresponding to de Broglie wavelengths of 60 nm.
Curiosities and fun challenges about units

Not using SI units can be expensive. In 1999, NASA lost a satellite on Mars because some
software programmers had used provincial units instead of SI units in part of the code.
As a result, the Mars Climate Orbiter crashed into the planet, instead of orbiting it; the
loss was around 100 million euro.*
∗∗
The Planck length is roughly the de Broglie wavelength λB = h/m󰑣 of a man walking
Ref. 143 comfortably (m = 80 kg, 󰑣 = 0.5 m/s); this motion is therefore aptly called the ‘Planck
stroll.’
∗∗
The Planck mass is equal to the mass of about 1019 protons. This is roughly the mass of
a human embryo at about ten days of age.
∗∗
The most precisely measured quantities in nature are the frequencies of certain millisec-
Ref. 145 ond pulsars, the frequency of certain narrow atomic transitions, and the Rydberg con-
stant of atomic hydrogen, which can all be measured as precisely as the second is de-
fined. The caesium transition that defines the second has a finite linewidth that limits
the achievable precision: the limit is about 14 digits.
∗∗
The most precise clock ever built, using microwaves, had a stability of 10−16 during a
Ref. 144 running time of 500 s. For longer time periods, the record in 1997 was about 10−15 ; but
Ref. 146 values around 10−17 seem within technological reach. The precision of clocks is limited
for short measuring times by noise, and for long measuring times by drifts, i.e., by sys-
tematic effects. The region of highest stability depends on the clock type; it usually lies
between 1 ms for optical clocks and 5000 s for masers. Pulsars are the only type of clock
for which this region is not known yet; it certainly lies at more than 20 years, the time
elapsed at the time of writing since their discovery.
∗∗
The shortest times measured are the lifetimes of certain ‘elementary’ particles. In particu-
Ref. 147 lar, the lifetime of certain D mesons have been measured at less than 10−23 s. Such times
are measured using a bubble chamber, where the track is photographed. Can you esti-
* This story revived an old (and false) urban legend that states that only three countries in the world do not
use SI units: Liberia, the USA and Myanmar.
Challenge 147 s mate how long the track is? (This is a trick question – if your length cannot be observed
with an optical microscope, you have made a mistake in your calculation.)
∗∗
The longest times encountered in nature are the lifetimes of certain radioisotopes, over
1015 years, and the lower limit of certain proton decays, over 1032 years. These times are
thus much larger than the age of the universe, estimated to be fourteen thousand million
Ref. 148 years.
∗∗
Variations of quantities are often much easier to measure than their values. For example,
in gravitational wave detectors, the sensitivity achieved in 1992 was Δl/l = 3 ⋅ 10−19 for
Ref. 149 lengths of the order of 1 m. In other words, for a block of about a cubic metre of metal
it is possible to measure length changes about 3000 times smaller than a proton radius.
These set-ups are now being superseded by ring interferometers. Ring interferometers
measuring frequency differences of 10−21 have already been built; and they are still being
Ref. 150 improved.
Precision and accuracy of measurements

Measurements are the basis of physics. Every measurement has an error. Errors are due
to lack of precision or to lack of accuracy. Precision means how well a result is reproduced
when the measurement is repeated; accuracy is the degree to which a measurement cor-
responds to the actual value. Lack of precision is due to accidental or random errors; they
are best measured by the standard deviation, usually abbreviated σ; it is defined through
1 n
σ2 = 󵠈(xi − x̄)2 , (111)
n − 1 i=1
where x̄ is the average of the measurements xi . (Can you imagine why n − 1 is used in
Challenge 148 s the formula instead of n?)
For most experiments, the distribution of measurement values tends towards a nor-
mal distribution, also called Gaussian distribution, whenever the number of measure-
ments is increased. The distribution, shown in Figure 226, is described by the expression Copyright © Christoph Schiller November 1997–July 2010
(󰑥−󰑥̄)2
N(x) ≈ e− 2󰜎 2 . (112)
The square σ 2 of the standard deviation is also called the variance. For a Gaussian distri-
Challenge 149 e bution of measurement values, 2.35σ is the full width at half maximum.
Lack of accuracy is due to systematic errors; usually these can only be estimated. This
estimate is often added to the random errors to produce a total experimental error, some-
Ref. 151 times also called total uncertainty.
The tables below give the values of the most important physical constants and particle
properties in SI units and in a few other common units, as published in the standard
Ref. 152 references. The values are the world averages of the best measurements made up to the
N
number of measurements
standard deviation
full width at half maximum

(FWHM)
limit curve for a large number

of measurements
x x
average value measured values
F I G U R E 76 A precision experiment and its measurement distribution
present. As usual, experimental errors, including both random and estimated systematic
errors, are expressed by giving the standard deviation in the last digits; e.g. 0.31(6) means
– roughly speaking – 0.31 ± 0.06. In fact, behind each of the numbers in the following
Ref. 153 tables there is a long story which is worth telling, but for which there is not enough room
here.
Limits to precision
What are the limits to accuracy and precision? There is no way, even in principle, to
measure a length x to a precision higher than about 61 digits, because the ratio between
the largest and the smallest measurable length is Δx/x > lPl /dhorizon = 10−61 . (Is this
Challenge 150 e ratio valid also for force or for volume?) In the final volume of our text, studies of clocks
Vol. VI, page 82 and metre bars strengthen this theoretical limit.
But it is not difficult to deduce more stringent practical limits. No imaginable machine
can measure quantities with a higher precision than measuring the diameter of the Earth
within the smallest length ever measured, about 10−19 m; that is about 26 digits of preci-
sion. Using a more realistic limit of a 1000 m sized machine implies a limit of 22 digits.
If, as predicted above, time measurements really achieve 17 digits of precision, then they
are nearing the practical limit, because apart from size, there is an additional practical
restriction: cost. Indeed, an additional digit in measurement precision often means an
additional digit in equipment cost.
Physical constants
Ref. 152 In principle, all quantitative properties of matter can be calculated with quantum the-
ory. For example, colour, density and elastic properties can be predicted using the val-
ues of the following constants using the equations of the standard model of high-energy
Vol. V, page 193 physics.
TA B L E 11 Basic physical constants
Q ua n t i t y Symbol Va l u e i n S I u n i t s U n c e r t. a
number of space-time dimensions 3+1 0b

vacuum speed of lightc c 299 792 458 m/s 0
vacuum permeabilityc μ0 4π ⋅ 10−7 H/m 0
= 1.256 637 061 435 ... μH/m 0
vacuum permittivityc ε0 = 1/μ0 c 2 8.854 187 817 620 ... pF/m 0
original Planck constant h 6.626 068 76(52) ⋅ 10−34 Js 7.8 ⋅ 10−8
reduced Planck constant ħ 1.054 571 596(82) ⋅ 10−34 Js 7.8 ⋅ 10−8
3.9 ⋅ 10−8
positron charge e 0.160 217 646 2(63) aC
Boltzmann constant k 1.380 650 3(24) ⋅ 10−23 J/K 1.7 ⋅ 10−6
gravitational constant G 6.673(10) ⋅ 10−11 Nm2 /kg2 1.5 ⋅ 10−3
gravitational coupling constant κ = 8πG/c 4 2.076(3) ⋅ 10−43 s2 /kg m 1.5 ⋅ 10−3
2
fine structure constant, d α = 4πεe ħc 1/137.035 999 76(50) 3.7 ⋅ 10−9
0
e.m. coupling constant = αem (me2 c 2 ) = 0.007 297 352 533(27) 3.7 ⋅ 10−9
Fermi coupling constant, d GF /(ħc)3 1.166 39(1) ⋅ 10−5 GeV−2 8.6 ⋅ 10−6
weak coupling constant αw (MZ ) = дw2 /4π 1/30.1(3) 1 ⋅ 10−2
weak mixing angle sin2 θW (MS) 0.231 24(24) 1.0 ⋅ 10−3
weak mixing angle sin2 θW (on shell) 0.2224(19) 8.7 ⋅ 10−3
= 1 − (mW /mZ )2
strong coupling constant d αs (MZ ) = дs2 /4π 0.118(3) 25 ⋅ 10−3
a. Uncertainty: standard deviation of measurement errors.

b. Only down to 10−19 m and up to 1026 m.
c. Defining constant.
d. All coupling constants depend on the 4-momentum transfer, as explained in the section on
Page 92 renormalization. Fine structure constant is the traditional name for the electromagnetic coupling con-
stant α in the case of a 4-momentum transfer of Q 2 = me2 c 2 , which is the smallest one possible. At higher
momentum transfers it has larger values, e.g., αem (Q 2 = MW2 2
c ) ≈ 1/128. In contrast, the strong coupling
constant has lover values at higher momentum transfers; e.g., αs (34 GeV) = 0.14(2).
Why do all these constants have the values they have? For any constant with a dimen-
sion, such as the quantum of action ħ, the numerical value has only historical meaning.
It is 1.054 ⋅ 10−34 Js because of the SI definition of the joule and the second. The ques-
tion why the value of a dimensional constant is not larger or smaller therefore always
requires one to understand the origin of some dimensionless number giving the ratio be-
tween the constant and the corresponding natural unit that is defined with c, G, ħ and α.
Understanding the sizes of atoms, people, trees and stars, the duration of molecular and
atomic processes, or the mass of nuclei and mountains, implies understanding the ratios
between these values and the corresponding natural units. The key to understanding na-
ture is thus the understanding of all ratios, and thus of all dimensionless constants. The
quest of understanding all ratios, all dimensionless constants, including the fine structure
constant α itself, is completed only in the final volume of our adventure.
The basic constants yield the following useful high-precision observations.
TA B L E 12 Derived physical constants
Q ua n t i t y Symbol Va l u e i n S I u n i t s U n c e r t.
Vacuum wave resistance Z0 = 󵀄μ0 /ε0 376.730 313 461 77... Ω 0

Avogadro’s number NA 6.022 141 99(47) ⋅ 1023 7.9 ⋅ 10−8
Rydberg constant a R∞ = me cα 2 /2h 10 973 731.568 549(83) m−1 7.6 ⋅ 10−12
conductance quantum G0 = 2e 2 /h 77.480 916 96(28) μS 3.7 ⋅ 10−9
magnetic flux quantum φ0 = h/2e 2.067 833 636(81) pWb 3.9 ⋅ 10−8
Josephson frequency ratio 2e/h 483.597 898(19) THz/V 3.9 ⋅ 10−8
von Klitzing constant h/e 2 = μ0 c/2α 25 812.807 572(95) Ω 3.7 ⋅ 10−9
Bohr magneton μB = eħ/2me 9.274 008 99(37) yJ/T 4.0 ⋅ 10−8
cyclotron frequency fc /B = e/2πme 27.992 4925(11) GHz/T 4.0 ⋅ 10−8
of the electron
classical electron radius re = e 2 /4πε0 me c 2 2.817 940 285(31) fm 1.1 ⋅ 10−8
Compton wavelength λc = h/me c 2.426 310 215(18) pm 7.3 ⋅ 10−9
of the electron λ c = ħ/me c = re /α 0.386 159 264 2(28) pm 7.3 ⋅ 10−9
Bohr radius a a∞ = re /α 2 52.917 720 83(19) pm 3.7 ⋅ 10−9
nuclear magneton μN = eħ/2mp 5.050 783 17(20) ⋅ 10−27 J/T 4.0 ⋅ 10−8
proton–electron mass ratio mp /me 1 836.152 667 5(39) 2.1 ⋅ 10−9
Stefan–Boltzmann constant σ = π2 k 4 /60ħ3 c 2 56.704 00(40) nW/m2 K4 7.0 ⋅ 10−6
Wien’s displacement constant b = λmax T 2.897 768 6(51) mmK 1.7 ⋅ 10−6
bits to entropy conversion const. 1023 bit = 0.956 994 5(17) J/K 1.7 ⋅ 10−6
TNT energy content 3.7 to 4.0 MJ/kg 4 ⋅ 10−2
a. For infinite mass of the nucleus.
Some useful properties of our local environment are given in the following table.
TA B L E 13 Astronomical constants
Q ua n t i t y Symbol Va l u e
tropical year 1900 a a 31 556 925.974 7 s

tropical year 1994 a 31 556 925.2 s
mean sidereal day d 23h 56󳰀 4.090 53󳰀󳰀
astronomical unit b AU 149 597 870.691(30) km
light year al 9.460 528 173 ... Pm
parsec pc 30.856 775 806 Pm = 3.261 634 al
Earth’s mass M♁ 5.973(1) ⋅ 1024 kg
Geocentric gravitational constant GM 3.986 004 418(8) ⋅ 1014 m3 /s2
Earth’s gravitational length l♁ = 2GM/c 2 8.870 056 078(16) mm
TA B L E 13 (Continued) Astronomical constants
Earth’s equatorial radius c R♁eq 6378.1366(1) km

c
Earth’s polar radius R♁p 6356.752(1) km
c
Equator–pole distance 10 001.966 km (average)
Earth’s flattening c e♁ 1/298.25642(1)
Earth’s av. density ρ♁ 5.5 Mg/m3
Earth’s age T♁ 4.50(4) Ga = 142(2) Ps
Moon’s radius Rv 1738 km in direction of Earth
Moon’s radius Rh 1737.4 km in other two directions
Moon’s mass M 7.35 ⋅ 1022 kg
Moon’s mean distance d
d 384 401 km
Moon’s distance at perigee d typically 363 Mm, historical minimum
359 861 km
Moon’s distance at apogee d typically 404 Mm, historical maximum
406 720 km
Moon’s angular size e
average 0.5181∘ = 31.08󳰀 , minimum 0.49∘ ,
maximum - shortens line 0.55∘
Moon’s average density ρ 3.3 Mg/m3
Jupiter’s mass M 1.90 ⋅ 1027 kg
Jupiter’s radius, equatorial R 71.398 Mm
Jupiter’s radius, polar R 67.1(1) Mm
Jupiter’s average distance from Sun D 778 412 020 km
Sun’s mass M⊙ 1.988 43(3) ⋅ 1030 kg
Sun’s gravitational length l⊙ = 2GM⊙ /c 2.953 250 08 km
2
Sun’s luminosity L⊙ 384.6 YW

Solar equatorial radius R⊙ 695.98(7) Mm
Sun’s angular size 0.53∘ average; minimum on fourth of July
(aphelion) 1888󳰀󳰀 , maximum on fourth of
January (perihelion) 1952󳰀󳰀
Sun’s average density ρ⊙ 1.4 Mg/m3
Sun’s average distance AU 149 597 870.691(30) km
Sun’s age T⊙ 4.6 Ga

Solar velocity 󰑣⊙g 220(20) km/s
around centre of galaxy
Solar velocity 󰑣⊙b 370.6(5) km/s
against cosmic background
Distance to Milky Way’s centre 8.0(5) kpc = 26.1(1.6) kal
Milky Way’s age 13.6 Ga
Milky Way’s size c. 1021 m or 100 kal
Milky Way’s mass 1012 solar masses, c. 2 ⋅ 1042 kg
Most distant galaxy cluster known SXDF-XCLJ 9.6 ⋅ 109 al
TA B L E 13 (Continued) Astronomical constants
0218-0510
a. Defining constant, from vernal equinox to vernal equinox; it was once used to define the second. (Remem-
ber: π seconds is about a nanocentury.) The value for 1990 is about 0.7 s less, corresponding to a slowdown
Challenge 151 s of roughly 0.2 ms/a. (Watch out: why?) There is even an empirical formula for the change of the length of
Ref. 154 the year over time.
b. Average distance Earth–Sun. The truly amazing precision of 30 m results from time averages of signals
sent from Viking orbiters and Mars landers taken over a period of over twenty years.
c. The shape of the Earth is described most precisely with the World Geodetic System. The last edition dates
from 1984. For an extensive presentation of its background and its details, see the www.wgs84.com website.
The International Geodesic Union refined the data in 2000. The radii and the flattening given here are those
for the ‘mean tide system’. They differ from those of the ‘zero tide system’ and other systems by about 0.7 m.
The details constitute a science in itself.
d. Measured centre to centre. To find the precise position of the Moon at a given date, see the www.
fourmilab.ch/earthview/moon_ap_per.html page. For the planets, see the page www.fourmilab.ch/solar/
solar.html and the other pages on the same site.
e. Angles are defined as follows: 1 degree = 1∘ = π/180 rad, 1 (first) minute = 1󳰀 = 1∘ /60, 1 second (minute)
= 1󳰀󳰀 = 1󳰀 /60. The ancient units ‘third minute’ and ‘fourth minute’, each 1/60th of the preceding, are not in
use any more. (‘Minute’ originally means ‘very small’, as it still does in modern English.)
Some properties of nature at large are listed in the following table. (If you want a chal-
Challenge 152 s lenge, can you determine whether any property of the universe itself is listed?)
TA B L E 14 Astrophysical constants
gravitational constant G 6.672 59(85) ⋅ 10−11 m3 /kg s2

cosmological constant Λ c. 1 ⋅ 10−52 m−2
age of the universe a
t0 4.333(53) ⋅ 1017 s = 13.73(0.17) ⋅ 109 a
(determined from space-time, via expansion, using general relativity)
age of the universe a t0 over 3.5(4) ⋅ 1017 s = 11.5(1.5) ⋅ 109 a
(determined from matter, via galaxies and stars, using quantum theory)
Hubble parameter a H0 2.3(2) ⋅ 10−18 s−1 = 0.73(4) ⋅ 10−10 a−1
= h0 ⋅ 100 km/s Mpc = h0 ⋅ 1.0227 ⋅ 10−10 a−1

reduced Hubble parameter a h0 0.71(4)
deceleration parameter q0 = −(̈a/a)0 /H0 −0.66(10)
2
universe’s horizon distance a d0 = 3ct0 40.0(6) ⋅ 1026 m = 13.0(2) Gpc

universe’s topology trivial up to 1026 m
number of space dimensions 3, for distances up to 1026 m
critical density ρc = 3H0 /8πG
2
h02 ⋅ 1.878 82(24) ⋅ 10−26 kg/m3
of the universe = 0.95(12) ⋅ 10−26 kg/m3
(total) density parameter a
Ω0 = ρ0 /ρc 1.02(2)
baryon density parameter a
ΩB0 = ρB0 /ρc 0.044(4)
TA B L E 14 (Continued) Astrophysical constants
cold dark matter density parameter a ΩCDM0 = ρCDM0 /ρc 0.23(4)

neutrino density parameter a Ω󰜈0 = ρ󰜈0 /ρc 0.001 to 0.05
dark energy density parameter a ΩX0 = ρX0 /ρc 0.73(4)
dark energy state parameter 󰑤 = pX /ρX −1.0(2)
baryon mass mb 1.67 ⋅ 10−27 kg
baryon number density 0.25(1) /m3
luminous matter density 3.8(2) ⋅ 10−28 kg/m3
stars in the universe ns 1022±1
baryons in the universe nb 1081±1
microwave background temperature b T0 2.725(1) K
photons in the universe nγ 1089
photon energy density ργ = π2 k 4 /15T04 4.6 ⋅ 10−31 kg/m3
photon number density 410.89 /cm3 or 400 /cm3 (T0 /2.7 K)3
density perturbation amplitude 󵀂S 5.6(1.5) ⋅ 10−6
gravity wave amplitude 󵀂T < 0.71󵀂S
mass fluctuations on 8 Mpc σ8 0.84(4)
scalar index n 0.93(3)
running of scalar index dn/d ln k -0.03(2)
Planck length lPl = 󵀆 ħG/c 3 1.62 ⋅ 10−35 m
Planck time tPl = 󵀆 ħG/c 5 5.39 ⋅ 10−44 s
Planck mass mPl = 󵀄 ħc/G 21.8 μg
instants in history a t0 /tPl 8.7(2.8) ⋅ 1060
space-time points N0 = (R0 /lPl )3 ⋅ 10244±1
inside the horizon a (t0 /tPl )
mass inside horizon M 1054±1 kg
a. The index 0 indicates present-day values.

b. The radiation originated when the universe was 380 000 years old and had a temperature of about 3000 K;
Vol. II, page 206 the fluctuations ΔT0 which led to galaxy formation are today about 16 ± 4 μK = 6(2) ⋅ 10−6 T0 .
Useful numbers
π 3.14159 26535 89793 23846 26433 83279 50288 41971 69399 375105
e 2.71828 18284 59045 23536 02874 71352 66249 77572 47093 699959
γ 0.57721 56649 01532 86060 65120 90082 40243 10421 59335 939923
Ref. 155
ln 2 0.69314 71805 59945 30941 72321 21458 17656 80755 00134 360255
ln 10 2.30258 50929 94045 68401 79914 54684 36420 76011 01488 628772
󵀂10 3.16227 76601 68379 33199 88935 44432 71853 37195 55139 325216
If the number π is normal, i.e., if all digits and digit combinations in its decimal expan-
sion appear with the same limiting frequency, then every text ever written or yet to be
written, as well as every word ever spoken or yet to be spoken, can be found coded in its
sequence. The property of normality has not yet been proven, although it is suspected to
hold. Does this mean that all wisdom is encoded in the simple circle? No. The property is
nothing special: it also applies to the number 0.123456789101112131415161718192021...
Challenge 153 s and many others. Can you specify a few examples?
By the way, in the graph of the exponential function ex , the point (0, 1) is the only
point with two rational coordinates. If you imagine painting in blue all points on the
plane with two rational coordinates, the plane would look quite bluish. Nevertheless, the
graph goes through only one of these points and manages to avoid all the others.
Appendix B
NUMBERS AND VECTOR SPAC ES
“
A mathematician is a machine that transforms
”
coffee into theorems.
Paul Erdős (b. 1913 Budapest, d. 1996 Warsaw)
M
athematical concepts can all be expressed in terms of ‘sets’ and ‘relations.’
any fundamental concepts were presented in the last chapter. Why does
athematics, given this simple basis, grow into a passion for certain people? The
Ref. 156 following pages present a few more advanced concepts as simplyand vividly as possible,
for all those who want to smell the passion for mathematics.
In particular, in this appendix we shall introduce the simplest algebraic structures. The
appendix in the next volume will present some more involved algebraic structures and
the most important topological structures; the third basic type of mathematical struc-
tures, order structures, are not so important in physics.
Mathematicians are concerned not only with the exploration of concepts, but also
with their classification. Whenever a new mathematical concept is introduced, mathe-
maticians try to classify all the possible cases and types. This has been achieved most
spectacularly for the different types of numbers, for finite simple groups, and for many
types of spaces and manifolds.
Numbers as mathematical structures

A person who can solve x 2 − 92y 2 = 1 in less
Challenge 154 ny
“ than a year is a mathematician.

Brahmagupta (b. 598 Sindh, d. 668) (implied:
”
solve in integers)
We start with a short introduction to the vocabulary. Any mathematical system with the
same basic properties as the natural numbers is called a semi-ring. Any mathematical
system with the same basic properties as the integers is called a ring. (The term is due to
David Hilbert. Both structures can also be finite rather than infinite.) More precisely, a
ring (R, +, ⋅) is a set R of elements with two binary operations, called addition and multi-
plication, usually written + and ⋅ (the latter may simply be understood without notation),
for which the following properties hold for all elements a, b, c ∈ R:
— R is a commutative group with respect to addition, i.e.
a + b ∈ R, a + b = b + a, a + 0 = a, a + (−a) = a − a = 0 and a + (b + c) = (a + b) + c;
— R is closed under multiplication, i.e., ab ∈ R;
numbers and vector spaces 191
— multiplication is associative, i.e., a(bc) = (ab)c;

— distributivity holds, i.e., a(b + c) = ab + ac and (b + c)a = ba + ca.
Many authors add the axiom
— a multiplicative unit exists, i.e., 1a = a1 = a.
Defining properties such as these are called axioms. Note that axioms are not basic beliefs,
as is often stated; axioms are the basic properties used in the definition of a concept: in
this case, of a ring. With the last axiom, one also speaks of a unital ring.
A semi-ring is a set satisfying all the axioms of a ring, except that the existence of
neutral and negative elements for addition is replaced by the weaker requirement that if
a + c = b + c then a = b. Sloppily, a semi-ring is a ring ‘without’ negative elements.
To incorporate division and define the rational numbers, we need another concept. A
field K is a ring with
— a multiplicative identity 1, such that all elements a obey 1a = a;
— at least one element different from zero; and most importantly
— a (multiplicative) inverse a−1 for every element a ̸= 0.
A ring or field is said to be commutative if the multiplication is commutative. A non-
commutative field is also called a skew field. Fields can be finite or infinite. (A field or a
ring is characterized by its characteristic p. This is the smallest number of times one has
to add 1 to itself to give zero. If there is no such number the characteristic is set to 0. p is
always a prime number or zero.) All finite fields are commutative. In a field, all equations
of the type cx = b and xc = b (c ̸= 0) have solutions for x; there is a unique solution
if b ̸= 0. To sum up sloppily by focusing on the most important property, a field is a set
of elements for which, together with addition, subtraction and multiplication, a division
(by non-zero elements) is also defined. The rational numbers are the simplest field that
incorporates the integers.
The system of the real numbers is the minimal extension of the rationals which is
complete and totally ordered.*
However, the concept of ‘number’ is not limited to these examples. It can be general-
Ref. 157 ized in several ways. The simplest generalization is achieved by extending the real num-
bers to manifolds of more than one dimension.
* A set is mathematically complete if physicists call it continuous. More precisely, a set of numbers is complete
if every non-empty subset that is bounded above has a least upper bound.
A set is totally ordered if there exists a binary relation ⩽ between pairs of elements such that for all
elements a and b
— if a ⩽ b and b ⩽ c, then a ⩽ c;
— if a ⩽ b and b ⩽ a, then a = b;
— a ⩽ b or b ⩽ a holds.
In summary, a set is totally ordered if there is a binary relation that allows to say about any two elements
which one is the predecessor of the other in a consistent way.
192 b numbers and vector spaces
ic
ih = − iab
c
F I G U R E 77 A property of triangles easily provable with

a 0 b complex numbers
Complex numbers
A complex number is defined by z = a + ib, where a and b are real numbers, and i is a
new symbol. Under multiplication, the generators of the complex numbers, 1 and i, obey
⋅ 1 i
1 1 i (113)
i i −1
often summarized as i = +󵀂−1 .

The complex conjugate z ∗ , also written z̄, of a complex number z = a + ib is defined
as z ∗ = a − ib. The absolute value |z| of a complex number is defined as |z| = 󵀂zz ∗ =
󵀂z ∗ z = 󵀄a2 + b2 . It defines a norm on the vector space of the complex numbers. From
|󰑤z| = |󰑤| |z| follows the two-squares theorem
(a12 + a22 )(b12 + b22 ) = (a1 b1 − a2 b2 )2 + (a1 b2 + a2 b1 )2 (114)
valid for all real numbers ai , bi . It was already known, in its version for integers, to Dio-
phantus of Alexandria.
Complex numbers can also be written as ordered pairs (a, A) of real numbers, with
their addition defined as (a, A) + (b, B) = (a + b, A + B) and their multiplication defined
as (a, A) ⋅ (b, B) = (ab − AB, aB + bA). This notation allows us to identify the complex
numbers with the points on a plane or, if we prefer, to arrows in a plane. Translating the
definition of multiplication into geometrical language allows us to rapidly prove certain
Challenge 155 e geometrical theorems, such as the one of Figure 77.
Complex numbers a + ib can also be represented as 2 × 2 matrices
a b
󶀥 󶀵 witha, b ∈ ℝ . (115)
−b a
Matrix addition and multiplication then correspond to complex addition and multipli-
cation. In this way, complex numbers can be represented by a special type of real matrix.
Challenge 156 s What is |z| in matrix language?

The set ℂ of complex numbers with addition and multiplication as defined above
Page 201 forms both a commutative two-dimensional field and a vector space over ℝ. In the field
of complex numbers, quadratic equations az 2 + bz + c = 0 for an unknown z always have
Challenge 157 ny two solutions (for a ̸= 0 and counting multiplicity).
Complex numbers can be used to describe the points of a plane. A rotation around the
origin can be described by multiplication by a complex number of unit length. Other two-
dimensional quantities can also be described with complex numbers. Electrical engineers
use complex numbers to describe quantities with phases, such as alternating currents or
electrical fields in space.
Writing complex numbers of unit length as cos θ + i sin θ is a useful method for re-
Challenge 158 e membering angle addition formulae. Since one has cos nθ + i sin nθ = (cos θ + i sin θ)n ,
one can easily deduce formulae cos 2θ = cos2 θ − sin2 θ and sin 2θ = 2 sin θ cos θ.
The complex exponential function is periodic in 2πi; in other words, one has
e1 = e1+2πi . (116)
If one uses this equation twice, one gets

2 2
e1 = (e1+2πi )1+2πi = e(1+2πi)(1+2πi) = e1−4π +4πi = e1−4π . (117)
Challenge 159 e Oops, that would imply π = 0! What is wrong here?

By the way, there are exactly as many complex numbers as there are real numbers.
Challenge 160 s Can you show this?
Challenge 161 e The unit complex numbers form the group SO(2)=U(1).
“ ”
Love is complex: it has real and imaginary parts.
Anonymous
Quaternions
The positions of the points on a line can be described by real numbers. Complex numbers
can be used to describe the positions of the points of a plane. It is natural to try to gen-
eralize the idea of a number to higher-dimensional spaces. However, it turns out that no
useful number system can be defined for three-dimensional space. A new number system,
the quaternions, can be constructed which corresponds the points of four-dimensional

space, but only if the commutativity of multiplication is sacrificed. No useful number
system can be defined for dimensions other than 1, 2 and 4.
The quaternions were discovered by several mathematicians in the nineteenth century,
among them Hamilton,* who studied them for much of his life. In fact, Maxwell’s theory
of electrodynamics was formulated in terms of quaternions before three-dimensional
Ref. 158 vectors were used.
Vol. V, page 281 Under multiplication, the quaternions ℍ form a 4-dimensional algebra over the reals
* William Rowan Hamilton (b. 1805 Dublin, d. 1865 Dunsink ), Irish child prodigy and famous mathemati-
cian, named the quaternions after an expression from the Vulgate (Acts. 12: 4).
with a basis 1, i, j, k satisfying
⋅ 1 i j k
1 1 i j k
i i −1 k − j . (118)
j j −k −1 i
k k j −i −1
These relations are also often written i 2 = j 2 = k 2 = −1, i j = − ji = k, jk = −k j = i,

ki = −ik = j. The quaternions 1, i, j, k are also called basic units or generators. The lack of
symmetry across the diagonal of the table shows the non-commutativity of quaternionic
multiplication. With the quaternions, the idea of a non-commutative product appeared
for the first time in mathematics. However, the multiplication of quaternions is associa-
tive. As a consequence of non-commutativity, polynomial equations in quaternions have
many more solutions than in complex numbers: just search for all solutions of the equa-
Challenge 162 s tion X 2 + 1 = 0 to convince yourself of it.
Every quaternion X can be written in the form
X = x0 + x1 i + x2 j + x3 k = x0 + 󰑣 = (x0 , x1 , x2 , x3 ) = (x0 , 󰑣) , (119)
where x0 is called the scalar part and 󰑣 the vector part. The multiplication is thus defined
as (x, 󰑣)(y, 󰑤) = (x y−󰑣⋅󰑤, x󰑤+ y󰑣+󰑣×󰑤). The multiplication of two general quaternions
can be written as
(a1 , b1 , c1 , d1 )(a2 , b2 , c2 , d2 ) = (a1 a2 − b1 b2 − c1 c2 − d1 d2 , a1 b2 + b1 a2 + c1 d2 − d1 c2 ,

a1 c2 − b1 d2 + c1 a2 + d1 b2 , a1 d2 + b1 c2 − c1 b2 + d1 a2 ) .
(120)
The conjugate quaternion X is defined as X = x0 − 󰑣, so that XY = Y X. The norm |X| of

a quaternion X is defined as |X|2 = X X = X X = x02 + x12 + x22 + x32 = x02 + 󰑣 2 . The norm
is multiplicative, i.e., |XY | = |X| |Y |.
Unlike complex numbers, every quaternion is related to its complex conjugate by
X = − 12 (X + iXi + jX j + kXk) . (121)

No relation of this type exists for complex numbers. In the language of physics, a complex
number and its conjugate are independent variables; for quaternions, this is not the case.
As a result, functions of quaternions are less useful in physics than functions of complex
variables.
The relation |XY | = |X| |Y | implies the four-squares theorem
(a12 + a22 + a32 + a42 )(b12 + b22 + b32 + b42 )

= (a1 b1 − a2 b2 − a3 b3 − a4 b4 )2 + (a1 b2 + a2 b1 + a3 b4 − a4 b3 )2
+ (a1 b3 + a3 b1 + a4 b2 − a2 b4 )2 + (a1 b4 + a4 b1 + a2 b3 − a3 b2 )2 (122)
α/2
l π − γ/2
β/2
m F I G U R E 78 Combinations of rotations
valid for all real numbers ai and bi , and thus also for any set of eight integers. It was
discovered in 1748 by Leonhard Euler (1707–1783) when trying to prove that each integer
is the sum of four squares. (That fact was proved only in 1770, by Joseph Lagrange.)
Hamilton thought that a quaternion with zero scalar part, which he simply called a vec-
tor (a term which he invented), could be identified with an ordinary three-dimensional
translation vector; but this is wrong. Such a quaternion is now called a pure, or homo-
geneous, or imaginary quaternion. The product of two pure quaternions V = (0, 󰑣) and
W = (0, 󰑤) is given by VW = (−󰑣 ⋅ 󰑤, 󰑣 × 󰑤), where ⋅ denotes the scalar product and ×
denotes the vector product. Note that any quaternion can be written as the ratio of two
pure quaternions.
In reality, a pure quaternion (0, 󰑣) does not behave like a translation vector under
coordinate transformations; in fact, a pure quaternion represents a rotation by the angle
Challenge 163 ny π or 180° around the axis defined by the direction 󰑣 = (󰑣x , 󰑣 y , 󰑣z ).
It turns out that in three-dimensional space, a general rotation about the origin can
be described by a unit quaternion Q, also called a normed quaternion, for which |Q| = 1.
Such a quaternion can be written as (cos θ/2, n sin θ/2), where n = (nx , n y , nz ) is the
normed vector describing the direction of the rotation axis and θ is the rotation angle.
Such a unit quaternion Q = (cos θ/2, n sin θ/2) rotates a pure quaternion V = (0, 󰑣) into
another pure quaternion W = (0, 󰑤) given by
W = QVQ ∗ . (123)
Thus, if we use pure quaternions such as V or W to describe positions, we can use unit
quaternions to describe rotations and to calculate coordinate changes. The concatenation
of two rotations is then given by the product of the corresponding unit quaternions. In-
deed, a rotation by an angle α about the axis l followed by a rotation by an angle β about
the axis m gives a rotation by an angle γ about the axis n, with the values determined by
(cos γ/2, sin γ/2n) = (cos β/2, sin β/2m)(cos α/2, sin α/2l) . (124)
One way to show the result graphically is given in Figure 78. By drawing a triangle on a
unit sphere, and taking care to remember the factor 1/2 in the angles, the combination
i
j
1 k
j i
palm back
of right of right
hand hand F I G U R E 79 The top and back of the right hand,
and the quaternions
of two rotations can be simply determined.

The interpretation of quaternions as rotations is also illustrated, in a somewhat differ-
Ref. 159 ent way, in the motion of any hand. To see this, take a green marker and write the letters 1,
i, j and k on your hand as shown in Figure 79. Defining the three possible 180° rotation
axes as shown in the figure and taking concatenation as multiplication, the motion of
Challenge 164 e the right hand follows the same ‘laws’ as those of pure unit quaternions. (One still needs
to distinguish +i and −i, and the same for the other units, by the sense of the arm twist.
And the result of a multiplication is that letter that can be read by a person facing you.)
Challenge 165 s You can show that i 2 = j 2 = k 2 = −1, that i 4 = 1, and all other quaternion relations.)
The model also shows that the rotation angle of the arm is half the rotation angle of the
corresponding quaternion. In other words, quaternions can be used to describe the belt
trick, if the multiplication VW of two quaternions is taken to mean that rotation V is
Page 111 performed after rotation W. Quaternions, like human hands, thus behave like a spin 1/2
particle. Quaternions and spinors are isomorphic.
The reason for the half-angle behaviour of rotations can be specified more precisely
using mathematical language. The rotations in three dimensions around a point form the
‘special orthogonal group’ in three dimensions, which is called SO(3). But the motions
of a hand attached to a shoulder via an arm form a different group, isomorphic to the Lie
Vol. V, page 292 group SU(2). The difference is due to the appearance of half angles in the parametriza-
tion of rotations; indeed, the above parametrizations imply that a rotation by 2π corre-
sponds to a multiplication by −1. Only in the twentieth century was it realized that there
exist fundamental physical observables that behaves like hands attached to arms: they
Page 111 are called spinors. More on spinors can be found in the section on permutation symme-
try, where belts are used as an analogy as well as arms. In short, the group SU(2) of the
Ref. 160 quaternions is the double cover of the rotation group SO(3).
The simple representation of rotations and positions with quaternions is used by com-
puter programmes in robotics, in astronomy and in flight simulation. In the software
used to create three-dimensional images and animations, visualization software, quater-
nions are often used to calculate the path taken by repeatedly reflected light rays and thus
give surfaces a realistic appearance.

The algebra of the quaternions is the only associative, non-commutative, finite-dimen-
sional normed algebra with an identity over the field of real numbers. Quaternions form
a non-commutative field, i.e., a skew field, in which the inverse of a quaternion X is
X/|X|. We can therefore define division of quaternions (while being careful to distin-
guish XY −1 and Y −1 X). Therefore quaternions are said to form a division algebra. In fact,
the quaternions ℍ, the complex numbers ℂ and the reals ℝ are the only three finite-
dimensional associative division algebras. In other words, the skew-field of quaternions
is the only finite-dimensional real associative non-commutative algebra without divisors
of zero. The centre of the quaternions, i.e., the set of quaternions that commute with all
other quaternions, is just the set of real numbers.
Quaternions can be represented as matrices of the form
a b c d
A B −b a −d c
󶀥 ∗ 󶀵 with A, B ∈ ℂ, or as 󶀫 󶀻 with a, b, c, d ∈ ℝ, (125)
−B A∗ −c d a −b
−d −c b a
where A = a + ib, B = c + id and the quaternion X is X = A + B j = a + ib + jc +

kd; matrix addition and multiplication then corresponds to quaternionic addition and
multiplication.
The generators of the quaternions can be realized as
1 : σ0 , i : −iσ1 , j : −iσ2 , k : −iσ3 (126)
where the σn are the Pauli spin matrices.*

Real 4 × 4 representations are not unique, as the alternative representation
a b −d −c
−b a −c d
󶀫 󶀻 (128)
d c a b
c −d −b a
Challenge 166 ny shows; however, no representation by 3 × 3 matrices is possible.

These matrices contain real and complex elements, which pose no special problems.
In contrast, when matrices with quaternionic elements are constructed, care has to be
taken, because quaternionic multiplication is not commutative, so that simple relations
such as trAB = trBA are not generally valid.
* The Pauli spin matrices are the complex Hermitean matrices
1 0 0 1 0 −i 1 0
σ0 = 1 = 󶀥 󶀵 , σ1 = 󶀥 󶀵 , σ2 = 󶀥 󶀵 , σ3 = 󶀥 󶀵 (127)
0 1 1 0 i 0 0 −1
all of whose eigenvalues are ±1; they satisfy the relations [σi , σk ]+ = 2 δik and [σi , σk ] = 2i εikl σl . The linear
combinations σ± = 12 (σ1 ± σ2 ) are also frequently used. By the way, another possible representation of the
quaternions is i : iσ3 , j : iσ2 , k : iσ1 .
What can we learn from quaternions about the description of nature? First of all, we
see that binary rotations are similar to positions, and thus to translations: all are rep-
resented by 3-vectors. Are rotations the basic operations of nature? Is it possible that
translations are only ‘shadows’ of rotations? The connection between translations and
Vol. VI, page 160 rotations is investigated in the last volume of our mountain ascent.
When Maxwell wrote down his equations of electrodynamics, he used quaternion
Vol. III, page 64 notation. (The now usual 3-vector notation was introduced later by Hertz and Heaviside.)
The equations can be written in various ways using quaternions. The simplest is achieved
Ref. 158 when one keeps a distinction between 󵀂−1 and the units i, j, k of the quaternions. One
Challenge 167 s then can write all of electrodynamics in a single equation:
Q
dF = − (129)
ε0
where F is the generalized electromagnetic field and Q the generalized charge. These are
defined by
F = E + 󵀂−1 cB
E = iEx + jE y + kEz
B = iBx + jB y + kBz (130)
d = δ + 󵀂−1 ∂t /c
δ = i∂x + j∂ y + k∂z
Q = ρ + 󵀂−1 J/c
where the fields E and B and the charge distributions ρ and J have the usual meanings.
The content of equation (129) for the electromagnetic field is exactly the same as the usual
formulation.
Despite their charm, quaternions do not seem to be ready for the reformulation of
special relativity; the main reason for this is the sign in the expression for their norm.
Therefore, relativity and space-time are usually described using real numbers.
Octonions
In the same way that quaternions are constructed from complex numbers, octonions
can be constructed from quaternions. They were first investigated by Arthur Cayley
(1821–1895). Under multiplication, octonions (or octaves) are the elements of an eight-
dimensional algebra over the reals with the generators 1, in with n = 1 . . . 7 satisfying
⋅ 1 i1 i2 i3 i4 i5 i6 i7
1 1 i1 i2 i3 i4 i5 i6 i7
i1 i1 −1 i3 −i2 i5 −i4 i7 −i6
i2 i2 −i3 −1 i1 −i6 i7 i4 −i5
i3 i3 i2 −i1 −1 i7 i6 −i5 −i4 (131)
i4 i4 −i5 i6 −i7 −1 i1 −i2 i3
i5 i5 i4 −i7 −i6 −i1 −1 i3 i2
i6 i6 −i7 −i4 i5 i2 −i3 −1 i1
i7 i7 i6 i5 i4 −i3 −i2 −i1 −1
479 other, equivalent multiplication tables are also possible. This algebra is called the
Cayley algebra; it has an identity and a unique division. The algebra is non-commutative,
and also non-associative. It is, however, alternative, meaning that for all elements x and
y, one has x(x y) = x 2 y and (x y)y = x y 2 : a property somewhat weaker than associativ-
ity. It is the only 8-dimensional real alternative algebra without zero divisors. Because it
is not associative, the set 𝕆 of all octonions does not form a field, nor even a ring, so
that the old designation of ‘Cayley numbers’ has been abandoned. The octonions are the
most general hypercomplex ‘numbers’ whose norm is multiplicative. Its generators obey
(in im )i l = ±in (im i l ), where the minus sign, which shows the non-associativity, is valid
for combinations of indices, such as 1-2-4, which are not quaternionic.
Octonions can be represented as matrices of the form
A B
󶀥 ̄ ̄ 󶀵 where A, B ∈ ℍ , or as real 8 × 8 matrices. (132)
−B A
Matrix multiplication then gives the same result as octonionic multiplication.

The relation |󰑤z| = |󰑤| |z| allows one to deduce the impressive eight-squares theorem
(a12 + a22 + a32 + a42 + a52 + a62 + a72 + a82 )(b12 + b22 + b32 + b42 + b52 + b62 + b72 + b82 )
= (a1 b1 − a2 b2 − a3 b3 − a4 b4 − a5 b5 − a6 b6 − a7 b7 − a8 b8 )2
+ (a1 b2 + a2 b1 + a3 b4 − a4 b3 + a5 b6 − a6 b5 + a7 b8 − a8 b7 )2
+ (a1 b3 − a2 b4 + a3 b1 + a4 b2 − a5 b7 + a6 b8 + a7 b5 − a8 b6 )2
+ (a1 b4 + a2 b3 − a3 b2 + a4 b1 + a5 b8 + a6 b7 − a7 b6 − a8 b5 )2
+ (a1 b5 − a2 b6 + a3 b7 − a4 b8 + a5 b1 + a6 b2 − a7 b3 + a8 b4 )2
+ (a1 b6 + a2 b5 − a3 b8 − a4 b7 − a5 b2 + a6 b1 + a7 b4 + a8 b3 )2
+ (a1 b7 − a2 b8 − a3 b5 + a4 b6 + a5 b3 − a6 b4 + a7 b1 + a8 b2 )2
+ (a1 b8 + a2 b7 + a3 b6 + a4 b5 − a5 b4 − a6 b3 − a7 b2 + a8 b1 )2 (133)
valid for all real numbers ai and bi and thus in particular also for all integers. (There
are many variations of this expression, with different possible sign combinations.) The
theorem was discovered in 1818 by Carl Ferdinand Degen (1766–1825), and then redis-
covered in 1844 by John Graves and in 1845 by Arthur Cayley. There is no generalization
to higher numbers of squares, a fact proved by Adolf Hurwitz (1859–1919) in 1898.
The octonions can be used to show that a vector product can be defined in more than
three dimensions. A vector product or cross product is an operation × satisfying
u × 󰑣 = −󰑣 × u anticommutativity
(u × 󰑣)󰑤 = u(󰑣 × 󰑤) exchange rule. (134)
Using the definition

1
X × Y = (XY − Y X) , (135)
2
the ×-products of imaginary quaternions, i.e., of quaternions of the type (0, u), are again
imaginary, and correspond to the usual vector product, thus fulfilling (134). Interestingly,
Ref. 157 it is possible to use definition (135) for octonions as well. In that case, the product of
Challenge 168 e imaginary octonions is also imaginary, and (134) is again satisfied. In fact, this is the
only other non-trivial example of a vector product. Thus a vector product exists only in
three and in seven dimensions.
Other types of numbers

The process of constructing new systems of hypercomplex ‘numbers’ or real algebras
by ‘doubling’ a given one can be continued ad infinitum. However, octonions, sedenions
and all the following doublings are neither rings nor fields, but only non-associative al-
gebras with unity. Other finite-dimensional algebras with unit element over the reals,
once called hypercomplex ‘numbers’, can also be defined: they include the so-called ‘dual
numbers’, ‘double numbers’, ‘Clifford–Lifshitz numbers’ etc. They play no special role in
physics.
Mathematicians have also defined number fields which have ‘one and a bit’ dimen-
sions, such as algebraic number fields. There is also a generalization of the concept of
Ref. 161 integers to the complex domain: the Gaussian integers, defined as n + im, where n and m
are ordinary integers. Gauss even defined what are now known as Gaussian primes. (Can
Challenge 169 s you find out how?) They are not used in the description of nature, but are important in
number theory.
Physicists used to call quantum-mechanical operators ‘q-numbers.’ But this term has
now fallen out of fashion.

Another way in which the natural numbers can be extended is to include numbers
Ref. 162 larger infinite numbers. The most important such classes of transfinite number are the
Vol. III, page 204 ordinals, the cardinals and the surreals. The ordinals are essentially an extension of the
integers beyond infinity, whereas the surreals are a continuous extension of the reals, also
beyond infinity. Loosely speaking, among the transfinites, the ordinals have a similar role
as the integers have among the reals; the surreals fill in all the gaps between the ordinals,
like the reals do for integers. Interestingly, many series that diverge in ℝ converge in the
Challenge 170 ny surreals. Can you find one example?
The surreals include infinitely small numbers, as do the numbers of nonstandard an-
Ref. 157 alysis, also called hyperreals. In both number systems, in contrast to real numbers, the
numbers 1 and 0.999 999 9... (where an infinite, but hyperfinite string of nines is implied)
Page 206 do not coincide, but are separated by infinitely many other numbers.
Vector spaces
Vector spaces, also called linear spaces, are mathematical generalizations of certain as-
pects of the intuitive three-dimensional space. A set of elements any two of which can
be added together and any one of which can be multiplied by a number is called a vector
space, if the result is again in the set and the usual rules of calculation hold.
More precisely, a vector space over a number field K is a set of elements, called vectors,
for which a vector addition and a scalar multiplication is defined, such that for all vectors
a, b, c and for all numbers s and r from K one has
(a + b) + c = a + (b + c) = a + b + c associativity of vector addition
n+a=a existence of null vector
(−a) + a = n existence of negative vector (136)
1a = a regularity of scalar multiplication
(s + r)(a + b) = sa + sb + ra + rb complete distributivity of scalar multiplication
If the field K, whose elements are called scalars in this context, is taken to be the real (or
complex, or quaternionic) numbers, one speaks of a real (or complex, or quaternionic)
vector space. Vector spaces are also called linear vector spaces or simply linear spaces.
The complex numbers, the set of all real functions defined on the real line, the set of
all polynomials, the set of matrices with a given number of rows and columns, all form
vector spaces. In mathematics, a vector is thus a more general concept than in physics.
Challenge 171 ny (What is the simplest possible mathematical vector space?)
In physics, the term ‘vector’ is reserved for elements of a more specialized type of
vector space, namely normed inner product spaces. To define these, we first need the
concept of a metric space.
A metric space is a set with a metric, i.e., a way to define distances between elements.
A real function d(a, b) between elements is called a metric if
d(a, b) ⩾ 0 positivity of metric

d(a, b) + d(b, c) ⩾ d(a, c) triangle inequality (137)

d(a, b) = 0 if and only if a = b regularity of metric
A non-trivial example is the following. We define a special distance d between cities. If

the two cities lie on a line going through Paris, we use the usual distance. In all other
cases, we define the distance d by the shortest distance from one to the other travelling
Challenge 172 s via Paris. This strange method defines a metric between all cities in France.
A normed vector space is a linear space with a norm, or ‘length’, associated to each a
vector. A norm is a non-negative number ‖a‖ defined for each vector a with the properties
‖ra‖ = |r| ‖a‖ linearity of norm

‖a + b‖ ⩽ ‖a‖ + ‖b‖ triangle inequality (138)
‖a‖ = 0 only if a = 0 regularity
Challenge 173 ny Usually there are many ways to define a norm for a given space. Note that a norm can
always be used to define a metric by setting
d(a, b) = ‖a − b‖ (139)
so that all normed spaces are also metric spaces. This is the natural distance definition
(in contrast to unnatural ones like that between French cities).
The norm is often defined with the help of an inner product. Indeed, the most special
class of linear spaces are the inner product spaces. These are vector spaces with an inner
product, also called scalar product ⋅ (not to be confused with the scalar multiplication!)
which associates a number to each pair of vectors. An inner product space over ℝ satisfies
a⋅b =b⋅a commutativity of scalar product

(ra) ⋅ (sb) = rs(a ⋅ b) bilinearity of scalar product
(a + b) ⋅ c = a ⋅ c + b ⋅ c left distributivity of scalar product
a ⋅ (b + c) = a ⋅ b + a ⋅ c right distributivity of scalar product (140)
a⋅a⩾0 positivity of scalar product
a ⋅ a = 0 if and only if a = 0 regularity of scalar product
for all vectors a, b, c and all scalars r, s. A real inner product space of finite dimension
is also called a Euclidean vector space. The set of all velocities, the set of all positions, or
the set of all possible momenta form such spaces.
An inner product space over ℂ satisfies*
a⋅b =b⋅a=b⋅a Hermitean property

(ra) ⋅ (sb) = rs(a ⋅ b) sesquilinearity of scalar product
(a + b) ⋅ c = a ⋅ c + b ⋅ c left distributivity of scalar product
a ⋅ (b + c) = a ⋅ b + a ⋅ c right distributivity of scalar product (141)

a⋅a⩾0 positivity of scalar product
a ⋅ a = 0 if and only if a = 0 regularity of scalar product
for all vectors a, b, c and all scalars r, s. A complex inner product space (of finite dimen-
sion) is also called a unitary or Hermitean vector space. If the inner product space is
Page 191 complete, it is called, especially in the infinite-dimensional complex case, a Hilbert space.
* Two inequivalent forms of the sesquilinearity axiom exist. The other is (ra) ⋅ (sb) = rs(a ⋅ b). The term
sesquilinear is derived from Latin and means for ‘one-and-a-half-linear’.
The space of all possible states of a quantum system forms a Hilbert space.
All inner product spaces are also metric spaces, and thus normed spaces, if the metric
is defined by
d(a, b) = 󵀄(a − b) ⋅ (a − b) . (142)
Only in the context of an inner product spaces we can speak about angles (or phase
differences) between vectors, as we are used to in physics. Of course, like in normed
spaces, inner product spaces also allows us to speak about the length of vectors and to
define a basis, the mathematical concept necessary to define a coordinate system.
The dimension of a vector space is the number of linearly independent basis vectors.
Challenge 174 ny Can you define these terms precisely?
A Hilbert space is a real or complex inner product space that is also a complete met-
ric space. In other terms, in a Hilbert space, distances vary continuously and behave as
naively expected. Hilbert spaces can have an infinite number of dimensions.
Challenge 175 ny Which vector spaces are of importance in physics?
Mathematical curiosities and fun challenges

Mathematics provides many counter-intuitive results. Reading a book on the topic,
such as Bernard R. Gelbaum & John M. H. Olmsted, Theorems and Counter-
examples in Mathematics, Springer, 1993, can help you sharpen your mind.
∗∗
The distinction between one, two and three dimensions is blurred in mathematics. This is
well demonstrated in the text Hans Sagan, Space Filling Curves, Springer Verlag, 1994.
∗∗
Show that two operators A and B obey
1
eA eB = exp(A + B + [A, B] (143)
2
1 1
+ [[A, B], B] − [[A, B], A] (144)
12 12
1 1
− [B, [A, [A, B]]] − [A, [B, [A, B]]] (145)
48 48
+ ...
(146)
for most operators A and B. This result is often called the Baker–Campbell–Hausdorff
formula or the BCH formula.
CHALLENGE HINT S AND SOLUTIONS
“
Never make a calculation before you know the
”
answer.
John Wheeler’s motto
Challenge 1, page 9: Do not hesitate to be demanding and strict. The next edition of the text
will benefit from it.
Challenge 2, page 14: Classical physics fails in explaining any material property, such as colour
or softness. Material properties result from nature’s interactions; they are inevitably quantum. Ex-
planations of material properties require, without exception, the use of particles and their quan-
tum properties.
Challenge 3, page 16: Classical physics allows any observable to change smoothly with time.
There is no minimum value for any observable physical quantity.
Challenge 4, page 18: The higher the mass, the smaller the motion fuzziness induced by the
quantum of action, because action is mass times speed times distance: For a large mass, the speed
and distance variations are small.
Challenge 5, page 18: The simplest time is 󵀆2Għ/c 5 . The factor 2 is obviously not fixed; it is
changed later on. With the correct factor 4, the time is the shortest time measurable in nature.
Challenge 6, page 18: The electron charge is special to the electromagnetic interactions; it does
not take into account the nuclear interactions or gravity. It is unclear why the length defined with
the elementary charge e should be of importance for neutral systems or for the vacuum. On the
other hand, the quantum of action ħ is valid for all interactions and all observations. However,
one can also argue that the two options to define a fundamental length are not too different, as
the electron charge is related to the quantum of action by e = 󵀄4πε0 αcħ . The two length scales
defined by the two options differ only by a factor near 137.036.
Challenge 8, page 18: On purely dimensional grounds, the radius of an atom must be
ħ2 4πε0
r≈ , (147)
me 2
which is about 160 nm. Indeed, this guess is excellent: it is just π times the Bohr radius.
Challenge 9, page 19: Due to the quantum of action, atoms in all people, be they giants or
dwarfs, have the same size. This implies that giants cannot exist, as was shown already by Galileo.
Vol. I, page 266 The argument is based on the given strength of materials; and a same strength everywhere is
equivalent to the same properties of atoms everywhere. That dwarfs cannot exist is due to a sim-
ilar reason; nature is not able to make people smaller than usual (except in the womb) as this
would require smaller atoms.
Challenge 12, page 24: A disappearance of a mass m in a time Δt is an action change mc 2 Δt.
That is much larger than ħ for all objects of everyday life.
challenge hints and solu tions 205
Challenge 14, page 25: Tunnelling of a lion would imply action values S of the order of S =
100 kgm2 /s ≫ ħ. This cannot happen spontaneously.
Challenge 17, page 26: Yes! Many beliefs and myths are due to the neglect of quantum effects.
Challenge 18, page 26: Continuous flow is in contrast to the fuzziness of motion induced by
the quantum of action.
Challenge 19, page 27: The impossibility of following two particles along their path appears
when their mutual distance d is smaller than their position indeterminacy due to their relative
momentum p, thus when d < ħ/p. Check the numbers with electrons, atoms, molecules, bacteria,
people and galaxies.
Challenge 20, page 27: Also photons are indistinguishable. See page 54.
Challenge 22, page 31: The total angular momentum counts, including the orbital angular mo-
mentum. The orbital angular momentum L is given, using the radius and the linear momentum,
L = r × p.
Challenge 23, page 31: Yes, we could have!
Challenge 24, page 31: That is just the indeterminacy relation. Bohr expanded this idea to all
sort of other pairs of concepts, more in the philosophical domain, such as clarity and precision
of explanations: both cannot be high at the same time.
Challenge 25, page 32: Growth is not proportional to light intensity or to light frequency, but
shows both intensity and frequency thresholds. That is a quantum effects.
Challenge 26, page 32: All effects mentioned above, such as tunnelling, interference, decay,
transformation, non-emptiness of the vacuum, indeterminacy and randomness, are also ob-
served in the nuclear domain.
Challenge 27, page 32: This is not evident from what was said so far, but it turns out to be
correct. In fact, there is no other option, as you will see when you try to find one.
Challenge 29, page 34: The big bang cannot have been an event, for example.
Challenge 36, page 42: Photons are elementary because they realize the minimum action, be-
cause they cannot decay, because they cannot be deformed or split, because they have no mass,
no electric charge and no other quantum number, and because they appear in the Lagrangian of
quantum electrodynamics.
Challenge 40, page 49: To be observable to the eye, the interference fringes need to be visible
for around 0.1 s. That implies a maximum frequency difference between the two beams of around
10 Hz. This is achievable only if either a single beam is split into two or if the two beams come
from high-precision, stabilized lasers.
Ref. 32 Challenge 54, page 60: The calculation is not easy, but not too difficult either. For an initial
orientation close to the vertical, the fall time T turns out to be

1 8
T= T ln (148)
2π 0 α
where α is the starting angle, and a fall through angle π is assumed. Here T0 is the oscillation
time of the pencil for small angles. (Can you determine it?) The indeterminacy relation for the
tip of the pencil yields a minimum starting angle, because the momentum indeterminacy cannot
be made arbitrarily large. You should be able to provide an upper limit. Once this angle is known,
you can calculate the maximum time.
Challenge 55, page 62: Use the temperature to calculate the average kinetic energy, and thus
the average speed of atoms.
206 challenge hints and solu tions
Challenge 56, page 62: The atoms cannot be fully distinguished; they form a state of matter
with peculiar properties, called a condensate. The condensate is not at rest either; but due to its
large mass, its fluctuations are greatly reduced, compared to those of a single atom.
Challenge 58, page 66: Only variables whose product has the same units as physical action –
Js – can be complementary to each other.
Challenge 59, page 66: Use ΔE < E and a Δt < c.
Challenge 68, page 82: Terabyte chips would need to have small memory cells. Small cells im-
ply thin barriers. Thin barriers imply high probabilities for tunnelling. Tunnelling implies lack
of memory.
Challenge 75, page 92: The difficulties to see hydrogen atoms are due to their small size and
their small number of electrons. As a result, hydrogen atoms produce only weak contrasts in X-
ray images. For the same reasons it is difficult to image them using electrons; the Bohr radius of
hydrogen is only slightly larger than the electron Compton wavelength.
For the first time, in 2008, a research team claimed to have imaged hydrogen atoms ad-
sorbed on graphene with the help of a transmission electron microscope. For details, see
J. C. Meyer, C. O. Grit, M. F. Crommle & A. Z et ti, Imaging and dynamics of light
atoms and molecules on graphene, Nature 454, pp. 319–322, 2008. However, it seems that the
report has not been confirmed by another group yet.
Challenge 78, page 93: r = 86 pm, thus T = 12 eV. That compares to the actual value of
13.6 eV. The trick for the derivation of the formula is to use < ψ|rx2 |ψ >= 13 < ψ|rr|ψ >, a re-
lation valid for states with no orbital angular momentum. It is valid for all coordinates and also
for the three momentum observables, as long as the system is non-relativistic.
Challenge 80, page 94: Point particles cannot be marked; nearby point particles cannot be dis-
tinguished, due to the quantum of action.
Challenge 81, page 94: The solution is two gloves. In other words, if two men and two women
want to make love without danger and , they need only two condoms. You can deduce the proce-
dure by yourself.
Challenge 85, page 95: The Sackur–Tetrode formula is best deduced in the following way. We
start with an ideal monoatomic gas of volume V , with N particles, and total energy U . In phase
space, state sum Z is given by
VN 1
Z= . (149)
N ! Λ3N
We use Stirling’s approximation N ! ≈ N N /e N , and the definition of the entropy as S =
∂(kT ln Z)/∂T . Inserting the definition of Λ, this gives the Sackur–Tetrode equation.
Challenge 86, page 97: For a large number of particles, the interaction energy will introduce
errors. For very large numbers, the gravitational binding energy will do so as well.
Challenge 88, page 98: To write two particles on paper, one has to distinguish them, even if
the distinction is arbitrary.
Challenge 89, page 99: Trees, like all macroscopic objects, have a spin value that depends on
their angular momentum. Being classical objects whose phase can be observed, the spin value is
uncertain. (But even large multiples of ħ are too small to be measurable in everyday life.) Gener-
ally speaking, trees, mountains and people are spin 1 objects. The spin 1 value implies that these
objects are unchanged after a full rotation. (How does a block of silver, made of an odd number
of silver atoms, each with spin 1/2, relate to this answer?)
Challenge 93, page 104: Twins differ in the way their intestines are folded, in the lines of their
hands and other skin folds. Sometimes, but not always, features like black points on the skin are
mirror inverted on the two twins.
challenge hints and solu tions 207
Challenge 99, page 113: Three.

Challenge 100, page 114: Angels can be distinguished by name, can talk and can sing; thus
they are made of a large number of fermions. In fact, many angels are human sized, so that they
do not even fit on the tip of a pin.
Challenge 108, page 119: Ghosts, like angels, can be distinguished by name, can talk and can
be seen; thus they contain fermions. However, they can pass through walls and they are trans-
parent; thus they cannot be made of fermions, but must be images, made of bosons. That is a
contradiction.
Challenge 110, page 123: The loss of non-diagonal elements leads to an increase in the diago-
nal elements, and thus of entropy.
Challenge 113, page 128: The energy speed is given by the advancement of the outer two tails;
that speed is never larger than the speed of light.
Challenge 114, page 130: No, as taking a photo implies an interaction with a bath, which
would destroy the superposition.
Challenge 115, page 131: A photograph requires illumination; illumination is a macroscopic
electromagnetic field; a macroscopic field is a bath; a bath implies decoherence; decoherence
destroys superpositions.
Challenge 118, page 132: It depends. They can be due to interference or to intensity sums. In
the case of radio the effect is clearer. If at a particular frequency the signals changes periodically
from one station to another, one has a genuine interference effect.
Challenge 120, page 132: Such a computer requires clear phase relations between components;
such phase relations are extremely sensitive to outside disturbances. At present, they do not hold
longer than a microsecond, whereas long computer programs require minutes and hours to run.
Challenge 124, page 139: Any other bath also does the trick, such as the atmosphere, sound
vibrations, electromagnetic fields, etc.
Challenge 125, page 140: The Moon is in contact with baths like the solar wind, falling mete-
orites, the electromagnetic background radiation of the deep universe, the neutrino flux from
the Sun, cosmic radiation, etc.
Challenge 126, page 141: Spatially periodic potentials have the property. Decoherence then
leads to momentum diagonalization.
Challenge 138, page 167: This is a trick question. A change in α requires a change in c, ħ, e
or ε0 . None of these changes is possible or observable, as all our measurement apparatus are
based on these units. Speculations about change of α, despite their frequency in the press and in
scientific journals, are idle talk.
Challenge 139, page 167: A change of physical units such that ħ = c = e = 1 would change the
value of ε0 in such a way that 4πεo = 1/α = 137.036...

Challenge 145, page 179: Planck limits can be exceeded for extensive observables for which
many particle systems can exceed single particle limits, such as mass, momentum, energy or
electrical resistance.
Challenge 147, page 182: Do not forget the relativistic time dilation.
Challenge 148, page 182: The formula with n − 1 is a better fit. Why?
Challenge 152, page 187: No, only properties of parts of the universe are listed. The universe
Vol. VI, page 98 itself has no properties, as shown in the last volume..
Challenge 151, page 187: The slowdown goes quadratically with time, because every new slow-
down adds to the old one!
208 challenge hints and solu tions
Challenge 153, page 189: The double of that number, the number made of the sequence of all
even numbers, etc.
a b
Challenge 156, page 193: |z|2 is the determinant of the matrix z = 󶀥 󶀵.
−b a
Challenge 160, page 193: Use Cantor’s diagonal argument, as in challenge 230.
Challenge 162, page 194: Any quaternion X = ai + b j + ck with a2 + b2 + c 2 = 1 solves the
equation X 2 +1 = 0; the purely imaginary solutions +i and −i are thus augmented by a continuous
sphere of solutions in quaternion space.
Challenge 165, page 196: Any rotation by an angle 2π is described by −1. Only a rotation by
4π is described by +1; quaternions indeed describe spinors.
Challenge 167, page 198: Just check the result component by component. See also the men-
tioned reference.
Challenge 169, page 200: For a Gaussian integer n + im to be prime, the integer n2 + m2 must
be prime, and in addition, a condition on n mod 3 must be satisfied; which one and why?
Challenge 172, page 201: The metric is regular, positive definite and obeys the triangle in-
equality.
Challenge 177, page 222: This could be solved with a trick similar to those used in the irra-
tionality of each of the two terms of the sum, but nobody has found one.
Challenge 178, page 222: There are still many discoveries to be made in modern mathematics,
especially in topology, number theory and algebraic geometry. Mathematics has a good future.

BI BLIO GR APHY
“
No man but a blockhead ever wrote except for
”
money.
Samuel Johnson
1 Giuseppe Fumagalli, Chi l’ha detto?, Hoepli, Milano, 1983. Cited on page 14.
2 The quantum of action was introduced in Max Pl anck, Über irreversible Strahlungs-
vorgänge, Sitzungsberichte der Preußischen Akademie der Wissenschaften, Berlin pp. 440–
480, 1899. In the paper, Planck used the letter b for what nowadays is called h. Cited on
page 15.
3 Bohr explained the indivisibilty of the quantum of action in his famous Como lecture.
N. B ohr, Atomtheorie und Naturbeschreibung, Springer, 1931. On page 16 he writes: ‘No
more is it likely that the fundamental concepts of the classical theories will ever become su-
perfluous for the description of physical experience. The recognition of the indivisibility of
the quantum of action, and the determination of its magnitude, not only depend on an an-
alysis of measurements based on classical concepts, but it continues to be the application of
these concepts alone that makes it possible to relate the symbolism of the quantum theory
to the data of experience.’ He also writes: ‘...the fundamental postulate of the indivisibility
of the quantum of action is itself, from the classical point of view, an irrational element
which inevitably requires us to forgo a causal mode of description and which, because of
the coupling between phenomena and their observation, forces us to adopt a new mode of
description designated as complementary in the sense that any given application of classi-
cal concepts precludes the simultaneous use of other classical concepts which in a different
connection are equally necessary for the elucidation of the phenomena ...’ and ‘...the finite
magnitude of the quantum of action prevents altogether a sharp distinction being made
between a phenomenon and the agency by which it is observed, a distinction which under-
lies the customary concept of observation and, therefore, forms the basis of the classical
ideas of motion.’ Other statements about the indivisibility of the quantum of action can be
found in N. B ohr, Atomic Physics and Human Knowledge, Science Editions, 1961. See also
Max Jammer, The Philosophy of Quantum Mechanics, Wiley, first edition, 1974, pp. 90–91.
Cited on page 15.
4 For some of the rare modern publications emphasizing the quantum of action see
M. B. Mensky, The action uncertainty principle and quantum gravity, Physics Letters
A 162, p. 219, 1992, and M. B. Mensky, The action uncertainty principle in continuous
quantum measurements, Physics Letters A 155, pp. 229–235, 1991. Schwinger’s quantum-
action principle is also used in R ichard F. W. Bader, Atoms in Molecules – A Quantum
Theory, Oxford University Press, 1994.
There is a large number of general textbooks on quantum theory. There is one for every
taste.
210 biblio graphy
A well-known conceptual introduction is Jean-Marc L év y-L eblond &

Françoise Balibar, Quantique – Rudiments, Masson, 1997, translated into English
as Quantics, North-Holland, 1990.
One of the most beautiful books is Julian S chwinger, Quantum Mechanics – Sym-
bolism of Atomic Measurements, edited by Berthold-Georg Englert, Springer Verlag, 2001.
A modern approach with a beautiful introduction is Max S chubert & Ger-
hard Weber, Quantentheorie – Grundlagen und Anwendungen, Spektrum Akademischer
Verlag, 1993.
A standard beginner’s text is C. Cohen-Tannoudji, B. Diu & F. L aloë, Mé-
canique quantique I et II, Hermann, Paris, 1977. It is also available in several translations.
A good text is Asher Peres, Quantum Theory – Concepts and Methods, Kluwer, 1995.
For a lively approach, see Vincent Icke, The Force of Symmetry, Cambridge Univer-
sity Press, 1994.
New textbooks are published regularly around the world. Cited on page 15.
5 The best source for the story about the walk in the forest with Planck’s son Erwin is
Hans Ro os & Armin Hermann, editors, Max Planck – Vorträge, Reden, Erinnerungen,
Springer, 2001, page 125. As the text explains, the story was told by Erwin Planck to at least
two different people. Erwin Planck himself was part of the failed 1944 plot against Hitler
and was hanged in January 1945. Cited on page 18.
6 Max B orn, Zur Quantenmechanik der Stoßvorgänge (vorläufige Mitteilung), Zeitschrift
für Physik 37, pp. 863–867, 1926, Max B orn, Quantenmechanik der Stoßvorgänge,
Zeitschrift für Physik 38, pp. 803–827, 1926. Cited on page 21.
7 See for example the papers by Jan Hilgevo ord, The uncertainty principle for energy
and time, American Journal of Physics 64, pp. 1451–1456, 1996, and by Paul Busch, On
the time–energy uncertainty reaction, parts 1 & 2, Foundations of Physics 20, pp. 1–43, 1990.
A classic is the paper by Eugene P. Wigner, On the time–energy uncertainty relation, in
Abdus Sal am & Eugene P. Wigner, editors, Aspects of Quantum Theory, Cambridge
University Press, 1972. Cited on page 23.
8 See also the booklet by Cl aus Mat theck, Warum alles kaputt geht – Form und Versagen
in Natur und Technik, Forschungszentrum Karlsruhe, 2003. Cited on page 25.
9 R. Clifton, J. Bub & H. Halvorson, Characterizing quantum theory in terms of
information-theoretic constraints, arxiv.org/abs/quant-ph/0211089. Cited on page 31.
10 This way to look at cans of beans goes back to the text by Susan Hewit t & Ed-
ward Subitzky, A call for more scientific truth in product warning labels, Journal of
Irreproducible Results 36, nr. 1, 1991. Cited on page 32.
11 J. Malik, The yields of the Hiroshima and Nagasaki nuclear explosions, Technical Report
LA-8819, Los Alamos National Laboratory, September 1985. Cited on page 33.
12 The quotes on motion are found in chapter VI of F. Engels, Herrn Eugen Dührings
Umwälzung der Wissenschaft, Verlag für fremdsprachliche Literatur, Moskau, 1946. The
book is commonly called Anti-Dühring. Cited on pages 34 and 62.
13 Rodney Loud on, The Quantum Theory of Light, Oxford University Press, 2000. Cited
on page 35.
14 E. M. Brumberg & S. I. Vavilov, Izvest. Akad. Nauk. Omen Ser. 7, p. 919, 1933. Cited
on page 35.
15 F. R ieke & D. A. Baylor, Single-photon detection by rod cells of the retina, Reviews of
Modern Physics 70, pp. 1027–1036, 1998. They also mention that the eye usually works at
photon fluxes between 108 /μm2 s (sunlight) and 10−2 /μm2 s (starlight). The cones, in the
biblio graphy 211
retina detect, in colour, light intensities in the uppermost seven or eight decades, whereas
the rods detect, in black and white, the lower light intensities. Cited on page 39.
16 E. Fischbach, H. Klo or, R. A. L angel, A. T. Y. Lui & M. Pered o, New geomag-
netic limit on the photon mass and on long-range forces coexisting with electromagnetism,
Physical Review Letters 73, pp. 514–517, 1994. Cited on page 40.
17 A. H. Compton, The scattering of X-rays as particles, American Journal of Physics 29,
pp. 817–820, 1961. This is a pedagogical presentation of the discoveries he made in 1923.
Cited on page 40.
18 The famous paper is R. Hanbury Brown & R. Q. Twiss, Nature 178, p. 1046, 1956.
They got the idea to measure light in this way from their earlier work, which used the same
method with radio waves: R. Hanbury Brown & R. Q. Twiss, Nature 177, p. 27, 1956,
Cited on page 46.
19 J. Gl anz, First light from a space laser, Science 269 p. 1336, 8 September 1995. Cited on
page 46.
20 A. Einstein, Über einen die Erzeugung und Umwandlung des Lichtes betreffenden
heuristischen Standpunkt, Annalen der Physik 17, pp. 132–184, 1905. Cited on page 47.
21 See the summary by P. W. Milonni, Answer to question 45: What (if anything) does the
photoelectric effect teach us?, American Journal of Physics 65, pp. 11–12, 1997. Cited on page
47.
22 For a detailed account, See J. J. Prentis, Poincaré ’s proof of the quantum discontinu-
ity of nature, American Journal of Physics 63, pp. 339–350, 1995. The original papers are
Henri Poincaré, Sur la théorie des quanta, Comptes Rendus de l’Académie des Sciences
(Paris) 153, pp. 1103–1108, 1911, and Henri Poincaré, Sur la théorie des quanta, Journal
de Physique (Paris) 2, pp. 5–34, 1912. Cited on page 47.
23 J. Jacobson, G. Björk, I. Chang & Y. Yamamoto, Photonic de Broglie waves, Phys-
ical Review Letters 74, pp. 4835–4838, 1995. The first measurement was published by
E. J. S. Fonseca, C. H. Monken & S. de Pádua, Measurement of the de Broglie wave-
length of a multiphoton wave packet, Physical Review Letters 82, pp. 2868–2671, 1995. Cited
on page 48.
24 For the three-photon state, see M. W. Mitchell, J. S. Lundeen & A. M. Steinberg,
Super-resolving phase measurements with a multiphoton entangled state, Nature 429,
pp. 161–164, 2004, and for the four-photon state see, in the same edition, P. Walther,
J. -W. Pan, M. Aspelmeyer, R. Ursin, S. Gasparoni & A. Z eilinger, De Broglie
wavelength of a non-local four-photon state, Nature 429, pp. 158–161, 2004. Cited on page
48.
25 For an introduction to spueezed light, see L. Mandel, Non-classical states of the electro-
magnetic field, Physica Scripta T 12, pp. 34–42, 1986. Cited on page 48.
26 The famous quote on single-photon interference is found on page 9 of P. A. M. Dirac,
The Principles of Quantum Mechanics, Clarendon Press, 1930. It is also discussed, in a some-
what confused way, in the otherwise informative article by H. Paul, Interference between
independent photons, Reviews of Modern Physics 58, pp. 209–231, 1986. Cited on page 51.
27 The original papers on coherent states are three: R. J. Gl auber, The quantum theory of
optical coherence, Physical Review 130, pp. 2529–2539, 1963, J. R. Kl auder, Continuous-
representation theory, I and II, Journal of Mathematical Physics 4, pp. 1055–1058, 1963, and
E. C. G. Sudarshan, Equivalence of semiclassical and quantum mechanical descriptions
of statistical light beams, Physical Review Letters 10, p. 227, 1963. Cited on page 53.
212 biblio graphy
28 R ichard P. Feynman, QED - The Strange Theory of Light and Matter, pp. 73–75, Prince-
ton University Press, 1988, or R ichard P. Feynman & Steven Weinberg, Elemen-
tary Particles and the Laws of Physics, p. 23, Cambridge University Press 1987. Cited on
page 54.
29 Wolf gang Tit tel, J. Brendel, H. Z binden & N. Gisin, Violation of Bell inequal-
ities by photons more than 10 km apart, Physical Review Letters 81, pp. 3563–3566, 26 Oc-
tober 1998. Cited on page 55.
30 N. B ohr & L. Rosenfeld, Zur Frage der Meßbarkeit der elektromagnetischen Feld-
größen, Mat.-fys. Medd. Danske Vid. Selsk. 12, p. 8, 1933. The results were later published
in English as N. B ohr & L. Rosenfeld, Field and charge measurements in quantum
electrodynamics, Physical Review 78, pp. 794–798, 1950. Cited on page 55.
31 R. Kidd, J. Aedini & A. Anton, Evolution of the modern photon, American Journal of
Physics 57, pp. 27–35, 1989, Cited on page 58.
32 E. I. Bu tikov, The rigid pendulum – an antique but evergreen physical model, European
Journal of Physics 20, pp. 429–441, 1999. Cited on page 205.
33 The whole bunch of atoms behaves as one single molecule; one speaks of a Bose–Einstein
condensate. The first observations, worthy of a Nobel prize, were by M.H. Anderson
& al., Observation of Bose–Einstein condensation in a dilute atomic vapour, Science 269,
pp. 198–201, 1995, C. C. Bradley, C. A. Sacket t, J. J. Tollet t & R. G. Hulet, Evi-
dence of Bose–Einstein condensation in an atomic gas with attractive interactions, Physical
Review Letters 75, pp. 1687–1690, 1995, K. B. Davis, M. -O. Mewes, M. R. Andrews,
N. J. van Dru ten, D. S. D urfee, D. M. Kurn & W. Ket terle, Bose–Einstein con-
densation in a gas of sodium atoms, Physical Review Letters 75, pp. 3969–3973, 1995. For a
simple introduction, see W. Ket terle, Experimental studies of Bose–Einstein condensa-
tion, Physics Today pp. 30–35, December 1999. Cited on page 62.
34 J. L. Costa-Krämer, N. Garcia, P. García-Mo chales & P. A. Serena,
Nanowire formation in macroscopic metallic contacts: a universal property of metals,
Surface Science Letters 342, pp. L1144–L1152, 1995. See also J. L. Costa-Krämer,
N. Garcia, P. A. Serena, P. García-Mo chales, M. Marqués & A. Correia,
Conductance quantization in nanowires formed in macroscopic contacts, Physical Review
B p. 4416, 1997. Cited on page 62.
35 The beautiful undergraduate experiments made possible by this discovery are desribed in
E. L. Foley, D. Candel a, K. M. Martini & M. T. Tuominen, An undergraduate
laboratory experiment on quantized conductance in nanocontacts, American Journal of
Physics 67, pp. 389–393, 1999. Cited on pages 62 and 63.
36 L. de Bro glie, Comptes rendus de l’Académie des Sciences 177, pp. 507–510, 1923. Cited
on page 63.
37 C. Jönsson, Interferenz von Elektronen am Doppelspalt, Zeitschrift für Physik 161,
pp. 454–474, 1961, C. Jönsson, Electron diffraction at multiple slits, American Journal of
Physics 42, pp. 4–11, 1974. Because of the charge of electons, this experiment is not easy to
perform: any parts of the set-up that are insulating get charged and distort the picture. That
is why the experient was performed much later with electrons than with atoms, neutrons
and molecules. Cited on page 65.
38 M. Arndt, O. Nairz, J. Vos-Andreae, C. Keller, van der Z ouw &
A. Z eilinger, Wave–particle duality of C60 molecules, Nature 401, pp. 680–682, 14
October 1999. See also the observation for tetraphenyleprophyrin and C60 F48 by the same
team, published as L. Hackermüller & al., Wave nature of biomolecules and fluoro-
fullerenes, Physical Review Letters 91, p. 090408, 2003.
biblio graphy 213
No phenomoenon of quantum theory has been experimentally studied as much as

quantum intereference. The transition from interference to non-interference has also been
explored, as in P. Facchi, A. Mariano & S. Pascazio, Mesoscopic interference,
Recent Developments in Physics 3, pp. 1–29, 2002. Cited on page 65.
39 G. Papini, Shadows of a maximal acceleration, arxiv.org/abs/gr-qc/0211011. Cited on page
66.
40 J. Perrin, Nobel Prize speech, found at www.nobel.se. H. Nagaoka, Kinetics of a sys-
tem of particles illustrating the line and the band spectrum and the phenomena of radioac-
tivity, Philosophical Magazine S6, 7, pp. 445–455, March 1904. Cited on page 67.
41 N. B ohr, On the constitution of atoms and molecules: Introduction and Part I – binding
of electrons by positive nuclei, Philosophical Magazine 26, pp. 1–25, 1913, On the consti-
tution of atoms and molecules: Part II – systems containing only a single nucleus, ibid.,
pp. 476–502, On the constitution of atoms and molecules: Part III, ibid., pp. 857–875. Cited
on page 67.
42 Robert H. Dicke & James P. Wit tke, Introduction to Quantum Theory, Addison-
Wesley, Reading, Massachusetts, 1960. See also Stephen Gasiorowicz, Quantum
Physics, John Wiley & Sons, 1974. Cited on page 68.
43 P. Carru thers & M. M. Nieto, Phase and angle variables in quantum mechanics, Re-
view of Modern Physics 40, pp. 411–440, 1968. Cited on page 69.
44 The indeterminacy relation for rotational motion is well explained by W. H. Louisell,
Amplitude and phase uncertainty relations, Physics Letters 7, p. 60, 1963. Cited on page 69.
45 S. Franke-Arnold, S. M. Barnet t, E. Yao, J. L each, J. Courtial &
M. Pad get t, Uncertainty principle for angular position and angular momentum, New
Journal of Physics 6, p. 103, 2004. This is a freely accessible online journal. Cited on page
69.
46 W. Gerl ach & O. Stern, Der experimentelle Nachweis des magnetischen Moments des
Silberatoms, Zeitschrift für Physik 8, p. 110, 1921. See also the pedagogical explanation by
M. Hannou t, S. Hoyt, A. Kryowonos & A. Wid om, Quantum measurement and
the Stern–Gerlach experiment, American Journal of Physics 66, pp. 377–379, 1995. Cited on
page 70.
47 E. S chrödinger, Quantisierung als Eigenwertproblem I, Annalen der Physik 79, pp. 361–
376, 1926, and Quantisierung als Eigenwertproblem II, Annalen der Physik 79, pp. 489–527,
1926. Cited on page 77.
48 C. G. Gray, G. Karl & V. A. Novikov, From Maupertius to Schrödinger. Quantiza-
tion of classical variational principles, American Journal of Physics 67, pp. 959–961, 1999.
Cited on page 77.

49 Y. Aharonov & D. B ohm, Significance of electromagnetic potentials in the quantum
theory, Physical Review 115, pp. 485–491, 1959. Cited on page 83.
50 R. Colell a, A. W. Overhauser & S. A. Werner, Observation of gravitationally in-
duced quantum interference, Physical review Letters 34, pp. 1472–1474, 1975. Cited on page
85.
51 G. E. Uhlenbeck & S. G oudsmit, Ersetzung der Hypothese vom unmechanischen
Zwang durch eine Forderung bezüglich des inneren Verhaltens jedes einzelnen Elektrons,
Naturwissenschaften 13, pp. 953–954, 1925. Cited on page 87.
52 L. Thomas, The motion of the spinning electron, Nature 117, p. 514, 1926. Cited on page
88.
214 biblio graphy
53 K. von Meyenn & E. S chucking, Wolfgang Pauli, Physics Today pp. 43–48, February
54 L. L. Foldy & S. A. Wou thuysen, On the Dirac theory of spin 1/2 particles and its
nonrelativistic limit, Physical Review 78, pp. 29–36, 1950. Cited on page 89.
55 John P. Costell a & Bruce H. J. McKell ar, The Foldy–Wouthuysen transforma-
tion, American Journal of Physics 63, pp. 1119–1121, 1995. Cited on page 90.
56 For an account of the first measuremnt of the g-factor of the electron, see HR Crane, How
we happended to measure g-2: a tale of serendipity, Physics in Perspective 2, pp. 135–140,
2000. The most interesting part is how the experimentalists had to overcome the conviction
of almost all theorists that the measurement was impossible in principle. Cited on page 90.
57 The д-factors for composite nuclei are explained briefly on en.wikipedia.org/wiki/
Nuclear_magnetic_moment and measured values are found at www-nds.iaea.org. See
also Hans Dehmelt, Is the electron a composite particle?, Hyperfine Interactions 81,
pp. 1–3, 1993. Cited on page 91.
58 J. P. Woerdman, G. Nienhuis, I. Kuščer, Is it possible to rotate an atom?, Op-
tics Communications 93, pp. 135–144, 1992. We are talking about atoms rotating around
their centre of mass; atoms can of course rotate around other bodies, as discussed by
M. P. Silverman, Circular birefringence of an atom in uniform rotation: the classical
perspective, American Journal of Physics 58, pp. 310–317, 1990. Cited on page 72.
59 The nearest anyone has come to an image of a hydrogen atom is found in A. Yazdani,
Watching an atom tunnel, Nature 409, pp. 471–472, 2001. The experiments on Bose–Einstein
condensates are also candidates for images of hydrogen atoms. The company Hitachi made
a fool of itself in 1992 by claiming in a press release that its newest electron microscope
could image hydrogen atoms. Cited on page 92.
60 J. S chmiedmayer, M. S. Chapman, C. R. Ekstrom, T. D. Hammond,
S. Wehinger & D. E. Pritchard, Index of refraction of various gases for sodium
matter waves, Physical Review Letters 74, p. 1043-1046, 1995. Cited on page 72.
61 A. M. Wolsky, Kinetic energy, size, and the uncertainty principle, American Journal of
Physics 42, pp. 760–763, 1974. Cited on page 93.
62 See the paper by Martin Gardner, Science fiction puzzle tales, Clarkson Potter, 67,
pp. 104–105, 1981, or his book Aha! Insight, Scientific American & W.H. Freeman, 1978. The
rabbit story is from A. Hajnal & P. Lovász, An algorithm to prevent the propagation of
certain diseases at minimum cost, in Interfaces Between Computer Science and Operations
Research, edited by J. K. L enstra, A. H. G. R inno oy Kan & P. Van Emde B oas,
Mathematisch Centrum, Amsterdam 1978, whereas the computer euphemism is used by
A. Orlitzky & L. Shepp, On curbing virus propagation, Technical memorandum, Bell
Labs 1989. Cited on page 94.

63 A complete discussion of the problem can be found in chapter 10 of Il an Vardi, Computa-
tional Recreations in Mathematica, Addison Wesley, Redwood City, California, 1991. Cited
on page 94.
64 On Gibbs’ paradox, see your favourite text on thermodynamics or statistical mechanics.
See also W. H. Zurek, Algorithmic randomness and physical entropy, Physical Review A
40, pp. 4731–4751, 1989. Zurek shows that the Sackur–Tetrode formula can be derived from
algorithmic entropy considerations. Cited on page 95.
65 S. N. B ose, Plancks Gesetz und Lichtquantenhypothese, Zeitschrift für Physik 26, pp. 178–
181, 1924. The theory was then expanded in A. Einstein, Quantentheorie des einatomigen
idealen Gases, Sitzungsberichte der Preussischen Akademie der Wissenschaften zu Berlin 22,
biblio graphy 215
pp. 261–267, 1924, A. Einstein, Quantentheorie des einatomigen idealen Gases. Zweite
Abhandlung, Sitzungsberichte der Preussischen Akademie der Wissenschaften zu Berlin 23,
pp. 3–14, 1925, A. Einstein, Zur Quantentheorie des idealen Gases, Sitzungsberichte der
Preussischen Akademie der Wissenschaften zu Berlin 23, pp. 18–25, 1925. Cited on page 99.
66 C. K. Hong, Z. Y. O u & L. Mandel, Measurement of subpicosecond time intervals be-
tween two photons by interference, Physical Review Letters 59, pp. 2044–2046, 1987. Cited
on page 99.
67 M. S chellekens, R. Hoppeler, A. Perrin, J. Viana G omes, D. B oiron,
C. I. Westbro ok & A. Aspect, Hanbury Brown Twiss effect for ultracold quan-
tum gases, Science 310, p. 648, 2005, preprint at arxiv.org/abs/cond-mat/0508466.
J. Viana G omes, A. Perrin, M. S chellekens, D. B oiron, C. I. Westbro ok
& M. Belsley, Theory for a Hanbury Brown Twiss experiment with a ballistically ex-
panding cloud of cold atoms, Physical Review A 74, p. 053607, 2006, preprint at arxiv.org/
abs/quant-ph/0606147. T. Jeltes, J. M. McNamara, W. Ho gervorst, W. Vassen,
V. Krachmalnicoff, M. S chellekens, A. Perrin, H. Chang, D. B oiron,
A. Aspect & C. I. Westbro ok, Comparison of the Hanbury Brown-Twiss effect for
bosons and fermions, Nature 445, p. 402, 2007, preprint at arxiv.org/abs/cond-mat/0612278.
Cited on page 100.
68 The experiment is described in E. R amberg & G. A. Snow, Experimental limit on a
small violation of the Pauli principle, Physics Letters B 238, pp. 438–441, 1990. Other exper-
imental tests are reviewed in O. W. Greenberg, Particles with small violations of Fermi
or Bose statistics, Physical Review D 43, pp. 4111–4120, 1991. Cited on page 102.
69 The original no-cloning theorem is by D. Dieks, Communication by EPR devices, Physics
Letters A 92, pp. 271–272, 1982, and by W. K. Wo ot ters & W. H. Zurek, A single quan-
tum cannot be cloned, Nature 299, pp. 802–803, 1982. For a discussion of photon and
multiparticle cloning, see N. Gisin & S. Massar, Optimal quantum cloning machines,
Physics Review Letters 79, pp. 2153–2156, 1997. The whole topic has been presented in detail
by V. Buzek & M. Hillery, Quantum cloning, Physics World 14, pp. 25–29, November
70 The most recent experimental and theoretical results on physical cloning are described
in A. L amas-L inares, C. Simon, J. C. Howell & D. B ouwmeester, Experimen-
tal quantum cloning of single photons, Science 296, pp. 712 – 714, 2002, D. Collins
& S. Popescu, A classical analogue of entanglement, preprint arxiv.org/abs/quant-ph/
0107082, 2001, and A. Daffertshofer, A. R. Pl astino & A. Pl astino, Classical
no-cloning theorem, Physical Review Letters 88, p. 210601, 2002. Cited on page 104.
71 E. Wigner, On unitary representations of the inhomogeneous Lorentz group, Annals of
Mathematics 40, pp. 149–204, 1939. This famous paper summarises the work which later
brought him the Nobel Prize in Physics. Cited on page 105.
72 For a full list of isotopes, see R. B. Firestone, Table of Isotopes, Eighth Edition, 1999 Up-
date, with CDROM, John Wiley & Sons, 1999. Cited on page 107.
73 G. Gabrielse, H. Dehmelt & W. Kells, Observation of a relativistic, bistable hys-
teresis in the cyclotron motion of a single electron, Physical Review Letters 54, pp. 537–540,
74 This is deduced from the д − 2 measurements, as explained in his Nobel-prize talk by
Hans Dehmelt, Experiments with an isolated subatomic particle at rest, Reviews of Mod-
ern Physics 62, pp. 525–530, 1990, and in Hans Dehmelt, Is the electron a composite
particle?, Hyperfine Interactions 81, pp. 1–3, 1993. Cited on page 109.
216 biblio graphy
75 Wolf gang Pauli, The connection between spin and statistics, Physical Review 58,
pp. 716– 722, 1940. Cited on page 110.
76 This famous explanation is given, for example, on page 1148 in C. W. Misner,
K. S. Thorne & J. A. Wheeler, Gravitation, Freeman, 1973. It is called the scissor
trick on page 43 of volume 1 of R. Penrose & W. R indler, Spinors and Spacetime, 1984.
It is also cited and discussed by R. G ould, Answer to question #7, American Journal of
Physics 63, p. 109, 1995. Cited on page 111.
77 M. V. Berry & J. M. Robbins, Indistinguishability for quantum particles: spin, statis-
tics and the geometric phase, Proceedings of the Royal Society in London A 453, pp. 1771–
1790, 1997. See also the comments to this result by J. Twamley, Statistics given a
spin, Nature 389, pp. 127–128, 11 September 1997. Their newer results are M. V. Berry
& J. M. Robbins, Quantum indistinguishability: alternative constructions of the trans-
ported basis, Journal of Physics A (Letters) 33, pp. L207–L214, 2000, and M. V. Berry &
J. M. Robbins, in Spin-Statistics, eds. R. Hilborn & G. Tino, (American Institute of
Physics), 2000, pp. 3–15. See also Michael Berry’s home page on the www.phy.bris.ac.uk/
staff/berry_mv.html website. Cited on page 112.
78 R. W. Hartung, Pauli principle in Euclidean geometry, American Journal of Physics 47,
pp. 900–910, 1979. Cited on page 113.
79 The issue is treated in his Summa Theologica, in question 52 of the first part. The complete
text, several thousand pages, can be found on the www.newadvent.org website. Cited on
page 114.
80 The point that spin can be seen as a rotation was already made by F. J. Belinfante, On
the spin angular momentum of mesons, Physica 6, p. 887, 1939, and taken up again by
Hans C. Ohanian, What is spin?, American Journal of Physics 54, pp. 500–505, 1986.
See also E. D uran & A. Erschow, Physikalische Zeitschrift der Sowjetunion 12, p. 466,
81 Generalizations of bosons and fermions are reviewed in the (serious!) paper by
O. W. Greenberg, D. M. Greenberger & T. V. Greenbergest, (Para)bosons,
(para)fermions, quons and other beasts in the menagerie of particle statistics, at arxiv.org/
abs/hep-th/9306225. A newer summary is O. W. Greenberg, Theories of violation of
statistics, electronic preprint available at arxiv.org/abs/hep-th/0007054, 2000. Cited on
page 119.
82 Gell-Mann wrote this for the 1976 Nobel Conference (not for the Nobel speech; he is the
only winner who never published it.) M. Gell-Mann, What are the building blocks of
matter?, in D. Huff & O. Prewit t, editors, The Nature of the Physical Universe, New
York, Wiley, 1979, p. 29. Cited on page 120.
83 See e.g. the reprints of his papers in the standard collection by John A. Wheeler &
Wojciech H. Zurek, Quantum Theory and Measurement, Princeton University Press,
84 H. D. Z eh, On the interpretation of measurement in quantum theory, Foundations of
Physics 1, pp. 69–76, 1970. Cited on page 121.
85 L inda R eichl, A Modern Course in Statistical Physics, Wiley, 2nd edition, 1998. An excel-
lent introduction into thermodynamics. Cited on page 122.
86 E. Jo os & H. D. Z eh, The emergence of classical properties through interactions with the
environment, Zeitschrift für Physik B 59, pp. 223–243, 1985. See also Erich Jo os, Decoher-
ence and the appearance of a classical world in quantum theory, Springer Verlag, 2003. Cited
on page 123.
biblio graphy 217
87 M. Tegmark, Apparent wave function collapse caused by scattering, Foundation of

Physics Letters 6, pp. 571–590, 1993, preprint at arxiv.org/abs/gr-qc/9310032. See also his pa-
per that shows that the brain is not a quantum computer, M. Tegmark, The importance
of quantum decoherence in brain processes, Physical Review E 61, pp. 4194–4206, 2000,
preprint at arxiv.org/abs/quant-ph/9907009. Cited on page 124.
88 The decoherence time is bound from above by the relaxation time. See A. O. Caldeira &
A. J. L egget t, Influence of damping on quantum interference: an exactly soluble model,
Physical Review A 31, 1985, pp. 1059–1066. This is the main reference about effects of deco-
herence for a harmonic oscillator. The general approach to relate decoherence to the influ-
ence of the environment is due to Niels Bohr, and has been pursued in detail by Hans Dieter
Zeh. Cited on page 125.
89 G. L indbl ad, On the generators of quantum dynamical subgroups, Communications in
Mathematical Physics 48, pp. 119–130, 1976. Cited on page 125.
90 Wojciech H. Zurek, Decoherence and the transition from quantum to classical, Physics
Today pp. 36–44, October 1991. An easy but somewhat confusing article. His reply to the
numerous letters of response in Physics Today, April 1993, pp. 13–15, and pp. 81–90, exposes
his ideas in a clearer way and gives a taste of the heated discussions on this topic. Cited on
pages 126 and 130.
91 John Bardeen, explained this regularly in the review talks he gave at the end of his life,
such as the one I heard in Tokyo in 2000. Cited on page 126.
92 The first decoherence measurement was performed by M. Brune & al., Observing pro-
gressive decoherence of the “meter” in a quantum measurement, Physical Review Letters 77,
pp. 4887–4890, 9 December 1996. Cited on page 126.
93 Later experiments confirming the numerical predictions from decoherence were published
by C. Monroe, D. M. Meekhof, B. E. King & D. J. Winel and, A “Schrödinger
cat” superposition state of an atom, Science 272, pp. 1131–1136, 1996, W. P. S chleich,
Quantum physics: engineering decoherence, Nature 403, pp. 256–257, 2000, C. J. Myat t,
B. E. King, Q. A. Turchet te, C. A. Sacket t, D. Kielpinski, W. M. Itano,
C. Monroe & D. J. Winel and, Decoherence of quantum superpositions through
coupling to engineered reservoirs, Nature 403, pp. 269–273, 2000. See also the summary
by W. T. Strunz, G. Alber & F. Haake, Dekohärenz in offenen Quantensystemen,
Physik Journal 1, pp. 47–52, November 2002. Cited on page 126.
94 L. Hackermüller, K. Hornberger, B. Brezger, A. Z eilinger & M. Arndt,
Decoherence of matter waves by thermal emission of radiation, Nature 427, pp. 711–714,
95 K. Baumann, Quantenmechanik und Objektivierbarkeit, Zeitschrift für Naturforschung
25a, pp. 1954–1956, 1970. Cited on page 127.

96 See for example D. St yer, Physics Today p. 11, September 2000. Cited on page 129.
97 David B ohm, Quantum Theory, Prentice-Hall, 1951, pp. 614–622. Cited on page 129.
98 A. Einstein, B. Pod olsky & N. Rosen, Can quantum-mechanical description of re-
ality be considered complete?, Physical Review 48, pp. 696–702, 1935. Cited on page 129.
99 A. Aspect, J. Dalibard & G. Ro ger, Experimental tests of Bell’s inequalities using
time-varying analyzers, Physical Review Letters 49, pp. 1804–1807, 1982, Cited on page 130.
100 G. C. Hergerfeldt, Causality problems for Fermi’s two-atom system, Physical Review
Letters 72, pp. 596–599, 1994. Cited on page 130.
101 On the superposition of magnetization in up and down directions there are numerous
papers. Recent experiments on the subject of quantum tunnelling in magnetic systems
218 biblio graphy
are described in D. D. Awschalom, J. F. Smith, G. Grinstein, D. P. DiVicenzo

& D. Loss, Macroscopic quantum tunnelling in magnetic proteins, Physical Review Let-
ters 88, pp. 3092–3095, 1992, and in C. Paulsen & al., Macroscopic quantum tunnelling
effects of Bloch walls in small ferromagnetic particles, Europhysics Letters 19, pp. 643–648,
102 For example, superpositions were observed in Josephson junctions by R. F. Voss &
R. A. Webb, Macroscopic quantum tunnelling in 1 mm Nb Josephson junctions, Physical
Review Letters 47, pp. 265–268, 1981, Cited on page 131.
103 S. Haro che, Entanglement, decoherence and the quantum-classical transition, Physics
Today 51, pp. 36–42, July 1998. An experiment putting atom at two places at once, distant
about 80 nm, was published by C. Monroe & al., Science 272, p. 1131, 1996. Cited on page
131.
104 M.R. Andres & al., Observations of interference between two Bose condensates, Science
275, pp. 637–641, 31 January 1997. See also the www.aip.org/physnews/special.htm website.
Cited on page 131.
105 An experimental measurement of superpositions of left and right flowing currents with
1010 electrons was J. E. Mo oij, T. P. Orl and o, L. L evitov, L. Tian, van der Wal
& S. L loyd, Josephson persistent-current qubit, Science 285, pp. 1036–1039, 1999. In the
year 2000, superpositions of 1 μA clockwise and anticlockwise have been detected; for more
details, see J.R. Friedman & al., Quantum superposition of distinct macroscopic states,
Nature 406, p. 43, 2000. Cited on page 131.
106 A clear discussion can be found in S. Haro che & J. -M. R aimond, Quantum comput-
ing: dream or nightmare?, Physics Today 49, pp. 51–52, 1996, as well as the comments in
Physics Today 49, pp. 107–108, 1996. Cited on page 132.
107 The most famous reference on the wave function collapse is chapter IV of the book by
Kurt G ot tfried, Quantum Mechanics, Benjamin, New York, 1966, It was the favorite
reference by Victor Weisskopf, and cited by him on every occasion he talked about the topic.
Cited on page 133.
108 The prediction that quantum tunnelling could be observable when the dissipative interac-
tion with the rest of the world is small enough was made by Leggett; the topic is reviewed
in A. J. L egget t, S. Chahravart y, A. T. D orsey, M. P. A. Fisher, A. Garg &
W. Zwerger, Dynamics of dissipative 2-state systems, Review of Modern Physics 59, pp. 1–
85, 1987. Cited on page 134.
109 Collapse times have been measured for the first time by the group of Serge Haroche in Paris.
See M. Brune & al., Observing the progressive decoherence of the “meter” in a quantum
measurement, Physical Review Letters 77, pp. 4887–4890, 1996, Cited on page 137.
110 S. Ko chen & E. P. Specker, The problem of hidden variables in quantum mechanics,
Journal of Mathematics and Mechanics 17, pp. 59–87, 1967. Cited on page 137.
111 J. F. Cl auser, M. A. Horne, A. Shimony & R. A. Holt, Physical Review Letters 23,
p. 880, 1969. The more general and original result is found in J. S. Bell, Physics 1, p. 195,
1964, Cited on page 138.
112 D. M. Greenberger, M. A. Horne & A. Z eilinger, Going beyond Bell’s theo-
rem, postprint of the 1989 paper at arxiv.org/abs/0712.0912. The first observation was
D. B ouwmeester, J. -W. Pan, M. Daniell, H. Weinfurter & A. Z eilinger,
Observation of three-photon Greenberger-Horne–Zeilinger entanglement, preprint at
arxiv.org/abs/quant-ph/9810035. Cited on page 139.
biblio graphy 219
113 Bryce de Wit t & Neill Graham, eds., The Many–Worlds Interpretation of Quantum
Mechanics, Princeton University Press, 1973. This interpretation talks about entities which
cannot be observed, namely the many worlds, and often assumes that the wave function of
the universe exists. Both habits are beliefs and in contrast with facts. Cited on page 140.
114 ‘On the other had I think I can safely say that nobody understands quantum mechanics.’
From R ichard P. Feynman, The Character of Physical Law, MIT Press, Cambridge,
1965, p. 129. He repeatedly made this statement, e.g. in the introduction of his otherwise
excellent QED: the Strange Theory of Light and Matter, Penguin Books, 1990. Cited on page
141.
115 M. Tegmark, The importance of quantum decoherence in brain processes, Physical Re-
view D 61, pp. 4194–4206, 2000, or also arxiv.org/abs/quant-ph/9907009. Cited on page
141.
116 Connections between quantum theory and information theory can be followed in the Inter-
national Journal of Quantum Information. Cited on page 142.
117 J. A. Wheeler, pp. 242–307, in Batelle Recontres: 1967 Lectures in Mathematics and
Physics, C. DeWit t & J. A. Wheeler, editors, W.A. Benjamin, 1968. For a pedagogical
explanation, see John W. Norbury, From Newton’s laws to the Wheeler-DeWitt equa-
tion, arxiv.org/abs/physics/980604 or European Journal of Physics 19, pp. 143–150, 1998.
Cited on page 143.
118 An informative account of the world of psychokinesis and the paranormal is given by the
famous professional magician James R andi, Flim-flam!, Prometheus Books, Buffalo 1987,
as well as in several of his other books. See also the www.randi.org website. Cited on page
171.
119 The most fascinating book on the topic is by Kurt Nassau, The Physics and Chemistry of
Color – the Fifteen Causes of Color, 1983, and the excellent webexhibits.org/causesofcolour
website. Cited on page 145.
120 Y. Ruiz-Morales & O. C. Mullins, Measured and Simulated Electronic Ab-
sorption and Emission Spectra of Asphaltenes, Energy & Fuels 23, pp. 1169–1177,
2009. U. Bergmann, H. Groenzin, O. C. Mullins, P. Gl atzel, J. Fetzer &
S. P. Cramer, Carbon K-edge X-ray Raman spectroscopy supports simple, yet power-
ful description of aromatic hydrocarbons and asphaltenes, Chemical Physics Letters 369,
pp. 184–191, 2003. Cited on page 145.
121 Two excellent reviews with numerous photographs are E. Grotewohl, The genetics and
biochemistry of floral pigments, Annual Reviews of Plant Biology 57, pp. 761–780, 2006, and
Y. Tanaka, N. Sasaki & A. Ohmiya, Biosynthesis of plant pigments: anthocyanins,
betalains and carotenoids, The Plant Journal 54, pp. 733–749, 2008. Cited on page 153.
122 L. Pérez-Rodriguez & J. Viñuda, Carotenoid-based bill and eye coloration as hon-
est signals of condition: an experimental test in the red-legged partridge (Alectoris rufa),
Naturwissenschaften 95, pp. 821–830, 2008, Cited on page 153.
123 R. Pello, D. S chaerer, J. R ichard, J. -F. L e B orgne & J. -P. Kneib, ISAAC/VLT
observations of a lensed galaxy at z=10.0, Astronomy and Astrophysics 416, p. L35, 2004.
Cited on page 156.
124 A pedagogical introduction is given by L. J. Curtis & D. G. Ellis, Use of the Einstein–
Brillouin–Keller action quantization, American Journal of Physics 72, pp. 1521–1523, 2004.
See also the introduction of A. Klein, WKB approximation for bound states by Heisen-
berg matrix mechanics, Journal of Mathematical Physics 19, pp. 292–297, 1978. Cited on
pages 157 and 159.
220 biblio graphy
125 J. Neukammer & al., Spectroscopy of Rydberg atoms at n ∼ 500, Physical Review Letters
59, pp. 2947–2950, 1987. Cited on page 158.
126 Mark P. Silverman, And Yet It Moves: Strange Systems and Subtle Questions in Physics,
Cambridge University Press 1993. A beautiful book by an expert on motion. Cited on pages
158, 165, and 166.
127 This is explained by J. D. Hey, Mystery error in Gamow’s Tompkins reappears, Physics To-
day pp. 88–89, May 2001. Cited on page 158.
128 L. L. Foldy, The electromagnetic properties of Dirac particles, Physical Review 83,
pp. 688–693, 1951. L. L. Foldy, The electron–neutron interaction, Physical Review 83,
pp. 693–696, 1951. L. L. Foldy, Electron–neutron interaction, Review of Modern Physics
30, pp. 471–481, 1952. Cited on page 162.
129 H. Euler & B. Ko ckel, Naturwissenschaften 23, p. 246, 1935, H. Euler, Annalen der
Physik 26, p. 398, 1936, W. Heisenberg & H. Euler, Folgerung aus der Diracschen The-
orie des Electrons, Zeitschrift für Physik 98, pp. 714–722, 1936. Cited on page 163.
130 A single–atom laser was built in 1994 by K. An, J. J. Childs, R. R. Dasari &
M. S. Feld, Microlaser: a laser with one atom in an optical resonator, Physical Review
Letters 73, p. 3375, 1994. Cited on page 166.
131 J.P. Briand & al., Production of hollow atoms by the excitation of highly charged ions
in interaction with a metallic surface, Physical Review Letters 65, pp. 159–162, 1990. See
also G. Marowsky & C. R hodes, Hohle Atome und die Kompression von Licht in Plas-
makanälen, Physikalische Blätter 52, pp. 991–994, Oktober 1996. Cited on page 166.
132 An introduction is given by P. Pinkse & G. R empe, Wie fängt man ein Atom mit einem
Photon?, Physikalische Blätter 56, pp. 49–51, 2000. Cited on page 166.
133 For the atomic case, see P. L. G ould, G. A. Ruff & D. E. Pritchard, Diffraction of
atoms by light: the near resonant Kapitza–Dirac effect, Physical Review Letters 56, pp. 827–
830, 1986. Many early experimental attempts to observe the diffraction of electrons by light,
in particular those performed in the 1980s, were controversial; most showed only the deflec-
tion of electrons, as explained by H. Batel aan, Contemporary Physics 41, p. 369, 2000.
Later on, he and his group performed the newest and most spectacular experiment, demon-
strating real diffraction, including interference effects; it is described in D. L. Freimund,
K. Afl ato oni & H. Batel aan, Observation of the Kapitza–Dirac effect, Nature 413,
pp. 142–143, 2001. Cited on page 165.
134 G. Gabrielse, D. Hanneke, T. Kinoshita, M. Nio & B. Od om, New determina-
tion of the fine structure constant from the electron g value and QED, Physical Review Let-
ters 97, p. 030802, 2006. Cited on page 167.
135 Wolf gang Pauli, Exclusion principle and quantum mechanics, Nobel lecture, 13 De-
cember 1946, in Nobel Lectures, Physics, Volume 3, 1942–1962, Elsevier, 1964. Cited on page
167.
136 Le Système International d’Unités, Bureau International des Poids et Mesures, Pavillon de
Breteuil, Parc de Saint Cloud, 92310 Sèvres, France. All new developments concerning SI
units are published in the journal Metrologia, edited by the same body. Showing the slow
pace of an old institution, the BIPM launched a website only in 1998; it is now reachable at
www.bipm.fr. See also the www.utc.fr/~tthomass/Themes/Unites/index.html website; this
includes the biographies of people who gave their names to various units. The site of its
British equivalent, www.npl.co.uk/npl/reference, is much better; it provides many details
as well as the English-language version of the SI unit definitions. Cited on page 175.
biblio graphy 221
137 The bible in the field of time measurement is the two-volume work by J. Vanier &
C. Aud oin, The Quantum Physics of Atomic Frequency Standards, Adam Hilge, 1989. A
popular account is Tony Jones, Splitting the Second, Institute of Physics Publishing, 2000.
The site opdaf1.obspm.fr/www/lexique.html gives a glossary of terms used in the field.
For precision length measurements, the tools of choice are special lasers, such as mode-
locked lasers and frequency combs. There is a huge literature on these topics. Equally large
is the literature on precision electric current measurements; there is a race going on for the
best way to do this: counting charges or measuring magnetic forces. The issue is still open.
On mass and atomic mass measurements, see page 64. On high-precision temperature mea-
surements, see page 425. Cited on page 176.
138 The unofficial prefixes were first proposed in the 1990s by Jeff K. Aronson of University of
Oxford, and might come into general usage in the future. Cited on page 177.
139 For more details on electromagnetic unit systems, see the standard text by
John David Jackson, Classical Electrodynamics, 3rd edition, Wiley, 1998. Cited
on page 180.
140 G. J. Stoney, On the physical units of nature, Philosophical Magazine 11, pp. 381–391, 1881.
No citations.
141 D.J. Bird & al., Evidence for correlated changes in the spectrum and composition of
cosmic rays at extremely high energies, Physical Review Letters 71, pp. 3401–3404, 1993.
Cited on page 180.
142 P. J. Hakonen, R. T. Vuorinen & J. E. Martikainen, Nuclear antiferromagnetism
in rhodium metal at positive and negative nanokelvin temperatures, Physical Review Letters
70, pp. 2818–2821, 1993. See also his article in Scientific American, January 1994. Cited on
page 181.
143 A. Z eilinger, The Planck stroll, American Journal of Physics 58, p. 103, 1990. Can you
Challenge 176 e find another similar example? Cited on page 181.
144 The most precise clock built in 2004, a caesium fountain clock, had a precision of one
part in 1015 . Higher precision has been predicted to be possible soon, among others by
M. Takamoto, F. -L. Hong, R. Higashi & H. Katori, An optical lattice clock, Na-
ture 435, pp. 321–324, 2005. Cited on page 181.
145 An overview of this fascinating work is given by J. H. Taylor, Pulsar timing and relativis-
tic gravity, Philosophical Transactions of the Royal Society, London A 341, pp. 117–134, 1992.
Cited on page 181.
146 J. Bergquist, ed., Proceedings of the Fifth Symposium on Frequency Standards and Metrol-
ogy, World Scientific, 1997. Cited on page 181.
147 See the information on D±s mesons from the particle data group at pdg.web.cern.ch/pdg.
Cited on page 181.

148 About the long life of tantalum 180, see D. Belic & al., Photoactivation of 180 Tam and its
implications for the nucleosynthesis of nature’s rarest naturally occurring isotope, Physical
Review Letters 83, pp. 5242–5245, 20 December 1999. Cited on page 182.
149 See the review by L. Ju, D. G. Bl air & C. Z hao, The detection of gravitational waves,
Reports on Progress in Physics 63, pp. 1317–1427, 2000. Cited on page 182.
150 See the clear and extensive paper by G. E. Stedman, Ring laser tests of fundamental
physics and geophysics, Reports on Progress in Physics 60, pp. 615–688, 1997. Cited on page
182.
151 The various concepts are even the topic of a separate international standard, ISO 5725, with
the title Accuracy and precision of measurement methods and results. A good introduction is
222 biblio graphy
John R. Taylor, An Introduction to Error Analysis: the Study of Uncertainties in Physical

Measurements, 2nd edition, University Science Books, Sausalito, 1997. Cited on page 182.
152 P. J. Mohr & B. N. Taylor, CODATA recommended values of the fundamental physical
constants: 1998, Reviews of Modern Physics 59, p. 351, 2000. This is the set of constants result-
ing from an international adjustment and recommended for international use by the Com-
mittee on Data for Science and Technology (CODATA), a body in the International Council
of Scientific Unions, which brings together the International Union of Pure and Applied
Physics (IUPAP), the International Union of Pure and Applied Chemistry (IUPAC) and other
organizations. The website of IUPAC is www.iupac.org. Cited on pages 182 and 183.
153 Some of the stories can be found in the text by N. W. Wise, The Values of Precision,
Princeton University Press, 1994. The field of high-precision measurements, from which
the results on these pages stem, is a world on its own. A beautiful introduction to it
is J. D. Fairbanks, B. S. Deaver, C. W. Everit t & P. F. Michaelson, eds., Near
Zero: Frontiers of Physics, Freeman, 1988. Cited on page 183.
154 The details are given in the well-known astronomical reference, Kenneth Seidelmann,
Explanatory Supplement to the Astronomical Almanac, 1992. Cited on page 187.
155 For information about the number π, and about some other mathematical constants, the
website oldweb.cecm.sfu.ca/pi/pi.html provides the most extensive information and refer-
ences. It also has a link to the many other sites on the topic, including the overview at
mathworld.wolfram.com/Pi.html. Simple formulae for π are
∞
n 2n
π+3= 󵠈 (150)
n=1 󶀡 n 󶀱
2n
or the beautiful formula discovered in 1996 by Bailey, Borwein and Plouffe

∞
1 4 2 1 1
π=󵠈 󶀤
n 8n + 1
− − − 󶀴 . (151)
n=0 16 8n + 4 8n + 5 8n +6
The mentioned site also explains the newly discovered methods for calculating specific bi-
nary digits of π without having to calculate all the preceding ones. The known digits of π
pass all tests of randomness, as the mathworld.wolfram.com/PiDigits.html website explains.
However, this property, called normality, has never been proven; it is the biggest open ques-
tion about π. It is possible that the theory of chaotic dynamics will lead to a solution of this
puzzle in the coming years.
Another method to calculate π and other constants was discovered and published by
D. V. Chudnovsky & G. V. Chudnovsky, The computation of classical constants, Pro-
ceedings of the National Academy of Sciences (USA) 86, pp. 8178–8182, 1989. The Chud-
nowsky brothers have built a supercomputer in Gregory’s apartment for about 70 000 euros,
and for many years held the record for calculating the largest number of digits of π. They
have battled for decades with Kanada Yasumasa, who held the record in 2000, calculated on
an industrial supercomputer. However, the record number of (consecutive) digits in 2010
was calculated in 123 days on a simple desktop PC by Fabrice Bellard, using a Chudnovsky
formula. Bellard calculated over 2.7 million million digits, as told on bellard.org. New for-
mulae to calculate π are still occasionally discovered.
For the calculation of Euler’s constant γ see also D. W. DeTemple, A quicker conver-
gence to Euler’s constant, The Mathematical Intelligencer, pp. 468–470, May 1993.
Note that little is known about the basic properties of some numbers; for example, it is
Challenge 177 r still not known whether π + e is a rational number or not! (It is believed that it is not.) Do
Challenge 178 s you want to become a mathematician? Cited on page 188.
biblio graphy 223
156 A good reference is the Encyclopedia of Mathematics, in 10 volumes, Kluwer Academic Pub-
lishers, 1988−1993. It explains most concepts used in mathematics. Spending an hour with
it looking up related keywords is an efficient way to get an introduction into any part of
mathematics, especially into the vocabulary and the main connections.
The opposite approach, to make things as complicated as possible, is taken in the de-
lightful text by Carl E. L inderholm, Mathematics Made Difficult, 1971. Cited on page
190.
157 An excellent introduction into number systems in mathematics, including hyperreal (or
nonstandard) numbers, quaternions, octonions, p-adic numbers and surreal numbers,
is the book H. -D. Ebbinghaus, H. Hermes, F. Hirzebruch, M. Koecher,
K. Mainzer, J. Neukirch, A. Prestel & R. R emmert, Zahlen, 3rd edition,
Springer Verlag, 1993. It is also available in English, under the title Numbers, Springer
Verlag, 1990. For a book explaining how to use hyperreals in secondary school, see Hel-
mu t Wunderling, Analysis als Infinitesimalrechnung, Duden Paetec Schulbuchverlag,
2007. Cited on pages 191, 200, and 201.
158 A. Waser, Quaternions in Electrodynamics, 2001. The text can be downloaded from vari-
ous websites. Cited on pages 193 and 198.
159 S. L. Altman, Rotations, Quaternions and Double Groups, Clarendon Press, 1986, and also
S. L. Altman, Hamilton, Rodriguez and the quaternion scandal, Mathematical Magazine
62, pp. 291–308, 1988. See also J. C. Hart, G. K. Francis & L. H. Kauffman, Visu-
alzing quaternion rotation, ACM Transactions on Graphics 13, pp. 256–276, 1994. The latter
can be downloaded in several places via the internet. Cited on page 196.
160 See the fine book by Louis H. Kauffman, Knots and Physics, World Scientific, 2nd edi-
tion, 1994, which gives a clear and visual introduction to the mathematics of knots and their
main applications to physics. Cited on page 196.
161 Gaussian integers are explored by G. H. Hardy & E. M. Wright, An Introduction to
the Theory of Numbers, 5th edition, Clarendon Press, Oxford, 1979, in the sections 12.2 ‘The
Rational Integers, the Gaussian Integers, and the Integers’, pp. 178–180, and 12.6 ‘Properties
of the Gaussian Integers’ pp. 182–183. For challenges relating to Gaussian integers, look at
www.mathpuzzle.com/Gaussians.html. Cited on page 200.
162 About transfinite numbers, see the delightful paperback by Rudy Rucker, Infinity and
the Mind – the Science and Philosophy of the Infinite, Bantam, 1983. Cited on page 200.

C R E DI T S
Many people who have kept their gift of curiosity alive have helped to make this project come
true. Most of all, Saverio Pascazio has been – present or not – a constant reference for this project.
Fernand Mayné, Anna Koolen, Ata Masafumi, Roberto Crespi, Serge Pahaut, Luca Bombelli, Her-
man Elswijk, Marcel Krijn, Marc de Jong, Martin van der Mark, Kim Jalink, my parents Peter
and Isabella Schiller, Mike van Wijk, Renate Georgi, Paul Tegelaar, Barbara and Edgar Augel, M.
Jamil, Ron Murdock, Carol Pritchard, Richard Hoffman, Stephan Schiller and, most of all, my
wife Britta have all provided valuable advice and encouragement.
Many people have helped with the project and the collection of material. Most useful was the
help of Mikael Johansson, Bruno Barberi Gnecco, Lothar Beyer, the numerous improvements by
Bert Sierra, the detailed suggestions by Claudio Farinati, the many improvements by Eric Shel-
don, the detailed suggestions by Andrew Young, the continuous help and advice of Jonatan Kelu,
the corrections of Elmar Bartel, and in particular the extensive, passionate and conscientious
help of Adrian Kubala.
Important material was provided by Bert Peeters, Anna Wierzbicka, William Beaty, Jim Carr,
John Merrit, John Baez, Frank DiFilippo, Jonathan Scott, Jon Thaler, Luca Bombelli, Douglas
Singleton, George McQuarry, Tilman Hausherr, Brian Oberquell, Peer Zalm, Martin van der
Mark, Vladimir Surdin, Julia Simon, Antonio Fermani, Don Page, Stephen Haley, Peter Mayr,
Allan Hayes, Norbert Dragon, Igor Ivanov, Doug Renselle, Wim de Muynck, Steve Carlip, Tom
Bruce, Ryan Budney, Gary Ruben, Chris Hillman, Olivier Glassey, Jochen Greiner, squark, Mar-
tin Hardcastle, Mark Biggar, Pavel Kuzin, Douglas Brebner, Luciano Lombardi, Franco Bagnoli,
Lukas Fabian Moser, Dejan Corovic, Paul Vannoni, John Haber, Saverio Pascazio, Klaus Finken-
zeller, Leo Volin, Jeff Aronson, Roggie Boone, Lawrence Tuppen, Quentin David Jones, Arnaldo
Uguzzoni, Frans van Nieuwpoort, Alan Mahoney, Britta Schiller, Petr Danecek, Ingo Thies, Vi-
taliy Solomatin, Carl Offner, Nuno Proença, Elena Colazingari, Paula Henderson, Daniel Darre,
Wolfgang Rankl, John Heumann, Joseph Kiss, Martha Weiss, Antonio González, Antonio Mar-
tos, André Slabber, Ferdinand Bautista, Zoltán Gácsi, Pat Furrie, Michael Reppisch, Enrico Pasi,
Thomas Köppe, Martin Rivas, Herman Beeksma, Tom Helmond, John Brandes, Vlad Tarko, Na-
dia Murillo, Ciprian Dobra, Romano Perini, Harald van Lintel, Andrea Conti, François Belfort,
Dirk Van de Moortel, Heinrich Neumaier, Jarosław Królikowski, John Dahlman, Fathi Namouni,
Paul Townsend, Sergei Emelin, Freeman Dyson, S.R. Madhu Rao, David Parks, Jürgen Janek,
Daniel Huber, Alfons Buchmann, William Purves, Pietro Redondi, Damoon Saghian, plus a num-
ber of people who wanted to remain unnamed.
The software tools were refined with extensive help on fonts and typesetting by Michael Zedler
and Achim Blumensath and with the repeated and valuable support of Donald Arseneau; help
came also from Ulrike Fischer, Piet van Oostrum, Gerben Wierda, Klaus Böhncke, Craig Upright,
Herbert Voss, Andrew Trevorrow, Danie Els, Heiko Oberdiek, Sebastian Rahtz, Don Story, Vin-
cent Darley, Johan Linde, Joseph Hertzlinger, Rick Zaccone, John Warkentin, Ulrich Diez, Uwe
Siart, Will Robertson, Joseph Wright Enrico Gregorio, Rolf Niepraschk and Alexander Grahn.
I also thank the lawmakers and the taxpayers in Germany, who, in contrast to most other
credits 225
countries in the world, allow residents to use the local university libraries.
The typesetting and book design is due to the professional consulting of Ulrich Dirr. The ty-
pography was improved with the help of Johannes Küster. The design of the book and its website
owe also much to the suggestions and support of my wife Britta.
Since May 2007, the electronic edition and distribution of the Motion Mountain text is
generously supported by the Klaus Tschira Foundation.
Film credits
The hydrogen orbital image and animation of page 67 were produced with a sponsored copy of
Dean Dauger’s software package Atom in a Box, available at daugerresearch.com. The coloured an-
imations of wave functions on page 76, page 80, page 81, page 84, page 92, page 161 and page 164
are copyright and courtesy by Bernd Thaller; they can be found on his splendid website vqm.
uni-graz.at and in the CDs that come with his two beautiful books, Bernd Thaller, Visual
Quantum Mechanics Springer, 2000, and Bernd Thaller, Advanced Visual Quantum Mechan-
ics Springer, 2004. These books are the best one can read to get an intuitive understanding for
wave functions and their evolution. The animation of the belt trick on page 110 is copyright and
courtesy by Greg Egan; it can be found on his website www.gregegan.net/APPLETS/21/21.html.
Image credits
The mountain photograph on the front cover is courtesy and copyright by Dave Thompson and
found on his website www.daveontrek.co.uk. The photograph of a glow worm on page 13 is copy-
right and courtesy of John Tyler, and found on his beautiful website at www.johntyler.co.uk/
gwfacts.htm. The photograph of a glass butterfly on page 15 is copyright and courtesy of Linda de
Volder and found on her site at www.flickr.com/photos/lindadevolder. The photograph of a train
window on page 28 is copyright and courtesy of Greta Mansour and found at her website www.
flickr.com/photos/wireful/. The graphics of the colour spectrum on page 36 is copyright and
courtesy of Andrew Young and explained on his website mintaka.sdsu.edu/GF/explain/optics/
rendering.html. The images of photographic film on page 37 are copyright and courtesy of Rich
Evans. The images of photomultipliers on page 37 are copyright and courtesy of Hamamatsu
Photonics. The pictures of the low-intensity photon interference experiment of page 38 are copy-
right of the Delft University of Technology, courtesy of Silvania Pereira, and found on the web-
site www.optica.tn.tudelft.nl/education/photons.asp. The photograph of the Compton effect ap-
paratus on page 40 was taken by Helene Hoffmann and is courtesy of Arne Gerdes from the
University of Göttingen; it is found at the physics teaching website lp.uni-goettingen.de. The
photograph of the Mach–Zehnder interferometer on page 45 is copyright and courtesy of Félix
Dieu and Gaël Osowiecki and found on theor web sites www.flickr.com/photos/felixdieu/sets/
72157622768433934/ and www.flickr.com/photos/gaeloso/sets/72157623165826538/. The blue sky
photograoh on page 57 is courtesy and copyright of Giorgio di Iorio, and found on his website
www.flickr.com/photos/gioischia/. The images about the wire contact experiment on page 57 is
courtesy and copyright of José Costa-Krämer and AAPT . The spookfish photograoh on page 124
is courtesy and copyright of Tamara Frank, and found on his website www.flickr.com/photos/
gioischia/. The famous photograph of electron diffraction on page 64 is copyright and courtesy
of Claus Jönsson. The almost equally famous image that shows the build-up of electron diffrac-
tion on page 64 is courtesy and copyright of Tonomura Akira/Hitachi: it is found on the www.
hqrd.hitachi.co.jp/em/doubleslit.cfm website. The photographs of the Aharonov–Bohm effect on
page 84 are copyright and courtesy of Doru Cuturela. The experiment pictures of the bunching
and antibunching of 3 He and 4 He on page 100 are from the website atomoptic.iota.u-psud.fr/
226 credits
research/helium/helium.html and courtesy and copyright of Denis Boiron and Jerome Chatin.
The molten metal photograph on page 146 is courtesy and copyright of Graela and found
at flickr.com/photos/alaig. The sparkler photograph on page 146 is courtesy and copyright of
Sarah Domingos and found at her flickr.com website. The reactor core photograph on page 146
is courtesy NASA and found on the grin.hq.nasa.gov website. The discharge lamp photographs
on page 146 are courtesy and copyright of Pslawinski and found at www.wikimedia.org. The au-
rora photograph on page 146 is courtesy and copyright of Jan Curtis and found at his climate.gi.
alaska.edu/Curtis/curtis.html website. The coloured flames photograph on page 146 is courtesy
and copyright of Philip Evans and found at his community.webshots.com/user/hydrogen01 web-
site. The iceberg photograph on page 147 is courtesy and copyright of Marc Shandro and found at
his flickr.com/photos/mshandro website. The shadow mask photograph on page 147 is courtesy
and copyright of Planemad and found at the www.wikimedia.org website. The narcissus photo-
graph on page 148 is courtesy and copyright of Thomas Lüthi and found at his website www.
tiptom.ch/album/blumen/. The photograph with a finger with blood on page 148 is courtesy and
copyright of Ian Humes and found at his website www.flickr.com/photos/ianhumes. The berries
photograph on page 148 is courtesy and copyright of Nathan Wall and found at his website www.
flickr.com/photos/ozboi-z. The photograph of a red-haired woman on page 148 is by dusdin and
courtesy of Wikimedia. The rare photograph of a living angler fish on page 148 is courtesy and
copyright of Steve Haddock and found at his website www.lifesci.uscb.edu/~biolum/. The desert
photograph on page 149 is copyright of Evelien Willemsen, courtesy Raf Verbeelen and found at
www.flickr.com/photos/verbeelen. The tenor saxophone photograph on page 149 is courtesy and
copyright of Selmer at www.selmer.fr. The photograph of a red-haired woman on page 149 is by
Walkerma and courtesy of Wikimedia. The fluorescing quantum dot photograph on page 149 is
courtesy and copyright of Andrey Rogach, Center for Nanoscience, München. The zirconia pho-
tograph on page 150 is courtesy and copyright of Gregory Phillips and found at the commons.
wikimedia.org website. The Tokyo sunset on page 150 is courtesy and copyright of Altus Plunkett
and found at his www.flickr.com/photos/altus website. The blue quartz photograph on page 150
is courtesy and copyright of Andreas Kostner and found at his www.thulescientific.com website.
The snowman photograph on page 151 is courtesy and copyright of Altus Plunkett and found
at his www.flickr.com/photos/bytesinmotion website. The endangered blue poison frog photo-
graph on page 151 is courtesy and copyright of Lee Hancock and found at the www.treewalkers.
org website. The ruby glass photograph on page 151 is courtesy and copyright of the Murano Glass
Shop and is found at their murano-glass-shop.it website. The photograph of a ring laser with sec-
ond harmonic generation on page 151 is courtesy and copyright of Jeff Sherman and found at his
flickr.com/photos/fatllama website. The abalone photograph on page 151 is courtesy and copy-
right of Anne Elliot and found at her flickr.com/photos/annkelliot website. The photograph of
polarization colours on page 151 is by Nevit Dilmen and courtesy of Wikimedia. The mallard
duck photograph on page 152 is courtesy and copyright of Simon Griffith and found at his www.
pbase.com/simon2005 website. The opal photograph on page 152 is courtesy and copyright of
Opalsnopals and found at his www.flickr.com website. The aeroplane condensation photograph
on page 152 is courtesy and copyright of Franz Kerschbaum and found at the epod.usra.edu web-
site. The CD photograph on page 152 is courtesy and copyright of Alfons Reichert and found at
his www.chemiephysikskripte.de/artikel/cd.htm website. The liquid crystal pattern on page 152
is courtesy and copyright of Ingo Dierking and Wiley/VCH; it is found in his wonderful book
Ingo Dierking, Textures of Liquid Crystals, Wiley-VCH, 2003. See also his website reynolds.
ph.man.ac.uk/people/staff/dierking/gallery. The measured colour spectrum on page 154 is copy-
right and courtesy of Nigel Sharp, NOAO, FTS, NSO, KPNO, AURA and NSF. The photograph of a
hydrogen discharge on page 155 is copyright and courtesy of Jürgen Bauer and found at the beau-
tiful website www.smart-elements.com.
credits 227
The historical portraits of physicists in the text do not have copyright, except where men-
tioned. The photograph on the back cover, of a basilisk running over water, is courtesy and copy-
right by the Belgian group TERRA vzw and found on their website www.terra.vzw.org. All draw-
ings are copyright by Christoph Schiller. If you suspect that your copyright is not correctly given
or obtained, this has not done on purpose; please contact me in this case.
NAME I NDEX
A Page numbers in italic typeface refer to pages where the person is presented in more detail.
AAPT A Bartel, Elmar 224 Boone, Roggie 224
AAPT 63, 225 Batelaan, H. 220 Borgne, J.-F. Le 219
Aedini, J. 212 Bauer, Jürgen 155, 226 Born, Max 19, 23, 76, 210
Aflatooni, K. 220 Baumann, K. 217 Bose, S.N. 214
Aharonov, Y. 213 Bautista, Ferdinand 224 Bose, Satyenra Nath 99
Alber, G. 217 Baylor, D.A. 210 Bouwmeester, D. 215, 218
Altman, S.L. 223 Beaty, William 224 Bradley, C.C. 212
An, K. 220 Beeksma, Herman 224 Brahmagupta 190
Anderson, Carl 163 Belfort, François 224 Brandes, John 224
Anderson, M.H. 212 Belic, D. 221 Brebner, Douglas 224
Andres, M.R. 218 Belinfante, F.J. 216 Brendel, J. 212
Andrews, M.R. 212 Bell, J.S. 218 Brezger, B. 217
Anton, A. 212 Bell, John 138 Briand, J.P. 166, 220
Aquinas, Thomas 114 Bellard, Fabrice 222 Brillouin, Léon 156
Arndt, M. 212, 217 Belsley, M. 215 Broglie, L. de 212
Aronson, Jeff 224 Bergmann, U. 219 Broglie, Louis de 29, 63
Aronson, Jeff K. 221 Bergquist, J. 221 Brown, R. Hanbury 211
Arseneau, Donald 224 Berry, M.V. 216 Bruce, Tom 224
Aspect, A. 215, 217 Berry, Michael 216 Brumberg, E.M. 35, 210
Aspect, Alain 130 Bessel 154 Brune, M. 217, 218
Aspelmeyer, M. 211 Beutelspacher, Albrecht 142 Bub, J. 210
Ata Masafumi 224 Beyer, Lothar 224 Buchmann, Alfons 224
Audoin, C. 221 Biggar, Mark 224 Budney, Ryan 224
Augel, Barbara and Edgar 224 Bird, D.J. 221 Bunsen, Robert 154
Awschalom, D.D. 218 Björk, G. 211 Busch, Paul 210
Blair, D.G. 221 Butikov, E.I. 212

B Blumensath, Achim 224 Buzek, V. 215
Babinet, Jacques 176 Boas, P. Van Emde 214 Böhncke, Klaus 224
Bader, Richard F.W. 209 Bohm, D. 213
Baez, John 224 Bohm, David 217 C
Bagnoli, Franco 224 Bohr, N. 209, 212, 213 Caldeira, A.O. 217
Balibar, Françoise 210 Bohr, Niels 15, 15, 31, 55, 66, 67, Candela, D. 212
Balmer, Johann 155 120, 140, 156, 217 Carlip, Steve 224
Barberi Gnecco, Bruno 224 Boiron, D. 215 Carr, Jim 224
Bardeen, John 217 Boiron, Denis 100, 226 Carruthers, P. 213
Barnett, S.M. 213 Bombelli, Luca 224 Cato, Marcus Porcius 145
name index 229
Cayley, Arthur 198, 200 Degen, Carl Ferdinand 200 Erschow, A. 216
Center for Nanoscience, Dehmelt, H. 215 Euler, H. 220
München 226 Dehmelt, Hans 214, 215 Euler, Hans 163
Chahravarty, S. 218 Delft University of Euler, Leonhard 195
Chang, H. 215 Technology 38, 225 Evans, Philip 146, 226
Chang, I. 211 DeTemple, D.W. 222 Evans, Rich 37, 225
Chapman, M.S. 214 DeWitt, C. 219 Everitt, C.W. 222
Chatin, Jerome 100, 226 Dicke, Robert H. 213
Childs, J.J. 220 Dieks, D. 215 F
Chu, Steven 166 Dieks, Dennis 103 Facchi, P. 213
Chudnovsky, D.V. 222 Dierking, Ingo 152, 226 Fairbanks, J.D. 222
C Chudnovsky, G.V. 222
Cicero, Marcus Tullius 60
Dieu, Félix 45, 225
Diez, Ulrich 224
Farinati, Claudio 224
Feld, M.S. 220
Clauser, J.F. 218 DiFilippo, Frank 224 Fermani, Antonio 224
Cayley
Clifton, R. 210 Dilmen, Nevit 151, 226 Fermi, Enrico 99
Cohen-Tannoudji, C. 210 Diophantus of Alexandria 192 Fetzer, J. 219
Cohen-Tannoudji, Claude 166 Dirac 165 Feynman, Richard 141
Colazingari, Elena 224 Dirac, P.A.M. 211 Feynman, Richard P. 212
Colella, R. 213 Dirac, Paul 51, 111, 160 Feynman,
Collins, D. 215 Dirr, Ulrich 225 Richard (‘Dick’) Phillips 51
Compton, A.H. 211 Diu, B. 210 Finkenzeller, Klaus 224
Compton, Arthur 40 DiVicenzo, D.P. 218 Firestone, R.B. 215
Conti, Andrea 224 Dobra, Ciprian 224 Fischbach, E. 211
Corovic, Dejan 224 Domingos, Sarah 146, 226 Fischer, Ulrike 224
Correia, A. 212 Dorsey, A.T. 218 Fisher, M.P.A. 218
Costa-Krämer, J.L. 212 Dragon, Norbert 224 Foldy, L.L. 89, 214, 220
Costa-Krämer, José 62, 63, 225 Druten, N.J. van 212 Foley, E.L. 212
Costella, John P. 214 Duran, E. 216 Fonseca, E.J.S. 211
Courtial, J. 213 Durfee, D.S. 212 Francis, G.K. 223
Cramer, S.P. 219 dusdin 148, 226 Frank, Tamara 225
Crane, HR 214 Dyson, Freeman 224 Franke-Arnold, S. 213
Crespi, Roberto 224 Fraunhofer, Joseph 154
Crommle, M.F. 206 E Freimund, D.L. 220
Curtis, Jan 146, 226 Ebbinghaus, H.-D. 223 Friedman, J.R. 218
Curtis, L.J. 219 Egan, Greg 110, 225 Fumagalli, Giuseppe 209
Cuturela, Doru 84, 225 Einstein, A. 211, 214, 215, 217 Furrie, Pat 224
Einstein, Albert 26, 47, 49, 89,
D 99, 156 G
Daffertshofer, A. 215 Ekstrom, C.R. 214 Gabrielse, G. 215, 220

Dahlman, John 224 Elliot, Anne 151, 226 Galilei, Galileo 18, 21
Dalibard, J. 217 Ellis, D.G. 219 Galle, Johann Gottfried 154
Danecek, Petr 224 Els, Danie 224 Garcia, N. 212
Daniell, M. 218 Elswijk, Herman B. 224 García-Mochales, P. 212
Darley, Vincent 224 Emelin, Sergei 224 Gardner, Martin 214
Darre, Daniel 224 Engels, F. 210 Garg, A. 218
Dasari, R.R. 220 Engels, Friedrich 34, 62 Gasiorowicz, Stephen 213
Dauger, Dean 67, 225 Englert, Berthold-Georg 210 Gasparoni, S. 211
Davis, K.B. 212 Epicurus 35, 38 Gauß, Carl-Friedrich 200
Deaver, B.S. 222 Erdős, Paul 190 Gell-Mann, M. 216
230 name index
Gell-Mann, Murray 120, 141 Hamilton, William Rowan 193 Howell, J.C. 215
Georgi, Renate 224 Hammond, T.D. 214 Hoyt, S. 213
Gerdes, Arne 225 Hanbury Brown, Robert 46 Huber, Daniel 224
Gerlach, W. 213 Hancock, Lee 151, 226 Huff, D. 216
Gerlach, Walther 70 Hanneke, D. 220 Hulet, R.G. 212
Gibbs, Josiah Willard 95 Hannout, M. 213 Humes, Ian 148, 226
Gisin, N. 212, 215 Hardcastle, Martin 224 Hurwitz, Adolf 200
Glanz, J. 211 Hardy, G.H. 223
Glassey, Olivier 224 Haroche, S. 218 I
Glatzel, P. 219 Haroche, Serge 126, 218 Icke, Vincent 210
Glauber, R.J. 211 Hart, J.C. 223 Iorio, Giorgio di 57, 225
G Glauber, Roy 53
González, Antonio 224
Hartung, R.W. 216
Hausherr, Tilman 224
Itano, W.M. 217
Ivanov, Igor 224
Gottfried, Kurt 218 Hayes, Allan 224
Gell-Mann
Goudsmit, S. 213 Heaviside 198 J
Goudsmit, Samuel 87 Hegerfeldt, Gerhard 130 Jacobson, J. 211
Gould, P.L. 220 Heisenberg, W. 220 Jalink, Kim 224
Gould, R. 216 Heisenberg, Werner 21, 22, 23, Jamil, M. 224
Graela 146, 226 66, 163 Jammer, Max 209
Graham, Neill 219 Helmond, Tom 224 Janek, Jürgen 224
Grahn, Alexander 224 Henderson, Paula 224 Janssen, Jules 154
Graves, John 200 Hentig, Hartmut von 7 Jeltes, T. 215
Gray, C.G. 213 Hergerfeldt, G.C. 217 Johansson, Mikael 224
Greenberg, O.W. 215, 216 Hermann, Armin 210 Johnson, Samuel 209
Greenberg, Oscar 119 Hermes, H. 223 Jones, Quentin David 224
Greenberger, D.M. 216, 218 Hertz 198 Jones, Tony 221
Greenbergest, T.V. 216 Hertz, Heinrich 47, 160 Jong, Marc de 224
Gregorio, Enrico 224 Hertzlinger, Joseph 224 Joos, E. 216
Greiner, Jochen 224 Hess, Victor 77 Joos, Erich 216
Griffith, Simon 152, 226 Heumann, John 224 Jordan, Pascual 21, 23
Grinstein, G. 218 Hewitt, Susan 210 Joseph Bohm, David 129
Grit, C.O. 206 Hey, J.D. 220 Joyce, James 128
Groenzin, H. 219 Higashi, R. 221 Ju, L. 221
Grotewohl, E. 219 Hilbert, David 190 Jönsson, C. 212
Gácsi, Zoltán 224 Hilborn, R. 216 Jönsson, Claus 64, 65, 225
Hilgevoord, Jan 210
H Hillery, M. 215 K
Haake, F. 217 Hillman, Chris 224 Köppe, Thomas 224
Haas, Arthur Erich 19 Hirzebruch, F. 223 Kan, A.H.G. Rinnooy 214

Haber, John 224 Hitachi 64, 225 Kanada Yasumasa 222
Hackermüller, L. 212 Hitler, Adolf 15 Kapitza 165
Hackermüller, L. 217 Hoffmann, Helene 40, 225 Karl, G. 213
Haddock, Steve 148, 226 Hogervorst, W. 215 Katori, H. 221
Hajnal, A. 214 Holt, R.A. 218 Kauffman, L.H. 223
Hakonen, P.J. 221 Hong, C.K. 215 Keller, C. 212
Haley, Stephen 224 Hong, F.-L. 221 Keller, Joseph 156
Halvorson, H. 210 Hoppeler, R. 215 Kells, W. 215
Hamamatsu Photonics 37 Hornberger, K. 217 Kelu, Jonatan 224
Hamilton 195 Horne, M.A. 218 Kerschbaum, Franz 152, 226
name index 231
Ketterle, W. 212 Lovász, P. 214 Myatt, C.J. 217

Kidd, R. 212 Luca Bombelli 224
Kielpinski, D. 217 Lui, A.T.Y. 211 N
King, B.E. 217 Lundeen, J.S. 211 Nagaoka Hantaro 67
Kinoshita, T. 220 Lynch, David 150 Nagaoka, H. 213
Kirchhoff, Gustav 154 Lévy-Leblond, Jean-Marc 210 Nairz, O. 212
Kiss, Joseph 224 Lüthi, Thomas 148, 226 Namouni, Fathi 224
Klauder, J.R. 211 NASA 146, 226
Klaus Tschira Foundation 225 M Nassau, Kurt 219
Klein, A. 219 Mahoney, Alan 224 Neukammer, J. 220
Klein, Oskar 163 Mainzer, K. 223 Neukirch, J. 223
K Kloor, H. 211
Kneib, J.-P. 219
Malik, J. 210
Mandel, L. 211, 215
Neumaier, Heinrich 224
Neumann, John von 121, 137
Kochen, S. 137, 218 Mansour, Greta 28, 225 Neumann, János von 137
Ketterle
Kockel, B. 220 Mariano, A. 213 Newton 58
Koecher, M. 223 Mark, Martin van der 224 Nienhuis, G. 214
Koolen, Anna 224 Marowsky, G. 220 Niepraschk, Rolf 224
Kostner, Andreas 151, 226 Marqués, M. 212 Nieto, M.M. 213
Krachmalnicoff, V. 215 Martikainen, J.E. 221 Nieuwpoort, Frans van 224
Krijn, Marcel 224 Martini, K.M. 212 Nio, M. 220
Kronig, Ralph 88 Martos, Antonio 224 Novikov, V.A. 213
Kryowonos, A. 213 Massar, S. 215
Królikowski, Jarosław 224 Mattheck, Claus 210 O
Kubala, Adrian 224 Maxwell 198 Oberdiek, Heiko 224
Kurn, D.M. 212 Mayné, Fernand 224 Oberquell, Brian 224
Kuzin, Pavel 224 Mayr, Peter 224 Odom, B. 220
Kuščer, I. 214 McKellar, Bruce H.J. 214 Offner, Carl 224
Küster, Johannes 225 McNamara, J.M. 215 Ohmiya, A. 219
McQuarry, George 224 Oostrum, Piet van 224
L Meekhof, D.M. 217 Opalsnopals 152, 226
Lagrange, Joseph 195 Mensky, M.B. 209 Orlando, T.P. 218
Laloë, F. 210 Merrit, John 224 Orlitzky, A. 214
Lamas-Linares, A. 215 Mewes, M.-O. 212 Osowiecki, Gaël 45, 225
Langel, R.A. 211 Meyenn, K. von 214 Ou, Z.Y. 215
Lavinsky, Rob 147, 150 Meyer, J.C. 206 Overhauser, A.W. 213
Leach, J. 213 Michaelson, P.F. 222
Leggett, A.J. 217, 218 Milonni, P.W. 211 P
Lenstra, J.K. 214 Misner, C.W. 216 Pádua, de 48
Leonardo da Vinci 68 Mitchell, M.W. 211 Padgett, M. 213

Levitov, L. 218 Mohr, P.J. 222 Page, Don 224
Lindblad, G. 217 Monken, C.H. 211 Pahaut, Serge 72, 224
Linde, Johan 224 Monroe, C. 217, 218 Pan, J.-W. 211, 218
Lintel, Harald van 224 Mooij, J.E. 218 Papini, G. 213
Lloyd, S. 218 Moortel, Dirk Van de 224 Parks, David 224
Lockyer, Joseph 154 Moser, Lukas Fabian 224 Pascal, Blaise 41
Lombardi, Luciano 224 Mullins, O.C. 219 Pascazio, S. 213
Loss, D. 218 Murdock, Ron 224 Pascazio, Saverio 224
Loudon, Rodney 210 Murillo, Nadia 224 Pasi, Enrico 224
Louisell, W.H. 213 Muynck, Wim de 224 Paul, H. 211
232 name index
Pauli, Wolfgang 23, 52, 88, 110, Rempe, G. 220 Serena, P.A. 212
166, 216, 220 Renselle, Doug 224 Shandro, Marc 147, 226
Paulsen, C. 218 Reppisch, Michael 224 Sharp, Nigel 154, 226
Peeters, Bert 224 Rhodes, C. 220 Shaw, George Bernard 60
Pello, R. 219 Richard, J. 219 Sheldon, Eric 224
Penrose, R. 216 Rieke, F. 210 Shepp, L. 214
Peredo, M. 211 Rindler, W. 216 Sherman, Jeff 151, 226
Pereira, Silvania 225 Rivas, Martin 224 Shimony, A. 218
Peres, Asher 210 Robbins, J.M. 216 Siart, Uwe 224
Perini, Romano 224 Robertson, Will 224 Sierra, Bert 224
Perrin, A. 215 Rogach, Andrey 149, 226 Silverman, M.P. 214
P Perrin, J. 213
Perrin, Jean 67
Roger, G. 217
Roos, Hans 210
Silverman, Mark 165
Silverman, Mark P. 220
Philips, William 166 Rosen, N. 217 Simon, C. 215
Pauli
Phillips, Gregory 150, 226 Rosenfeld 55 Simon, Julia 224
Photonics, Hamamatsu 225 Rosenfeld, L. 212 Singleton, Douglas 224
Pinkse, P. 220 Ruben, Gary 224 Slabber, André 224
Planck, Erwin 210 Rucker, Rudy 223 Smith, J.F. 218
Planck, Max 15, 18, 42, 47, 89, Ruff, G.A. 220 Snow, G.A. 102, 215
209 Ruiz-Morales, Y. 219 Solomatin, Vitaliy 224
Planemad 147, 226 Rydberg, Johannes 155 Sommerfeld, Arnold 158–160,
Plastino, A. 215 167
Plastino, A.R. 215 S Specker, E.P. 137, 218
Plunkett, Altus 150, 226 S.R. Madhu Rao 224 Stedman, G.E. 221
Podolsky, B. 217 Sackett, C.A. 212, 217 Steinberg, A.M. 211
Poincaré, Henri 47, 211 Sackur, Otto 95 Stern, O. 213
Popescu, S. 215 Sagan, Hans 203 Stern, Otto 70
Prentis, J.J. 211 Saghian, Damoon 224 Stoney, G.J. 221
Prestel, A. 223 Salam, Abdus 210 Story, Don 224
Prewitt, O. 216 Sasaki, N. 219 Strunz, W.T. 217
Pritchard, Carol 224 Schaerer, D. 219 Styer, D. 217
Pritchard, D.E. 214, 220 Schellekens, M. 215 Subitzky, Edward 210
Pritchard, David 72 Schiller, Britta 224, 225 Sudarshan, E.C.G. 211
Proença, Nuno 224 Schiller, Christoph 227 Surdin, Vladimir 224
Pslawinski 146, 226 Schiller, Friedrich 35
Purves, William 224 Schiller, Isabella 224 T
Pádua, S. de 211 Schiller, Peter 224 Takamoto, M. 221
Pérez-Rodriguez, L. 219 Schiller, Stephan 224 Tanaka, Y. 219
Schleich, W.P. 217 Tarko, Vlad 224

R Schmiedmayer, J. 214 Taylor, B.N. 222
Rahtz, Sebastian 224 Schrödinger, Erwin 156 Taylor, J.H. 221
Raimond, J.-M. 218 Schrödinger, E. 213 Tegelaar, Paul 224
Ramberg, E. 102, 215 Schrödinger, Erwin 31, 77 Tegmark, M. 217, 219
Randi, James 219 Schubert, Max 210 Tetrode, Hugo 95
Rankl, Wolfgang 224 Schucking, E. 214 Thaler, Jon 224
Redondi, Pietro 224 Schwenk, Jörg 142 Thaller, Bernd 76, 80, 81, 84,
Reichert, Alfons 152, 226 Schwinger, Julian 86, 87, 210 92, 161, 164, 225
Reichl, Linda 216 Scott, Jonathan 224 Thies, Ingo 224
Remmert, R. 223 Selmer 149, 226 Thomas, L. 213
name index 233
Thomas, Llewellyn 88 Volin, Leo 224 Wineland, D.J. 217

Thompson, Dave 225 Voltaire 169 Wise, N.W. 222
Thorne, K.S. 216 Vos-Andreae, J. 212 Wittke, James P. 213
Tian, L. 218 Voss, Herbert 224 Woerdman, J.P. 214
Tino, G. 216 Voss, R.F. 218 Wolfenstätter, Klaus-Dieter
Tittel, Wolfgang 212 Vuorinen, R.T. 221 142
Tollett, J.J. 212 Wolfsried, Stephan 147, 149
Tonomura Akira 64, 225 W Wollaston, William 154
Townsend, Paul 224 Walkerma 149, 226 Wolsky, A.M. 214
Trevorrow, Andrew 224 Wall, Nathan 148, 226 Wootters, W.K. 215
Tschira, Klaus 225 Walther, P. 211 Wootters, W.L. 103
T Tuominen, M.T. 212
Tuppen, Lawrence 224
Warkentin, John 224
Waser, A. 223
Wouthuysen, S.A. 89, 214
Wright, E.M. 223
Turchette, Q.A. 217 Webb, R.A. 218 Wright, Joseph 224
Thomas
Twamley, J. 216 Weber, Gerhard 210 Wunderling, Helmut 223
Twiss, R.Q. 211 Wehinger, S. 214
Twiss, Richard 46 Weinberg, Steven 91, 212 Y
Tyler, John 14, 225 Weinfurter, H. 218 Yamamoto, Y. 211
Weiss, Martha 224 Yao, E. 213
U Weisskopf, Victor 218 Yazdani, A. 214
Uguzzoni, Arnaldo 224 Werner, S.A. 213 Young, Andrew 36, 224, 225
Uhlenbeck, G.E. 213 Westbrook, C.I. 215
Uhlenbeck, George 87 Weyl, Hermann 110 Z
Upright, Craig 224 Wheeler, J.A. 216, 219 Zaccone, Rick 224
Ursin, R. 211 Wheeler, John 204 Zalm, Peer 224
Widom, A. 213 Zbinden, H. 212
V Wierda, Gerben 224 Zedler, Michael 224
Vanier, J. 221 Wierzbicka, Anna 224 Zeh, H.D. 216
Vannoni, Paul 224 Wiesner, Steve 103 Zeh, Hans Dieter 121, 217
Vardi, Ilan 214 Wigner, E. 215 Zeilinger, A. 211, 212, 217, 218,
Vassen, W. 215 Wigner, Eugene 105 221
Vavilov, S.I. 35, 210 Wigner, Eugene P. 210 Zetti, A. 206
Verbeelen, Raf 226 Wijk, Mike van 224 Zhao, C. 221
Vico, Giambattista 139 Wikimedia 226 Zurek, W.H. 103, 214, 215
Viñuda, J. 219 Wiley/VCH 226 Zuse, Konrad 137
Volder, Linda de 15, 225 Willemsen, Evelien 149, 226 Zwerger, W. 218 Copyright © Christoph Schiller November 1997–July 2010
SUBJECT I NDEX
A Page numbers in italic typeface refer to pages where the keyword is defined or presented in detail.
The subject index thus acts as a glossary.
acausal
A asphaltenes 145 Bennett–Brassard protocol
acausal 128 astrology 171 104
acceleration, maximum 66 astronomy 196 bioluminescence 148
acceleration, quantum limit atom rotation 72 biphoton 48
66 atom size 167 BIPM 175, 176
accuracy 182 atomic 181 bits to entropy conversion 185
accuracy, limits to 183 atomic mass unit 109 blasphemies 60
action, EBK 159 atoms 14 blood colour 148
action, quantum of 16 atoms, hollow 166 blue colour of the sea 165
addition 190 atoms, single 166 bodies, rigid 32
ageing 26 atto 177 Bohr radius 158, 185
Aharonov–Bohm effect 83 aurora 146 Boltzmann constant 125, 184
Aharonov–Casher effect 85 average 125 Boltzmann’s constant 15
Alectoris rufa 219 Avogadro’s number 185 Bose–Einstein condensate 61,
algebra, alternative 199 axioms 191 214
ampere 175 azimuthal quantum number bosons 54, 99, 102
angels 114, 170 158 bottom quark 109
angular momentum 116, 172 bottomness 109
angular momentum, B brain 82
indeterminacy relation 69 Baker–Campbell–Hausdorff breaking 25
annihilation operator 102 formula 203 Bremsstrahlung 146
anti-bunching 46 baryon number density 188 Brillouin scattering 58
anticommutator bracket 102 base units 175 bromine 147
antimatter 163 basic units 194 bulge 101
anyons 118, 118 basis 203 Bureau International des

aphelion 186 bath, physical 122 Poids et Mesures 175
apogee 186 BCH formula 203
apparatus, classical 139 beans, dangers of 32 C
apparatus, irreversible 139 beauty 109 candela 175
arc lamp 146 becquerel 177 candle colour 146
argon lamp 146 bell 114 cans of beans, dangers of 32
arm 111, 196 Bell’s equation 138 car theft 82
arrow, rotating 75 Bell’s inequality 138 cardinals 200
arrows, rotating 75 belt 196 cars on highways 65
artefact 176 belt trick 111, 196 Casimir effect 172
subject index 235
causality 136 Commission Internationale creation 164

Cayley algebra 199 des Poids et Mesures 175 creation operator 102
Cayley numbers 199 commutation of Hamiltonian cross product 200
centi 177 and momentum operator cryptoanalysis 142
centre, quaternion 197 92 cryptography 142
Čerenkov radiation 146 commutation, lack of 31 cryptography, quantum 104
CERN 180 commutative 191 cryptology 142
challenge classification 9 commute, observables do 31 cryptology, quantum 104, 142
challenge level 9 complementarity 66
challenges 9, 14, 16, 18–20, 22, complementarity principle 31, D
24–27, 31, 32, 34, 39–43, 45, 66 daemons 170
C 48–50, 53–57, 59, 60, 62, 63,
66, 67, 71, 72, 80, 82, 85,
completeness property of sets
191
damping 124
day, sidereal 185
90–95, 97–99, 101, 103, 104, complex conjugate 192 day, time unit 177
causality
111–119, 122, 123, 125, 128, complex numbers 192 de Broglie wavelength 176
130–132, 135, 138–141, 145, complex numbers as arrows death 26, 126
156, 157, 159, 163, 166, 167, 192 deca 177
170, 177–180, 182, 183, 187, composed 98 decay 171
189, 190, 192–198, 200–203, compositeness 90 deci 177
221, 222 Compton (wave)length 91 decoherence 31, 122, 131
characteristic 191 Compton scattering 58 decoherence process 132
charge inversion 106 Compton wavelength 172, 185 decoherence time 123, 125
charm quark 109 computer science and degree Celsius 177
chimaera 104 quantum theory 30 degree, angle unit 177
chlorine 147 computer, universe as 143 Dendrobates azureus 151
classical apparatus 139 concepts, classification of 190 density functional 121
classical electron radius 185 condensate 206 density matrix 121
classification of concepts 190 condom problem 94 detachable 127
cleveite 154 conductance quantum 185 detector 134
clocks 23 cones, in the retina 210 detectors of motion 14
clone 103 Conférence Générale des determinism 140
clones, biological 104 Poids et Mesures 175 devil 166, 167
clouds 67 configuration space 112 devils 170
clouds as results of quantum Conférence Générale des different 51
theory 72 Poids et Mesures 176 diffraction 51, 53
clouds in quantum theory 67 consciousness 140 diffraction as colour cause 152
CODATA 222 Convention du Mètre 175 diffraction of matter by light
coherence 65, 121, 131 copy, perfect 103 165
coherence length 46 copying machine 103 diffraction of quantum states

coherence time 46 copying machines 31, 103, 104 79
coherent 127 Coriolis acceleration in atoms diffraction pattern 132
collapse 133, 137 165 dimension 203
collapse of the wave function corrected Planck units 179 dimensionless 184
128 cosmological constant 187 dimensions, three spatial 113
collapse of the wavefunction coulomb 177 disentanglement 122
79 Coulomb gauge 129 disentanglement process 132
colour 39, 109 coupling minimal 162 disinformation 34
colour centres 150 CPT 88 dispersion 150
colours 167 cream, whipped 19 dispersion of wave functions
236 subject index
80 escape velocity 158 G

distribution, Gaussian 182 Euclidean vector space 202 g-factor 90
distribution, normal 182 eumelanin 148 G-parity 109
division 191 europium 147 д-value 88
division algebra 197 evanescent waves 82 Göttingen 21
Doppler effect 155 evolution equation, first order gas lasers 146
double cover 196 79 gases 95
double numbers 200 Exa 177 gauge, Coulomb 129
down quark 109 excitations in gases 146 Gaussian distribution 182
dyadic product 122 explanation 140 Gaussian integers 200
eye and the detection of Gaussian primes 200
D E
Earth’s age 186
photons 35 Gaussian unit system 180
gelatine 170
Earth’s average density 186 F generators 194
distribution
Earth’s gravitational length fall, free 18 genius 47
185 farad 177 Geocentric gravitational
Earth’s radius 186 femto 177 constant 185
EBK quantization 156 fencing 116 ghosts 119, 170
eigenstates 75 Fermi coupling constant 184 Gibbs’ paradox 95
eigenvalue 75, 135 fermions 99, 102 Giga 177
eigenvalues 89, 133 field, mathematical 191 glove problem 94
eigenvector 75 film 15 gloves 104
eigenvectors 75 fine structure 160 gloves, difference with
eight-squares theorem 199 fine structure constant 158, quantum systems 105
Einstein–Podolsky–Rosen 159, 160, 173, 178, 184, 184 glow-worms 148
paradox 129 fine-structure constant 167, gluon 108
Ekert protocol 104 167 goddess 170
electrodynamics 193 fire colour 146 gods 139, 170
electromagnetic unit system firework colour 146 graphics, three-dimensional
180 first property of quantum 196
electromagnetism, strength of measurements 133 grating 53
166 flashlamp colour 146 gratings of light 165
electron 108 flight simulation 196 gravitational coupling
electron radius 109, 116 floor, why does it not fall 114 constant 184
electronvolt 180 flows are made of particles 62 graviton 107
electrostatic unit system 180 flows must vary 62 gravitons 30
elementary particles 105 fluctuations 122 gray 177
elementary quantum particle four-momentum 106 ground state 158
170 four-squares theorem 194 group 190

energy levels 157 fractals 32 group velocity 79
energy width 109 fractals do not appear in growth 27
ensemble 95 nature 72 Gulliver’s travels 19
entangled systems 31 Fraunhofer lines 154 gyromagnetic ratio 90
entanglement 31, 127, 129 friction 124 gyromagnetic ratio of the
entanglement, degree of 132 full width at half maximum electron 172
environment 123 182
EPR 104, 129 function collapse, wave 128 H
equilibrium 122 fuzziness, fundamental 61 H2 O 19
errors in measurements 182 half-life 109
subject index 237
Hall effect, indeterminacy relations 22, 65 kilogram 175

fractional quantum 118 indistinguishable 95, 171 kilotonne 33
Hamiltonian 77 indoctrination 34 Klitzing, von – constant 185
Hanbury Brown-Twiss information science and knocking on table and
experiment 54 quantum theory 30 fermionic character of
Hanbury Brown–Twiss effect inhomogeneous Lorentz matter 114
48 group 105 knocking on tables 62
hand, for quaternion inner product 202 Korteweg–de Vries equation
visualization 196 inner product spaces 202 92
Heaviside–Lorentz unit inorganic charge transfer 149
system 180 integers 190 L
H hecto 177
Heisenberg picture 120, 130
interference 121
interference and photons 51
Lagrangian operator 86
Lamb shift 172
Heisenberg’s indeterminacy interference as colour cause Lampyris noctiluca 14
Hall
relations 65 151 laser cavities 42
helicity 41, 106 interference fringes 49 laser sword 116
helium 90, 100, 126, 154 interference, quantum 79 lasers 46
helium, discovery of 154 interferometer 44, 65 lava colour 146
hemoglobin 153 interferometers 182 lawyers 34
henry 177 intermediate bosons 91 learning, best method for 8
Hermitean vector space 202 International Geodesic Union Lego 14
hertz 177 187 length scales 171
hidden variables 137 interpenetration 116 lifetime 109
Higgs 109 interpenetration of matter 114 lifetime, atomic 172
Hilbert space 75, 76, 202, 203 interpretation of quantum light 41
Hiroshima 33 mechanics 121, 140 light gratings 165
Hitachi 214 intrinsic angular momentum light quanta 35
homogeneous 195 70 light quantum 30
horseshoe 32 intrinsic properties 170 light source 46
hour 177 inverse Raman scattering 58 light tunnelling 82
Hubble parameter 187 iodine 147 light year 185
human observer 140 ionization energy 158 light, macroscopic 131
hydrogen atoms 214 irreducible representation 105 light, made of bosons 116, 117
hydrogen atoms, existence of irreversible 124 lightbulb 146
92 isotopes 103 Lilliput 171
hyperreals 201 IUPAC 222 limits to precision 183
IUPAP 222 linear spaces 201
I linear vector spaces 201
ice colour 147 J linearity of quantum

ice, blue 147 Josephson effect 85, 176 mechanics 120
images 171 Josephson frequency ratio 185 litre 177
imaginary 195 joule 177 locality 136
immediate 129 Journal of Irreproducible lumen 177
impenetrability of matter 25, Results 210 luminary movement 41
116, 171 Jupiter’s mass 186 luminous bodies 41
impenetrability, of matter 118 lux 177
incandescence 16, 146 K Lyman-alpha line 155
indeterminacy relation for kelvin 175
angular momentum 69 kilo 177
238 subject index
M minimal coupling 162 nuclear warhead 33

macroscopic system 127 minute 177, 187 nucleus 70
macroscopically distinct 121 mirror 51 number theory 200
magic 173 mirrors 82 numbers, hypercomplex 199,
magma colour 146 mixed state 122 200
magnetic flux quantum 185 mole 103, 175 nymphs 170
magnetic resonance force molecular vibrations and
microscope 88 rotations 147 O
magnetite 149 molecule size 172 object 127
magneton 90, 185 molecules as bosons 99 objects are made of particles
many worlds interpretation Moon’s mass 186 169
M 140
Mars 181
Moon’s mean distance 186
Moon’s radius 186
observables 75
observation 134
Maslov index 157 Moore’s law 32 observations 132
macroscopic
material properties 166 motion and measurement octaves 198
material research 166 units 176 octonions 198
materials, dense optically 53 motion backwards in time 24 ohm 177
measured 135 motion inversion 106 open link 119
measurement 178 motion is fundamental 176 operator, adjoint 102
measurement apparatus 139 mozzarella 20 operators 75
measurement errors 182 multiplication 190 orbits inside atoms 155
measurement precision, no muon 108 order, total 191
infinite 61 muon neutrino 108 ordinals 200
measurement results 75 myoglobin 153 organic compounds 148
measurements 75, 132, 175 original Planck constant 184
measurements disturb 140 N
Mega 177 nano 177 P
megatonne 33 nanoscopic 21 π = 3.141592... 222
melanin 153 NASA 181 π, normality of 189
memory 82, 132 nature 139 pair creation 172
mercury lamp 146 nature and computer science paradox, EPR 129
mesoscopic 21 30 parity 109
metallic bands 149 neon lamp 146 parsec 185
metre 175 neutrino 124 particle 100
metre rules 23 neutrino, electron 108 particle counting, limits to
metric space 201 neutron 90 164
micro 177 new age 141 particle, elementary 105
microscope 21 newton 177 particle, real 165
microscopic 21 no-cloning theorem 103, 215 particle, virtual 55, 165

microscopic system 127 non-classical 46 particles 98
microwave background non-classical light 43, 48 pascal 177
temperature 188 non-local 128 path integral formulation 86
Mie scattering 58 non-unitarity 140 paths 29
mile 178 nonstandard analysis 200 Pauli equation 88
Milky Way’s age 186 norm 192, 194, 202 Pauli exclusion principle 103,
Milky Way’s mass 186 normality 222 110, 113, 115, 116
Milky Way’s size 186 normality of π 189 Pauli pressure 114
milli 177 North Pole 69 Pauli spin matrices 197
mind 140 nuclear magneton 185 Pauli’s exclusion principle 113
subject index 239
penetrability of matter 25 praesodymium 147 qubits 132

perfect copy 103 precision 182, 183
perigee 186 precision, limits to 183 R
perihelion 186 precision, no infinite radian 176
periodic decimal expansion measurement 61 radiation, observers made of
201 prefixes 177, 221 142
permanence 24, 141 prefixes, SI 177 radiative decay 172
permanganate 149 principle of least action 86 radicals 99
permutation symmetry 98 prison 34 radioactivity 96
Peta 177 probability 133 rainbows and the elements in
phase 29 probability amplitude 139 the Sun 154
P phase of wave function 83
phase, thermodynamic 96
probability distribution 67
proton 90
RAM 82
Raman scattering 58
pheomelanin 148 proton radius 90 random errors 182
penetrability
philosophers 41 proton–electron mass ratio random-access memory 82
photochromism 150 185 randomness, experimental 133
photon 30, 108 protonvolt 180 rational coordinates 189
photon as elementary particle pure 195 rational numbers 191
42 pure state 121 Rayleigh scattering 58
photon cloning 215 reactions 27
photon number density 188 Q real numbers 191
photon, position of 44 q-numbers 200 real particle 165
photon-photon scattering 172 QED 162 recognition 101
photons 35, 37, 41, 170 quanti, piccolissimi 21 record 132
photons and interference 51 quantization 39 reduced Planck constant 184
photons and naked eye 35 quantization, EBK 156 reflection 51
photons as arrows 49 quantons 41, 63, 170 reflection, total 166
photons, entangled 48 quantum action 86 refraction 53, 150
photons, eye detection of quantum action principle 87 refraction and photons 56
single 39 quantum computers 104 refraction of matter waves 72
photons, spin of 41 quantum computing 132, 218 relaxation 124
photons, virtual 54 quantum cryptology 104, 142 representation 192, 197, 199
pico 177 quantum electrodynamics 162 representations, irreducible
Planck stroll 181 quantum interference 79 105
Planck units, corrected 179 quantum mechanical system reservoir 122, 123
Planck’s (unreduced) constant 127 rest 17
15 quantum mechanics applied rest does not exist 170
Planck’s constant 16, 39 to single events 140 rest, no 61
Planck’s natural units 178 quantum numbers 107, 109 rigid bodies 32
plankton 165 quantum of action 15, 16 ring 190
plate trick 111 quantum of change 16 ring interferometers 182
pointer 135 quantum particles 171 robotics 196
polarization 54, 150 quantum principle 16 rods in retina 211
polarization of light 40 quantum state 77 rotation 116, 195
police 82 quantum states 75 rotation of atoms 72
position 141 quantum theory and rotational axis 69
positron 163 computer science 30 ruby glass 151
positron charge 184 quaternion, conjugate 194 Rydberg atoms 158
potential, spherical 157 quaternions 193 Rydberg constant 155, 172, 185
240 subject index
S sparkler colour 146 system, classical 23

Sackur–Tetrode formula 95 sparks 146 system, cloning of
samarium 147 spatial parity 106 macroscopic 104
sapphire 149 special orthogonal group 196 system, macroscopic 22
scalar 201 spectrum of hot objects 172 systematic errors 182
scalar multiplication 201 sperm 20
scalar part of a quaternion 194 spin 70, 87, 106, 171 T
scalar product 202 spin 1/2 and quaternions 196 tachyons 24
scattering 150 spin and rotation 116 tau 108
Schrödinger picture 120 spin myth 116 tau neutrino 108
Schrödinger’s cat 120, 128 spin–statistics theorem 115 tax collection 175
S Schrödinger’s equation of
motion 77
spinor 113, 163
spinors 196, 196
TB 82
telekinesis 171
Schwarzschild radius as spirits 170 teleportation 104, 132, 171
Sackur–Tetrode length unit 180
spreading of wave function 79 tensor product 122
science fiction 116 squark 224 Tera 177
scissor trick 111, 216 squeezed light 43, 48 terabyte 82
sea water 165 standard deviation 182 tesla 177
sea, bluest 165 star colours 146 thermal de Broglie
second 175, 177, 187 state 120, 143 wavelength 126
second property of quantum state function 139 thermal light 43
measurements: 133 state sum 206 thermodynamics, third ‘law’
sedenions 200 state, quantum 77 of 60
semi-ring 190, 191 states 74 third ‘law’ of
semiconductor bands 149 states are rotating arrows 75 thermodynamics 60
senses 14 steel, hot 146 Thomas precession 88
separability 127 Stefan–Boltzmann black body Thomson scattering 58
sesquilinear 202 radiation constant 172 time of collapse 137
sex 94 Stefan–Boltzmann constant time scales 171
sexuality 27 185 time travel 24
shape 19 steradian 176 TNT 33
shapes 67 Stern–Gerlach experiment 71 TNT energy content 185
short 24 stone 30 Tom Thumb 32
SI units 175, 182 stones 26, 53, 63, 113, 170 tonne, or ton 177
SI units, supplementary 176 strange quark 109 top quark 109
siemens 177 strength of electromagnetism topness 109
sievert 177 166 total reflection and light
single atom 131 strong coupling constant 184 amplification 166
single events in quantum Sun’s age 186 trace 122

mechanics 140 Sun’s luminosity 186 train windows 28
sizes of atoms 167 Sun’s mass 186 transfinite number 200
sizes of tings 167 supernatural phenomena 171 transition metal compounds
skew field 191 superposition, coherent 121 147
Smekal–Raman scattering 58 superposition, incoherent 122 transition metal impurities
SO(3) 106 surreals 200 147
sodium 72 symmetry 105 tree, noise of falling 140
sodium street lamps 146 Système International trick, belt 111
soliton 92 d’Unités (SI) 175 trick, scissor 111
soul 170 system 120, 127 tropical year 185
subject index 241
truth 109 vacuum polarization 164 wave function, spreading of

tunnelling 82, 171 vacuum state 101 79
tunnelling effect 25, 80 vacuum wave resistance 185 wave function, symmetry of
tunnelling of light 82 value, absolute 192 98
tunnelling rate 172 vanishing 122 wave–particle duality 41
TV tube 82 variance 182 waves, evanescent 82
two-squares theorem 192 Vavilov–Čerenkov radiation weak charge 109
Tyndall scattering 58 146 weak isospin 109
vector 195, 201 weak mixing angle 184
U vector part of a quaternion weber 177
udeko 177 194 weko 177
T Udekta 177
uncertainty relations 22, 65
vector product 200
vector space 201
Wekta 177
Wheeler–DeWitt equation 143
uncertainty, total 182 vector space, Euclidean 202 Wien’s displacement constant
truth
understanding quantum vector space, Hermitean 202 172, 185
theory 34 vector space, unitary 202 windows in trains 28
unit 175, 195 vendeko 177 wine 60, 125
unital ring 191 Vendekta 177 World Geodetic System 187
unitarity 136, 140 viewpoint changes 75
unitary vector space 202 virtual particle 55, 165 X
units, astronomical 185 virtual particles 97 X-ray scattering 58
units, non-SI 178 virtual photons 54 X-rays 40
units, provincial 178, 181 vitamin C 99 xenno 177
units, SI 175 volt 177 Xenta 177
units, true natural 179 von Neumann equation 122
universe’s initial conditions Y
do not exist 144 W yocto 177
universe, wave function of 143 W boson 108 Yotta 177
up quark 108 water colour 147
watt 177 Z
V wave equation 78 Z boson 108
vacuoles 153 wave function 77, 77, 139 zepto 177
vacuum 97, 165, 165 wave function as rotating zero-point fluctuations 61
vacuum permeability 184 cloud 83 Zetta 177
vacuum permittivity 184 wave function, phase of 83
MOTION MOUNTAIN
The Adventure of Physics – Vol. IV
Quantum Theory: The Smallest Change
Why do change and motion exist?

How does a rainbow form?
What is the most fantastic voyage possible?
Is ‘empty space’ really empty?
How can one levitate things?
At what distance between two points does it become
impossible to find room for a third one in between?
What does ‘quantum’ mean?
Which problems in physics are unsolved?
Answering these and other questions on motion,

this series gives an entertaining and mind-twisting
introduction into modern physics – one that is
surprising and challenging on every page.
Starting from everyday life, the adventure provides
an overview of the recent results in mechanics, pdf file available free of charge at www.motionmountain.net
thermodynamics, electrodynamics, relativity,
quantum theory, quantum gravity and unification.
It is written for undergraduate students and for
anybody interested in physics.
Christoph Schiller, PhD Université Libre de Bruxelles,

is a physicist with more than 25 years of experience
in the presentation of physical topics.
ISBN 978-3-00-021946-7
9 783000 219467

Motionmountain Volume4

Uploaded by

Copyright:

Available Formats

Motionmountain Volume4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Motionmountain Volume4

Uploaded by

Copyright:

Available Formats

Christoph Schiller

The Adventure of Physics

Edition 24.1, available as free pdf at

Proprietas scriptoris © Christophori Schiller

Omnia proprietatis iura reservantur et vindicantur.

Twenty-third edition, ISBN 978-300-021946-7.

Copyright © 2009 by Christoph Schiller,

This pdf file is licensed under the Creative Commons

Munich, 27 July 2010.

PHYSICS: Unified description of motion Why does motion

Galilean physics, heat and electricity

Advice for learners

Using this book

Copyright © Christoph Schiller November 1997–July 2010

209 Biblio graphy

In our quest to understand how things move,

MINIMUM ACTION – QUANTUM

Ref. 1 * ‘Nature [in its workings] makes no jumps.’

F I G U R E 3 Max Planck (1858–1947)

principle can thus rephrased as

TA B L E 1 How to convince yourself and others that there is a minimum

The action value ħ is check all observations

The effects of the quantum of action on rest

* ‘Nature makes jumps.’

Inertial motion is an approximation! An object can move in straight, uniform motion

F I G U R E 5 An artist’s impression of a water molecule

all because of the existence of the quantum of action.

System and change Action Motion

Single DNA duplication step during cell division c. 100 ħ quantum

F I G U R E 6 Max Born (1882–1970)

F I G U R E 7 Werner Heisenberg (1901–1976)

they are useless.

Quantum limits and surprises

Transformation, life and Democritus

F I G U R E 8 Hills are never high

Challenge 15 ny impossible. Can you explain it?

F I G U R E 13 A particle and a screen with two nearby slits

Waves – a consequence of the quantum of action

We are thus led to ascribe a frequency and a wavelength to a quantum system:

Particles – a consequence of the quantum of action

Curiosities and fun challenges about the quantum of action

The dangers of buying a can of beans

Warning: care should be taken when touching this product:

Warning: care should be taken when using this product:

The impression of a certain paranoid side to quantum physics is purely coincidental.

A summary: quantum physics, the law and indoctrination

Copyright © Christoph Schiller November 1997–July 2010

LIGHT – THE STR ANGE

lamp strong shutter head, after

F I G U R E 16 Exposed photographic ﬁlm at increasing magniﬁcation (© Rich Evans)

F I G U R E 19 An atom radiating one

Whenever a radiating atom is surrounded by many detectors, only a single detector is

ħ = 1.1 ⋅ 10−34 Js , (9)

the expression p = ħ k. In short, the quantum of action determines the momentum of

particles. Single quantons often behave like classical particles.

The size of photons

Are photons countable? – Squeezed light

cleverly placed inside such a cavity.

Thermal light Laser light Non-classical light