Particle Physics-A Los Alamos Primer
Particle Physics-A Los Alamos Primer
Particle Physics-A Los Alamos Primer
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
i' --<br />
I i<br />
LA-UR-88-9114
<strong>Particle</strong> <strong>Physics</strong><br />
A <strong>Los</strong> <strong>Alamos</strong> <strong>Primer</strong>
C./<br />
<strong>Particle</strong> Phvsics<br />
J<br />
A <strong>Los</strong> Alarnos <strong>Primer</strong><br />
Edited by Necia Grant Cooper and Geoffrey B. West<br />
<strong>Los</strong> <strong>Alamos</strong> National Laboratory<br />
CAMBRIDGE UNIVERSITY PRESS<br />
Cain bridge<br />
New York New Rochelle Melboitrne Sydney
Published by the Press Syndicate of the University of Cambridge<br />
The Pit1 Building. Trurnpington Street. Cambridge CB2 IRP<br />
32 East 57th Street. New York. NY 10022. USA<br />
10 Svarnford Road, Oakleigh. Melbourne 3166. Australia<br />
0 Cambridge University Press 1988<br />
First publi:ihed 1988<br />
Printed in the United States of America<br />
Lihrury cf Congress Cat~tlo~iii~-iii-Puhlic~itioii Dota<br />
Panicle physics<br />
An updated version of <strong>Los</strong> <strong>Alamos</strong> science, no. I I<br />
(surnrner/fiill 1984).<br />
Includes index.<br />
I. Panicles (Nuclear physics) I. Cooper. Necia<br />
Grant. II. West, Geoffrey B.<br />
QC793.P358 1987 539.7'21 87-10858<br />
British Librrirv Cotologuiiig irr Puhliccition Dcttct<br />
<strong>Particle</strong> physics : a <strong>Los</strong> Alarnos primer.<br />
I. <strong>Particle</strong>s (Nuclear physics)<br />
1. Cooper. Necia Grant 2. West. Geoffrey B.<br />
539.7'21 QC793.2<br />
ISBN 0-52 1-34542- I hard covers<br />
ISBN 0-521-34780-7 paperback
General Editors<br />
Editor<br />
Associate Editors<br />
Designer<br />
Illustration and Production<br />
Necia Grant Cooper<br />
Geoffrey B. West<br />
Necia Grant Cooper<br />
Roger Eckhardt<br />
Nancy Shera<br />
Gloria Sharp<br />
Jim Cmz<br />
Anita Flores<br />
John Flower<br />
Judy Gibes<br />
Jim E. Lovato<br />
Lenny Martinez<br />
LeRoy Sanchez<br />
Mary Stovall<br />
Chris West
Contents<br />
Prefae to <strong>Los</strong> <strong>Alamos</strong> Science, Number 11, Summer/Falll984 -<br />
Introduction<br />
vi ii<br />
ix<br />
Theoretical Framework<br />
Scale and Dimension-From Animals to Quarks 2<br />
by Geofrey B. West<br />
Fundamental Constants and the Rayleigh-Riabouchinsky Paradox 12<br />
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model 22<br />
by Stuart Raby, Richard C. Slansky, and Geoflrey B. West<br />
QCD on a Cray: The Masses of Elementary <strong>Particle</strong>s ~<br />
by Gerald Guralnik, Tony Warnock, and Charles Zemach<br />
41<br />
Lecture Notes-From Simple Field Theories to the Standard Model 54<br />
by Richard C. Slansky<br />
Toward a Unified Theory: An Essay on the Role of Supergravity in the Search for Unification 72<br />
by Richard C. Slansky<br />
Fields and Spins in Higher Dimensions 86<br />
Supersymmetry at 100 GeV 98<br />
by Stuart Raby<br />
vi<br />
Supersymmetry in Quantum Mechanics 102
The Family Problem 114<br />
by T. Goldman and Michael Martin Nieto<br />
Addendum: CP Violation in Heavy-Quark Systems 124<br />
Experimental Developments<br />
Experiments to Test Unification Schemes 128<br />
by Gary H. Sanders<br />
An Experimentalist’s View of the Standard Model 130<br />
Addendum: An Experimental Update 149<br />
The March toward Higher Energies 150<br />
by S. Peter Rosen<br />
Addendum: The Next Step in Energy<br />
LAMPF I1 and the High-Intensity Frontier<br />
by Henry A. Thiessen<br />
156<br />
158<br />
The SSC-An Engineering Challenge 164<br />
by Mahlon T. Wilson<br />
Science Underground-The Search for Rare Events 166<br />
by L. M. Simmons, Jr.<br />
Personal Perspectives<br />
Quarks and Quirks among Friends 180<br />
A round table on the history and future of particle physics with Peter A. Carruthers, Stuart Raby,<br />
Richard C. Slansky, Geoflrey B. West, and George Zweig<br />
Index<br />
196<br />
vii
Preface<br />
0<br />
n the cover a mandala of the laws of physics floats in the<br />
cosmos of reality. It symbolizes the interplay between the<br />
inner world of abstract creation and the outer realms of<br />
measurable truth. The tension between these two is the magic and the<br />
challenge of fundamental physics.<br />
According to Jung, the “squaring ofthe circle” (the mandala) is the<br />
archetype of wholeness, the totality of the self. Such images are<br />
sometimes created spontaneously by individuals attempting to integrate<br />
what seem to be irreconcilable differences within themselves.<br />
Here the mandala displays the modern attempt by particle physicists<br />
to bring together the basic forces of nature in one theoretical<br />
framework.<br />
The content of this so-called standard model is summarized by the<br />
mysterious-looking symbols labeling each force: U( 1) for electromagnetism,<br />
SU(2) for weak interactions, SU(3)c for strong interactions,<br />
and SL(2C) for gravity; each symbol stands for an invariance,<br />
or symmetry, of nature. Symmetries tell us what remains<br />
constant through the changing universe. They are what give order to<br />
the world. There are many in nature, but those listed on the mandala<br />
are special. Each is a local symmetry, that is, it manifests independently<br />
at every space-time point and therefore implies the<br />
existence of a separate force. In other words, local symmetries<br />
determine all the forces of nature. This discovery is the culmination<br />
of physics over the last century. It is a simple idea, and it turns out to<br />
describe all phenomena so far observed.<br />
Where does particle physics go from here? The major direction of<br />
present research (and a major theme of this issue) is represented by<br />
the spiral that starts at electromagnetism and turns into the center at<br />
gravity. It suggests that the separate symmetries may be encompassed<br />
in one larger symmetry that governs the entire universe-one symmetry,<br />
one principle, one theory. The spiral also suggests that including<br />
gravity in such a theory involves understanding the structure of<br />
space-time at unimaginably small distance scales.<br />
Julian Schwinger, whose seminal idea led to the modern unification<br />
of electromagnetic and weak interactions, regards the present<br />
emphasis on unification with skepticism: “It’s nothing more than<br />
another symptom of the urge that amicts every generation of<br />
physicists-the itch to have all the fundamental questions answered<br />
in their own lifetime.”* To others the goal seems tantalizingly close,<br />
an achievement that may be reached, if not this year-then maybe<br />
the next.. . .<br />
The hope of unification depends on a second theme of this issue,<br />
symbolized by the ants and elephants walking round the mandala.<br />
These creatures are our symbol of scaling, the sizing up and sizing<br />
down of physical systems. Strength (or any other quality, for that<br />
matter) may look different on different scales. But if we look hard<br />
enough, we can find certain invariances to changes in scale that<br />
define the correct variables for describing a problem. Why do ants<br />
appear stronger than elephants? Why does the strong force look weak<br />
at high energies? How could all the forces of nature be manifestations<br />
of a single theory? These are the questions explored in “Scale and<br />
Dimension-from animals to quarks,” a seductively playful article<br />
that leads us to one of the most important contributions to modem<br />
physics, the renormalization group equations of quantum field theory.<br />
The insights about scaling gained from these equations are<br />
important not only to elementary particle physics but also to phase<br />
transition theory and the dynamics of complex systems.<br />
All the articles in this issue were written by scientists who care to<br />
tell not only about their own research but about the whole field of<br />
particle physics, its stunning achievements and its probing questions.<br />
Outsiders lo this field hear the names of the latest new particles, the<br />
buzz words such as grand unification or supersymmetry, and the<br />
plans for the United States to regain its leadership in this glamorous,<br />
high tech area of big science. But what is the real progress? Why does<br />
this field continue to attract the best minds in science? Why is it a<br />
major achievement of human thought? From a distance it may be<br />
hard to tell-except that it satisfies some deep urge to undcrstand<br />
how the world works. But if one could be given a closer look at the<br />
technical content of this field, its depth and richness would become<br />
apparent. That is the aim ofthe present issue.<br />
The hardest job was defining the technical level. How could the<br />
framework of the standard model be appreciated by someone unfamiliar<br />
with symmetry principles? How could modern particle<br />
physics research, all of which builds on the standard model, be<br />
understood by someone unfamiliar with what everyone in the field<br />
takes for granted? We hope we have solved this problem by presenting<br />
some of the major concepts on several levels and in several<br />
different places. We even include our own reference material, a<br />
remarkably clear and friendly set of lecture notes prepared especially<br />
for this issue.<br />
As one who was trained in this field, I returned to it with some<br />
trepidation-to deal with the subject matter, which had been so<br />
difficult, and with the personalitites competing in the field, who<br />
sometimes ride roughshod over each other as they battle these unruly<br />
abstractions. Much to my delight and the delight of the <strong>Los</strong> 241amos<br />
Science staff, the experience of preparing this issue was immensely<br />
enjoyable and rewarding. The authors were enthusiastic about explaining<br />
and re-explaining, about considering the essence of each<br />
point one more time to make sure that the readers too would be able<br />
to grasp it. Their generosity and interest made it fun for us to learn.<br />
May this presentation also be a treat for you.<br />
*This quote appeared in “How the Universe Works” by Robert P. Crease and<br />
Charles C. Mann (The Atlantic Monthly, August, 1984), a fast-paced article<br />
about the history of the electoweak theory.<br />
Necia Grant Cooper<br />
1984<br />
viii
Introduction<br />
B<br />
eginning with the dramatic discovery of the J/y particle in<br />
1974, particle physics has gone through a remarkably productive<br />
and exciting period. Quantum field theory, developed<br />
during the 1940s and ‘50s but abandoned in the O OS, was reestablished<br />
as the language for formulating theoretical concepts. The<br />
unification of the weak and electromagnetic interactions via a socalled<br />
non-Abelian gauge theory could only be understood in this<br />
framework. A similar theory of the strong interactions, quantum<br />
chromodynamics, was also constructed during this period, and<br />
nowadays one refers to the total package of the strong and electroweak<br />
theories as “the standard model.” Over the last decade the<br />
predictions of the standard model have been spectacularly confirmed,<br />
so much so that it is now almost taken for granted as<br />
embodying all physics below about 100 GeV. The culmination of this<br />
exuberant period was the inevitable discovery in 1983 of the W* and<br />
Zo particles, the massive bosons predicted by the standard model to<br />
mediate the weak interactions. Although the masses of these particles<br />
were precisely those predicted by the SU(2) X U(l) electroweak<br />
theory, their discovery was almost anticlimactic, so accepting had the<br />
particle-physics community become of the standard model. Indeed,<br />
future research in particle physics is often referred to as “physics<br />
beyond the standard model,” an implicit tribute to the progress of the<br />
past decade.<br />
The development of the standard model during the 1970s brought<br />
with it a lexicon of new words and concepts-quark, gluon, charm,<br />
color, spontaneous symmetry breaking, and asymptotic freedom, to<br />
name a few. Supersymmetry, preons, strings, and worlds of ten<br />
dimensions are among the buzz words added in the ’80s. While<br />
scientists, engineers, and even many lay people will recognize some<br />
subset of these words, only a few have more than a superficial<br />
understanding of the profound achievement they denote. Add to this<br />
the demand by particle physicists for several billion dollars to build a<br />
super-accelerator in order to explore “physics beyond the standard<br />
model,” and one can sense the gap between the particle physicist and<br />
his “public” reaching irreparable proportions. On the other hand<br />
there remains an endless wonder and fascination in the public’s eye<br />
for such speculative conceptual ideas, which are more usually associated<br />
with the literature of science fiction than with Physical<br />
Review.<br />
It was with some of these thoughts in mind that a group of us at<br />
<strong>Los</strong> <strong>Alamos</strong> National Laboratory decided to put together a series of<br />
pedagogical articles explaining in relatively elementary scientific<br />
language the accomplishments, successes, and projected future of<br />
high-energy physics. The articles, intended for a wide scientific<br />
audience, originally appeared in a 1984 issue of <strong>Los</strong> <strong>Alamos</strong> Science,<br />
a technical publication of the Laboratory. Since that time they have<br />
been used as a teaching tool in particle-physics courses and as a<br />
reference source by experimentalists in the field.<br />
<strong>Particle</strong> <strong>Physics</strong>-A <strong>Los</strong> <strong>Alamos</strong> <strong>Primer</strong> is basically an updated<br />
version of the original <strong>Los</strong> <strong>Alamos</strong> Science issue. We believe it will<br />
continue to help educate undergraduate and graduate students as<br />
well as bridge the gap between experimentalists and theorists. We are<br />
also confident that it will help non-experts to develop a good feel for<br />
the subject.<br />
The text consists of eight “chapters,” the first five devoted to the<br />
concepual fiamework of modem particle physics and the last three to<br />
experiments and accelerators. Each is written by a separate author, or<br />
group of authors, and is to a large extent self-contained. In addition,<br />
we have included a round table among several particle physicists that<br />
addresses some of the broader issues facing the field. This discussion<br />
is in some ways a unique evaluation of the present status of particle<br />
physics. It is quite personal and idiosyncratic, sometimes irreverent,<br />
and occasionally controversial. For the nonexpert it is probably the<br />
place to begin!<br />
The first article addresses the question of scaling. In its broadest<br />
sense this lies at the heart of any attempt to unify into one theory the<br />
fundamental forces of Nature-forces seemingly so very different in<br />
strength. “Scale and Dimension-From Animals to Quarks” begins<br />
by reviewing in an elementary and somewhat whimsical fashion the<br />
whole question of scale in classical physics and then introduces the<br />
more sophisticated concept of the renormalization group. The renormalization<br />
group is really no more than a generalization of<br />
classical dimensional analysis to the area of quantum field theory: it<br />
answers the seminal question of how a physical system responds to a<br />
change in scale. The concept plays a central role in the modem view<br />
of quantum field theory and has been particularly successful in<br />
elucidating the nature of phase transitions. Indeed, it is from this<br />
vantage point that the intimate relationship between particle and<br />
condensed-matter physics has developed. Clearly, the manner in<br />
which physics evolves from one energy or length scale to another is of<br />
fundamental importance.<br />
The second article, “<strong>Particle</strong> <strong>Physics</strong> and the Standard Model,”<br />
addresses the question of unification with an elementary yet comprehensive<br />
discussion of how the famous electroweak theory is<br />
constructed and works. The role of internal symmetries and their<br />
incorporation into a principle of local gauge invariance and subsequent<br />
manifestation as a non-Abelian gauge field theory are explained<br />
in a pedagogical fashion. The other component of the<br />
standard model, namely quantum chromodynamics (QCD), the<br />
theory of the strong interactions, is similarly treated in this article.<br />
Again, the discussion is rather elementary, beginning with an exposition<br />
of the “old” SU(3) of the “Eightfold Way” and finishing with<br />
the field theory of quarks and gluons. For the more ambitious reader<br />
we have included a set of “lectures” by Richard Slansky that give
some of the technical details necessary in going “from simple field<br />
theories to the standard model.” Crucial concepts such as local gauge<br />
invariance, spontaneous symmetry breaking, and emergence of the<br />
Higgs particles that give rise to the masses of elementary particles are<br />
expressed in the mathematical language of field theory and should be<br />
readily accessible to the serious student of the field. These lectures<br />
are very clear and provide the reader with the explicit equations<br />
embodying the physics discussed in the article on the standard<br />
model.<br />
Following this review of accepted lore, we begin our journey into<br />
“physics beyond the standard model” with an essay on supergravity<br />
by Slansky entitled “Toward a Unified Theory.” In it he discusses<br />
some of the speculative ideas that gained popularity in the late 1970s.<br />
Among them are supersymmetry (a proposed symmetry between<br />
fermions and bosons) and the embedding of our four-dimensional<br />
space-time world in a larger number of dimensions. Supergravity, a<br />
theory that encompasses both of these ideas, was the first serious<br />
attempt to include Einstein’s gravity in the unification scheme. This<br />
article also includes a description of superstring theory, which has<br />
gained tremendous popularity just in the last year or so. Slansky<br />
explains how the shortcomings of the supergravity scenario are<br />
circumvented by basing a unified theory on elementary fibers, or<br />
strings, rather than on point particles. This area of research is in a<br />
state of flux at the moment, and it is still far from clear whether<br />
strings really will form the basis of the “final” theory. The problems<br />
are both conceptual and technical. Conceptually there is still no hint<br />
as to what principles are to replace the equivalence principle and<br />
general coordinate invariance, which form the bases of Einstein’s<br />
gravity. Technically, the mathematics of string theory is beyond the<br />
usual expertise of the theoretical physicist; indeed it is on the<br />
forefront of mathematical research itself. This may be the first time<br />
for a hundred years or more that research in physics and<br />
mathematics has coincided. Some may view this as a bad omen,<br />
others as the dawning of a new exciting age leading to the equations<br />
of the universe! Only time will tell.<br />
A less ambitious use of supersymmetry has been in the attempts<br />
to unify, without gravity, the electroweak and strong theories of the<br />
standard model. Stuart Raby, in his article “Supersymmetry at<br />
100 GeV,” discusses some of these efforts by concentrating on the<br />
phenomenological implications of a world in which every boson<br />
has a fermion partner and vice versa. These include a possible<br />
explanation for why proton decay, certainly one of the more<br />
dramatic predictions of grand unified theories, has not yet been<br />
seen. Supersymmetric phenomenology has served as an important<br />
guide for speculating about what can be seen at new accelerators. A<br />
special feature of this article is the selfcontained section “Supersymmetry<br />
in Quantum Mechanics,” in which Raby explains this<br />
novel space-time symmetry in a setting stripped of all field-theoretic<br />
baggage.<br />
One of the more mysterious problems in particle physics is “the<br />
family problem” described in an article of that title by Terry<br />
Goldman and Michael Nieto. The apparent replication of the<br />
electron and its neutrino in at least two more families differing only<br />
in their mass scales has remained a mystery ever since the discovery<br />
of the muon. This replication, exhibited also by the quarks, can be<br />
accommodated in unified theories, though no satisfactory explanation<br />
of the family structure, nor even a prediction of the total<br />
number of families, has been advanced. The phenomenology of this<br />
problem as well as some attempts to understand it are carefully<br />
reviewed. An addendum to the original article presents a slightly<br />
more technical discussion of how experiments involving the third<br />
quark family might extend our knowledge of CP violation. This<br />
symmetry violation remains perhaps the most mysterious aspect of<br />
the known particle phenomenology.<br />
The next three articles concern the experimental side of particle<br />
physics. Although the choice of <strong>Los</strong> <strong>Alamos</strong> experiments to illustrate<br />
certain points does reflect some parochial interests of the authors,<br />
these articles succeed in providing a broad overview of experimental<br />
methodology. In this era of elaborate detection techniques requiring<br />
extensive collaboration, it is often difficult for the uninitiated to<br />
unravel the complicated machinations that are involved in the<br />
experimental process. In “Experiments to Test Unification Schemes”<br />
Gary Sanders presents a very clear exposition of the physics input to<br />
this process. Indeed, as if to emphasize the departure from the world<br />
of theory, he has included a brief page-and-a-half pr6cis subtitled “An<br />
Experimentalist’s View of the Standard Model.” For the beginner this<br />
might be read immediately following the round table! Sanders describes<br />
in some detail four experiments designed specifically to test<br />
the standard model, all being conducted at <strong>Los</strong> <strong>Alamos</strong>. Each is a<br />
“high-precision” experiment in which, say, a specific decay rate is<br />
measured and compared with the value predicted by the standard<br />
model. These experiments are prototypical of the kind that have been<br />
and will continue to be done at accelerators around the world to push<br />
the theory to its limits. Most exciting, or course, would be the<br />
observation of some deviation from the standard model that could be<br />
associated with grand unification. However, in an addendum Sanders<br />
reports that no such deviations were seen in the data from the <strong>Los</strong><br />
<strong>Alamos</strong> experiments and others. So far the standard model has stood<br />
the test of time.<br />
The following article by Peter Rosen, “The March toward Higher<br />
Energies,” surveys the high-energy accelerator landscape beginning<br />
with a historical perspective and finishing with a glimpse into what<br />
we might expect in the not-toodistant future. The emphasis here is<br />
on tests of the standard model and searches for new and exotic<br />
particles not included in it. The traditional methodology is quite<br />
simple: go for the highest energy possible. This has certainly been<br />
successhl in the past, and we have no reason to believe that it won’t<br />
be successful in the future. Thus, there is a push to build a giant<br />
superconducting supercollider (SSC) that could probe mass scales in<br />
excess of 20 TeV, or 2 X loi3 eV. We have also included a brief<br />
X
eport by Mahlon Wilson, an accelerator physicist, on some of the<br />
problems peculiar to the gigantic scale of the SSC.<br />
An alternative technique for probing high mass scales is to perform<br />
very accurate experiments in search of deviations from expected<br />
results, such as those described in Sanders’ article. Obviously, highintensity<br />
beams are the desired tool in this approach. A high-intensity<br />
machine has been proposed for b s <strong>Alamos</strong>, and another brief report<br />
by Henry Thiessen, also an accelerator physicist, describes that<br />
machine and some of the questions it might answer. The reports on<br />
the SSC and LAMPF I1 provide an idea of what is involved in<br />
designing tomorrow’s accelerators.<br />
The final article is a review by Mike Simmons on “science underground.”<br />
In it he discusses what particle physics can be learned from<br />
experiments performed deep underground to isolate rare events of<br />
interest. The most famous of these is the search for proton decay.<br />
Other experiments measure the flux of neutrinos from the sun and<br />
search for exotic particles (such as magnetic monopoles) in cosmic<br />
rays. These essential fishing expeditions use “beams” from the<br />
biggest accelerator of them all, namely the universe!<br />
<strong>Particle</strong> <strong>Physics</strong>-A LQS <strong>Alamos</strong> <strong>Primer</strong> thus provides the reader<br />
with a comprehensive, up-to-date introduction to the field of particle<br />
physics. Our belief is that it will be a useful educational guide to both<br />
the student and professional worker in the field as well as provide the<br />
general scientist with an insight into some of the recent accomplishments<br />
in understanding the fundamental structure of the universe.<br />
In conclusion we would like to thank the staff of <strong>Los</strong> <strong>Alamos</strong><br />
Science for their invaluable help in making this primer lively and<br />
accessible to a wide audience.<br />
Geoffrey B. West<br />
1986<br />
xi
<strong>Particle</strong> Phvsics<br />
A <strong>Los</strong> <strong>Alamos</strong> <strong>Primer</strong>
“1 have multiplied visions and used similitudes.” - Hosea 7:lO<br />
In his marvelous book Dialogues Concerning<br />
Two New Sciences there is a remarkably clear<br />
discussion on the effects of scaling up the<br />
dimensions of a physical object. Galileo realized<br />
that ifone simply scaled up its size, the<br />
weight of an animal would increase significantly<br />
faster than its strength, causing it ultimately<br />
to collapse. As Galileo says (in the<br />
words of Salviati during the discorso of the<br />
second day), “. . . you can plainly see the<br />
impossibility of increasing the size of structures<br />
to vast dimensions . . . if his height be<br />
increased inordinately, he will fall and be<br />
crushed under his own weight.” The simple<br />
scaling up of an insect to some monstrous<br />
size is thus a physical impossibility, and we<br />
can rest assured that these old sci-fi images<br />
are no more than fiction! Clearly, to create a<br />
giant one “must either find a harder and<br />
stronger material . . . or admit a diminution<br />
of strength,” a fact long known to architects.<br />
It is remarkable thal. so many years before<br />
its deep significance could be appreciated,<br />
Galileo had investigated one of the most<br />
fundamental questions of nature: namely,<br />
what happens to a physical system when one<br />
changes scale? Nowadays this is the seminal<br />
question for quantum field theory, phase<br />
transition theory, the dynamics of complex<br />
systems, and attempts to unify all forces in<br />
nature. Tremendous progress has been made<br />
in these areas during the past fifteen years<br />
based upon answers to this question, and I<br />
shall try in the latter part of this article to give<br />
some flavor of what has been accomplished.<br />
However, I want first to remind the reader of<br />
the power of dimensional analysis in<br />
classical physics. Although this is stock-intrade<br />
to all physicists, it is useful (and, more<br />
pertinently, fun) to go through several examples<br />
that explicate the basic ideas. Be warned,<br />
there are some surprises.<br />
Classical Scaling<br />
Let us first re-examine Galileo’s original<br />
analysis. For similar structures* (that is,<br />
structures having the same physical<br />
characteristics such as shape, density, or<br />
chemical composition) Galileo perceived<br />
that weight Wincreases linearly with volume<br />
V, whereas strength increases only like a<br />
cross-sectional area A. Since for similar<br />
structures V a l3 and A 12, where 1 is some<br />
characteristic length (such as the height ofthe<br />
structure), we conclude that<br />
-_<br />
Strength A 1 1<br />
tc-a-a -<br />
Weight V I WV~’<br />
Thus, as Galileo noted, smaller animals “appear”<br />
stronger than larger ones. (It is amusing<br />
that Jerome Siege1 and Joe Shuster, the<br />
creators of Superman, implicitly appealed to<br />
such an argument in one of the first issues of<br />
their comic.+ They rationalized his super<br />
strength by drawing a rather dubious analogy<br />
with “the lowly ant who can support weights<br />
hundreds of times its own” (sic!).) Incidentally,<br />
the above discussion can be used to<br />
understand why the bones and limbs of<br />
larger animals must be proportionately<br />
stouter than those of smaller ones, a nice<br />
example of which can be seen in Fig. 1.<br />
Arguments of this sort were used extensively<br />
during the late 19th century to un-<br />
4
Scale and Dimension<br />
Fig. 1. Two extinct mammals: (a) Neohipparion, a small American horse and (b)<br />
Mastodon, a large, elephant-like animal, illustrating that the bones of heavier<br />
animals are proportionately stouter and thus proportionately stronger.<br />
derstand the gross features of the biological<br />
world; indeed, the general size and shape of<br />
animals and plants can be viewed as nature’s<br />
way of responding to the constraints of gravity,<br />
surface phenomena, viscous flow, and<br />
the like. For example, one can understand<br />
why man cannot fly under his own muscular<br />
power, why small animals leap as high as<br />
larger ones, and so on.<br />
A classic example is the way metabolic<br />
rate varies from animal to animal. A<br />
measure B of metabolic rate is simply the<br />
heat lost by a body in a steady inactive state,<br />
which can be expected to be dominated by<br />
the surface effects of sweating and radiation.<br />
Symbolically, therefore, one expects<br />
B a W213. The data (plotted logarithmically<br />
in Fig. 2) show that metabolic rate does<br />
*The concept of similitude is usually attributed to<br />
Newton, who first spelled it out in the Principia<br />
when dealing with gravitational attraction. On<br />
reading the appropriate section it is clear that this<br />
was introduced only as a passing remark and does<br />
not have the same profound content as the remarks<br />
of Galileo.<br />
?This amusing observation was brought to my attention<br />
by Chris Llewellyn Smith.<br />
*This relationship with a slope of 3/4 is known as<br />
Kleiber’s law (M. Kleiber, Hilgardia 6(1932):315),<br />
whereas the area law is usually attributed to Rubner<br />
(M. Rubner, Zeitschrift fur Biologie (Munich)<br />
19(1883):535).<br />
indeed scale, that is, all animals lie on a<br />
single curve in spite of the fact that an<br />
elephant is neither a blown-up mouse nor a<br />
blown-up chimpanzee. However, the slope of<br />
the best-fit curve (the solid line) is closer to<br />
3/4 than to 2/3, indicating that effects other<br />
than the pure geometry of surface dependence<br />
are at work.#<br />
It is not my purpose here to discuss why<br />
this is so but rather to emphasize the importance<br />
of a scaling curve not only for establishing<br />
the scaling phenomenon itself but for<br />
revealing deviations from some naive<br />
prediction (such as the surface law shown as<br />
the dashed line in Fig. 2). Typically, deviations<br />
from a simple geometrical or<br />
kinematical analysis reflect the dynamics of<br />
the system and can only be understood by<br />
examining it in more detail. Put slightly differently,<br />
one can view deviations from naive<br />
scaling as a probe of the dynamics.<br />
The converse of this is also true: generally,<br />
one cannot draw conclusions concerning<br />
dynamics from naive scaling. As an illustration<br />
of this I now want to discuss some<br />
simple aspects of birds’ eggs. I will focus on<br />
the question of breathing during incubation<br />
and how certain physical variables scale<br />
from bird to bird. Figure 3, adapted from a<br />
Scientijic American article by Hermann<br />
Rahn, Amos Ar, and Charles V. Paganelli<br />
entitled “How Bird Eggs Breathe,” shows the<br />
dependence of oxygen conductance K and<br />
pore length I (that is, shell thickness) on egg<br />
mass W. The authors, noting the smaller<br />
slope for /, conclude that “pore length<br />
probably increases slower because the eggshell<br />
must be thin enough for the embryo to<br />
hatch.” This is clearly a dynamical conclusion!<br />
However, is it warranted?<br />
From naive geometric scaling one expects<br />
that for similar eggs I a W”3, which is in<br />
reasonable agreement with the data: a best fit<br />
(the straight line in the figure) actually gives /<br />
oc @.4. Since these data for pore length agree<br />
reasonably well with geometric scaling, no<br />
dynamical conclusion (such as the shell being<br />
thin enough for the egg to hatch) can be<br />
drawn. Ironically, rather than showing an<br />
anomalously slow growth with egg mass, the<br />
data for / actually manifest an anomalously<br />
fast growth (0.4 versus 0.33), not so dissimilar<br />
from the example of the metabolic<br />
rate!<br />
What about the behavior of the conductance,<br />
for which K 0: @.9? This relationship<br />
can also be understood on geometric<br />
grounds. Conductance is proportional to the<br />
totd available pore area and inversely<br />
proportional to pore length. However, total<br />
pore area is made up of two factors: the total<br />
number of pores times the area of individual<br />
pores. If one assumes that the number of<br />
pores per unit area remains constant from<br />
bird to bird (a reasonable assumption consistent<br />
with other data), then we have two<br />
factors that scale like area and one that<br />
5
IO0 1 o2 io4<br />
Body Weight (kg)<br />
scales inversely as length. One thus expects<br />
K a ( W2/3)2/ W1I3 = W, again in reasonable<br />
agreement with the data.<br />
Dimensional Analysis. The physical content<br />
of scaling is very often formulated in<br />
terms of the language of dimensional analysis.<br />
The seminal idea seems to be due to<br />
Fourier. He is, ofcourse, most famous for the<br />
invention of “Fourier analysis,” introduced<br />
in his great treatise Theorie Analpique de la<br />
Chaleur, first published in Paris in 1822.<br />
However, it is generally not appreciated that<br />
this same book contains another great contribution,<br />
namely, the use of dimensions for<br />
physical quantities. It is the ghost of Fourier<br />
that is the scourge of all freshman physics<br />
majors, for it was he who first realized that<br />
every physical quantity “has one dimension<br />
proper to itself, and that the terms of one and<br />
the same equation could not be compared, if<br />
they had not the same exponent of<br />
dimension.” He goes on: “We have introduced<br />
this consideration . . . to verify the<br />
analysis. .. it is the equivalent of the fundamental<br />
lemmas which the Greeks have left us<br />
without proof.” Indeed it is! Check the<br />
dimensions!-the rallying call of all<br />
physicists (and, hopefully, all engineers).<br />
However, it was only much later that<br />
physicists began to use the “method of<br />
dimensions” to solve physical problems. In a<br />
famous paper on the subject published in<br />
Fig. 2. Metabolic rate, measured as heat produced by the body in a steady state,<br />
plotted logarithmically against body weight. An analysis based on a surface<br />
dependence for the rate predicts a scaling curve with slope equal to 2/3 (dashed<br />
line) whereas the actual scaling curve has a slope equal to 3/4. Such deviation from<br />
simple geometrical scaling is indicative of other effects at work. (Figure based on<br />
one by Thomas McMahon, Science 179(1973):1201-1204 who, in turn, adapted it<br />
from M. Kleiber, Hilgardia 6(1932):315.)<br />
L- 5 I<br />
I I ‘ e<br />
Oxygen Conductance<br />
L<br />
B loo/<br />
m<br />
*<br />
’13<br />
L 10-<br />
a2<br />
a.<br />
(Slope = 0.9) .<br />
(Slope * 0.4)<br />
I=<br />
QI<br />
P<br />
3 0.01 I I I I<br />
1 10 100 1000<br />
Egg Mass (9)<br />
00<br />
-<br />
E<br />
-5<br />
en<br />
C<br />
3<br />
E! 0<br />
p.<br />
.I<br />
Fig. 3. Logarithmic plot of two parameters relevant to the breathing of birds’ eggs<br />
during incubation: the conductance of oxygen through the shell and thepore length<br />
(or shell thickness) as a function of egg mass. Both plots have slopes close to those<br />
predicted by simple geometrical scaling analyses. (Figure adapted from H. Rahn,<br />
A. Ar, and C. V. Paganelli, Scientific American 24O(February 1979):46-55.)<br />
.01<br />
6
Scale and Dimension<br />
Nature in 19 15, Rayleigh indignantly begins:<br />
“I have often been impressed by the scanty<br />
attention paid even by original workers in<br />
the field to the great principle of similitude.<br />
It happens not infrequently that results in the<br />
form of ‘laws’ are put forward as novelties on<br />
the basis of elaborate experiments, which<br />
might have been predicted a priori after a few<br />
minutes consideration!” He then proceeds to<br />
set things right by giving several examples of<br />
the power of dimensional analysis. It seems<br />
to have been from about this time that the<br />
method became standard fare for the<br />
physicist. I shall illustrate it with an amusing<br />
example.<br />
Most of us are familiar with the traditional<br />
Christmas or Thanksgiving problem of how<br />
much time to allow for cooking the turkey or<br />
goose. Many (inferior) cookbooks simply say<br />
something like “20 minutes per pound,” implying<br />
a linear relationship with weight.<br />
However, there exist superior cookbooks,<br />
such as the Better Homes and Gardens<br />
Cookbook, that recognize the nonlinear<br />
nature of this relationship.<br />
Figure 4 is based on a chart from this<br />
cookbook showing how cooking time t varies<br />
with the weight ofthe bird W. Let us see how<br />
Fig. 4. The cooking time for a turkey or<br />
goose as a logarithmic function of its<br />
weight. (Based on a table in Better<br />
Homes and Gardens Cookbook, Des<br />
MoinexMeridith Corp., Better Homes<br />
and Gardens Books, 1962, p. 272.)<br />
one can understand this variation using “the<br />
great principle of similitude.” Let T be the<br />
temperature distribution inside the turkey<br />
and To the oven temperature (both measured<br />
relative to the outside air temperature). T<br />
satisfies Fourier’s heat diffusion equation:<br />
arjat = KV~T, where K is the diffusion coeficient.<br />
Now, in general, for the dimensional<br />
quantities in this problem, there will be a<br />
functional relationship of the form<br />
where p is the bird’s density. However,<br />
Fourier’s basic observation that the physics<br />
be independent of the choice of units, imposes<br />
a constraint on the form of the solution,<br />
which can be discerned by writing it in terms<br />
of dimensionless quantities. Only two independent<br />
dimensionless quantities can be<br />
constructed: T/To and p (~t)~/~/ W. If we use<br />
the first of these as the dependent variable,<br />
the solution, whatever its form, must be<br />
expressible in terms of the other. The relationship<br />
must therefore have the structure<br />
T<br />
To<br />
(3)<br />
The important point is that, since the lefthand<br />
side is dimensionless, the “arbitrary”<br />
function f must be a dimensionless function<br />
of a dimensionless variable. Equation 3, unlike<br />
the previous one, does not depend upon<br />
the choice of units since dimensionless quantities<br />
remain invariant to changes in scale.<br />
Let us now consider different but<br />
geometrically similar birds cooked to the<br />
same temperature distribution at the same<br />
oven temperature. Clearly, for all such birds<br />
there will be a scaling law<br />
p(Kt)3‘2<br />
=constant.<br />
(4)<br />
If the birds have the same physical<br />
characteristics (that is, the same p and K), Eq.<br />
4 reduces to<br />
t = constant x w2I3,<br />
(5)<br />
reflecting, not surprisingly, an area law, As<br />
can be seen from Fig. 4, this agrees rather<br />
well with the “data.”<br />
This formal type of analysis could also, of<br />
course, have been camed out for the<br />
metabolic rate and birds’ eggs problems. The<br />
advantage of such an analysis is that it delineates<br />
the assumptions made in reaching<br />
conclusions like B W2I3 since, in principle,<br />
it focuses upon all the relevant variables.<br />
Naturally this is crucial in the discussion of<br />
any physics problem. For complicated systems,<br />
such as birds’ eggs, with a very large<br />
number of variables, some prior insight or<br />
intuition must be used to decide what the<br />
important variables are. The dimensions of<br />
these variables are determined by the fundamental<br />
laws that they obey (such as the diffusion<br />
equation). Once the dimensions are<br />
known, the structure of the relationship between<br />
the variables is determined by<br />
Fourier’s principle. There is therefore no<br />
magic in dimensional analysis, only the art of<br />
choosing the “right” variables, ignoring the<br />
irrelevant, and knowing the physical laws<br />
they obey.<br />
As a simple example, consider the classic<br />
problem of the drag force F on a ship moving<br />
through a viscous fluid of density p. We shall<br />
choose F, p, the velocity v, the viscosity of the<br />
fluid p, some length paraTeter of the ship I,<br />
and the acceleration due to gravity g as our<br />
7
_- v<br />
,<br />
variables. Notice that we exclude other<br />
variables, such as the wind velocity and the<br />
amplitude of the sea waves because, under<br />
calm conditions, these are of secondary importance.<br />
Our conclusions may therefore not<br />
be valid for sailing ships!<br />
The physics of the problem is governed by<br />
the Navier-Stokes equation (which incorporates<br />
Newton’s law of viscous drag,<br />
telling us the dimensions of p) and the gravitational<br />
force law (telling us the dimensions<br />
of g). Using these dimensions automatically<br />
incorporates the appropriate physics. Since<br />
we have limited the variables to a set of six,<br />
which must be expressible in terms of three<br />
basic units (mass M, length L, and time T),<br />
there will only be three independent<br />
dimensionless combinations. These are<br />
chosen to be P = F/p$12 (the pressure coeficient),<br />
R = vlp/p (Reynold’s number), and<br />
N, = $/lg (Froude’s number). Although any<br />
three similar combinations could have been<br />
chosen, these three are special because they<br />
delineate the physics. For example, Reynold‘s<br />
number R relates to the viscous drag<br />
on a body moving through a fluid, whereas<br />
Froude’s number NF relates to the forces<br />
involved with waves and eddies generated on<br />
the surface of the fluid by the movement.<br />
Thus the rationale for the combinations R<br />
and NF is to separate the role of the viscous<br />
forces from that of the gravitational: R does<br />
not depend on g, and F does not depend on<br />
p. Furthermore, Pdoes not depend on either!<br />
Dimensional analysis now requires that<br />
the solution for the pressure coefficient,<br />
Libster 1924<br />
0 Allan 1900<br />
1, Cottingen 1921<br />
a Cottingen 1926<br />
--- Results of higher pressures 1922-23<br />
1 o-2 1 oo 1 o2 lo4<br />
Reynolds Number R<br />
lo6<br />
Fig. 5. The scaling curve for the motion of a sphere through a<br />
fluid that results when data from a variety of experiments<br />
are plotted in terms of two dimensionless variables: the<br />
pressure or drag coefficent P versus Reynolds number R.<br />
(Figure adapted from AIP Handbook of <strong>Physics</strong>, 2nd edition<br />
(1963):section II, p. 253.)<br />
8
Scale and Dimension<br />
whatever it is, must be expressible in the<br />
dimensionless form<br />
The actual drag force F can easily be obtained<br />
from this equation by re-expressing it<br />
in terms of the dimensional variables (see<br />
Eq. 8 below).<br />
First, however, consider a situation where<br />
surface waves generated by the moving object<br />
are unimportant (an extreme example is<br />
a submarine). In this case g will not enter the<br />
solution since it is manifested as tht restoring<br />
force for surface waves. N,c can then be<br />
dropped from the solution, reducing Eq. 6 to<br />
the simple form<br />
P=f(R). (7)<br />
In terms of the original dimensional<br />
variables, this is equivalent to<br />
Historically, these last equations have been<br />
well tested by measuring the speed of different<br />
sizes and types of balls moving<br />
through different liquids. If the data are<br />
plotted using the dimensionless variables,<br />
that is, Pversus R, then all the data should lie<br />
on just one curve regardless of the size of the<br />
ball or the nature of the liquid. Such a curve<br />
is called a scaling curve, a wonderful example<br />
of which is shown in Fig. 5 where one sees a<br />
scaling phenomenon that varies over seven<br />
orders of magnitude! It is important to recognize<br />
that if one had used dimensional<br />
variables and plotted F versus I, for example,<br />
then, instead of a single curve, there would<br />
have been many different and apparently<br />
unrelated curves for the different liquids.<br />
Using carefully chosen dimensionless<br />
variables (such as Reynold’s number) is not<br />
only physically more sound but usually<br />
greatly simplifies the task of representing the<br />
data.<br />
A remarkable consequence of this analysis<br />
is that, for similar bodies, the ratio of drag<br />
Fig. 6. The time needed for a rowing boat to complete a 2000-meter course in calm<br />
coxditions as a function of the number of oarsmen. Data were taken from several<br />
international rowing championship events and illustrate the surprisingly slow<br />
dropoff predicted by modeling theory. (Adapted from T. A. McMahon, Science<br />
173(1971):349-351.)<br />
force to weight decreases as the size of the<br />
structure increases. From Archimedes’ principle<br />
the volume ofwater displaced by a ship<br />
is proportional to its weight, that is, W a l3<br />
(this, incidentally, is why there is no need to<br />
include W as an independent variable in<br />
deriving these equations). Combined with<br />
Eq. 8 this leads to the conclusion that<br />
F *<br />
- cc-. (9)<br />
W I<br />
This scaling law was extremely important in<br />
the 19th century because it showed that it<br />
was cost efective to build bigger ships,<br />
thereby justifying the use of large iron steamboats!<br />
The great usefulness of scaling laws is also<br />
illustrated by the observation that the<br />
behavior of P for large ships (I- W) can be<br />
derived from the behavior of small ships<br />
moving very fast (v- w). This is so because<br />
both limits are controlled by the same<br />
asymptotic behavior off(R) =f(vlp/p). Such<br />
observations form the basis of modeling theory<br />
so crucial in the design of aircraft, ships,<br />
buildings, and so forth.<br />
Thomas McMahon, in an article in Science,<br />
has pointed out another, somewhat<br />
more amusing, consequence to the drag force<br />
equation. He was interested in how the speed<br />
of a rowing boat scales with the number of<br />
oarsmen n and argued that, at a steady velocity,<br />
the power expended by the oarsmen E to<br />
overcome the drag force is given by Fv. Thus<br />
9
Using Archimedes’ principle again and the<br />
fact that both E and W should be directly<br />
proportional to n leads to the remarkable<br />
scaling law<br />
which shows a very slow growth with n.<br />
Figure 6 exhibits data collected by McMahon<br />
from various rowing events for the time t (a<br />
I/v) taken to cover a fixed 2000-meter course<br />
under calm conditions. One can see quite<br />
plainly the verification of his predicted<br />
law-a most satisfying result!<br />
There are many other fascinating and<br />
exotic examples of the power of dimensional<br />
analysis. However, rather than belaboring<br />
the point, I would like to mention a slightly<br />
different application of scaling before I turn<br />
to the mathematical formulation. All the examples<br />
considered so far are ofa quantitative<br />
nature based on well-known laws of physics.<br />
There are, however, situations where the<br />
qualitative observation of scaling can be<br />
used to scientific advantage to reveal phenomenological<br />
“laws.”<br />
A nice example (Fig. 7), taken from an<br />
article by David Pilbeam and Stephen Jay<br />
Gould, shows how the endocranial volume V<br />
(loosely speaking, the brain size) scales with<br />
body weight W for various hominids and<br />
pongids. The behavior for modern pongids is<br />
typical of most species in that the exponent<br />
a, defined by the phenomenological relationship<br />
V 0: W, is approximately 1/3 (for<br />
mammals a varies from 0.2 to 0.4). It is very<br />
satisfying that a similar behavior is exhibited<br />
by australopithecines, extinct cousins of our<br />
lineage that died out over a million years ago.<br />
However, as Pilbean and Gould point out,<br />
I<br />
our homo sapiens lineage shows a strikingly<br />
different behavior, namely: a % 5/3. Notice<br />
that neither this relationship nor the “standard”<br />
behavior (a % 1/3) is close to the naive<br />
geometrical scaling prediction of a = 1.<br />
These data illustrate dramatically the<br />
qualitative evolutionary advance in the<br />
brain development of man. Even though the<br />
reasons for a = 1/3 may not be understood,<br />
this value can serve as the “standard” for<br />
revealing deviations and provoking speculation<br />
concerning evolutionary progress: for<br />
example, what is the deep significance of a<br />
brain size that grows linearly with height<br />
versus a brain size that grows like its fifth<br />
power? I shall not enter into such questions<br />
here, tempting though they be.<br />
Such phenomenological scaling laws<br />
(whether for brain volume, tooth area, or<br />
some other measurable parameter of the fos-<br />
- 6<br />
0<br />
-<br />
1250<br />
1000<br />
0,<br />
E 75c<br />
3<br />
0<br />
><br />
c.<br />
m<br />
I-<br />
S<br />
!?<br />
0” 500<br />
0 c<br />
w<br />
350<br />
a= 1.73 / /r<br />
/<br />
I<br />
?‘<br />
/<br />
A<br />
I<br />
/<br />
sil) can also be used as corroborative<br />
evidence for assigning a newly found fossil of<br />
some large primate to a particular lineage.<br />
The fossil’s location on such curves can, in<br />
principle, be used to distinguish an australopithecine<br />
from a homo. Notice, however,<br />
that implicit in all this discussion is knowledge<br />
of body weight; presumably,<br />
anthropologists have developed verifiable<br />
techniques for estimating this quantity. Since<br />
they necessarily work with fragments only,<br />
some further scaling assumptions must be<br />
involved in their estimates!<br />
Relevant Variables. As already emphasized,<br />
the most important and artful aspect of the<br />
method of dimensions is the choice of<br />
variables relevant to the problem and their<br />
grouping into dimensionless combinations<br />
that delineate the physics. In spite of the<br />
/ 0 Australopithecines<br />
/ A Homo Lineage<br />
Pongids<br />
30 40 50 75 100<br />
Body Weight (kg)<br />
Fig. 7, Scaling curves for endocranial volume (or brain size) as a function of body<br />
weight. The slope of the curve for our homo sapiens lineage (dashed line) is<br />
markedly different from those for australopithecines, extinct cousins of the homo<br />
lineage, and for modern pongids, which include the chimpanzee, gorilla and<br />
orangutan. (Adapted from D. Pilbeam and S. J. Gould, Science<br />
I86(19 74):892-901.)<br />
10
Scale and Dimension<br />
relative simplicity of the method there are<br />
inevitably paradoxes and pitfalls, a famous<br />
case of which occurs in Rayleigh’s 1915<br />
paper mentioned earlier. His last example<br />
concerns the rate of heat lost H by a conductor<br />
immersed in a stream of inviscid fluid<br />
moving past it with velocity v (“Boussinesq’s<br />
problem”). Rayleigh showed that, if K is the<br />
heat conductivity, C the specific heat of the<br />
fluid, 0 the temperature difference, and 1<br />
some linear dimension of the conductor,<br />
then, in dimensionless form,<br />
Approximately four months after Rayleigh’s<br />
paper appeared, Nature published an<br />
eight line comment (half column, yet!) by a<br />
D. Riabouchinsky pointing out that Rayleigh’s<br />
result assumed that temperature was a<br />
dimension independent from mass, length,<br />
and time. However, from the kinetic theory<br />
of gases we know that this is not so: temperature<br />
can be defined as the mean kinetic<br />
energy of the molecules and so is not an<br />
independent unit! Thus, according to<br />
Riabouchinsky, Rayleigh’s expression must<br />
be replaced by an expression with an additional<br />
dimensionless variable:<br />
a much less restrictive result.<br />
Two weeks later, Rayleigh responded to<br />
Riabouchinsky saying that “it would indeed<br />
be a paradox if thefurther knowledge of the<br />
nature of heat afforded by molecular theory<br />
put us in a worse position than before in<br />
dealing with a particular problem. . . . It<br />
would be well worthy of discussion.” Indeed<br />
it would; its resolution, which no doubt the<br />
reader has already discerned, is left as an<br />
exercise (for the time being)! Like all<br />
paradoxes, this one cautions us that we occasionally<br />
make casual assumptions without<br />
quite realizing that we have done so (see<br />
“Fundamental Constants and the Rayleigh-<br />
Riabouchinsky Paradox”).<br />
Scale Invariance<br />
Let us now turn our attention to a slightly<br />
more abstract mathematical formulation<br />
that clarifies the relationship of dimensional<br />
analysis to scale invariance. By scale invariance<br />
we simply mean that the structure<br />
of physical laws cannot depend on the choice<br />
of units. As already intimated, this is automatically<br />
accomplished simply by employing<br />
dimensionless variables since these<br />
clearly do not change when the system of<br />
units changes. However, it may not be immediately<br />
obvious that this is equivalent to<br />
the form invariance of physical equations.<br />
Since physical laws are usually expressed in<br />
terms of dimensional variables, this is an<br />
important point to consider: namely, what<br />
are the general constraints that follow from<br />
the requirement that the laws of physics look<br />
the same regardless of the chosen units. The<br />
crucial observation here is that implicit in<br />
any equation written in terms of dimensional<br />
variables are the “hidden” fundamental<br />
scales of mass M, length L, time T, and so<br />
forth that are relevant to the problem. Of<br />
course, one never actually makes these scale<br />
parameters explicit precisely because of form<br />
invariance.<br />
Our motivation for investigating this<br />
question is to develop a language that can be<br />
generalized in a natural way to include the<br />
subtleties of quantum field theory. Hopefully<br />
classical dimensional analysis and scaling<br />
will be sufficiently familiar that its generalization<br />
to the more complicated case will<br />
be relatively smooth! This generalization has<br />
been named the renormalization group since<br />
its origins b’e in the renormalization program<br />
used to ma f, e sense out of the infinities inherent<br />
in quantum field theory. It turns out<br />
that renormalization requires the introduction<br />
of a new arbitrary “hidden”. scale that<br />
plays a role similar to the role of the scale<br />
parameters implicit in any dimensional<br />
equation. Thus any equation derived in<br />
quantum field theory that represents a physical<br />
quantity must not depend upon this<br />
choice of hidden scale. The resulting con-<br />
straint will simply represent a generalization<br />
of ordinary dimensional analysis; the only<br />
reason that it is different is that variables in<br />
quantum field theory, such as fields, change<br />
in a much more complicated fashion with<br />
scale than do their classical counterparts.<br />
Nevertheless, just as dimensional analysis<br />
allows one to learn much about the behavior<br />
of a system without actually solving the<br />
dynamical equations, so the analogous constraints<br />
of the renormalization group lead to<br />
powerful conclusions about the behavior ofa<br />
quantum field theory without actually being<br />
able to solve it. It is for this reason that the<br />
renormalization group has played such an<br />
important part in the renaissance of quantum<br />
field theory during the past decade or so.<br />
Before describing how this comes about, I<br />
shall discuss the simpler and more familiar<br />
case of scale change in ordinary classical<br />
systems.<br />
To begin, consider some physical quantity<br />
F that has dimensions; it will, of course, be a<br />
function of various dimensional variables<br />
x,: F(xl,x2,. . .,x,,). An explicit example is<br />
given by Eq. 2 describing the temperature<br />
distribution in a cooked turkey or goose.<br />
11
Fundamental Constants and the<br />
L<br />
et us examine Riabouchinsky’s paradox a little more carefully<br />
and show how its resolution is related to choosing a system of<br />
units where the “fundamental constants” (such as Planck’s<br />
constant h and the speed of light c) can be set equal to unity.<br />
The paradox had to do with whether temperature could be used as<br />
an independent dimensional unit even though it can be defined as the<br />
mean kinetic energy of the molecular motion. Rayleigh had chosen<br />
five physical variables (length I, temperature difference 8, velocity v,<br />
specific heat C‘, and heat conductivity K) to describe Boussinesq’s<br />
problem and had assumed that there were four independent<br />
dimensions (energy E, length L, time T, and temperature 6). Thus<br />
the solution for T/T, necessarily is an arbitrary function of one<br />
dimensionless combination. To see this explicitly, let us examine the<br />
dimensions of the five physical variables:<br />
[h = L, [e] = e, = LT’, [q = EL-~B-’,<br />
and [K] = EL-’ TI@-’ .<br />
Clearly the combination chosen by Rayleigh, IvCIK, is dimensionless.<br />
Although other dimensionless combinations can be formed, they<br />
are not independent of the two combinations (1vCIK and TIT,)<br />
selected by Rayleigh.<br />
Now suppose, along with Riabouchinsky, we use our knowledge of<br />
the kinetic theory to define temperature “as the mean kinetic energy<br />
of the molecules” so that 6 is no longer an independent dimension.<br />
This means there are now only three independent dimensions and the<br />
solution will depend on an arbitrary function of two dimensionless<br />
combinations. With 6 a E, the dimensions of the physical variables<br />
become:<br />
[O= L, [e] = E, [v] = LT’, [CJ = LP3, and [K] = L-’T’.<br />
It is clear that, in addition to Rayleigh’s dimensionless variable, there<br />
is now a new independent combination, C13 for example, that is<br />
dimensionless. To reiterate Rayleigh: “it would indeed be a paradox<br />
if the further knowledge of the nature of heat . . . put us in a worse<br />
position than before . . . it would be well worthy of discussion.”<br />
Like almost all paradoxes, there is a bogus aspect to the argument.<br />
It is certainly true that the kinetic theory allows one to express an<br />
energy as a temperature. However, this is only useful and appropriate<br />
for situations where the physics is dominated by molecular considerations.<br />
For macroscopic situations such as Boussinesq’s problem, the<br />
molecular nature of the system is irrelevant; the microscopic<br />
variables have been replaced by macroscopic averages embodied in<br />
phenomenological properties such as the specific heat and conductivity.<br />
To make Riabouchinsky’s identification of energy with temperature<br />
is to introduce irrekvant physics into the problem.<br />
Exploring this further, we recall that such an energy-temperature<br />
identification implicitly involves the introduction of Boltzmann’s<br />
factor k. By its very nature, k will only play an explicit role in a<br />
physical problem that directly involves the molecular nature of the<br />
system; otherwise it will not enter. Thus one could describe the<br />
system from the molecular viewpoint (so that k is involved) and then<br />
take a macroscopic limit. Taking the limit is equivalent to setting<br />
k = 0 the presence of a finite k indicates that explicit effects due to<br />
the kinetic theory are important.<br />
With this in mind, we can return to Boussinesq’s problem and<br />
derive Riabouchinsky’s result in a somewhat more illuminating<br />
fashion. Let us follow Rayleigh and keep E, L, T, and 6 as the<br />
Each of these variables, including F itself, is<br />
always expressible in terms of some standard<br />
set of independent units, which can be<br />
chosen to be mass M, length L, and time T.<br />
These are the hidden scale parameters. Obviously,<br />
other combinations could be used.<br />
There could even be other independent<br />
units, such as temperature (but remember<br />
Riabouchinsky!), or more than one independent<br />
length (say, transverse and longitudinal).<br />
In this discussion, we shall simply<br />
use the conventional M, L, and T. Any<br />
generalization is straightforward.<br />
In terms of this standard set of units, the<br />
magnitude of each x, is given by<br />
x, = Mal L ~I TYI (15)<br />
The numbers a,, PI, and y, will be recognized<br />
12<br />
as “the dimensions” of xi. Now suppose we<br />
change the system of units by some scale<br />
transformation of the form<br />
A4 -+<br />
M‘ = hMM ,<br />
L- L’=hLL,<br />
and<br />
T +<br />
T‘ = hTT.<br />
Each variable then responds as follows:<br />
x, - XI’ = Z,(h)X, , (17)<br />
where<br />
and h is shorthand for h ~ h , ~ and , h ~. Since<br />
Fis itself a dimensional physical quantity, it<br />
transforms in an identical fashion under this<br />
scale change:<br />
where<br />
Here a, p, and y are the dimensions of F.<br />
There is, however, an alternate but equivalent<br />
way to transform from F to F‘, namely,<br />
by transforming each of the variables xi<br />
separately. Explicitly we therefore also have
Scale and Dimension<br />
adox<br />
independent dimensions but add k (with ons E@-’) as a new<br />
physical variable. The solution will now rbitrary function of<br />
two independent dimensionless variables: IvC/K and kCl? When<br />
Riabouchinsky chose to make C13 his other dimensionless variable,<br />
he, in effect, chose a system of units where k= 1. But that was a<br />
temble thing to do here since the physics dictates that k = O! Indeed,<br />
if k = 0 we regain Rayleigh’s original result, that is, we have only one<br />
dimensionless variable. It is somewhat ironic that Rayleigh’s remarks<br />
miss the point: “further knowledge oft<br />
molecular theory” does not put one in a<br />
the problem-rather, it leads to a micros<br />
C. The important point pertinent to the<br />
that knowledge of the molecular theory is irrelevant and k must not<br />
enter.<br />
The lesson here is an important one because it illustrates the role<br />
played by the fundamental constants. Consider Planck’s constant<br />
h h/2~ it would be completely inappropriate to introduce it into a<br />
problem of classical dynamics. For e any solution of the<br />
scattering of two billiard balls will dep acroscopic variables<br />
such as thc masses, velocities, friction coefficients, and so on. Since<br />
billiard balls are made of protons, it might be tempting to the purist<br />
to include as a dependent variable the proton-proton total cross<br />
section, which. of course. involves fi. This would clearly be totally<br />
inappropriate but is analogous to what Riabouchinsky did in<br />
Boussinesq’s problem.<br />
Obviously, if the scattering is between two microscopic “atomic<br />
billiard balls” then h must he included. In thiscase it is not only quite<br />
legitimate but often convenient to choose a system of units where<br />
h = I . However, having done so one cannot directly recover the<br />
g to h - 0. With h = 1, one is stuck in<br />
with k = I , one is stuck in kinetic theory.<br />
A similar situation obviously occurs in relativity: the velocity of<br />
light c must not occur in the classical Newtonian limit. However, in a<br />
relativistic situation one is quite at liberty to choose units where<br />
c= 1. Making that choice, though, presumes the physics involves<br />
relativity.<br />
The core of particle physics, relativistic quantum field theory, is a<br />
hanics and relativity. For this reason,<br />
a system of units in which h = c = 1 is<br />
manifesto that quantum mechanics and<br />
relativity are the basic physical laws governing their area of physics.<br />
In quantum mechanics, momentum p and wavelength k are related<br />
by the de Broglie relation: p= 27thlh; similarly, energy E and frequency<br />
w are related by Planck‘s formula: E = ha. In relativity we<br />
have the famous Einstein relation: E = mc?. Obviously if we choose<br />
= c = 1, all energies, masses, and momenta have the same units<br />
example, electron<br />
ev)), and these are the same as inverse<br />
r energies and momenta inevitab1.v<br />
correspond to shorter times and lengths.<br />
Using this choice of units automatically incorporates the profound<br />
physics of the uncertainty principle: to probe short space-time intervals<br />
one needs large energies. A useful number to remember is that<br />
IO-13 centimeter, or I fermi (fm), equals the reciprocal of 200 MeV.<br />
We then find that the electron mass (e 1/2 MeV) corresponds to a<br />
length of e 400 fm-its Compton wavelength. Or the 20 TeV<br />
(2 X lo’ MeV) typically proposed for a possible future facility<br />
corresponds to a length of lo-’* centimeter. This is the scale distance<br />
that such a machine will probe!<br />
Equating these two different ways ofeffecting<br />
a scale change leads to the identity<br />
of taking a/& and then setting h = 1. For and x[ = xi, so that Eq. 23 reduces to<br />
example, if we were to consider changes in<br />
the mass scale, we would use a/ahM and the<br />
chain rule for partial differentiation to amve<br />
at ax, 8x2 ax3<br />
az, aF<br />
az<br />
Ex;- 1=- F.<br />
r=l a7\M axi dhM<br />
aF aF aF<br />
alxl - + azxz - + a3x3 - + . . .<br />
aF<br />
+ a,x, - = aF.<br />
ax”<br />
As a concrete example, consider the equation<br />
E = mc?. To change scale one can either<br />
transform E directly or transform m and c<br />
separately and multiply the results appropriately-obviously<br />
the final result must<br />
be the same.<br />
We now want to ensure that the resulting<br />
form of the equation does not depend on h.<br />
This is best accomplished using Euler’s trick<br />
When we set AM = I , differentiation of Eqs.<br />
18 and 20 yields<br />
Obviously this can be repeated with hL<br />
and AT to obtain a set of three coupled partial<br />
differential equations expressing the fundamental<br />
scale invariance ofphysical laws (that<br />
is, the invariance of the physics to the choice<br />
of units) implicit in Fourier’s original work.<br />
These equations can be solved without too<br />
much difficulty; their solution is, in fact, a<br />
speciai case of the solution to the re-<br />
13
normalization group equation (given explicitly<br />
as Eq. 35 below). Not too surprisingly,<br />
one finds that the solution is precisely<br />
equivalent to the constraints of dimensional<br />
analysis. Thus there is never any explicit<br />
need to use these rather cumbersome equations:<br />
ordinary dimensional analysis takes<br />
care of it for you!<br />
Quantum Field Theory<br />
We have gone through this little mathematical<br />
exercise to illustrate the well-known<br />
relationship of dimensional analysis to scale<br />
and form invariance. I now want to discuss<br />
how the formalism must be amended when<br />
applied to quantum field theory and give a<br />
sense of the profound consequences that follow.<br />
Using the above chain of reasoning as a<br />
guide, I shall examine the response of a<br />
quantum field theoretic system to a change<br />
in scale and derive a partial differential equation<br />
analogous to Eq. 25. This equation is<br />
known as the renormalization group equation<br />
since its origins lay in the somewhat<br />
arcane area of the renormalization procedure<br />
used to tame the infinities of quantum field<br />
theory. I shall therefore have to digress<br />
momentarily to give a brief rCsumC of this<br />
subject before returning to the question of<br />
scale change.<br />
Renormalization. Perhaps the most unnerving<br />
characteristic of quantum field theory for<br />
the beginning student (and possibly also for<br />
the wise old men) is that almost all calculations<br />
of its physical consequences naively<br />
lead to infinite answers. These infinities stem<br />
from divergences at high momenta associated<br />
with virtual processes that are<br />
always present in any transition amplitude.<br />
The renormalization scheme, developed by<br />
Richard 1’. Feynman, Julian S. Schwinger,<br />
Sin-Itiro Tomonaga, and Freeman Dyson,<br />
was invented to make sense out of this for<br />
quantum electrodynamics (QED).<br />
To get a feel for how this works I shall<br />
focus on the photon, which cames the force<br />
associated with the electromagnetic field. At<br />
the classical limit the propagator* for the<br />
14<br />
photon represents the usual static I/r<br />
Coulomb potential. The corresponding<br />
Fourier transform (that is, the propagator’s<br />
representation in momentum space) in this<br />
limit is I/$, where q is the momentum carried<br />
by the photon. Now consider the<br />
“classical” scattering of two charged particles<br />
(represented by the Feynman diagram in Fig.<br />
8 (a)). For this event the exchange of a single<br />
photon gives a transition amplitude proportional<br />
to &/$, where eo is the charge (or<br />
coupling constant) occurring in the Lagrangian.<br />
A standard calculation results in<br />
the classical Rutherford formula, which can<br />
be extended relativistically to the spin-1/2<br />
case embodied in the diagram.<br />
A typical quantum-mechanical correction<br />
to the scattering formula is illustrated in Fig.<br />
8 (b). The exchanged photon can, by virtue of<br />
the uncertainty principle, create for a very<br />
short time a virtual electron-positron pair,<br />
which is represented in the diagram by the<br />
loop. We shall use k to denote the momentum<br />
carried around the loop by the two<br />
particles.<br />
There are, of course, many such corrections<br />
that serve to modify the I/$ single-<br />
*Roughly speaking, the photon propagator can be<br />
thought of as the Green’s function for the electromagnetic<br />
field. In the relativistically covariant<br />
Lorentz gauge, the classical Maxwell’s equations<br />
read<br />
0’ A(X) = j(x),<br />
where A(x) is the vector potential and j(x) is the<br />
current source term derived in QED from the motion<br />
of the electrons. (To keep things simple I am<br />
suppressing all space-time indices, thereby ignoring<br />
spin.) This equation can be solved in the standard<br />
way using a Green’s function:<br />
A(x) = Id‘x’<br />
with<br />
0 2 G(x) = 6(x) .<br />
G(x’ -x) j(x‘),<br />
Now a transition amplitude is proportional to the<br />
interaction energy, and this is given by<br />
HI = Id4 x j(x) A(x) =<br />
Id‘ x Id‘x’ j(x) G(x-x‘) j(x‘) ,<br />
photon behavior, and this is represented<br />
schematically in part (c). It is convenient to<br />
include all these corrections in a single multiplicative<br />
factor DO that represents deviations<br />
from the single-photon term. The “full”<br />
photon propagator including all possible<br />
radiative corrections is therefore Do/$. The<br />
reason for doing this is that DO is a<br />
dimensionless function that gives a measure<br />
of the polarization of the vacuum caused by<br />
the production of virtual particles. (The origin<br />
of the Lamb shift is vacuum polarization.)<br />
We now come to the central problem:<br />
upon evaluation it is found that contributions<br />
from diagrams like (b) are infinite because<br />
there is no restriction on the magnitude<br />
of the momentum k flowing in the loop!<br />
Thus, typical calculations lead to integrals of<br />
the form<br />
(26)<br />
which diverge logarithmically. Several<br />
prescriptions have been invented for making<br />
such integrals finite; they all involve “reg-<br />
illustrating how G “mediates” the force between<br />
two currents separated by a space-time interval<br />
(x-x’). It is usually more convenient to work with<br />
Fourier transforms of these quantities (that is, in<br />
momentum space). For example, the momentum<br />
space solution for G is C(q) = I/q2, and this is<br />
usually called the free photon propagator since it<br />
is essentially classical. The corresponding<br />
“classical” transition amplitude in momentum<br />
space is just j(q)(I/q’Jj(q), which is represented<br />
by the Feynman graph in Fig. 8 (a).<br />
In quantum field theory, life gets much more<br />
complicated because of radiative corrections as<br />
discussed in the text and illustrated in (b) and (c)<br />
of Fig. 8. The definition of the propagator is<br />
generally in terms of a correlation function in<br />
which a photon is created at point x out of the<br />
vacuum for a period x-x‘ and then returns to the<br />
vacuum’at point x’. Symbolically, this is represented<br />
by<br />
G(x-x’) - (vaclA(x’) A(x)lvac).<br />
During propagation, anything allowed by the<br />
uncertainty principle can happen-these are the<br />
radiative corrections that moke an exact calculation<br />
of G almost impossible.
Scale and Dimension<br />
ularizing” the integrals by introducing some<br />
large mass parameter A. A standard technique<br />
is the so-called Pauli-Villars scheme in<br />
which a factor A2/(k2+A2) is introduced<br />
into the integrand with the understanding<br />
that A is to be taken to infinity at the end of<br />
the calculation (notice that in this limit the<br />
regulating factor approaches one). With this<br />
prescription, the above integral is therefore<br />
replaced by<br />
A2<br />
= In -<br />
aq2<br />
The integral can now be evaluated and its<br />
divergence expressed in terms of the (infinite)<br />
mass parameter A. All the infinities<br />
arising from quantum fluctuations can be<br />
dealt with in a similar fashion with the result<br />
that the following series is generated:<br />
In this way the structure of the infinite<br />
divergences in the theory are parameterized<br />
in terms of A, which can serve as a finite<br />
cutoffin the integrals over virtual momenta.*<br />
The remarkable triumph of the renormalization<br />
program is that, rather than<br />
imposing such an arbitrary cutoff, all these<br />
divergences can be swallowed up by an infinite<br />
rescalingof the fields and coupling con-<br />
Fig. 8. Feynman diagrams for (a) the ChZSSiCal scattering Of two particles Of<br />
charge e,, (b) a typical correction that must be made to that scattering-here<br />
because of the creation of a virtual electron-positron pair-and (c) a diagram<br />
representing all such possible corrections. The matrix element is proportional for<br />
(a) to e:/q2 and for (c) to D,/q2 where D,.includes all corrections.<br />
*In this discussion I assumed, for simplicity,<br />
that the original Lagrangian was massless; that<br />
~ ~ ~ ~ t ~<br />
plicate the discussion unnecessarily without giving<br />
any new insights.<br />
15
stants. Thus, afinite propagator D, that does<br />
not depend on A, can be derived from DO by<br />
rescaling if, at the same time, one rescales the<br />
charge similarly. These rescalings take the<br />
form<br />
D = Zflo and e = Zgo . (29)<br />
The crucial property of these scaling factors<br />
is that they are independent of the physical<br />
momenta (such as q) but depend on A in<br />
such a way that when the cutoff is removed,<br />
D and e remain finite. In other words, when<br />
A - m, ZD and Z, must develop infinities of<br />
their own that precisely compensate for the<br />
infinities of Do and eo. The original so-called<br />
bare parameters in the theory calculated<br />
from the Lagrangian (DO and eo) therefore<br />
have no physical meaning-only the renormalized<br />
parameters (D and e) do.<br />
Now let us apply some ordinary dimensional<br />
analysis to these remarks. Because<br />
they are simply scale factors, the 2’s must be<br />
dimensionless. However, the 2’s are functions<br />
of ,4 but not of q. But that is very<br />
peculiar: a dimensionless function cannot<br />
depend on a single mass parameter! Thus, in<br />
order to express the Z’s in dimensionless<br />
form, a new jnite mass scale p must be<br />
introduced so that one can write<br />
Z = Z(A2/p2,eo). An immediate consequence<br />
of renormalization is therefore to induce a<br />
mass scale not maniyest in the Lagrangian.<br />
This is extremely interesting because it<br />
provides a possible mechanism for generating<br />
mass even though no mass parameter<br />
appears in the Lagrangian. We therefore<br />
have the exciting possibility of being able to<br />
calculate the masses of all the elementary<br />
particles in terms ofjust oneofthem. Similar<br />
considerations for the dimensionless Ds<br />
clearly require that they be expressible as<br />
DO = Do(q2/A2,eo), as in Eq. 28, and<br />
D = D(q2/p2,e). (The dream of particle<br />
theorists is to write down a Lagrangian with<br />
no mass parameter that describes all the<br />
interations in terms ofjust onecoupling constant.<br />
The mass spectrum and scattering<br />
amplitudes for all the elementary particles<br />
16<br />
would then be calculable in terms of the<br />
value of this single coupling at some given<br />
scale! A wonderful fantasy.)<br />
To recapitulate, the physical finite renormalized<br />
propagator 12 is related to its bare<br />
and divergent counterpart DO (calculated<br />
from the Lagrangian using a cutoff mass) by<br />
an infinite rescaling:<br />
Similarly, the physical finite charge e is given<br />
by an infinite rescaling of the bare charge eo<br />
that occurs in the Lagrangian<br />
e = lim zc( $ , eo) eo<br />
A-n.<br />
Notice that the physical coupling e now depends<br />
implicitly on the renormalization<br />
scale parameter p. Thus, in QED, for example,<br />
it is not strictly sufficient to state that the<br />
fine structure constant a =: 11137; rather,<br />
one must also specifj the corresponding<br />
scale. From this point of view there is<br />
nothing magic about the particular number<br />
137 since a change of scale would produce a<br />
different value.<br />
At this stage, some words of consolation to<br />
a possibly bewildered reader are in order. It is<br />
not intended to be obvious how such infinite<br />
rescalings of infinite complex objects lead to<br />
consistent finite results! An obvious question<br />
is what happens with more complicated<br />
processes such as scattering amplitudes and<br />
particle production? These are surely even<br />
more divergent than the relatively simple<br />
photon propagator. How does one know that<br />
a similar rescaling procedure can be carried<br />
through in the general case?<br />
The proof that such a procedure does indeed<br />
work consistently for any transition<br />
amplitude in the theory was a real tour de<br />
force. A crucial aspect of this proof was the<br />
remarkable discovery that in QED only a<br />
finite number (three) of such rescalings was<br />
necessary to render the theory finite. This is<br />
tembly important because it means that<br />
once we have renormalized a few basic entities,<br />
such as eo, all further rescalings of<br />
more complicated quantities are completely<br />
determined. Thus, the theory retains predictive<br />
power-in marked contrast to the highly<br />
unsuitable scenario in which each transition<br />
amplitude would require its own infinite<br />
rescaling to render it finite. Such theories,<br />
termed nonrenormalizable, would apparently<br />
have no predictive power. High<br />
energy physicists have, by and large, restricted<br />
their attention to renormalizable theories<br />
just because all their consequences can, in<br />
principle, be calculated and predicted in<br />
terms of just a few parameters (such as the<br />
physical charge and some masses).<br />
I should emphasize the phrase “in principle”<br />
since in practice there are very few<br />
techniques available for actually carrying out<br />
honest calculations. The most prominent of<br />
these is perturbation theory in the guise of<br />
Feynman graphs. Most recently a great deal<br />
of effort, spurred by the work of K. G.<br />
Wilson, has gone into trying to adapt quantum<br />
field theory to the computer using lattice<br />
gauge theories.* In spite of this it remains<br />
sadly true that perturbation theory is our<br />
only “global” calculational technique. Certainly<br />
its success in QED has been nothing<br />
less than phenomenal.<br />
Actually only a very small class of renormalizable<br />
theories exist and these are<br />
characterized by dimensionless coupling<br />
constants. Within this class are gauge theories<br />
like QED and its non-Abelian extension<br />
in which the photon interacts with<br />
itself. All modern particle physics is based<br />
upon such theories. One of the main reasons<br />
for their popularity, besides the fact they are<br />
renormalizable, is that they possess the property<br />
of being asymptotically free. In such<br />
theories one finds that the renormalization<br />
group constraint, to be discussed shortly,<br />
requires that the large momentum behavior<br />
*In recent years there has been some effort to<br />
come to grips analytically with the<br />
nonperturbative aspects of gauge theories.
Scale and Dimension<br />
be equivalent to the small coupling limit;<br />
thus for large momenta the renormalized<br />
coupling effectively vanishes thereby allowing<br />
the use of perturbation theory to calculate<br />
physical processes.<br />
This idea was of paramount importance in<br />
substantiating the existence of quarks from<br />
deep inelastic electron scattering experiments.<br />
In these experiments quarks behaved<br />
as if they were quasi-free even though they<br />
must be bound with very strong forces (since<br />
they are never observed as free particles).<br />
Asymptotic freedom gives a perfect explanation<br />
for this: the effective coupling, though<br />
strong at low energies, gets vanishingly small<br />
as 4’ becomes large (or equivalently, as distance<br />
becomes small).<br />
In seeing how this comes about we will be<br />
led back to the question of how the field<br />
theory responds to scale change. We shall<br />
follow the exact same procedure used in the<br />
classical case: first we scale the hidden parameter<br />
(p, in this case) and see how a typical<br />
transition amplitude, such as a propagator,<br />
responds. A partial differential equation,<br />
analogous to Eq. 25, is then derived using<br />
Euler’s trick. This is solved to yield the general<br />
constraints due to renormalization<br />
analogous to the constraints of dimensional<br />
analysis. I will then show how these constraints<br />
can be exploited, using asymptotic<br />
freedom as an example.<br />
The Renormalization Group Equation. As<br />
already mentioned, renormalization makes<br />
the bare parameters occumng in the Lagrangian<br />
effectively irrelevant; the theory has<br />
been transformed into one that is now specified<br />
by the value of its physical coupling<br />
constants at some mass scale p. In this sense<br />
p plays the role of the hidden scale parameter<br />
M in ordinary dimensional analysis by setting<br />
the scale of units by which all quantities<br />
are measured.<br />
This analogy can be made almost exact by<br />
considering a scale change for the arbitrary<br />
parameter p in which p - h’/*p. This change<br />
allows us to rewrite Eq. 30 in a form that<br />
expresses the response of D to a scale change:<br />
(From now on I will use g to denote the<br />
coupling rather than e because e is usually<br />
reserved for the electric charge in QED.)<br />
The scale factor Z(h), which is independent<br />
of q2 and g, must, unlike the Zs of Eqs.<br />
30 and 3 I , befinite since it relates two finite<br />
quantities. Notice that all explicit reference<br />
to the bare quantities has now been<br />
eliminated. The structure of this equation is<br />
identical to Eq. 22, the scaling equation derived<br />
for the classical case; the crucial dij<br />
ference is that Z(h) no longer has the simple<br />
power law behavior expressed in Eq. 18. In<br />
fact, the general structure of Z(1) and g(p) are<br />
not known in field theories of interest.<br />
Nevertheless we can still learn much by converting<br />
this equation to the differential form<br />
analogous to Eq. 25 that expresses scale invariance.<br />
As before we simply take a/ah and<br />
set h = I, thereby deriving the so-called renormalization<br />
group equation:<br />
where<br />
and<br />
(33)<br />
(34)<br />
(35)<br />
Comparing Eq. 33 with the scaling equation<br />
of classical dimensional analysis (Eq. 25), we<br />
see that the role ofthe dimension is played by<br />
y. For this reason, and to distinguish it from<br />
ordinary dimensions, y is usually called the<br />
anomalous dimension of D, a phrase originally<br />
coined by Wilson. (We say anomalous<br />
because, in terms of ordinary dimensions<br />
and again by analogy with Eq. 25, D is actually<br />
dimensionless!) It would similarly have<br />
been natural to call p(g)/g the anomalous<br />
dimension of g, however, conventionally,<br />
one simply refers to p(g) as the p-function.<br />
Notice that p(g) characterizes the theory as a<br />
whole (as does g itself since it represents the<br />
coupling) whereas y(g) is a property of the<br />
particular object or field one is examining.<br />
The general solution of the renormalization<br />
group equation (Eq. 33) is given by<br />
D (f,g) = eA(g)j( f , (36)<br />
where<br />
and<br />
(37)<br />
The arbitrary function f is, in principle, fixed<br />
by imposing suitable boundary conditions.<br />
(Equation 25 can be viewed as a special and<br />
rather simple case of Eq. 33. If this is done,<br />
17
the analogues of y(g) and p(g)/g are constants,<br />
resulting in trivial integrals for A and<br />
K. One can then straightforwardly use this<br />
general solution (Eq. 36) to verify the claim<br />
that the scaling equation (Eq. 22) is indeed<br />
exactly equivalent to using ordinary dimensional<br />
analysis.) The general solution reveals<br />
what is perhaps the most profound consequence<br />
of the renormalization group,<br />
namely, that in quantum field theory the<br />
momentum variables and the coupling constant<br />
are inextricably linked. The photon<br />
propagator (D/q2), for instance, appears at<br />
first sight to depend separately on the<br />
momentum 4’ and the coupling constant g.<br />
Actually, however, the renormalizability of<br />
the theory constrains it to depend effectively,<br />
as shown in Eq. 36, on only one variable<br />
($dC(g)/bi2). This, of course, is exactly what<br />
happens in ordinary dimensional analysis.<br />
For example, recall the turkey cooking problem.<br />
The temperature distribution at first<br />
sight depended on several different variables:<br />
however, scale invariance, in the guise of<br />
dimensional analysis, quickly showed that<br />
there was in fact only a single relevant<br />
variable.<br />
The observation that renormalization introduces<br />
an arbitrary mass scale upon which<br />
no physical consequences must depend was<br />
first made in 1953 by E. Stueckelberg and A.<br />
Peterman. Shortly thereafter Murray Gell-<br />
Mann and F. Low attempted to exploit this<br />
idea to understand the high-energy structure<br />
of QED and, in so doing, exposed the intimate<br />
connection between g and $. Not<br />
much use was made of these general ideas<br />
until the pioneering work of Wilson in the<br />
late 1960s. I shall not review here his seminal<br />
work on phase transitions but simply remark<br />
that the scaling constraint implicit in the<br />
renormalization group can be applied to correlation<br />
functions to learn about critical exponents.*<br />
Instead I shall concentrate on the<br />
*Since the photon propagator is defined as the<br />
correlation function of two electromagnetic<br />
fields in the vacuum it is not diffielt to imagine<br />
that the formalism discussed here can be directly<br />
applied to the correlation functions of statistical<br />
physics.<br />
18<br />
particle physics successes, including<br />
Wilson’s, that led to the discovery that non-<br />
Abelian gauge theories were asymptotically<br />
free. Although the foci of particle and condensed<br />
matter physics are quite different,<br />
they become unified in a spectacular way<br />
through the language of field theory and the<br />
renormalization group. The analogy with dimensional<br />
analysis is a good one, for, as we<br />
saw in the first part of this article, its constraints<br />
can be applied to completely diverse<br />
problems to give powerful and insightful results.<br />
In a similar fashion, the renormalization<br />
group can be applied to any problem<br />
that can be expressed as a field theory (such<br />
as particle physics or statistical physics).<br />
Often in physics, progress is made by examining<br />
the system in some asymptotic regime<br />
where the underlying dynamics<br />
simplifies sufficiently for the general structure<br />
to become transparent. With luck,<br />
having understood the system in some extreme<br />
region, one can work backwards into<br />
the murky regions of the problem to understand<br />
its more complex structures. This is<br />
essentially the philosophy behind bigger and<br />
bigger accelerators: k.eep pushing to higher<br />
energies in the hope that the problem will<br />
crack, revealing itself in all its beauty and<br />
simplicity. ’Tis indeed a faithful quest for the<br />
holy grail. As I shall now demonstrate, the<br />
paradigm of looking first for simplicity in<br />
asymptotic regimes is strongly supported by<br />
the methodology of the renormalization<br />
group.<br />
In essence, we use the same modelingtheory<br />
scaling technique used by ship designers.<br />
Going back to Eq. 36, one can see<br />
immediately that the high-energy or shortdistance<br />
limit ($- m with g fixed) is iden-<br />
tical to keeping 4’ fixed while taking K +<br />
m.<br />
However, from its definition (Eq. 38), K<br />
diverges whenever p(g) has a zero. Similarly,<br />
the low-energy or long-distance limit (42 - 0<br />
while g is fixed) is equivalent to K +<br />
-m,<br />
which also occurs when p - 0. Thus knowledge<br />
of the zeros of 0, the so-called fixed<br />
points of the equation, determines the highand<br />
low-energy behaviors of the theory.<br />
If one assumes that for small coupling<br />
quantum field theory is governed by ordinary<br />
perturbation theory, then the p-function<br />
has a zero at zero coupling (g- 0). In<br />
this limit one typically finds p(g) = -b$<br />
where b is a calculable coefficient. Of course,<br />
p might have other zeroes, but, in general,<br />
this is unknown. In any case, for small g we<br />
find (using Eq. 38) that K(g) = (2b$)-’,<br />
which diverges to either +m or -4, depending<br />
on the sign of b. In QED, the case originally<br />
studied by Cell-Mann and Low, b < 0 so that<br />
K - -m, which is equivalent to the lowenergy<br />
limit. One can think of this as an<br />
explanation of why perturbation theory<br />
works so well in the low-energy regime of<br />
QED: the smaller the energy, the smaller the<br />
effective coupling constant.<br />
Quantum Chromodynamics. It appears that<br />
some non-Abelian gauge theories and, in<br />
particular, QCD (see “<strong>Particle</strong> <strong>Physics</strong> and<br />
the Standard Model”) possess the unique<br />
property of having a positive b. This<br />
marvelous observation was first made by H.<br />
D. Politzer and independently by D. J. Gross<br />
and F. A. Wilczek in 1973 and was crucial in<br />
understanding the behavior of quarks in the<br />
famous deep inelastic scattering experiments<br />
at the Stanford Linear Accelerator Center. As<br />
a result, it promoted QCD to the star position<br />
of being a member of “the standard<br />
model.” With b > 0 the high-energy limit is<br />
I
Scale and Dimension<br />
related to perturbation theory and is therefore<br />
calculable and understandable. I shall<br />
now give an explicit example of how this<br />
comes about.<br />
First we note that no boundary conditions<br />
have yet been imposed on the general solution<br />
(Eq. 36). The one boundary condition<br />
that must be imposed is the known free field<br />
theory limit (g= 0). For the photon in QED,<br />
or the gluon in QCD, the propagator G<br />
(=D/$) in this limit is just I/$. Thus<br />
D($/p2,0) = 1. Imposing this on Eq. 36 gives<br />
Now when g - 0, y(g) = -a$, where a is a<br />
calculable coeficient. Combining this with<br />
the fact that B(g) = -62 leads, by way of Eq.<br />
37, to A(g) = (a/@ In g. Since K(g) =<br />
(26$)-‘, the boundary condition (Eq. 39)<br />
gives<br />
Defining the dimensionless variable in the<br />
functionfas<br />
it can be shown that with b > 0 Eq. 40 is<br />
equivalent to<br />
lim fix) = (26 In x)alZb. (42)<br />
X-m<br />
An important point here is that the x - m<br />
limit can be reached either by lettingg- 0 or<br />
by taking 42 - m<br />
calculable, so is the 2 - m<br />
. Since the g - 0 limit is<br />
limit. The free<br />
field (g- 0) boundary condition therefore<br />
determines the large x behavior offlx), and,<br />
once again, the “modeling technique” can be<br />
used-here to determine the large q2<br />
behavior of the propagator G.<br />
In fact, combining Eq. 36 with Eq. 42 leads<br />
to the conclusion that<br />
(43)<br />
This is the generic structure that finally<br />
emerges: the high-energy or large-$ behavior<br />
of the propagator G = D/$ is given by free<br />
field theory (I/$) modulated by calculable<br />
powers of logarithms. The wonderful miracle<br />
that has happened is that all the’powers of<br />
ln(A2/$) originally generated from the<br />
divergences in the “bare” theory (as illustrated<br />
by the series in Eq. 28) have been<br />
summed by the renormalization group to<br />
give the simple expression of Eq. 43. The<br />
amazing thing about this “exact” result is<br />
that is is far easier to calculate than having to<br />
sum an infinite number of individual terms<br />
in a series. Not only does the methodology<br />
do the summing, but, more important, it<br />
justifies it!<br />
I have already mentioned that asymptotic<br />
freedom (that is, the equivalence of vanishingly<br />
small coupling with increasing<br />
momentum) provides a natural explanation<br />
of the apparent paradox that quarks could<br />
appear free in high-energy experiments even<br />
though they could not be isolated in the<br />
laboratory. Furthermore, with lepton probes,<br />
where the theoretical analysis is least ambiguous,<br />
the predicted logarithmic modulation<br />
of free-field theory expressed in Eq. 43<br />
has, in fact, been brilliantly verified. Indeed,<br />
this was the main reason that QCD was<br />
accepted as the standard model for the strong<br />
interactions.<br />
There is, however, an even more profound<br />
consequence of the application of the renormalization<br />
group to the standard model<br />
that leads to interesting speculations con-<br />
cerning unified field theories. As discussed in<br />
“<strong>Particle</strong> <strong>Physics</strong> and the Standard Model,”<br />
QED and the weak interactions are partially<br />
unified into the electroweak theory. Both of<br />
these have a negative b and so are not<br />
asymptotically free; their effective couplings<br />
grow with energy rather than decrease. By the<br />
same token, the QCD coupling should grow<br />
as the energy decreases, ultimately leading to<br />
the confinement of quarks. Thus as energy<br />
increases, the two small electroweak couplings<br />
grow and the relatively large QCD<br />
coupling decreases. In 1974, Georgi, Quinn,<br />
and Weinberg made the remarkable observation<br />
that all fhree couplings eventually became<br />
equal at an energy scale of about IOl4<br />
GeV! The reason that this energy turns out to<br />
be so large is simply due to the very slow<br />
logarithmic variation of the couplings. This<br />
is a very suggestive result because it is extremely<br />
tempting to conjecture that beyond<br />
IOl4 GeV (that is, at distances below<br />
cm) all three interactions become unified<br />
and are governed by the same single coupling.<br />
Thus, the strong, weak, and electromagnetic<br />
forces, which at low energies<br />
appear quite disparate, may actually be<br />
manifestations of the same field theory. The<br />
search for such a unified field theory (and its<br />
possible extension to gravity) is certainly one<br />
of the central themes of present-day particle<br />
physics. It has proven to be a very exciting<br />
but frustrating quest that has sparked the<br />
imagination of many physicists. Such ideas<br />
are, of course, the legacy of Einstein, who<br />
devoted the last twenty years of his life to the<br />
search for a unified field theory. May his<br />
dreams become reality! On this note of fantasy<br />
and hope we end our brief discourse<br />
about the role of scale and dimension in<br />
understanding the world-or even the universe-around<br />
us. The seemingly innocuous<br />
investigations into the size and scale of<br />
animals, ships, and buildings that started<br />
with Galileo have led us, via some minor<br />
diversions, into baked turkey, incubating<br />
eggs, old bones, and the obscure infinities of<br />
Feynman diagrams to the ultimate question<br />
of unified field theories. Indeed, similitudes<br />
have been used and visions multiplied. H<br />
19
Geoffrey B. West was born in the county town of Taunton in Somerset,<br />
England. He received his B.A. from Cambridge University in 1961 and<br />
his Ph.D. from Stanford in 1966. His thesis, under the aegis of Leonard<br />
Schiff, dealt mostly with the electromagnetic interaction, an interest he<br />
has sustained throughout his career. He was a postdoctoral fellow at<br />
Cornell and Harvard before returning to Stanford in 1970 as a faculty<br />
member. He came to <strong>Los</strong> <strong>Alamos</strong> in 1974 as Leader of what was then<br />
called the High-Energy <strong>Physics</strong> Group in the Theoretical Division, a<br />
position he held until I98 I when he was made a Laboratory Fellow. His<br />
present interests revolve around the structure and consistency of quantum<br />
field theory and, in particular, its relevance to quantum<br />
chromodynamics and unified field theories. He has served on several<br />
advisory panels and as a member of the executive committee of the<br />
Division of <strong>Particle</strong>s and Fields of the American Physical Society.<br />
Further Reading<br />
The following are books on the classical application of dimensional analysis:<br />
Percy Williams Bridgman. Dimensional Analysis. New Haven: Yale University Press, 1963.<br />
Leonid lvanovich Sedov. Similarity and Dimensional Methods in Mechanics. New York: Academic Press,<br />
1959.<br />
Garrett Birkhoff. Hydrodynamics: A Study in Logic, Fact and Similitude. Princeton: Princeton University<br />
Press, 1960.<br />
DArcy Wentworth Thompson. On Growth and Form. Cambridge: Cambridge University Press, 1.917. This<br />
book is, in some respects, comparable to Galileo’s and should be required reading for all budding young<br />
scientists.<br />
Benoit B. Mandelbrot. The Fractal GeometryofNafure. New York: W. H. Freeman, 1983. This recent, very<br />
interesting book represents a modern evolution ofthe subject into the area of fractals; in principle, the book<br />
deals with related problems, though I find it somewhat obscure in spite of its very appealing format.<br />
Examples of classical scaling were drawn from the following:<br />
Thomas McMahon. “Size and Shape in Biology.” Science 179( 1973):1201-1204.<br />
Hermann Rahn, Amos Ar, and Charles V. Paganelli. “How Bird Eggs Breathe.” Scientific American<br />
24O(February 1979):46-55.<br />
20
Scale and Dimension<br />
Thomas A. McMahon. “Rowing: a Similarity Analysis.” Science 173( 1971):349-351.<br />
David Pilbeam and Stephen Jay Could. “Size and Scaling in Human Evolution.” Science 186(1974):<br />
892-901.<br />
The Rayleigh-Riabouchinsky exchange is to be found in:<br />
Rayleigh. “The Principle of Similitude.” Nature 95( 19 I5):66-68.<br />
D. Riabouchinsky. “Letters to Editor.” Nature 95( 191 5):591.<br />
Rayleigh. “Letters to Editor.” Nature 95( 19 I5):644.<br />
Books on quantum electrodynamics (QED) include:<br />
Julian Schwinger, editor. Selected Papers on Quanfurn Elecfrodynarnics. New York Dover, 1958. This book<br />
gives a historical perspective and general review.<br />
James D. Bjorken and Sidney D. Drell. Relativistic Quantum Mechanics. New York: McGraw-Hill, 1964.<br />
N. N. Bogoliubov and D. V. Shirkov. Introduction to the Theory of Quantized Fields. New York:<br />
Interscience, 1959.<br />
H. David Politzer. “Asymptotic Freedom: An Approach to Strong Interactions.” <strong>Physics</strong> Reports<br />
14( 1974): 129- 180. This and the previous reference include a technical review of the renormalization group.<br />
Claudio Rebbi. “The Lattice Theory of Quark Confinement.” Scientific American 248(February<br />
1983):54-65. This reference is also a nontechnical review of lattice gauge theories.<br />
For a review of the deep inelastic electron scattering experiments see:<br />
Henry W. Kendall and Wolfgang K. H. Panofsky. “The Structure of the Proton and the Neutron.” Scientific<br />
American 224(June 1971):60-76.<br />
Geoffrey B. West. “Electron Scattering from Atoms, Nuclei and Nucleons.” <strong>Physics</strong> Reports<br />
18( 1975):263-323.<br />
References dealing with detailed aspects of renormalization and its consequences are:<br />
Kenneth G. Wilson. “Non-Lagrangian Models of Current Algebra.” Physical Review l79( 1969): 1499-1 512.<br />
Geoffrey B. West. “Asymptotic Freedom and the Infrared Problem: A Novel Solution to the Renormalization-Group<br />
Equations.” Physical Review D 27( 1983): 1402-1405.<br />
E. C. G. Stueckelberg and A. Petermann. “La Normalisation des Constantes dans la Theorie des Quanta.”<br />
Helveticu Physica Acta 26( I953):499-520.<br />
M. Cell-Mann and F. E. Low. “Quantum Electrodynamics at Small Distances.” Physical Review<br />
95( 1954): 1300- I3 12.<br />
H. David Politzer. “Reliable Perturbative Results for Strong Interactions?’ Physical Review Letters<br />
30( 1973): 1346- 1349.<br />
David J. Gross and Frank Wilczck. “Ultraviolet Behavior of Non-Abelian Gauge Theories.” Physical<br />
Review Letters 30( 1973): 1343-1 346.<br />
21
c<br />
<strong>Particle</strong> <strong>Physics</strong><br />
and the<br />
Standard Model<br />
by Stuart A. Raby, Richard C. Slansky, and Geoffrey B. West<br />
U<br />
ntil the 1930s all natural phenomena were<br />
presumed to have their origin in just two<br />
basic forces-gravitation and electromagnetism.<br />
Both were described by classical<br />
fields that permeated all space. These fields extended<br />
out to infinity from well-defined sources,<br />
mass in the one case and electric charge in the<br />
other. Their benign rule over the physical universe<br />
seemed securely established.<br />
As atomic and subatomic phenomena were explored,<br />
it became apparent that two completely<br />
novel forces had to be added to the list; they were<br />
dubbed the weak and the strong. The strong force<br />
was necessary in order to understand how the<br />
nucleus is held together: protons bound together in<br />
a tight nuclear ball ( centimeter across) must<br />
be subject to a force much stronger than electromagentism<br />
to prevent their flying apart. The<br />
weak force was invoked to understand the transmutation<br />
of a neutron in the nucleus into a proton<br />
during the particularly slow form of radioactive<br />
decay known as beta decay.<br />
Since neither the weak force nor the strong force<br />
is directly observed in the macroscopic world,<br />
both must be very short-range relative to the more<br />
familiar gravitational and electromagnetic forces.<br />
Furthermore, the relative strengths of the forces<br />
associated with all four interactions are quite different,<br />
as can be seen in Table 1. It is therefore not<br />
too surprising that for a very long period these<br />
interactions were thought to be quite separate. In<br />
spite of this, there has always been a lingering<br />
suspicion (and hope) that in some miraculous<br />
fashion all four were simply manifestations of one<br />
source or principle and could therefore be described<br />
by a single unified field theory.<br />
The color force among quarks and gluons is described by a generalization of the Lagrangian 6p of quantum<br />
electrodynamics shown above. The large interaction vertex dominating these pages is a common feature of the<br />
strong, the weak, and the electromagnetic forces. A feature unique to the strong force, the self-interaction of<br />
colored gluons, is suggested by the spiral in the background.
Table 1<br />
basic interactions are ob
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Fig. 1. The main features of the standard model. The strong<br />
force and the electroweak force are each induced by a local<br />
symmetry group, SU(3) and SU(2) X U(l), respectively.<br />
These two symmetries are entirely independent of each other.<br />
SU(3) symmetry (called the color symmetry) is exact and<br />
therefore predicts conservation of color charge. The SU(2) X<br />
U(1) symmetry of the electroweak theory is an exact sym-<br />
metry of the Lagrangian of the theory but not of the solutions<br />
to the theory. The standard model ascribes this symmetry<br />
breaking to the Higgs particles, particles that create a<br />
nonzero weak charge in the vacuum (the lowest energy state<br />
of the system). The only conserved quantity that remains<br />
after the symmetry breaking is electric charge.<br />
The spectacular progress in particle physics<br />
over the past ten years or so has renewed<br />
this dream; many physicists today believe<br />
that we are on the verge of uncovering the<br />
structure of this unified theory. The theoretical<br />
description of the strong, weak, and electromagnetic<br />
interactions is now considered<br />
well established, and, amazingly enough, the<br />
theory shows these forces to be quite similar<br />
despite their experimental differences. The<br />
weak and strong forces have sources<br />
analogous to, but more complicated than,<br />
electric charge, and, like the electromagnetic<br />
force, both can be described by a special type<br />
of field theory called a local gauge theory.<br />
This formulation has been so successful at<br />
explaining all known phenomenology up to<br />
energies of 100 GeV (1 GeV = io9 electron<br />
volts) that it has been coined “the standard<br />
model” and serves as the point of departure<br />
for discussing a grand unification of all<br />
forces, including that of gravitation.<br />
The elements of the standard model are<br />
summarized in Fig. I. In this description the<br />
basic constituents of matter are quarks and<br />
leptons, and these constituents interact with<br />
each other through the exchange of gauge<br />
particles (vector bosons), the modern<br />
analogue of force fields. These so-called local<br />
gauge interactions are inscribed in the language<br />
of Lagrangian quantum field theory,<br />
whose rich formalism contains mysteries<br />
that escape even its most faithful practitioners.<br />
Here we will introduce the central<br />
themes and concepts that have led to the<br />
standard model, emphasizing how its formalism<br />
enables us to describe all<br />
phenomenology of the strong, weak, and<br />
electromagnetic interactions as different<br />
manifestations of a single symmetry principle,<br />
the principle of local symmetry. As we<br />
shall see, the standard model has many<br />
arbitrary parameters and leaves unanswered<br />
a number of important questions. It can<br />
hardly be regarded as a thing of great<br />
beauty-unless one keeps in mind that it<br />
embodies a single unifying principle and<br />
therefore seems to point the way toward a<br />
grander unification.<br />
For those readers who are more<br />
mathematically inclined, the arguments here<br />
are complemented by a series of lecture notes<br />
immediately following the main text and<br />
entitled “From Simple Field Theories to the<br />
Standard Model.” The lecture notes introduce<br />
Lagrangian formalism and stress the<br />
symmetry principles underlying construc-<br />
25
tion of the standard model. The main<br />
emphasis is on the classical limit of the<br />
model, but indications of its quantum generalizations<br />
are also included.<br />
Unification and Extension<br />
Two central themes of physics that have<br />
led to the present synthesis are “unification”<br />
and “extension.” By “unification” we mean<br />
the coherent description of phenomena that<br />
are at first sight totally unrelated. This takes<br />
the form of a mathematical description with<br />
specific rules of application. A theory must<br />
not only describe the known phenomena but<br />
also make predictions of new ones. Almost<br />
all theories are incomplete in that they<br />
provide a description of phenomena only<br />
within a specific range of parameters. Typically,<br />
a theory changes as it is extended to<br />
explain phenomena over a larger range of<br />
parameters, and sometimes it even<br />
simplifies. Hence, the second theme is called<br />
extension-and refers in particular to the<br />
extension of theories to new length or energy<br />
scales. It is usually extension and the resulting<br />
simplification that enable unification.<br />
Perhaps the best-known example of extension<br />
and unification is Newton’s theory of<br />
gravity (1666), which unifies the description<br />
of ordinary-sized objects falling to earth with<br />
that ofthe planets revolving around the sun.<br />
It describes phenomena over distance scales<br />
ranging from a few centimeters up to<br />
102’ centimeters (galactic scales). Newton’s<br />
theory is superceded by Einstein’s theory of<br />
relativity only when one tries to describe<br />
phenomena at extremely high densities<br />
and/or velocities or relate events over cosmological<br />
distance and time scales.<br />
The other outstanding example of unification<br />
in classical physics is Maxwell’s theory<br />
of electrodynamics, which unifies electricity<br />
with magnetism. Coulomb (1785) had established<br />
the famous inverse square law for the<br />
force between electrically charged bodies,<br />
and Biot and Savart (1820) and Ampere<br />
(1820-1825) had established the law relating<br />
the magnetic field B to the electric current as<br />
well as the law for the force between two<br />
26<br />
electric currents. Thus it was known that<br />
static charges give rise to an electric field<br />
E and that moving charges give rise to a<br />
magnetic field B. Then in 183 1 Faraday discovered<br />
that the field itself has a life of its<br />
own, independent of the sources. A timedependent<br />
magnetic field induces an electric<br />
field. This was the first clear hint that electric<br />
and magnetic phenomena were manifestations<br />
of the same force field.<br />
Until the time of Maxwell, the basic laws<br />
of electricity and magnetism were expressed<br />
in a variety of different mathematical forms,<br />
all of which left the central role of the fields<br />
obscure. One of Maxwell’s great achievements<br />
was to rewrite these laws in a single<br />
formalism using the fields E and B as the<br />
fundamental physical entities, whose sources<br />
are the charge density p and the current<br />
density J, respectively. In this formalism the<br />
laws of electricity arid magnetism are expressed<br />
as differential equations that manifest<br />
a clear interrelationship between the two<br />
fields. Nowadays they are usually written in<br />
standard vector notation as follows.<br />
Coulomb’s law:<br />
Ampere’s law:<br />
V . E = 41tp/%;<br />
V X B = 4npoJ;<br />
Faraday’s law: v x E + aB/at = 0;<br />
and the absence of<br />
magnetic monopoles: V B = 0.<br />
The parameters EO and po are determined by<br />
measuring Coulomb’s force between two<br />
static charges and Ampere’s force between<br />
two current-carrying wires, respectively.<br />
Although these equations clearly “unite”<br />
E with B, they are incomplete. In 1865 Maxwell<br />
realized that the above equations were<br />
not consistent with the conservation of electric<br />
charge, which requires that<br />
v . J + ap/at = 0 ,<br />
This inconsistency can be seen from<br />
Ampere’s law, which in its primitive form<br />
requires that<br />
V*J=(4xpo)-’V*(VXB) 0.<br />
Maxwell obtained a consistent solution by<br />
amending Ampere’s law to read<br />
aE<br />
V X B=4xpoJ + kclo-. at<br />
With this new equation, Maxwell showec<br />
that both E and B satisfy the wave equation<br />
For example,<br />
at2<br />
This fact led him to propose the elec<br />
tromagnetic theory of light. Thus, from Max<br />
well’s unification of electric and magnetic<br />
phenomena emerged the concept of elec<br />
tromagnetic waves. Moreover, the speed c o<br />
the electromagnetic waves, or light, is give1<br />
by (~po)-’/~ and is thus determined unique]!<br />
in terms of purely static electric and magne<br />
tic measurements alone!<br />
It is worth emphasizing that apart fron<br />
the crucial change in Ampere’s law, Max<br />
well’s equations were well known to natura<br />
philosophers before the advent of Maxwell<br />
The unification, however, became manifes<br />
only through his masterstroke of expressin)<br />
them in terms of the “right” set of variables<br />
namely, the fields E and B.<br />
Extension to Small Distance<br />
Maxwell’s unification provides an ac<br />
curate description of large-scale elec<br />
tromagnetic phenomena such as radic<br />
waves, current flow, and electromagnets<br />
This theory can also account for the effects o<br />
a medium, provided macroscopic concepti<br />
such as conductivity and permeability arc<br />
introduced. However, ifwe try to extend it tc<br />
very short distance scales, we run intc<br />
trouble; the granularity, or quantum nature<br />
of matter and of the field itself become!<br />
important, and Maxwell’s theory must bt<br />
altered.<br />
Determining the physics appropriate tc<br />
each length scale is a crucial issue and ha!<br />
been known to cause confusion (see “Funda.<br />
mental Constants and the Rayleigh.
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Fig. 2. The wavelength of the probe must be smaller than the scale of the structure<br />
one wants to resolve. Viruses, which are approximately 1r5 centimeter in extent,<br />
cannot be resolved with visible light, the average wavelength of which is 5 X l(T5<br />
centimeter. However, electrons with momentum p of about 20 eV/c have de Broglie<br />
wavelengths short enough to resolve them.<br />
Riabouchinsky Paradox”). For example, the<br />
structure of the nucleus is completely irrelevant<br />
when dealing with macroscopic distances<br />
of, say, 1 centimeter, so it would be<br />
absurd to try to describe the conductivity of<br />
iron over this distance in terms of its quark<br />
and lepton structure. On the other hand, it<br />
would be equally absurd to extrapolate<br />
Ohm’s law to distance intervals of<br />
centimeter to determine the flow of electric<br />
current. Relevant physics changes with scale!<br />
The thrust of particle physics has been to<br />
study the behavior of matter at shorter and<br />
shorter distance scales in hopes of understanding<br />
nature at its most fundamental<br />
level. As we probe shorter distance scales, we<br />
encounter two types of changes in the phys-<br />
ics. First there is the fundamental change<br />
resulting from having to use quantum mechanics<br />
and special relativity to describe<br />
phenomena at very short distances. According<br />
to quantum mechanics, particles have<br />
both wave and particle properties. Electrons<br />
can produce interference patterns as waves<br />
and can deposit all their energy at a point as a<br />
particle. The wavelength k associated with<br />
the particle of momentum p is given by the<br />
de Broglie relation<br />
h A=-<br />
P’<br />
where h is Planck‘s constant (h/2rr = h =<br />
1.0546 X erg . second). This relation is<br />
the basis of the often-stated fact that resolving<br />
smaller distances requires particles of<br />
greater momentum or energy. Notice, incidentally,<br />
that for sufficiently short wavelengths,<br />
one is forced to incorporate special<br />
relativity since the corresponding particle<br />
momentum becomes so large that Newtonian<br />
mechanics fails.<br />
The mamage of quantum mechanics and<br />
special relativity gave birth to quantum field<br />
theory, the mathematical and physical language<br />
used to construct theories of the<br />
elementary particles. Below we will give a<br />
brief review of its salient features. Here we<br />
simply want to remind the reader that quantum<br />
field theory automatically incorporates<br />
quantum ideas such as Heisenberg’s uncertainty<br />
principle and the dual wave-particle<br />
properties of all of matter, as well as the<br />
equivalence of mass and energy.<br />
Since the wavelength of our probe determines<br />
the size of the object that can be<br />
studied (Fig. 2), we need extremely short<br />
wavelength (high energy) probes to investigate<br />
particle phenomena. To gain some<br />
perspective, consider the fact that with visible<br />
light we can see without aid objects as<br />
small as an amoeba (about lo-* centimeter)<br />
and with an optical microscope we can open<br />
up the world of bacteria at about IO4 centimeter.<br />
This is the limiting scale of light<br />
probes because wavelengths in the visible<br />
spectrum are on the order of 5 X IO-’ centimeter.<br />
To resolve even smaller objects we can<br />
exploit the wave-like aspects of energetic<br />
particles as is done in an electron microscope.<br />
For example, with “high-energy” electrons<br />
(E = 20 eV) we can view the world of<br />
viruses at a length scale of about lo-’ centimeter.<br />
With even higher energy electrons<br />
we can see individual molecules (about I 0-7<br />
centimeter) and atoms (lo-* centimeter). To<br />
probe down to nuclear ( centimeter)<br />
and subnuclear scales, we need the particles<br />
available from high-energy accelerators. Today’s<br />
highest energy accelerators produce<br />
100-GeV particles, which probe distance<br />
scales as small as centimeter.<br />
This brings us to the second type of change<br />
27
in appropriate physics with change in scale,<br />
namely, changes in the forces themselves.<br />
Down to distances of approximately<br />
centimeter, electromagnetism is the dominant<br />
force among the elementary particles.<br />
However, at this distance the strong force,<br />
heretohre absent, suddenly comes into play<br />
and completely dominates the interparticle<br />
dynamics. The weak force, on the other<br />
hand, is present at all scales but only as a<br />
small effect. At the shortest distances being<br />
probed by present-day accelerators, the weak<br />
and electromagnetic forces become comparable<br />
in strength but remain several orders<br />
of magnitude weaker than the strong force. It<br />
is at this scale however, that the fundamental<br />
similarity of all three forces begins to emerge.<br />
Thus, as the scale changes, not only does<br />
each force itself change, but its relationship<br />
to the other forces undergoes a remarkable<br />
evolution. In our modern way of thinking,<br />
which has come from an understanding of<br />
the renormalization, or scaling, properties of<br />
quantum field theory, these changes in physics<br />
are in some ways analogous to the<br />
paradigm of phase transitions. To a young<br />
and naive child, ice, water, and steam appear<br />
to be quite different entities, yet rudimentary<br />
observations quickly teach that they are different<br />
manifestations of the same stuff, each<br />
associated with a different temperature scale.<br />
The modern lesson from renormalization<br />
group analysis, as discussed in “Scale and<br />
Dimension-From Animals to Quarks,” is<br />
that the physics of the weak, electromagnetic,<br />
and strong forces may well represent different<br />
aspects of the same unified interaction.<br />
This is the philosophy behind grand<br />
unified theories of all the interactions.<br />
Quantum Electrodynamics and<br />
Field Theory<br />
Let us now return to the subject of electromagnetism<br />
at small distances and describe<br />
quantum electrodynamics (QED), the<br />
relativstic quantum field theory, developed<br />
in the 1930s and 194Os, that extends Maxwell’s<br />
theory to atomic scales. We emphasize<br />
that the standard model is a generalization of<br />
Antiproton<br />
this first and most successful quantum field<br />
theory.<br />
In quantum field theory every particle has<br />
associated with it a mathematical operator,<br />
called a quantum field, that carries the particle’s<br />
characteristic quantum numbers.<br />
Probably the most familiar quantum number<br />
is spin, which corresponds to an intrinsic<br />
angular momentum. In classical mechanics<br />
angular momentum is a continuous variable,<br />
whereas in quantum mechanics it is restricted<br />
to multiples of ‘12 when measured in units<br />
of h. <strong>Particle</strong>s with 5’2-integral spin (1/2, 3/2,<br />
5/2, ...) are called fermions; particles with<br />
integral spin (0, I , 2, 3, ... ) are called bosons.<br />
Since no two identical fermions can occupy<br />
the same position at the same time (the<br />
famous Pauli exclusion principle), a collel<br />
tion of identical fermions must necessari<br />
take up some space. This special property I<br />
fermions makes it natural to associate the1<br />
with matter. Bosons, on the other hand, ca<br />
crowd together at a point in space-time 1<br />
form a classical field and are naturally ri<br />
garded as the mediators of forces.<br />
In the quantized version of Maxwell’s thc<br />
ory, the electromagnetic field (usually in tt<br />
guise of the vector potential A,,) is a boso<br />
field that carries the quantum numbers of tk<br />
photon, namely, mass m = 0, spin s = 1, an<br />
electric charge Q = 0. This quantized field, t<br />
the very nature of the mathematics, aut1<br />
matically manifests dual wave-partic<br />
properties. Electrically charged particle<br />
28
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
f<br />
ig. 3. (a) The force between two electrons is described classically by Coulomb’s<br />
aw. Each electron creates a forcefield (shown as lines emanating from the charge<br />
(e) that is felt by the other electron. The potential energy V is the energy needed to<br />
ring the two electrons to within a distance r of each other. (6) In quantum field<br />
heory two electrons feel each other’s presence by exchanging virtual photons, or<br />
virtual particles of light. Photons are the quanta of the electromagnetic field. The<br />
eynman diagram above represents the (lowest order, see Fig. 5) interaction<br />
etween two electrons (straight lines) through the exchange of a virtual photon<br />
I<br />
wavy line).<br />
such as electrons and positrons, are also represented<br />
by fields, and, as in the classical<br />
theory, they interact with each other through<br />
the electromagnetic field. In QED, however,<br />
the interaction takes place via an exchange of<br />
photons. Two electrons “feel” each other’s<br />
presence by passing photons back and forth<br />
between them. Figure 3 pictures the interaction<br />
with a “Feynman diagram”: the straight<br />
lines represent charged particles and the<br />
wavy line represents a photon. (In QED such<br />
diagrams correspond to terms in a<br />
perturbative expansion for the scattering between<br />
charged particles (see Fig. 5).<br />
Similarly, most Feynman diagrams in this<br />
issue represent lowest order contributions to<br />
the particle reactions shown.)<br />
These exchanged photons are rather<br />
special. A real photon, say in the light by<br />
which you see, must be massless since only a<br />
massless particle can move at the speed of<br />
light. On the other hand, consider the lefthand<br />
vertex of Fig. 3, where a photon is<br />
emitted by an electron; it is not difficult to<br />
convince oneself that if the photon is massless,<br />
energy and momentum are not conserved!<br />
This is no sin in quantum mechanics,<br />
however, as Heisenberg’s uncertainty principle<br />
permits such violations provided they<br />
occur over sufficiently small space-time intervals.<br />
Such is the case here: the violating<br />
photon is absorbed at the right-hand vertex<br />
by another electron in such a way that, overall,<br />
energy and momentum are conserved.<br />
The exchanged photon is “alive” only for a<br />
period concomitant with the constraints of<br />
the uncertainty principle. Such photons are<br />
referred to as virtual photons to distinguish<br />
them from real ones, which can, of course,<br />
live forever.<br />
The uncertainty principle permits all sorts<br />
of virtual processes that momentarily violate<br />
energy-momentum conservation. As illustrated<br />
in Fig. 4, a virtual photon being<br />
exchanged between two electrons can, for a<br />
very short time, turn into a virtual electronpositron<br />
pair. This conversion of energy into<br />
mass is allowed by the famous equation of<br />
special relativity, E = mc2. In a similar<br />
fashion almost anything that can happen will<br />
29
happen, given a sufficiently small space-time<br />
interval. It is the countless multitude of such<br />
virtual processes that makes quantum field<br />
theory :so rich and so difficult.<br />
Given the immense complexity of the theory,<br />
one wonders how any reliable calculation<br />
can ever be made. The saving grace of<br />
quantum electrodynamics, which has made<br />
its predictions the most accurate in all of<br />
physics, is the smallness of the coupling between<br />
the electrons and the photons. The<br />
coupling strength at each vertex where an<br />
electron spews out a virtual photon is just the<br />
electronic charge e, and, since the virtual<br />
photon must be absorbed by some other<br />
electron, which also has charge e, the<br />
probability for this virtual process is of magnitude<br />
e’. The corresponding dimensionless<br />
parameter that occurs naturally in this theory<br />
is denoted by a and defined as e2/4nh e. It is<br />
approximately equal to 11137. The<br />
probabilities of more complicated virtual<br />
processes involving many virtual particles<br />
are proportional to higher powers of a and<br />
are therefore very much smaller relative to<br />
the probabilities for simpler ones. Put<br />
slightly differently, the smallness of a implies<br />
that perturbation theory is applicable, and<br />
we can control the level of accuracy of our<br />
calcu1ai:ions by including higher and higher<br />
order virtual processes (Fig. 5). In fact, quantum<br />
electrodynamic calculations of certain<br />
atomic and electronic properties agree with<br />
experiment to within one part in a billion.<br />
As we will elaborate on below, the quantum<br />
field theories of the electroweak and the<br />
strong interactions that compose the standard<br />
model bear many resemblances to<br />
quantum electrodynamics. Not too surprisingly,<br />
the coupling strength of the weak interaction<br />
is also small (and in fact remains small<br />
at all energy or distance scales), so perturbation<br />
theory is always valid. However, the<br />
analogue of a for the strong interaction is not<br />
always small, and in many calculations<br />
perturbation theory is inadequate. Only at<br />
the high energies above 1 GeV, where the<br />
theory is said to be asymptotically free, is the<br />
analogue of a so small that perturbation theory<br />
is valid. At low and moderate energies<br />
Fig. 4. A virtualphoton being exchanged between two electrons can, for a very shor<br />
time, turn into a virtual electron-positron (e+-e-) pair. This virtualprocess is on<br />
of many that contribute to the electromagnetic interaction between electricall<br />
chargedparticles (see Fig. 5).<br />
(for example, those that determine the<br />
properties of protons and neutrons) the<br />
strong-interaction coupling strength is large,<br />
and analytic techniques beyond perturbation<br />
theory are necessary. So far such techniques<br />
have not been very :successful, and one has<br />
had to resort to the nasty business of numerical<br />
simulations!<br />
As discussed at the end of the previous<br />
section, these changes in coupling strengths<br />
with changes in scale are the origin of the<br />
changes in the forces that might lead to a<br />
unified theory. For an example see Fig. 3 in<br />
“Toward a Unified Theory.”<br />
Symmetries<br />
One cannot discuss the standard model<br />
without introducing the concept of symmetry.<br />
It has played a central role in classifying<br />
the known particle states (the ground<br />
states of 200 or so particles plus excited<br />
states) and in predicting new ones. Just as the<br />
chemical elements fall into groups in the<br />
periodic table, the particles fall into multiplets<br />
characterized by similar quantum<br />
numbers. However, the use of symmetry in<br />
particle physics goes well beyond mere<br />
classification. In the construction of the stai<br />
dard model, the special kind of symmeti<br />
known as local symmetry has become tl<br />
guiding dynamical principle; its aesthetic ii<br />
fluence in the search for unification is ren<br />
iniscent of the quest for beauty among tk<br />
ancient Greeks. Before we can discuss th<br />
dynamical principle, we must first review tl<br />
general concept of symmetry in partic<br />
physics.<br />
In addition to electric charge and mas<br />
particles are characterized by other quantui<br />
numbers such as spin, isospin, strangenes<br />
color, and so forth. These quantum numbe<br />
reflect the symmetries of physical laws an<br />
are used as a basis for classification an1<br />
ultimately, unification.<br />
Although quantum numbers such as spi<br />
and isospin are typically the distinguishir<br />
features of a particle, it is probably less we<br />
known that the mass of a particle is som<br />
times its only distinguishing feature. For e<br />
ample, a muon (p) is distinguished from a<br />
electron (e) only because its mass is 2C<br />
times greater that that of the electron. I1<br />
deed, when the muon was discovered 1<br />
1938, Rabi was reputed to have made tk<br />
remark. “Who ordered that?” And the ta<br />
30
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Electron Scattering<br />
Electromagnetic<br />
Interaction = eJpA,<br />
(Interaction)* a a<br />
(~nteraction)~ a a2<br />
(Interaction)6 a a3<br />
.<br />
where Jo = p<br />
A _ _<br />
Fig. 5. As shown above, the basic interaction<br />
vertex of quantum electrodynamics<br />
is an electron current J’<br />
interacting with the electromagnetic<br />
field A,,. Because the coupling strength<br />
a is small, the amplitude for processes<br />
involving such interactions can be approximated<br />
by a perturbation expansion<br />
on a free field theory. The<br />
terms in such an expansion, shown at<br />
left for electron scattering, are proportional<br />
to various powers of a. The largest<br />
contribution to the electron-scattering<br />
amplitude is proportional to a and<br />
is represented by a Feynmann diagram<br />
in which the interaction vertex appears<br />
twice. Successively smaller contributions<br />
arise from terms proportional to<br />
a’ with four interaction vertices, from<br />
terms proportional to a3 with six interaction<br />
vertices, and so on.<br />
(T), discovered in 1973, is 3500 times heavier<br />
than an electron yet again identical to the<br />
electron in other respects. One of the great<br />
unsolved mysteries of particle physics is the<br />
origin of this apparent hierarchy of mass<br />
among these leptons. (A lepton is a fundamental<br />
fermion that has no strong interactions.)<br />
Are there even more such particles? Is<br />
there a reason why the mass hierarchy among<br />
the leptons is paralleled (as we will describe<br />
below) by a similar hierarchy among the<br />
quarks? It is believed that when we understand<br />
the origin of fermion masses, we will<br />
also understand the origin of CP violation in<br />
nature (see box). These questions are frequently<br />
called the family problem and are<br />
discussed in the article by Goldman and<br />
Nieto.<br />
Groups and Group Multiplets. Whether or<br />
not the similarity among e, p, and ‘5 reflects a<br />
fundamental symmetry of nature is not<br />
known. However, we will present several<br />
possibilities for this family symmetry to introduce<br />
the language of groups and the<br />
significance of internal symmetries.<br />
Consider a world in which the three leptons<br />
have the same mass. In this world atoms<br />
with muons or taus replacing electrons<br />
would be indistinguishable: they would have<br />
identical electromagnetic absorption or<br />
emission bands and would form identical<br />
elements. We would say that this world is<br />
invariant under the interchange of electrons,<br />
muons, and taus, and we would call this<br />
invariance a symmetry ofnature. In the,real<br />
world these particles don’t have the same<br />
mass; therefore our hypothetical symmetry,<br />
if it exists, is broken and we can distinguish a<br />
muonic atom from, say, its electronic<br />
counterpart.<br />
We can describe our hypothetical invariance<br />
or family symmetry among the<br />
three leptons by a set of symmetry operations<br />
that form a mathematical construct called a<br />
group. One property of a group is that any<br />
two symmetry operations performed in succession<br />
also corresponds to a symmetry<br />
operation in that group. For example, replacingan<br />
electron with a muon, and then replacing<br />
a muon with a tau can be defined as two<br />
discrete symmetry operations that when<br />
performed in succession are equivalent to<br />
the discrete symmetry operation of replacing<br />
an electron with a tau. Another group property<br />
is that every operation must have an<br />
inverse. The inverse of replacing an electron<br />
with a muon is replacing a muon with an<br />
electron. This set of discrete operations on<br />
e, p, and T forms the discrete six-element<br />
group rr3 (with K standing for permutation).<br />
In this language e, p, and T are called a<br />
multiplet or representation of ~3 and are said<br />
to transform as a triplet under ~ 3 .<br />
Another possibility is that the particles e,<br />
p, and T transform as a triplet under a group<br />
of conlinuous symmetry operations. Consider<br />
Fig. 6, where e, p, and T are represented<br />
as three orthogonal vectors in an abstract<br />
31
three-dimensional space. The set of continuous<br />
rotations of the three vectors about three<br />
independent axes composes the group<br />
known as the three-dimensional rotation<br />
group and denoted by SO(3). As shown in<br />
Fig. 6, SO(3) has three independent transformations,<br />
which are represented by orthogonal<br />
3 X 3 matrices. (Note that n3 is a<br />
subset of SO(3).)<br />
Suppose that SO(3) were an unbroken<br />
family symmetry of nature and e, p, and r<br />
transformed as a triplet under this symmetry.<br />
How would it be revealed experimentally?<br />
The SO(3) symmetry would add an<br />
extra degree of freedom to the states that<br />
could be formed by e, p, and r. For example,<br />
the spatially symmetric ground state of<br />
helium, which ordinarily must be antisymmetric<br />
under the interchange of the two electron<br />
spins, could now be antisymmetric<br />
under the interchange of either the spin or<br />
the family quantum number of the two leptons.<br />
In particular, the ground state would<br />
have three different antisymmetric configurations<br />
and the threefold degeneracy<br />
might be split by spin-spin interactions<br />
among the leptons and by any SO(3) symmetric<br />
interaction. Thus the ground state of<br />
known helium would probably be replaced<br />
by sets of degenerate levels with small hyperfine<br />
energy splittings.<br />
In particle physics we are always interested<br />
in the largest group of operations that leaves<br />
all properties of a system unchanged. Since e,<br />
p, and T are described by complex fields, the<br />
largest group of operations that could act on<br />
this triplet is U(3) (the group of all unitary 3<br />
X 3 matrices Usatisfying UtU= I). Another<br />
possibility is SU(3), a subgroup ofU(3) satisfying<br />
the additional constraint that det U = I.<br />
This list of symmetries that may be<br />
reflected in the similarity of e, p, and 7 is not<br />
exhaustive. We could invoke a group of symmetry<br />
operations that acts on any subset of<br />
the three particles, such as SU(2) (the group<br />
of 2 X 2 unitary matrices with det U = I)<br />
acting, say, on e and p as a doublet and on r<br />
as a singlet. Any one of these possibilities<br />
may be realized in nature, and each possibility<br />
has different experimentally observable<br />
,<br />
i<br />
I t<br />
1<br />
Fig. 6. (a) The three leptons e, F, and T are represented as three orthogonal vecton<br />
in an abstract three-dimensional space. (b) The set of rotations about the thret<br />
orthogonal axes defines S0(3), the three-dimensional rotation group. SO(3) haJ<br />
three charges (or generators) associated with the infinitesimal transformatiom<br />
about the three independent axes. These generators have the same Lie algebra as the<br />
generators of the g,roup SU(2), as discussed in Lecture Note 4 following this article.<br />
32
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
onsequences. However, the known dif-<br />
:rences in the masses of e, p, and T imply<br />
iat any symmetry used to describe the<br />
imilarity among them is a broken symietry.<br />
Still, a broken symmetry will retain<br />
aces of its consequences (if the symmetry is<br />
broken by a small amount) and thus also<br />
provides useful predictions.<br />
Our hypothetical broken symmetry<br />
among e, p, and T is but one example of an<br />
approximate internal global symmetry. Another<br />
is the symmetry between, say, the neu-<br />
tron and the proton in strong interactions,<br />
which is described by the group known as<br />
strong-isospin SU(2). The neutron and<br />
proton transform as a doublet under this<br />
symmetry and the three pions transform as a<br />
triplet. We will discuss below the classifica-<br />
CP Violation<br />
T<br />
he faith of physicists in symmetries of<br />
nature, so shaken by the observation<br />
of parity violation in 1956, was soon<br />
restored by invocation of a new symmetry<br />
principle-CP conservation-to interpret<br />
parity-violating processes. This principle<br />
states that a process is indistinguishable from<br />
its mirror image provided<br />
mirror image are replaced<br />
cles. Alas, in 1964 this<br />
shattered with the results<br />
on the decay of neutral kaons.<br />
According to the classic analysis of M.<br />
Gell-Mann and A. Pais, neutral kaons cxist<br />
in two forms: Kg, with an even CP eigenvalue<br />
and decaying with a relatively short<br />
lifetime of IO-” second into two pions, and<br />
K:, with an odd CP eigenvalue and decaying<br />
with a lifetime of about 5 X lo-* second into<br />
three pions. CP conservati<br />
decay of the longer lived K<br />
But in an experiment at Brookhaven, J.<br />
Christenson, J. Cronin, V. Fitch, and R.<br />
Turlay found that about I in 500 Kf mesons<br />
decays into two pions. This first observation<br />
of CP violation has been confirmed in many<br />
other experiments on the neutral kaon system,<br />
but to date no other CP-violating effects<br />
have been found. The underlying mechanism<br />
of CP violation remains to be understood,<br />
and an implication of the phenomenon,<br />
the breakdown of time-reversal invariance<br />
(which is necessary to maintain<br />
CPT conservation), remains to be observed.<br />
30<br />
20<br />
10<br />
w<br />
0<br />
0.9996<br />
J<br />
494 < m* < 504<br />
I<br />
0.9998<br />
cos 8<br />
I<br />
1 .oooo<br />
Evidence for the CP-violating decay of Ki into two pions. Here the number<br />
of events in which the invariant mass (m *) of the decayproducts was in close<br />
proximity to the mass of the neutral kaon is plotted versus the cosine of the<br />
angle 0 between the Kf beam and the vector sum of the momenta of the<br />
decay products. The peak in the number of events at cos 0 z 1 (indicative of<br />
two-body decays) could only be explained as the decay of K; into two pions<br />
with a branchingratio of about 2 X (Adapted from “Evidence for the<br />
2rc Decay of the Kg Meson” by J. H. Christenson, J. W. Cronin, V. L. Fitch,<br />
and R. Turlay, Physical Review Letters 13(1964):138.)<br />
33
tion of strongly interacting particles into<br />
multiplets of SU(3), a scheme that combines<br />
strong isospin with the quantum number<br />
called strangeness, or strong hypercharge.<br />
(For a more complete discussion of continuous<br />
symmetries and internal global symmetries<br />
such as SU(2), see Lecture Notes 2<br />
and 4.)<br />
Exact., or unbroken, symmetries also play<br />
a fundamental role in the construction of<br />
theories: exact rotational invariance leads to<br />
the exact conservation of angular momentum,<br />
and exact translational invariance in<br />
space-time leads to the exact conservation of<br />
energy and momentum. We will now discuss<br />
how the exact phase invariance of electrodynamics<br />
leads to the exact conservation<br />
of electric charge.<br />
Global U(l) Invariance and Conservation<br />
Laws. In quantum field theory the dynamics<br />
of a system are encoded in a function of the<br />
fields called a Lagrangian, which is related to<br />
the energy of the system. The Lagrangian is<br />
the mosl. convenient means for studying the<br />
symmetries ofthe theory because it is usually<br />
a simple task to check if the Lagrangian<br />
remains unchanged under particular symmetry<br />
operations.<br />
An electron is described in quantum field<br />
theory by a complex field,<br />
and a positron is described by the complex<br />
conjugate of that field,<br />
Although the real fields wI and I+I~ are<br />
separately each able to describe a spin-%<br />
particle, the two together are necessary to<br />
describe a particle with electric charge.*<br />
The Lagrangian of quantum electrodynamics<br />
is unchanged by the continuous<br />
operation of multiplying the electron field by<br />
*The real fields VI and y12 are four-component<br />
Majorana fields that together make up the standard<br />
four-component complex Dirac spinor field.<br />
an arbitrary phase, that is, by the transformation<br />
w - eiAQW,<br />
where A is an arbitary real number and Q is<br />
the electric charge operator associated with<br />
the field. The eigenvalue of Q is -I for an<br />
electron and +1 for a positron. This set of<br />
phase transformations forms the global symmetry<br />
group U(1) (the set of unitary 1 X 1<br />
matrices). In QED this symmetry is unbroken,<br />
and electric charge is a conserved<br />
quantum number of the system.<br />
There are other global U(I) symmetries<br />
relevant in particle physics, and each one<br />
implies a conserved quantum number. For<br />
example, baryon number conservation is associated<br />
with a U(1) phase rotation of all<br />
baryon fields by an amount e', where B = 1<br />
for protons and neutrons, B = '1 for quarks,<br />
and B = 0 for leptons. Analogously, electron<br />
number is conserved if the field of the electron<br />
neutrino is assigned the same electron<br />
number as the field of the electron and all<br />
other fields are assigned an electron number<br />
of zero. The same holds true for muon number<br />
and tau number. Thus a global U(1)<br />
phase symmetry seems to operate on each<br />
type of lepton. (Possible violation of muonnumber<br />
conservation is discussed in "Experiments<br />
To Test Unification Schemes.")<br />
The Principle of Local Symmetry<br />
We are now ready to distinguish a global<br />
phase symmetry from a local one and examine<br />
the dynamical consequences that emerge<br />
from the latter. Figure 7 illustrates what hap-<br />
pens to the electron field under the global<br />
phase transformation w - e"Qy. For convenience,<br />
space-time is represented by a set<br />
of discrete points labeled by the index j. The<br />
phase of the electron field at each point is<br />
represented by an arrow that rotates about<br />
the point, and the kinetic energy of the field<br />
is represented by springs connecting the arrows<br />
at different space-time points. A global<br />
U( 1) transformation rotates every two-dimensional<br />
vector by the same arbitrary angle<br />
A: 0,- 0, + QA, where Q is the electric<br />
charge. In order for the Lagrangian to be<br />
invariant under this global phase rotation, it<br />
is clearly sufficient for it to be a function only<br />
ofthe phase differences (0, - 0,). Both the free<br />
electron terms and the interaction terms in<br />
the QED Lagrangian are invariant under this<br />
continuous global symmetry.<br />
A local U( I ) transformation, in contrast,<br />
rotates every two-dimensional vector by a<br />
different angle A,. This local transformation,<br />
unlike its global counterpart, does not leave<br />
the Lagrangian of the free electron invariant.<br />
As represented in Fig. 7 by the stretching and<br />
compressing of the springs, the kinetic<br />
energy of the electron changes under local<br />
phase transformations. Nevertheless, the full<br />
Lagrangian of quantum electrodynamics is<br />
invariant under these local U( I ) transformations.<br />
The electromagnetic field (A,)<br />
precisely compensates for the local phase<br />
rotation and the Lagrangian is left invariant.<br />
This is represented in Fig. 7 by restoring the<br />
stretched and compressed springs to their<br />
initial tension. Thus, the kinetic energy of the<br />
electron (the energy stored in the springs) ir<br />
the same before and after the local phase<br />
transformation.<br />
In our discrete notation, the full La-<br />
Fig. 7. Global versus local phase transformations. The arrows represent the phases<br />
of an electron field at four discrete points labeled by j = I, 2, 3, and 4. The springs<br />
represent the kinetic energy of the electrons. A global phase transformation does<br />
not change the tension in the springs and therefore costs no energy. A local phase<br />
transformation witlhout gauge interactions stretches and compresses the springs<br />
and thus does cost energy. However introduction of the gaugefield (represented by<br />
the white haze) exactly compensates for the local phase transformation of the<br />
electron field and the springs return to their original tension so that local phase<br />
transformations with gauge interactions do not cost energy.<br />
34
~<br />
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
__<br />
Phase<br />
I<br />
j=2 j'3 j=4<br />
Global Phase<br />
Transformation<br />
+(j) -+ eiQA $'(I)<br />
ei -+ ei + QA<br />
Q$' =+$<br />
Q= +I<br />
iQ A .<br />
Local Phase J/(j)-+e ' $'(i)<br />
Transformation<br />
Oi-+Oj+QAi Q=+I<br />
I I<br />
Gauge Field I I<br />
Compensates for<br />
Local Phase<br />
Transformation<br />
I I<br />
Ai, -+ Ai, - Ai + A,<br />
35
and is<br />
invariant under the simultaneous transformations<br />
grangian is a function of e, - ek + A,&<br />
r--<br />
I<br />
I<br />
The matrix with elements A,k is the discrete<br />
space-time analogue of the electromagnetic<br />
potential defined on the links between the<br />
points k and j. Thus, if one starts with a<br />
theory of free electrons with no interactions<br />
and demands that the physics remain invariant<br />
under local phase transformation of<br />
the electron fields, then one induces the standard<br />
electromagnetic interactions between<br />
the electron current .Ip and photon field A,,,<br />
as shown in Figs. 5 and 8. From this point of<br />
view, Maxwell’s equations can be viewed as a<br />
consequence of the local U(1) phase invariance.<br />
Although this local invariance was<br />
originally viewed as a curiosity of QED, it is<br />
now viewed as the guiding principle for constructing<br />
field theories. The invariance is<br />
usually termed gauge invariance, and the<br />
photon is referred to as a gauge particle since<br />
it mediates the U(1) gauge interaction. It is<br />
worth emphasizing that local U(1) invariance<br />
implies that the photon is massless<br />
because the term that would describe a<br />
massive photon is not itself invariant under<br />
local U(1) transformations.<br />
The local gauge invariance of QED is the<br />
prototype for theories of both the weak and<br />
the strong interactions. Obviously, since<br />
neither of these is a long-range interaction,<br />
some additional features must be at work to<br />
accouni. for their different properties. Before<br />
turning to a discussion of these features, we<br />
stress that in theories based on local gauge<br />
invariance, currents always play an important<br />
role. In classical electromagnetism the<br />
fundamental interaction takes place between<br />
the vector potential and the electron current;<br />
this is reflected in quantum electrodynamics<br />
by Feynman diagrams: the virtual photon<br />
(the gauge field) ties into the current<br />
produced by the moving electron (see Fig. 8).<br />
As will become clear below, a similar situation<br />
exists in the strong interaction and,<br />
more important, in the weak interaction.<br />
36<br />
Fig. 8. The U(I) local symmetry of QED implies the existence of a gauge field to<br />
compensate for the local phase transformation of the electrically charged matter<br />
fields. The generator of the U(I) local phase transformation is Q, the electric<br />
charge operator defined in the figure in terms of the current density Jo. The gauge<br />
field A,, interacts with the electrically charged matter fields through the current J p.<br />
The coupling strength is e, the charge of the electron.<br />
The Strong Interaction<br />
In an atom electrons are bound to the<br />
nucleus by the Coulomb force and occupy a<br />
region about IO-* centimeter in extent. The<br />
nucleus itself is a tightly bound collection of<br />
protons and neutrons confined to a region<br />
about IO-” centimeter across. As already<br />
emphasized, the force that binds the protons<br />
and neutrons together to form the nucleus is<br />
much stronger and considerably shorter in<br />
range than the electromagnetic force. Leptons<br />
do not feel this strong force; particles<br />
that do participate in the strong interactions<br />
are called hadrons.
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
a-<br />
I<br />
n 1961 M. Gell-Mann and independently<br />
Y. Ne’eman proposed a sys-<br />
tem for classifying the roughly one hundred<br />
baryons and mesons known at the t<br />
This “Eightfold Way” was based on<br />
SU(3) group, which has eight independent<br />
symmetry operations. According to this system,<br />
hadrons with the same baryon number,<br />
spin angular momentum, and parity and<br />
with electric charge, strangeness (or hypercharge),<br />
isotopic spin, and mass related by<br />
certain rules were grouped into large multiplets<br />
encompassing the already establis<br />
isospin multiplets, such as the neutron and<br />
proton doublet or the negative, neutral, and<br />
positive pion triplet. Most of the known<br />
hadrons fit quite neatly into octets. However,<br />
the decuplet partly tilled by the quartet of A<br />
baryons and the triplet of C( 1385) baryons<br />
lacked three members. Discovery of the<br />
E( 1520) doublet was announced in 1962, and<br />
these baryons satisfied the criteria for membership<br />
in the decuplet. This partial confirmation<br />
of the Eightfold Way motivated a<br />
search at Brookhaven for the remaining<br />
member, already named R- and predicted to<br />
be stable against strong and electromagnetic<br />
interactions, decaying (relatively slowly) by<br />
the weak interaction. Other properties<br />
predicted for this particle were a bary<br />
number of I, a spin angular momentum of<br />
3/2, positive parity, negative electric charge, a<br />
strangeness of-3, an isotopic spin of 0, and a<br />
mass ofabout 1676 MeV.<br />
A beam of 5-GeV negative kaons<br />
produced at the AGS was directed into a<br />
liquid-hydrogen bubble chamber, where the<br />
R- was to be produced by reaction of the<br />
kaons with protons. The tracks of the decay<br />
products of the new particle were then sought<br />
in the bubble-chamber photographs. In early<br />
1964 a candidate event was found for decay Analysis of the tracks for these two events<br />
of an R- into a x- and a So, one of three confirmed the predicted mass and strangepossible<br />
decay modes. Within several weeks, ness, and further studies confirmed the<br />
by coincidence and good fortune, another 0- predicted spin and panty. Discovery of the<br />
was found, this time decaying into a Ao and a R- estabIished the Eightfold Way as a viable<br />
K-, the mode now known to be dominant. description of hadronic states. W<br />
- -_<br />
The R- was first detected in the bubble-chamber photograph reproduced above.<br />
A K- entered the bubble chamber from the bottom (track 1) and collided with a<br />
proton. The collision produced an R- (track 3), a K+ (track 2), and a KO, which,<br />
being neutral, left no track and must have decayed outside the bubble chamber.<br />
The R- decayed into a IC- (track 4) and a 5’. The 5’ in turn decayed into a A’<br />
and a no. The A’ decayed into a IC- (track 5) and a proton (track 6), and the no<br />
veuy quickly decayed into two gamma rays, one of which (track 7) created an e--<br />
e’pair within the bubble chamber. (Photo courtesy of the Niels Bohr Library of<br />
the American Institute of <strong>Physics</strong> and Brookhaven National Laboratouy.)<br />
37
Table of “Elementary <strong>Particle</strong>s”<br />
BARYONS<br />
Spin-112 Octet<br />
Strong<br />
Isospin<br />
Mass<br />
,<br />
The mystery of the strong force and the<br />
structure of nuclei seemed very intractable as<br />
little as fifteen years ago. Studying the relevant<br />
distance scales requires machines that<br />
can accelerate protons or electrons to<br />
energies of 1 CeV and beyond. Experiments<br />
with less energetic probes during the 1950s<br />
revealed two very interesting facts. First, the<br />
strong force does not distinguish between<br />
protons and neutrons. (In more technical<br />
language, the proton and the neutron transform<br />
into each other under isospin rotations,<br />
and the Lagrangian of the strong interaction<br />
is invariant under these rotations.) Second,<br />
the structure of protons and neutrons is as<br />
rich as that of nuclei. Furthermore, many<br />
new hadrons were discovered that were apparently<br />
just as “elementary” as protons and<br />
neutrons.<br />
The table of “elementary particles” in the<br />
mid-1960s displayed much of the same complexity<br />
and symmetry as the periodic table of<br />
the elements. In 1961 both Cell-Mann and<br />
Ne’eman proposed that all hadrons could be<br />
classified in multiplets of the symmetry<br />
group called SU(3). The great triumph of this<br />
proposal was the prediction and subsequent<br />
discovery ofa new hadron, the omega minus.<br />
This hadron was needed to fill a vacant space<br />
in one of the SU(3) multiplets (Fig. 9).<br />
In spite of the SU(3) classification scheme,<br />
the belief that all of these so-called elementary<br />
particles were truly elementary became<br />
more and more untenable. The most contradictory<br />
evidence was the finite size of<br />
hadrons (about centimeter), which<br />
drastically contrasted with the point-like<br />
nature of the leptons. Just as the periodic<br />
table was eventually explained in terms of a<br />
few basic building blocks, so the hadronic<br />
zoo was eventually tamed by postulating the<br />
existence of a small number of “truly<br />
elementary point-like particles” called<br />
quarks. In 1963 Cell-Mann and, independently,<br />
Zweig realized that all hadrons<br />
could be constructed from three spin4 fermions,<br />
designated u, d, and s (up, down, and<br />
strange). The SU(3) symmetry that manifested<br />
itself in the table of “elementary particles”<br />
arose from an invariance of the La-<br />
38<br />
Y<br />
Y<br />
1.<br />
0.<br />
- 1.<br />
-1 -112 0 112 1<br />
’3<br />
in-312 DecuDlet<br />
-312 -1 -112 0 112 1 312<br />
-<br />
‘3<br />
IUS)<br />
1 I2<br />
0<br />
1<br />
1 12<br />
312<br />
1<br />
939<br />
A(1 f 16)<br />
C(1193)<br />
Z(1348)<br />
A( 1232)<br />
C *( 1385)<br />
1 I2 Z* (1 530)<br />
0 (1672)<br />
Fig. 9. The Eightfold Way classified the hadrons into multiplets of the<br />
symmetry group SU(3). <strong>Particle</strong>s of each SU(3) multiplet that lie on a<br />
horizontal line form strong-isospin (SU(2)) multiplets. Each particle is<br />
plotted according to the quantum numbers I, (the third component of strong<br />
isospin) and strong hypercharge Y (Y = S + B, where S is strangeness and B is<br />
baryon number). These quantum numbers correspond to the two diagonal<br />
generators of SU(3). The quantum numbers of each particle are easily<br />
understood in terms of its fundamental quark constituents. Baryons contain<br />
three quarks and mesons contain quark-antiquark pairs. Baryons in the spin-<br />
3/2 decuplet are obtained from baryons in the spin-% octet by changing the<br />
spin and SU(3) flavor quantum numbers of the three quark wave functions.<br />
For example, the three quarks that compose the neutron in the spin-% octet can<br />
reorient their spins to form the A’ in the spin-3/2 decuplet. Similar changes in<br />
the meson quark-antiquark wave functions change the spin-0 meson octet into<br />
the spin-I meson octet.
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
MESONS<br />
Mass<br />
K(495)<br />
K(495)<br />
K* (892)<br />
w (783)<br />
P (770)<br />
K*(892)<br />
Strong<br />
Isospin<br />
1 /2<br />
0<br />
1<br />
1 /2<br />
112<br />
0<br />
1<br />
112<br />
Spin-0 Octet<br />
K0(d9<br />
K'(u5)<br />
1 212 0 112 1<br />
13<br />
Spin-I Oclet<br />
*+ -<br />
K*'(d3 K (us)<br />
-1 -112 0 112 1<br />
Quarks<br />
'3<br />
Electric<br />
Name Symbol Charge Y<br />
UP U 213 1/3<br />
Down d -1J3 113<br />
Strange 5 -113 -2131<br />
11<br />
-1<br />
O Y<br />
grangian of the strong interaction to rotations<br />
among these three objects. This global<br />
symmetry is exact only if the u, d, and s<br />
quarks have identical masses, which implies<br />
that the particle states populating a given<br />
SU(3) multiplet also have the same mass.<br />
Since this is certainly not the case, SU(3) is a<br />
broken global symmetry. The dominant<br />
breaking is presumed to arise, as in the example<br />
of e, p, and 'I, from the differences in the<br />
masses of the u, d, and s quarks. The origin of<br />
these quark masses is one of the great unanswered<br />
questions. It is established, however,<br />
that SU(3) symmetry among the u, d,<br />
and s quarks is preserved by the strong interaction.<br />
Nowadays, one refers to this SU(3) as<br />
ayuvor symmetry, with u, d, and s representing<br />
different quark flavors. This nomen-\,<br />
clature is to distinguish it from another and<br />
quite different SU(3) symmetry dossessed by<br />
quarks, a local symmetry that is associated<br />
directly with the strong force and has become<br />
known as the SU(3) of color. The theory<br />
resulting from this symmetry is called quantum<br />
chromodynamics (QCD), and we now<br />
turn our attention to a discussion of its<br />
properties and structure.<br />
The fundamental structure of quantum<br />
chromodynamics mimics that of quantum<br />
electrodynamics in that it, too, is a gauge<br />
theory (Fig. IO). The role of electric charge is<br />
played by three "colors" with which each<br />
quark is endowed-red, green, and blue. The<br />
three color varieties of each quark form a<br />
triplet under the SU(3) local gauge symmetry.<br />
A local phase transformation of the<br />
quark field is now considerably extended<br />
since it can rotate the color and thereby<br />
change a red quark into a blue one. The local<br />
gauge transformations of quantum electrodynamics<br />
simply change the phase of an<br />
electron, whereas the color transformations<br />
of QCD actually change the particle. (Note<br />
that these two types of phase transformation<br />
are totally independent of each other.)<br />
We explained earlier that the freedom to<br />
change the local phase of the electron field<br />
forces the introduction of the photon field<br />
(sometimes called the gauge field) to keep the<br />
Lagrangian (and therefore the resulting phys-<br />
39
~<br />
,<br />
its) invariant under these local phase<br />
changes. This is the principle of local symmetry.<br />
A similar procedure applied to the<br />
quark field induces the so-called chromodynamic<br />
force. There are eight independent<br />
symmetry transformations that change the<br />
color of a quark and these must be compensated<br />
for by the introduction of eight<br />
gauge fields, or spin-I bosons (analogous to<br />
the single photon of quantum electrodynamics).<br />
Extension of the local U( 1)<br />
gauge invariance of QED to more complicated<br />
symmetries such as SU(2) and SU(3)<br />
was first done by Yang and Mills in 1954.<br />
These larger symmetry groups involve socalled<br />
non-Abelian, or non-commuting algebras<br />
(in which AB + BA), so it has become<br />
customary to refer to this class of theories as<br />
“non-Abelian gauge theories.” An alternative<br />
term is simply “Yang-Mills theories.”<br />
The eight gauge bosons of QCD are referred<br />
to by the bastardized term “gluon,”<br />
since they represent the glue that holds the<br />
physical hadrons, such as the proton,<br />
together. The interactions of gluons with<br />
quarks are depicted in Fig. 10. Although<br />
gluons are the counterpart to photons in that<br />
they have unit spin and are massless, they<br />
possess one crucial property not shared by<br />
photons: they themselves carry color. Thus<br />
they not only mediate the color force but also<br />
carry it; it is as if photons were charged. This<br />
difference (it is the difference between an<br />
Abelian arid a non-Abelian gauge theory) has<br />
many profound physical consequences. For<br />
example, becauie gluons carry color they can<br />
(unlike photons) interact with themselves<br />
(see Fig. IO) and, in effect, weaken the force<br />
of the color charge at short distances. The<br />
opposite effect occurs in quantum electrodynamics:<br />
screening effects weaken the<br />
effective electric charge at long distances. (As<br />
mentioned above, a virtual photon emanating<br />
from an electron can create a virtual<br />
electron-positron pair. This polarization<br />
screens, or effectively decreases, the electron’s<br />
charge.)<br />
The weakening of color charge at short<br />
distances goes by the name of asymptotic<br />
freedom. Asymptotic freedom was first ob-<br />
I<br />
r<br />
Gluon Self-I nteractions<br />
9<br />
‘b I<br />
3 Quark Colors 8 Colored Gluons<br />
Fig. 10. The SU(3) local color symmetry implies the existence of eight massless<br />
gaugefields (the gluons) to compensate for the eight independent local transformations<br />
of the colored quark fields. The subscripts r, g, and b on the gluon and quark<br />
fields correspond respectively to red, green, and blue color charges. The eight<br />
gluons carry color and obey the non-Abelian algebra of the SU(3) generators (see<br />
Lecture Note 4). The interactions induced by the local SU(3) color symmetry<br />
include a quark-gluon coupling as well as two types of gluon self interactions (one<br />
proportional to the couping g, and the other proportional to g:).<br />
I<br />
I<br />
I<br />
I<br />
I<br />
40
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
H<br />
ow can we extract answers fr<br />
QCD at energies below 1 Ge<br />
As noted in the text, the confinement<br />
of quarks suggests that weak-couphng<br />
perturbative methods arc not going to be<br />
successful at these energies. Nevertheless, if<br />
QCD is a valid theory it must explain the<br />
multiplicities, masses, and couplings of the<br />
experimentally observed strongly interacting<br />
particles. These would emerge from the theory<br />
as bound states and resonances of quarks<br />
and gluons. A valid theory must also account<br />
for the apparent absence of isolated quark<br />
states and might predict the existence and<br />
properties of particles (such as glueballs) that<br />
have not yet been seen.<br />
The most promising nonperturbative formulation<br />
of QCD exploits the Feynman path<br />
integral. Physical quantities are expressed as<br />
integrals of the quark and gluon fields over<br />
the space-time continuum with the QCD<br />
Lagrangian appearing in an exponential as a<br />
kind of Gibbs weight factor. This is directly<br />
analogous to the partition function formulation<br />
of statistical rnachanics. The path integral<br />
prescription for strong interaction<br />
dynamics becomes well defined mathematically<br />
when the space-time continuum is<br />
approximated by a discrete four-dimensional<br />
lattice of finite size and the integrals are<br />
evaluated by Monte Carlo sampling.<br />
The original Monte Carlo ideas of Metropolis<br />
and Ulam have now been applied to<br />
QCD by many researchers. These efforts<br />
have given credibility, but not confirmation,<br />
to the hope that computer simulations might<br />
indeed provide critical tests of QCD and<br />
significant numerical results. With considerable<br />
patience (on the order of many months<br />
ofcomputer tirne)a VAX 11/780can be used<br />
to study universes of about 3000 space-time<br />
points, Such a universe is barely large enough<br />
to contain a proton and not really adequate<br />
for a quantitative calculation. Consequently,<br />
with these methods, any result from a computer<br />
of VAX power is, at best, only an<br />
indication of what a well-done numerical<br />
simulation might produce.<br />
by Gerald Guralnik, Tony Warnock, and Charles Zemach<br />
physical and mathematical ingenuity to<br />
search out the best formulations of problems<br />
ming; and (3) a computer with the speed,<br />
memory, and input/output rate of the Cray<br />
Cray XMP. Using new meth<br />
with coworkers R. Gupta, J.<br />
A. Patel, we are examining gl<br />
renormalization group behavior, and the<br />
behavior of the theory on much larger lattices.<br />
The results to date support the belief<br />
that QCD describes interactions of the<br />
elementary particles and that these numerical<br />
methods are currently the most powerful<br />
means for extracting the predictive content<br />
of QCD.<br />
p meson 767<br />
Excited p<br />
I426<br />
6 meson 1154<br />
A, meson 1413<br />
Proton 989<br />
A baryon<br />
1 I99<br />
Couplings<br />
The calculations, which have two input<br />
parameters (the pion mass and the longrange<br />
quark-quark force constant in units of<br />
the lattice spacing), provide estimates of<br />
ble quantities. The accompaows<br />
some of our results on<br />
elementary particle masses and certain<br />
meson coupling strengths. These results represent<br />
several hundred hours of Cray timc.<br />
The quoted relative errors derive from the<br />
statistical analysis of the Monte Carlo calculation<br />
itself rather than from a comparison<br />
with experimental data. Significantly more<br />
computer time would significantly reducc<br />
the errors in the calculated masses and couplings.<br />
Our work would not have been possible<br />
without the support of C Division and many<br />
of its staff. We have received generous support<br />
from Cray Research and are particularly<br />
indebted to Bill Dissly and George Spix for<br />
contribution of their skills and their time.<br />
Calculated and experimental values for the masses and coupling<br />
strengths of some mesons and baryons.<br />
Masses<br />
fn<br />
fP<br />
Calculated Relative Experimental<br />
Value Error Value<br />
(MeV/c2) (O/O) (MeV/c2)<br />
18<br />
27<br />
15<br />
17<br />
23<br />
17<br />
769<br />
1300?<br />
983<br />
1275<br />
940<br />
1210<br />
121 21 93<br />
21 1 15 144<br />
41
I<br />
observable hadrons are necessarily colorless,<br />
whereas quarks and gluons are permanently<br />
confined. This is just as well since gluons are<br />
massless, and by analogy with the photon,<br />
unconfined massless gluons should give rise<br />
to a long-range, Coulomb-like, color force in<br />
the strong interactions. Such a force is clearly<br />
at variance with experiment! Even though<br />
color is confined, residual strong color forces<br />
can still “leak out” in the form of colorneutral<br />
pions or other hadrons and be responsible<br />
for the binding of protons and<br />
neutrons in nuclei (much as residual electromagnetic<br />
forces bind atoms together to<br />
form molecules).<br />
The success of QCD in explaining shortdistance<br />
behavior and its aesthetic appeal as<br />
a generalization of QED have given it its<br />
place in the standard model. However, confidence<br />
in this theory still awaits convincing<br />
calculations of phenomena at distance scales<br />
of lo-” centimeter, where the “strong”<br />
nature of the force becomes dominant and<br />
perturbation theory is no longer valid. (Lattice<br />
gauge theory calculations of the hadronic<br />
spectrum are becoming more and more reliable.<br />
See “QCD on a Cray: The Masses of<br />
E!ementary <strong>Particle</strong>s.”)<br />
The Weak Interaction<br />
I<br />
served in deep inelastic scattering experiments<br />
(see “Scaling in Deep Inelastic Scattering”).<br />
This phenomenon explains why<br />
hadrons at high energies behave as if they<br />
were made of almost free quarks even though<br />
one knows that quarks must be tightly bound<br />
together since they have never been experimentally<br />
observed in their free state. The<br />
weakening of the force at high energies<br />
means that we can use perturbation theory to<br />
calculate hadronic processes at these<br />
energies.<br />
42<br />
The self-interaction of the gluons also explains<br />
the apparently permanent confinement<br />
of quarks. At long distances it leads to<br />
such a proliferation of virtual gluons that the<br />
color charge effectively grows without limit,<br />
forbidding the propagation of all colored<br />
,particles. Only bleached, or color-neutral,<br />
states (such as baryons, which have equal<br />
proportions of red, blue, and green, or<br />
mesons which have equal proportions of redantired,<br />
green-antigreen, and blue-antiblue)<br />
are immune from this c:onfinement. Thus all<br />
alpha particles, beta rays, and gamma rays.<br />
These three are now associated with three<br />
quite different modes of decay. An alpha<br />
particle, itself a helium nucleus, is emitted<br />
during the strong-interaction decay mode<br />
known as fission. Large nuclei that are only<br />
loosely bound by the strong force (such as<br />
uranium-238) can split into two stable<br />
pieces, one of which is an alpha particle. A<br />
gamma ray is simply a photon with “high”<br />
energy (above a few MeV) and is emitted<br />
during the decay of an excited nucleus. A<br />
beta ray is an electron emitted when a neutron<br />
in a nucleus decays into a proton, an<br />
electron, and an electron antineutrino (n-p<br />
+e-+ Cc, see Fig. 1 1). The proton remains in
I-<br />
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Positive Weak Currents JLeak<br />
the nucleus, and the electron and its antineutrino<br />
escape. This decay mode is<br />
characterized as weak because it proceeds<br />
much more slowly than most electromagnetic<br />
decays (see Table I). Other<br />
baryons may also undergo beta decay.<br />
Beta decay remained very mysterious for a<br />
long time because it seemed to violate<br />
energy-momentum conservation. The free<br />
neutron was observed to decay into two<br />
particles, a proton and an electron, each with<br />
a spectrum of energies, whereas energymomentum<br />
conservation dictates that each<br />
should have a unique energy. To solve this<br />
dilemma, Pauli invoked the neutrino, a<br />
massless, neutral fermion that participates<br />
only in weak interactions.<br />
I<br />
Fig. ll. (a) Components of the charge-raising weak current Jtk<br />
are represented in<br />
the figure by Feynman diagrams in which a neutron changes into a proton, an<br />
electron into an electron neutrino, and a muon into a muon neutrino. The chargelowering<br />
current Jieak is represented by reversing the arrows. (b) Beta decay<br />
(shown in thefigure) and other low-energy weak processes are well described by the<br />
Fermi interaction Jieak X Jieak. The figure shows the Feynman diagram of the<br />
Fermi interaction for beta decay.<br />
The Fermi Theory. Beta decay is just one of<br />
many manifestations ofthe weak interaction.<br />
By the 1950s it was known that all weak<br />
processes could be concisely described in<br />
terms of the current-current interaction first<br />
proposed in 1934 by Fermi. The charged<br />
weak currents &ak and Jieak change the<br />
electric charge of a fermion by one unit and<br />
can be represented by the sum of the Feynman<br />
diagrams of Fig. I la. In order to describe<br />
the maximal parity violation, (that is,<br />
the maximal right-left asymmetry) observed<br />
in weak interactions, the charged weak current<br />
includes only left-handed fermion fields.<br />
(These are defined in Fig. 12 and Lecture<br />
Note 8.)<br />
Fermi’s current-current interaction is then<br />
given by all the processes included in the<br />
product (G~/fi)(J+weak X JGeak) where<br />
Jieak means all arrows in Fig. 1 la are reversed.<br />
This interaction is in marked contrast<br />
to quantum electrodynamics in which<br />
two currents interact through the exchange of<br />
a virtual photon (see Fig. 3). In weak<br />
processes two charge-changing currents appear<br />
to interact locally (that is, at a single<br />
point) without the help of such an intermediary.<br />
The coupling constant for this local<br />
interaction, denoted by GF and called the<br />
Fermi constant, is not dimensionless like the<br />
coupling parameter a in QED, but has the<br />
43
Left-Handed<br />
<strong>Particle</strong> State<br />
I<br />
2 +<br />
S<br />
Fig. 12. A Dirac spinor field can be decomposed into leftand<br />
right-handed pieces. A left-handed field creates two<br />
types of particle states at ultrarelativistic energies-u, a<br />
particle with spin opposite to the direction of motion, and urn<br />
an antiparticle with spin along the direction of motion. Only<br />
lef-handed fields contribute to the weak charged currents<br />
shown in Fig. 11. The left- and right-handedness (or<br />
chirality) of a field describes a Lorentz covariant decomposition<br />
of Dirac spinorfields.<br />
dimension of mass-’ or energy-’ . In units of<br />
energy, the measured value of GF”’ equals<br />
293 CeV. Thus the strength of the weak<br />
processes seems to be determined by a specific<br />
energy scale. But why?<br />
Predictions of the W boson. An explanation<br />
emerges if we postulate the existence of an<br />
intermediary for the weak interactions. Recall<br />
from Fig. 3 that the exchanged, or virtual,<br />
photons in QED basically correspond to<br />
the Coulomb potential a/r, whose Fourier<br />
transform is a/$, where q is the momentum<br />
of the virtual photon. It is tempting to suggest<br />
that the nearly zero range of the weak<br />
interaction is only apparent in that the two<br />
charged currents interact through a potential<br />
of the form a’[exp(-M~.r)]/r (a form originally<br />
proposed by Yukawa for the short-<br />
44<br />
range force between nucleons), where a’ is<br />
the analogue of a and the mass MI+/ is so large<br />
that this potential has essentially no range.<br />
The Fourier transform of this potential,<br />
a’/($ + ML), suggests that, if this idea is<br />
correct, the interaction between the weak<br />
currents is mediated by a “heavy photon” of<br />
mass Mw. Nowadays this particle is called<br />
the W boson; its existence explains the short<br />
range of the weak interactions.<br />
Notice that at low energies, or, equivalently,<br />
when M& >> $, the Fourier transform,<br />
or so-called propagator of the W<br />
boson, reduces to a’/(M&r), and since this<br />
factor multiplies the two currents, it must be<br />
proportional to Fermi’s constant. Thus the<br />
existence of the PV boson gives a natural<br />
explanation ofwhy GF is not dimensionless.<br />
Now, since both the weak and electro-<br />
magnetic interactions involve electric<br />
charge, these two might be manifestations ol<br />
the same basic force. If they were, then a’<br />
might be the same as a and GF would be<br />
proportional to a/M$. Thus the existence 01<br />
a very massive W boson can explain not onl)<br />
the short range but also the weakness ofweak<br />
interactions relative to electromagnetic interactions!<br />
This argument not only predict:<br />
the existence of a W boson but also yields a<br />
rough estimate of its mass:<br />
M~ = = 25 GeV/c2.<br />
I<br />
beyond reach of the existing accelerators.<br />
physicists that a theoretical unification o
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Table 2<br />
Multiplets and quantum numbers in the SU(2) X U(1) electroweak theory.<br />
Quarks<br />
--<br />
1 UL<br />
I<br />
SU(2) Doublet<br />
I<br />
SU(2) Singlets<br />
SU(2) Doublet<br />
SU(2)Singlet<br />
Leptons<br />
Gauge Bosons<br />
1-6 i<br />
UR __<br />
1- dR_i<br />
I(Vi)2<br />
I<br />
PL!<br />
- -<br />
eR I<br />
W+l<br />
SU(2) Triplet W3‘<br />
’ w-/<br />
SU(2) Singlet B- 1<br />
SU(2) Doublet<br />
c---<br />
Higgs Boson<br />
__-<br />
‘p+ ’<br />
Weak Weak Electric<br />
Isotopic Hypercharge Charge Q<br />
Charge I3 Y (=Z3 + ‘/2 Y)<br />
112<br />
- Y2<br />
0<br />
0<br />
‘/l<br />
- Y2<br />
0<br />
1<br />
0<br />
-1<br />
electromagnetic and weak interactions must<br />
be possible. Several attempts were made in<br />
the 1950s and 1960s, notably by Schwinger<br />
and his student Glashow and by Ward and<br />
Salam, to construct an “electroweak theory”<br />
in terms of a local gauge (Yang-Mills) theory<br />
that generalizes QED. Ultimately, Weinberg<br />
set forth the modern solution to giving<br />
0<br />
Y2<br />
--%<br />
‘/3<br />
Y3<br />
413<br />
-2/3<br />
-I<br />
-1<br />
-2<br />
0<br />
0<br />
0<br />
0<br />
1<br />
1<br />
- Y3<br />
2/3<br />
-%<br />
0<br />
-1<br />
-1<br />
1<br />
0<br />
-1<br />
masses to the weak bosons in 1967, although<br />
it was not accepted as such until ’t Hooft and<br />
Veltman showed in 1971 that it constituted a<br />
consistent quantum field theory. The success<br />
of the electroweak theory culminated in 1982<br />
with the discovery at CERN of the W boson<br />
at almost exactly the prediced mass. Notice,<br />
incidentally, that at sufficiently high<br />
0<br />
1<br />
0<br />
energies, where q2 >> ML, the Leak interaction<br />
becomes comparable in strength to the<br />
electromagnetic. Thus we see explicitly how<br />
the apparent strength of the interaction depends<br />
on the wavelength of the probe.<br />
The SU(2) X U(1) Electroweak Theory.<br />
Since quantum electrodynamics is a gauge<br />
theory based on local U(I) invariance; it is<br />
not too surprising that the theory unifying<br />
the electromagnetic and weak forces is also a<br />
gauge theory. Construction of such a theory<br />
required overcoming both technical and phenomenological<br />
problems.<br />
The technical problem concerned the fact<br />
that an electroweak gauge theory is<br />
necessarily a Yang-Mills theory (that is, a<br />
theory in which the gauge fields interact with<br />
each other); the gauge fields, namely the W<br />
bosons, must be charged to mediate the<br />
charge-changing weak interactions and therefore<br />
by definition must interact with each<br />
other electromagnetically through the<br />
photon. Moreover, the local gauge symmetry<br />
of the theory must be broken because an<br />
unbroken symmetry would require all the<br />
gauge particles to be massless like the photon<br />
and the gluons, whereas the W boson must<br />
be massive. A major theoretical difficulty<br />
was understanding how to break a Yang-<br />
Mills gauge symmetry in a consistent way.<br />
(The solution is presented below.)<br />
In addition to the technical issue, there<br />
was the phenomenological problem of choosing<br />
the correct local symmetry group. The<br />
most natural choice was SU(2) because the<br />
low-lying states (that is, the observed quarks<br />
and leptons) seemed to form doublets under<br />
the weak interaction. For example, a W-<br />
changes vc into e, v,, into p, or u into d (where<br />
all are left-handed fields), and the W+ effects<br />
the reverse operation. Moreover, the three<br />
gauge bosons required to compensate for the<br />
three independent phase rotations of a local<br />
SU(2) symmetry could be identified with the<br />
W’, the W-, and the photon. Unfortunately,<br />
this simplistic scenario does not<br />
work: it gives the wrong electric charge assignments<br />
for the quarks and leptons in the<br />
SU(2) doublets. Specifically, electric charge<br />
45
Q would be equal to the SU(2) charge f3, and<br />
the values of f 3 for a doublet are fY2. This is<br />
clearly the wrong charge. In addition, SU(2)<br />
would not distinguish the charges of a quark<br />
doublet (*/I and 4 3 ) from those of a lepton<br />
doublet (0 and -1).<br />
To get the correct charge assignments, we<br />
can either put quarks and leptons into SU(2)<br />
triplets (or larger multiplets) instead of<br />
doublets, or we can enlarge the local symmetry<br />
group. The first possibility requires<br />
the introduction of new heavy fermions to<br />
fill the multiplets. The second possibility<br />
requires the introduction of at least one new<br />
U( 1) symmetry (let's call it weak hypercharge<br />
Y), which yields the correct electric charge<br />
assignments if we define<br />
This is exactly the possibility that has been<br />
confirmed experimentally. Indeed, the electroweak<br />
theory of Glashow, Salam, and<br />
Weinberg is a local gauge theory with the<br />
symmetry group SU(2) X U(1). Table 2 gives<br />
the quark and lepton multiplets and their<br />
associated quantum numbers under SU(2) X<br />
U(I), and Fig. 13 displays the interactions<br />
defined by this local symmetry. There is one<br />
coupling associated with each factor of SU(2)<br />
X U( I), a couplingg for SU(2) and a coupling<br />
g'/2 for U( I).<br />
The addition of the local U( 1) symmetry<br />
introduces a new uncharged gauge particle<br />
into the theory that gives rise to the so-called<br />
neutral-current interactions. This new type<br />
of weak interaction, which allows a neutrino<br />
to interact with matter without changing its<br />
identity, had not been observed when the<br />
neutral weak boson was first proposed in<br />
1961 by Glashow. Not until 1973, after all<br />
the technical problems with the SU(2) X<br />
U( 1) theory had been worked out, were these<br />
interactions observed in data taken at CERN<br />
in 1969 (see Fig. 14).<br />
The physical particle that mediates the<br />
weak interaction between neutral currents is<br />
the massive Zo. The electromagnetic interaction<br />
between neutral currents is mediated by<br />
the familiar massless photon. These two<br />
46
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
physical particles are different from the two<br />
neutral gauge particles (Band W3) associated<br />
with the unbroken SU(2) X U(l) symmetry<br />
shown in Fig. 13. In fact, the photon and the<br />
Zo are linear combinations of the neutral<br />
gauge particles W3 and B:<br />
A=BcosBw+ W3sinBw<br />
and<br />
Zo = B sin Bw - W3 COS OW<br />
The mixing ofSU(2) and U( 1) gauge particles<br />
to give the physical particles is one result<br />
of the fact that the SU(2) X U( 1) symmetry<br />
must be a broken symmetry.<br />
Spontaneous Symmetry Breaking. The astute<br />
reader may well be wondering how a local<br />
gauge theory, which in QED required the<br />
photon to be massless, can allow the<br />
mediator of the weak interactions to be<br />
massive, especially since the two forces are to<br />
be unified. The solution to this paradox lies<br />
in the curious way in which the SU(2) X U( 1)<br />
symmetry is broken.<br />
As Nambu described so well, this breaking<br />
is very much analogous to the symmetry<br />
breaking that occurs in a superconductor. A<br />
superconductor has a local U( 1) symmetry,<br />
namely, electromagnetism. The ground state,<br />
however, is not invariant under this symmetry<br />
since it is an ordered state of bound<br />
electron-electron pairs (the so-called Cooper<br />
pairs) and therefore has a nonzero electric<br />
charge distribution. As a result of this asymmetry,<br />
photons inside the superconductor<br />
acquire an effective mass, which is responsible<br />
for the Meissner effect. (A magnetic field<br />
cannot penetrate into a superconductor; at<br />
the surface it decreases exponentially at a<br />
rate proportional to the effective mass of the<br />
photon.)<br />
In the weak interactions the symmetry is<br />
also assumed to be broken by an asymmetry<br />
of the ground state, which in this case is the<br />
“vacuum.” The asymmetry is due to an ordered<br />
state of electrically neutral bosons that<br />
carry the weak charge, the so-called Higgs<br />
bosons. They break the SU(2) X U(l) sym-<br />
47
metry to give the U( 1) of electromagnetism<br />
in such a way that the W' and the Zo obtain<br />
masses and the photon remains massless. As<br />
a result the charges 13 and Y associated with<br />
SU(2) X U(1) are not conserved in weak<br />
processes because the vacuum can absorb<br />
these quantum numbers. The electric charge<br />
Q associated with U( 1) of electromagnetism<br />
remains conserved.<br />
The asymmetry of the ground state is frequently<br />
referred to as spontaneous symmetry<br />
breaking; it does not destroy the symmetry of<br />
the Lagrangian but destroys only the symmetry<br />
of the states. This symmetry breaking<br />
mechanism allows the electroweak Lagrangian<br />
to remain invariant under the local<br />
symmetry transformations while the gauge<br />
particles become massive (see Lecture Notes<br />
3, 6, and 8 for details).<br />
In the spontaneously broken theory the<br />
electromagnetic coupling e is given by the<br />
expression e = gsin Ow, where<br />
sin2ew = g2/(g2 + gt2) .<br />
Thus, e and Ow are an alternative way of<br />
expressing the couplings g and g', and just as<br />
e is not determined in QED, the equally<br />
important mixing angle Ow is not determined<br />
by the electroweak theory. It is, however,<br />
measured in the neutral-current interactions.<br />
The experimental value is sin2 Ow = 0.224 k<br />
0.015. The theory predicts that<br />
These relations (which are changed only<br />
slightly by small quantum corrections) and<br />
the experimental value for the weak angle Ow<br />
predict masses for the W' and Zo that are in<br />
very good agreement with the 1983 observations<br />
of the W' and Zo at CERN.<br />
In the electroweak theory quarks and leptons<br />
also o.btain mass by interacting with the<br />
ordered vacuum state. However, the values<br />
of their masses are not predicted by the<br />
Fig. 14. Neutral-current interactions were first identified in 1973 in photographs<br />
taken with the CEUN Gargamelle bubble chamber. The figure illustrates the<br />
difference between neutral-current and charged-current interactions and shows the<br />
bubble-chamber signature of each. The bubble tracks are created by charged<br />
particles moving through superheated liquid freon. The incoming antineutrinos<br />
interact with protons in the liquid. A neutral-current interaction leaves no track<br />
from a lepton, only u track from the positivley charged proton and perhaps some<br />
tracks from pions. A charged-current interaction leaves a track from a positively<br />
charged muon only.<br />
48
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
rons or protons. U<br />
of the<br />
fortv-<br />
Beam<br />
49
theory but are proportional to arbitrary<br />
parameters related to the strength of the<br />
coupling of the quarks and leptons to the<br />
Higgs boson.<br />
The Higgs Boson. In the simplest version of<br />
the spontaneously broken electroweak<br />
model, the Higgs boson is a complex SU(2)<br />
doublet consisting of four real fields (see<br />
Table 2). These four fields are needed to<br />
transform massless gauge fields into massive<br />
ones. A massless gauge boson such as the<br />
photon has only two orthogonal spin components<br />
(both transverse to the direction of<br />
motion), whereas a massive gauge boson has<br />
three (two transverse and one longitudinal,<br />
that is, in the direction of motion). In the<br />
electroweak theory the W’, the W-, and the<br />
Zo absorb three of the four real Higgs fields<br />
to form their longitudinal spin components<br />
and in so doing become massive. In more<br />
picturesque language, the gauge bosons “eat”<br />
the Higgs boson and become massive from<br />
the feast. The remaining neutral Higgs field<br />
is not used up in this magic transformation<br />
from massless to massive gauge bosons and<br />
therefore should be observable as a particle<br />
in its own right. Unfortunately, its mass is<br />
not fixed by the theory. However, it can<br />
decay into quarks and leptons with a definite<br />
signature. It is certainly a necessary component<br />
of the theory and is presently being<br />
looked for in high-energy experiments at<br />
CERN. Its absence is a crucial missing link in<br />
the confirmation of the standard model.<br />
Open Problems. Our review of the standard<br />
model would not be complete without mention<br />
of some questions that it leaves unanswered.<br />
We discussed above how the three<br />
charged leptons (e, p, and T) may form a<br />
triplet under some broken symmetry. This is<br />
only part of the story. There are, in fact, three<br />
quark-lepton families (Table 3), and these<br />
three families may form a triplet under such<br />
a broken symmetry. (There is a missing state<br />
in this picture: conclusive evidence for the<br />
top quark / has yet to be presented. The<br />
bottom quark b has been observed in<br />
e’e-annihilation experiments at SLAC and<br />
so<br />
w-, w+, zo<br />
I<br />
n January 1983 two groups announced the results ofseparate searches at the CERN<br />
llider for the W- and W+ vector bosons of the electroweak<br />
headed by C. Rubbia and A. Astbury, reported definite<br />
identification, from among about a billion proton-antiproton collisions, of four W-<br />
decays and one W+ decay. The mass reported by this group (81 k 5 GeV/c2) agrees well<br />
predicted by the electro ode1 (82 k 2.4 GeVlc’). The other group,<br />
P. Darriulat and using nt detector, reported identification of four<br />
possible W* decays, again from among a billion events. The charged vector bosons<br />
were produced by annihilation of a quark inside a proton (uud) with an antiquark<br />
inside an antiproton (&a):<br />
and<br />
d+G-+ W-<br />
u+a+<br />
w+.<br />
Since these reactions have a threshold energy equal to the mass of the charged bosons,<br />
the colliding proton and antiproton beams were each accelerated to about 270 GeV to<br />
provide the quarks with an average center-of-mass energy slightly above the threshold<br />
energy. (Only one-half of the energy of a proton or antiproton is camed by its three<br />
quark constituents; the other half is carried by the gluons.) Rubbia’s group distinguished<br />
the two-body decay of the bosons (into a charged and neutral lepton pair<br />
such as P’v,) by two methods: selection of events in which the charged lepton<br />
possessed a large momentum transverse to the axis of the colliding beams, and<br />
selection of events in which a large amount of energy appeared to be missing,<br />
presumably camed off by the (undetected) neutrino. Both methods converged on the<br />
same events.<br />
By mid 1983 each of the two groups had succeeded also in finding Zo, the neutral<br />
vector boson of the electroweak model. They reported slightly different mass values<br />
(96.5 k 1.5 and 91.2 k 1.7 GeVfc’), both in agreement with the predicted value of 94.0<br />
2 2.5 GeVfc’. For Zo the production and decay processes are given by<br />
ui- ;(or d+ 2)- Zo -f e-<br />
+ e+ (or p-i- p+) .<br />
In addition, both groups reported an asymmetry in the angular distribution of<br />
charged leptons from the many more decays of W- and W+ that had been seen<br />
since their discovery. This parity violation confirmed that the particles observed<br />
are truly electroweak vector bosons. W
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Table 3<br />
The thre masses. Th note right- and left-handed<br />
particles<br />
Quark Mass<br />
(MeV/c2)<br />
Quarks Leptons Lepton Mass<br />
( MeV/c2)<br />
First Family<br />
5<br />
8<br />
UP<br />
down<br />
UL<br />
dL<br />
UR (v,}~ electron neutrino 0<br />
dR eL 6 electron 0.51 1<br />
Second Family<br />
1270<br />
175<br />
charm cL CR<br />
strange SL SR<br />
(v&<br />
PI+ PR<br />
muon neutrino<br />
muon<br />
0<br />
105.7<br />
Third Family<br />
45000 (?)<br />
4250<br />
top tL<br />
bottom bL<br />
tR (VJL tau neutrino 0<br />
bR TL TR tau 1784<br />
Cornell.) The standard model says nothing<br />
about why three identical families of quarks<br />
and leptons should exist, nor does it give any<br />
clue about the hierarchical pattern of their<br />
masses (the 7 family is heavier than the p<br />
family, which is heavier than the e family).<br />
This hierarchy is both puzzling and intriguing.<br />
Perhaps there are even more undiscovered<br />
families connected to the broken<br />
family symmetry. The symmetry could be<br />
global or local, and either case would predict<br />
new, weaker interactions among quarks and<br />
leptons.<br />
Table 3 brings up two other open questions.<br />
First, we have listed the neutrinos as<br />
being massless. Experimentally, however,<br />
there exist only upper limits on their possible<br />
masses. The most restrictive limit comes<br />
from cosmology, which rcquires the sum of<br />
neutrino masses to be less then 100 eV. It is<br />
known from astrophysical observations that<br />
most of the energy in the universe is in a<br />
form that does not radiate electromagnetically.<br />
If neutrinos have mass, they<br />
could, in fact, be the dominant form of<br />
energy in the universe today.<br />
Second, we have listed u and d, c and s,<br />
and t and b as doublets under weak SU(2).<br />
This is, however, only approximately true.<br />
As a result of the broken family symmetry,<br />
states with the same electric charge (the d, s,<br />
and b quarks or the u, c, and t quarks) can<br />
mix, and the weak doublets that couple to the<br />
W' bosons are actually given by IJ and d',<br />
c and s', and t and b'. A 3 X 3 unitary matrix<br />
known as the Kobayashi-Maskawa (K-M)<br />
matrix rotates the mass eigenstates (states of<br />
definite mass) d, s, and b into the weak<br />
doublet states d', s', and b'. The K-M matrix<br />
is conventionally written in terms of three<br />
mixing angles and an arbitrary phase. The<br />
largest mixing is between the d and s quarks<br />
and is characterized by the Cabibbo angle<br />
€IC (see Lecture Note 9),which is named for<br />
the man who studied strangeness-changing<br />
weak decays such as Co - p + e- + V,. The<br />
observed value of sin 8C is about 0.22. The<br />
other mixing angles are all at least an order of<br />
magnitude smaller. The structure of the K-M<br />
matrix, like the masses of the quarks and<br />
leptons, is a complete mystery.<br />
Conclusions<br />
Although many mysteries remain, the<br />
standard model represents an intriguing and<br />
compelling theoretical framework for our<br />
present-day knowledge of the elementary<br />
particles. Its great virtue is that all of the<br />
known forces can be described as local gauge<br />
theories in which the interactions are generated<br />
from the single unifying principle of<br />
local gauge invariance. The fact that in quantum<br />
field theory interactions can drastically<br />
change their character with scale is crucial to<br />
51
I<br />
and the electr<br />
v)<br />
f 103<br />
I]<br />
0<br />
C<br />
(c1<br />
C<br />
3.10 3.12 3.14<br />
Center-of-Mass Energy (GeV)<br />
Graph of the evidence for formation of J/yr in electron- graph of their evidence. (Photo courtesy of the Niels<br />
positron annihilations at SPEAR. (A rn SLAC Library of the American Instiru <strong>Physics</strong><br />
e, Volume 7, Number 11, Novem 6.)<br />
I<br />
52
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
T<br />
I<br />
n I977 a group led by L. Lederman provided evidence for a fifth, or bottom, quark<br />
with the discovery of”, a long-lived particle three times more massive than J/y. In<br />
an experiment similar to that of Ting and coworkers and performed at the<br />
Fermilab proton accelerator, the group determined the number of events giving rise to<br />
muon-antimuon pairs as a function of the mass ofthe parent particle and found a sharp<br />
increase at about 9.5 GeV/cZ. Like the J /~I system, the ‘I’ system has been elucidated in<br />
detail from experiments involving electron-positron collisions rather than proton<br />
collisions, in this case at Cornell’s electron sto<br />
The existence ofthe bottom quark, and of a sixth, or top, quark, was expected on the<br />
basis of the discovery of the tau lepton at SPEAR in 1975 and Glashow and Bjorken’s<br />
I964 argument of quark-lepton symmetry. Recent results from high-energy protonantiproton<br />
collision experiments at CERN have been interpreted as possible evidence<br />
for the top quark with a mass somewhere between 30 and 50 GeV/c2.<br />
this approach. The essence of the standard<br />
model is to put the physics of the apparently<br />
separate strong, weak, and electromagnetic<br />
interactions in the single language of local<br />
gauge field theories, much as Maxwell put<br />
the apparently separate physics of<br />
Coulomb’s, Ampere’s, and Faraday’s laws<br />
into the single language of classical field theory.<br />
It is very tempting to speculate that, because<br />
of the chameleon-like behavior of<br />
quantum field theory, all the interactions are<br />
simply manifestations of a single field theory.<br />
Just as the “undetermined parameters”<br />
EO and po were related to the velocity of light<br />
through Maxwell’s unification of electricity<br />
and magnetism, so the undetermined<br />
parameters of the standard model (such as<br />
quark and lepton masses and mixing angles)<br />
might be fixed by embedding the standard<br />
model in some grand unified theory.<br />
A great deal of effort has been focused on<br />
this question during the past few years, and<br />
some of the problems and successes are discussed<br />
in “Toward a Unified Theory” and<br />
“Supersymmetry at 100 GeV.” Although<br />
hints of a solution have emerged, it is fair to<br />
say that we arc still a long way from for-<br />
mulating an ultimate synthesis ofall physical<br />
laws. Perhaps one of the reasons for this is<br />
that the role of gravitation still remains mysterious.<br />
This weakest of all the forces, whose<br />
effects are so dramatic in the macroscopic<br />
world, may well hold the key to a truly deep<br />
understanding of the physical world. Many<br />
particle physicists are therefore turning their<br />
attention to the Einsteinian view in which<br />
geometry becomes the language of expression.<br />
This has led to many weird and<br />
wonderful speculations concerning higher<br />
dimensions, complex manifolds, and other<br />
arcane subjects.<br />
An alternative approach to these questions<br />
has been to peel yet another skin off the<br />
onion and suggest that the quarks and leptons<br />
are themselves composite objects made<br />
of still more elementary objects called<br />
preons. After all, the prolifcration of quarks,<br />
leptons, gauge bosons, and Higgs particles is<br />
beginning to resemble the situation in the<br />
early 1960s when the proliferation of the<br />
observed hadronic states made way for the<br />
introduction of quarks. Maybe introducing<br />
preons can account for the mystery of flavor:<br />
e, p, and T, for example, may simply be<br />
bound states of such objects.<br />
Regardless of whether the ultimate understanding<br />
of the structure of matter, should<br />
there be one, lies in the realm of preons,<br />
some single primitive group, higher<br />
dimensions, or whatever, the standard<br />
model represents the first great step in that<br />
direction. The situation appears ripe for<br />
some kind of grand unification. Where are<br />
you, Maxwell?<br />
I<br />
Further Reading<br />
Gerard ’I Hooft. “Gauge Theories of the Forces Between Elementary <strong>Particle</strong>s.” Scieniific American, June 1980, pp. 104-137.<br />
Howard Georgi. “A Unified Theory of Elementary <strong>Particle</strong>s and Forces.” Scienfijic American, April 1981, pp. 48-63.<br />
53
Lecture Notes<br />
fiom simple field theories to the standard model<br />
by Richard C. Slansky<br />
The standard model of electroweak and strong interactions<br />
consists of two relativistic quantum field theories, one to<br />
describe the strong interactions and one to describe the<br />
- electromagnetic and weak interactions. This model, which<br />
incorporates all the known phenomenology of these fundamental<br />
interactions, describes spinless, ~pin-~h, and spin-l fields interacting<br />
with one another in a manner determined by its Lagrangian. The<br />
theory is relativistically invariant, so the mathematical form of the<br />
Lagrangian is unchanged by Lorentz transformations.<br />
Although rather complicated in detail, the standard model Lagrangian<br />
is based on just two basic ideas beyond those necessary for a<br />
quantum field theory. One is the concept of local symmetry, which is<br />
encountered in its simplest form in electrodynamics. Local symmetry<br />
r -<br />
--<br />
determines the form of the interaction between particles, or fields,<br />
that carry the charge associated with the symmetry (not necessarily<br />
the electric charge). The interaction is mediated by a spin-I particle,<br />
the vector boson, or gauge particle. The second concept is spontaneous<br />
symmetry breaking, where the vacuum (the state with no<br />
particles) has a nonzero charge distnbution. In the standard model<br />
the nonzero weak-interaction charge distribution of the vacuum is<br />
the source of most masses of the particles in the theory. These two<br />
basic ideas, local symmetry and spontaneous symmetry breaking, are<br />
exhibited by simple field theones. We begin these lecture notes with a<br />
Lagrangian for scalar fields and then, through the extensions and<br />
generalizations indicated by the arrows in the diagram below, build<br />
up the formalism needed to construct the standard model.<br />
~-<br />
I<br />
i<br />
@) Quarks<br />
Future Theories ?<br />
See Article on<br />
Unified Theories<br />
54
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Fields, Lagrangians, 6~(tI,tz)= J2 [ G6q +<br />
and Equations<br />
of Motion<br />
We begin this introduction to field theory with one of the simplest<br />
theories, a complex scalar field theory with independent fields cp(x)<br />
and cpt(x). (cpt(x) is the complex conjugate of cp(x) if cp(x) is a classical<br />
field, and, if cp(x) is generalized to a column vector or to a quantum<br />
field, qt(x) is the Hermitian conjugate of cp(x).) Since cp(x) is a<br />
complex function in classical field theory, it assigns a complex<br />
number to each four-dimensional point x = (ct, x) of time and space.<br />
The symbol x denotes all four components. In quantum field theory<br />
cp(x) is an operator that acts on a state vector in quantum-mechanical<br />
Hilbert space by adding or removing elementary particles localized<br />
around the space-time point x.<br />
In this note we present the case in which cp(x) and cpt(x) correspond<br />
respectively to a spinless charged particle and its antiparticle of equal<br />
mass but opposite charge. The charge in this field theory is like<br />
electric charge, except it is not yet coupled to the electromagnetic<br />
field. (The word “charge” has a broader definition than just electric<br />
charge.) In Note 3 we show how this complex scalar field theory can<br />
describe a quite different particle spectrum: instead of a particle and<br />
its antiparticle of equal mass, it can describe a particle of zero mass<br />
and one of nonzero mass, each of which is its own antiparticle. Then<br />
the scalar theory exhibits the phenomenon called spontaneous symmetry<br />
breaking, which is important for the standard model.<br />
A complex scalar theory can be defined by the Lagrangian density,<br />
where d’cp = dcp/d9. (Upper and lower indices are related by the<br />
metric tensor, a technical point not central to this discussion.) The<br />
Lagrangian itself is<br />
The first term in Eq. la is the kinetic energy of the fields q(x) and<br />
cpt(s), and the last two terms are the negative of the potential energy.<br />
Terms quadratic in the fields, such as the -mzqtcp term in Eq. la.<br />
are called mass terms. If in2 > 0, then cp(s) describes a spinless<br />
particle and cpt(x) its antiparticle of identical mass. If rn2 < 0, the<br />
theory has spontaneous symmetry breaking.<br />
The equations of motion are derived from Eq. 1 by a variational<br />
method. Thus, let us change the fields and their derivatives by a small<br />
amount Scp(x) and Sd,q(s) = d,Scp(x). Then,<br />
‘2 a 2 a 2 a 2<br />
6qt + a(apcp)a~6~ -<br />
where the variation is defined with the restrictions Gcp(x,fl) = 6cp(x,tz)<br />
= Scpt(x,fl) =6cpt(x,f2) = 0, and Fcp(x) and 6cpt(x) are independent. The<br />
last two terms are integrated by parts, and the surface term is dropped<br />
since the integrand vanishes on the boundary. This procedure yields<br />
the Euler-Lagrange equations for cpt(x),<br />
and for cp(x). (The Euler-Lagrange equation for cp(x) is like Eq. 3<br />
except that cpt replaces cp. There are two equations because 6cp(x) and<br />
6cpt(x) are independent.) Substituting the Lagrangian density, Eq. la,<br />
into the Euler-Lagrange equations, we obtain the equations of motion,<br />
plus another equation of exactly the same form with cp(x) and<br />
cpt(x) exchanged.<br />
This method for finding the equations of motion can be easily<br />
generalized to more fields and to fields with spin. For example, a field<br />
theory that is incorporated into the standard model is electrodynamics.<br />
Its list of fields includes particles that carry spin. The<br />
electromagnetic vector potential A,(x) describes a “vector” particle<br />
with a spin of 1 (in units of the quantum of action h = 1.0546 X<br />
erg second), and its four spin components are enumerated by the<br />
space-time vector index p ( = 0, I, 2, 3, where 0 is the index for the<br />
time component and I, 2, and 3 are the indices for the three space<br />
components). In electrodynamics only two ofthe four components of<br />
A@) are independent. The electron has a spin of Y2, as does its<br />
antiparticle, the positron. Electrons and positrons of both spin projections,<br />
W 2 , are described by a field w(~), which is a column vector<br />
with four entries. Many calculations in electrodynamics are complicated<br />
by the spins of the fields.<br />
There is a much more difficult generalization of the Lagrangian<br />
formalism: if there are constraints among the fields, the procedure<br />
yielding the Euler-Lagrange equations must be modified, since the<br />
field variations are not all independent. This technical problem<br />
complicates the formulation of electrodynamics and the standard<br />
model, especially when computing quantum corrections. Our examination<br />
of the theory is not so detailed as to require a solution of<br />
the constraint problem.<br />
(3)<br />
55
Continuous<br />
Symmetries<br />
It is often possible to find sets of fields in the Lagrangian that can<br />
be rearranged or transformed in ways described below without<br />
changing the Lagrangian. The transformations that leave the Lagrangian<br />
unchanged (or invariant) are called symmetries. First, we<br />
will look at the form of such transformations, and then we will<br />
discuss implications of a symmetrical Lagrangian. In some cases<br />
symmetries imply the existence of conserved currents (such as the<br />
electromagnetic current) and conserved charges (such as the electric<br />
charge), which remain constant during elementary-particle collisions.<br />
The conservation of energy, momentum, angular momentum, and<br />
electric charge are all derived from the existence of symmetries.<br />
Let us consider a continuous linear transformation on three real<br />
spinless fields cp,(x) (where i = I, 2, 3) with cp;(x) = cp/(x). These three<br />
fields might correspond to the three pion states. As a matter of<br />
notation, cp(x) is a column vector, where the top entry is cpl(x), the<br />
second entry is cp2(x), and the bottom entry is cp3(x). We write the<br />
linear transformation of the three fields in terms of a 3-by-3 matrix<br />
U(E), where<br />
cP’(f) = U(E)(P(x) , (54<br />
The repeated index is summed from 1 to 3, and generalizations to<br />
different numbers or kinds of fields are obvious. The parameter E is<br />
continuous, and as E approaches zero, U(E) becomes the unit matrix.<br />
The dependence of x’ on x and E is discussed below. The continuous<br />
transformation U(E) is called linear since cp,(x) occurs linearly on the<br />
right-hand side of Eq. 5. (Nonlinear transformations also have an<br />
important role in particle physics, but this discussion of the standard<br />
model will primarily involve linear transformations except for the<br />
vector-boson fields, which have a slightly different transformation<br />
law, described in Note 5.) For N independent transformations, there<br />
will be a set of parameters E,, where the index a takes on values from<br />
1 to N.<br />
For these continuous transformations we can expand cp’(x‘) in a<br />
Taylor series about E, = 0; by keeping only the leading term in the<br />
expansion, Eq. 5 can be rewritten in infinitesimal form as<br />
Gcp(x) = cp’(x) - cp(x) = is”T,cp(x), (64<br />
where T,, is the first term in the Taylor expansion,<br />
with 6x = x‘ - x. The Tu are the “generators” of the symmetry<br />
transformations of cp(x). (We note that Gcp(x) in Eq. 6a is a small<br />
symmetry transformation, not to be confused with the field variations<br />
6cp in Eq. 2.)<br />
The space-time point x’ is, in general, a function of x. In the case<br />
where x’ = x, Eq. 5 is called an internal transformation. Although our<br />
primary focus will be on internal transformations, space-time symmetries<br />
have many applications. For example, all theories we describe<br />
here have Poincare symmetry, which means that these theories<br />
are invariant under transformations in which x‘ = Ax + b, where A is<br />
a 4-by-4 matrix representing a Lorentz transformation that acts on a<br />
four-component column vector x consisting of time and the three<br />
space components, and b is the four-component column vector of the<br />
parameters of a space-time translation. A spinless field transforms<br />
under Poincart transformations as cp’(x’) = cp (x) or 6cp = -bpd,cp(x).<br />
Upon solving Eq. 6b, we find the infinitesimal translation is represented<br />
by id,. The components of fields with spin are rearranged by<br />
Poincart transformations according to a matrix that depends on both<br />
the E’S and the spin of the field.<br />
We now restrict attention to internal transformations where the<br />
space-time point is unchanged; that is, 6xv = 0. If E, is an in-<br />
finitesimal, arbitrary function ofx, E&), then Eqs. 5 and 6a are called<br />
local transformations. If the ,E, are restricted to being constants in<br />
space-time, then the transformation is called global.<br />
Before beginning a lengthy development of the symmetries of<br />
various Lagrangians, we give examples in which each of these kinds<br />
of linear transformations are, indeed, symmetries of physical theories.<br />
An example of a global, internal symmetry is strong isospin, as<br />
discussed briefly in “<strong>Particle</strong> <strong>Physics</strong> and the Standard Model.”<br />
(Actually, strong isospin is not an exact symmetry of Nature, but it is<br />
still a good example.) All theories we discuss here have global Lorentz<br />
invariance, which is a space-time symmetry. Electrodynamics has a<br />
local phase symmetry that is an internal symmetry. For a charged<br />
spinless field the infinitesimal form of a local phase transformation is<br />
6cp(x) = i~(x)cp(x) and 6cpt(x) = -~E(X)~+(X), where cp(x) is a complex<br />
field. Larger sets of local internal symmetry transformations are<br />
fundamental in the standard model of the weak and strong interactions.<br />
Finally, Einstein’s gravity makes essential use of local spacetime<br />
Poincare transformations. This complicated case is not discussed<br />
here. It is quite remarkable how many types of transformations<br />
like Eqs. 5 and 6 are basic in the formulation of physical<br />
theoiies.<br />
Let us return to the column vector of three real fields cp(x) and<br />
suppose we have a Lagrangian that is unchanged by Eqs. 5 and 6,<br />
where we now restrict our attention to internal transformations. (One<br />
such Lagrangian is Eq. la, where cp(x) is now a column vector and<br />
cpt(x) is its transpose.) Not only the Lagrangian, but the Lagrangian<br />
density, too, is unchanged by an internal symmetry transformation.<br />
56
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Let us consider the infinitesimal transformation (Eq. 6a) and calculate<br />
62' in two different ways. First of all, 6 9 = 0 if 69 is a symmetry<br />
identified from the Lagrangian. Moreover, according to the rules of<br />
partial differentiation,<br />
Then, using the Euler-Lagrange equations (Eq. 3) for the first term<br />
and collecting terms, Eq. 7 can be written in an interesting way:<br />
The next step is to substitute Eq. 6a into Eq. 8. Thus, let us<br />
define the current JE(x) as<br />
Then Eq. 8 plus the requirement that 6cp is a symmetry imply the<br />
continuity equation,<br />
(9)<br />
Jt;(x) must fall off rapidly enough as 1x1 approaches infinity that the<br />
integral is finite.<br />
If Q"(t) is indeed a conserved quantity, then its value does not<br />
change in time, which means that its first time derivative is zero. We<br />
can compute the time derivative of go([) with the aid of Eq. IO:<br />
d<br />
3 Q"(t) = d3x - aJ8(x) /d3x V . J"(x) = /J' - dS = 0. (12)<br />
at<br />
The next to the last step is Gauss's theorem, which changes the<br />
volume integral of the divergence of a vector field into a surface<br />
integral. If J"(x) falls off more rapidly than I/lxl* as 1x1 becomes very<br />
large, then the surface integral must be zero. It is not a always true<br />
that J"(x) falls off so rapidly, but when it does, Q"(t) = Q' is a<br />
constant in time. One of the most important experimental tests of a<br />
Lagrangian is whether the conserved quantities it predicts are, indeed,<br />
conserved in elementary-particle interactions.<br />
The Lagrangian for the complex scalar field defined by Eq. I has an<br />
internal global symmetry, so let us practice the above steps and<br />
identify the conserved current and charge. It is easily verified that the<br />
global phase transformation<br />
apqx) = 0 . (10)<br />
We can gain intuition about Eq. 10 from electrodynamics, since the<br />
electromagnetic current satisfies a continuity equation. It says that<br />
charge is neither created nor destroyed locally: the change in the<br />
charge density, J&), in a small region of space is just equal to the<br />
current J(x) flowing out of the region. Equation IO generalizes this<br />
result of electrodynamics to other kinds of charges, and so J@) is<br />
called a current. In particle physics with its many continuous symmetries,<br />
we must be careful to identify which current we are talking<br />
about.<br />
Although the analysis just performed is classical, the results are<br />
usually correct in the quantum theory derived from a classical<br />
Lagrangian. In some cases, however, quantum corrections contribute<br />
a nonzero term to the right-hand side of Eq. IO; these terms are called<br />
anomalies. For global symmetries these anomalies can improve the<br />
predictions from Lagrangians that have too much symmetry when<br />
compared with data because the anomaly wrecks the symmetry (it<br />
was never there in the quantum theory, even though the classical<br />
Lagrangian had the symmetry). However, for local symmetries<br />
anomalies are disastrous. A quantum field theory is locally symmetric<br />
only if its currents satisfy the continuity equation, Eq. 10.<br />
Otherwise local symmetry transformations simply change the theory.<br />
(Some care is needed to avoid this kind of anomaly in the standard<br />
model.) We now show that Eq. IO can imply the existence of a<br />
conserved quantity called the global charge and defined by<br />
provided the integral over all space in Eq. 11 is well defined; that is,<br />
leaves the Lagrangian density invariant. For example, the first term<br />
of Eq. 1 by itself is unchanged: dpcptdrq becomes a,(e-"qt)ap(B"cp)<br />
= dpcptdhp, where the last equality follows only if E is constant in<br />
space-time. (The case of local phase transformations is treated in<br />
Note 5.) The next step is to write the infinitesimal form of Eq. 13 and<br />
substitute it into Eq. 9. The conserved current is<br />
Jp(x) = i[(apcpt)cp - @,cp)cp+] > (14)<br />
where the sum in Eq. 9 over the fields cp(x) and cpt(x) is written out<br />
explicitly.<br />
If m2 > 0 in Eq. 1, then all the charge can be localized in space and<br />
time and made to vanish as the distance from the charge goes to<br />
infinity. The steps in Eq. 12 are then rigorous, and a conserved charge<br />
exists. The calculation was done here for classical fields, but the same<br />
results hold for quantum fields: the conservation law implied by Eq.<br />
I2 yields a conserved global charge equal to the number of cp particles<br />
minus the number of cp antiparticles. This number must remain<br />
constant in any interaction. (We will see in Note 3 that if m2 < 0, the<br />
charge distribution is spread out over all space-time, so the global<br />
charge is no longer conserved even though the continuity equation<br />
remains valid.)<br />
Identifying the transformations of the fields that leave the Lagrangian<br />
invariant not only-salisfies our sense of symmetry but also<br />
leads to important predictions of the theory without solving the<br />
equations of motion. In Note 4 we will return to the example of three<br />
real scalar fields to introduce larger global symmetries, such as SU(2),<br />
that interrelate different fields.<br />
57
I _______<br />
~<br />
~<br />
Spontaneous<br />
Breaking of a<br />
Global Symmetry<br />
It is possible for the vacuum or ground state of a physical system to<br />
have less symmetry than the Lagrangian. This possibility is called<br />
spontaneous symmetry breaking, and it plays an important role in<br />
the standard model. The simplest example is the complex scalar field<br />
theory of Eq. la with tn' < 0.<br />
In order to identify the classical fields with particles in the quantum<br />
theory, the classical field must approach zero as the number of<br />
particles in the corresponding quantum-mechanical state approaches<br />
zero. Thus the quantum-mechanical vacuum (the state with no<br />
particles) corresponds to the classical solution cp(x) = 0. This might<br />
seem automatic, but it is not. Symmetry arguments do not<br />
necessarily imply that cp(x) = 0 is the lowest energy state of the<br />
system. However, ifwe rewrite cp(x) as a function of new fields that do<br />
vanish for the lowest energy state, then the new fields may be directly<br />
identified with particles. Although this prescription is simple, its<br />
I<br />
i<br />
I<br />
I<br />
t<br />
justification and analysis of its limitations require extensive use of nonzero Of q-<br />
the details of quantum field theory.<br />
The energy of the complex scalar theory is the sum of kinetic and<br />
potential energies of the cp(x) and cpt(x) fields, so the energy density is<br />
x = apcp+apcp + 'dcptcp + h(cp+cp)2 , (15)<br />
with h > 0. Note that dPcpt~,cp is nonnegative and is zero if cp is a<br />
constant. For cp = 0, .vV = 0. However, if m2 < 0, then there are<br />
nonzero values of cp(x) for which H < 0. Thus, there is a nonzero<br />
field configuration with lowest energy. A graph of .W as a function of<br />
lcpl is shown in Fig. 1. In this example ,W is at its lowest value when<br />
both the kinetic and potential energies ( V= m2cptcp + k(cptcp)') are at<br />
their lowest values. Thus, the vacuum solution for cp(s) is found by<br />
solving the equation a V/dcp = 0, or<br />
1<br />
________.I___<br />
i<br />
1<br />
I<br />
__._..__-,<br />
Fig- 1. The Hamiltonian ,#f defined by &. 15 has minima at<br />
!<br />
I<br />
I !<br />
,<br />
I<br />
Next we find new fields that vanish when Eq. 16 is satisfied. For<br />
example, we can set<br />
/<br />
where the real fields p(x) and n(x) are zero when the system is in the<br />
lowest energy state. Thus p(x) and n(x) may be associated with<br />
particles. Note, however, that cp0 is not completely specified; it may<br />
lie at any point on the circle in field space defined by Eq. 16, as shown<br />
in Fig. 2.<br />
Suppose 90 is real and given by<br />
cpo = (--tn2/h)'f2. (18)<br />
Then the Lagrangian is still invariant under the phase transformations<br />
in Eq. 13. but the choice of the vacuum field solution is changed<br />
Fig. 2. The closed curve is the location of the minimum of<br />
V in the field space 9.<br />
by the phase transformation. Thus, the vacuum solution is not<br />
invariant under the phase transformations, so the phase symmetry is<br />
spontaneously broken. The symmetry of the Lagrangian is not a<br />
symmetry of the vacuum. (For m2 > 0 in Eq. I, the vacuum and the<br />
Lagrangian both have the phase symmetry.)
This Lagrangian has the following features.<br />
0 The fields p(x) and ~(x) have standard kinetic energy terms.<br />
0 Since m2 < 0, the term m2p2 can be interpreted as the mass term for<br />
the p(x) field. The p(x) field thus describes a particle with masssquared<br />
equal to Im'I, not - Im'I.<br />
0 The n(x) field has no mass term. (This is obvious from Fig. 2,<br />
which shows that Y(p,n) has no curvature (that is, a2V/arc2 = 0) in<br />
the ~(x) direction.) Thus, n(x) corresponds to a massless particle.<br />
This result is unchanged when all the quantum effects are included.<br />
0 The phase symmetry is hidden in V when it is written in terms of<br />
p(x) and n(x). Nevertheless, 2' has phase symmetry, as is proved<br />
by working backward from Eq. 20 to Eq. 16 to recover Eq. la.<br />
0 In theories without gravity, the constant term V' 0: m4/h can be<br />
ignored, since a constant overall energy level is not measurable.<br />
The situation is much more complicated for gravitational theories,<br />
where terms of this type contribute to the vacuum energy-momentum<br />
tensor and, by Einstein's equations, modify the geometry of<br />
space-time.<br />
0 The p field interacts with the K field only through derivatives of n.<br />
The interaction terms in Eq. 20 may be pictured as in Fig. 3.<br />
Fig. 3. A graphic representation of the last four terms of Eq.<br />
20, the interaction terms. Solid lines denote the p field and<br />
dotted lines the nfield. The interaction of three p(x)fields at<br />
a singlepoint is shown as three solid lines emanating from a<br />
single point. In perturbation theory this so-called vertex<br />
represents the lowest order quantum-mechanical amplitude<br />
for one particle to turn into two. All possible configurations<br />
of these vertices represent the quantum-mechanical<br />
amplitudes defined by the theory.<br />
We now rewrite the Lagrangian in terms of the particle fields p(x)<br />
and ~(x) by substituting Eq. 17 into Eq. 1. The Lagrangian becomes<br />
To estimate the masses associated with the particle fields p(x) and<br />
~(x), we substitute Eq. 18 for the constant cpo and expand 9 in powers<br />
of the fields ~(x) and p(x), obtaining<br />
Although this model might appear to be an idle curiosity, it is an<br />
example of a very general result known as Goldstone's theorem. This<br />
theorem states that in any field theory there is a zero-mass spinless<br />
particle for each independent global continuous symmetry of the<br />
Lagrangian that is spontaneously broken. The zero-mass particle is<br />
called a Goldstone boson. (This general result does not apply to local<br />
symmetries, as we shall see.)<br />
There has been one very important physical application of spontaneously<br />
broken global symmetries in particle physics, namely,<br />
theories of pion dynamics. The pion has a surprisingly small mass<br />
compared to a nucleon, so it might be understood as a zero-mass<br />
particle resulting from spontaneous symmetry breaking of a global<br />
symmetry. Since the pion mass is not exactly zero, there must also be<br />
some small but explicit terms in the Lagrangian that violate the<br />
global symmetry. The feature of pion dynamics that justifies this<br />
procedure is that the interactions of pions with nucleons and other<br />
pions are similar to the interactions (see Fig. 3) of the n(x) field with<br />
the p(x) field and with itself in the Lagrangian of Eq. 20. Since the<br />
pion has three (electric) charge states, it must be associated with a<br />
larger global symmetry than the phase symmetry, one where three<br />
independent symmetries are spontaneously broken. The usual choice<br />
of symmetry is global SU(2) X SU(2) spontaneously broken to the<br />
SU(2) of the strong-interaction isospin symmetry (see Note 4 for a<br />
discussion of SU(2)). This description accounts reasonably well for<br />
low-energy pion physics.<br />
Perhaps we should note that only spinless fields can acquire a<br />
vacuum value. Fields carrying spin are not invariant under Lorentz<br />
transformations, so if they acquire a vacuum value, Lorentz invariance<br />
will be spontaneously broken, in disagreement with experiment.<br />
Spinless particles trigger the spontaneous symmetry breaking<br />
in the standard model.
T ncIwnnn1n<br />
1s with<br />
Larger Global<br />
Symmetries<br />
7<br />
proaches zero.<br />
TO identify the generators T, with matrix elements ( T~)~~, we use a<br />
specific Lagrangian,<br />
9<br />
1 1<br />
= -p<br />
‘pi~ll’p, - 7 m2’picpi - (Cplcpl)’<br />
L<br />
L<br />
In a theory with a single complex scalar field the phase transformation<br />
in Eq. 13 defines the “largest” possible internal symmetry since<br />
the only possible symmetries must relate q(x) to itself. Here we will<br />
discuss global symmetries that interrelate different fields and group<br />
them together into “symmetry multiplets.” Strong isospin, an approximate<br />
symmetry of the observed strongly interacting particles, is<br />
an example. It groups the neutron and the proton into an isospin<br />
doublet, reflecting the fact that the neutron and proton have nearly<br />
the same mass and share many similarities in the way that they<br />
interact with other particles. Similar comments hold for the three<br />
pion states (K’, no, and f), which form an isospin triplet.<br />
We will derive the structure of strong isospin symmetry by examining<br />
the invariance of a specific Lagrangian for the three real scalar<br />
fields q,(x) already described in Note 2. (Although these fields could<br />
describe the pions, the Lagrangian will be chosen for simplicity, not<br />
for its capability to describe pion interactions.)<br />
We are about to discover a symmetry by deriving it from a<br />
Lagrangian; however, in particle physics the symmetries are often<br />
discovered from phenomenology. Moreover, since there can be many<br />
Lagrangians with the same symmetry, the predictions following from<br />
the symmetry are viewed as more general than the predictions of a<br />
specific Lagrangian with the symmetry. Consequently, it becomes<br />
important to abstract from specific Lagrangians the general features<br />
ofa symmetry; see the comments later in this note.<br />
A general linear transformation law for the three real fields can be<br />
written<br />
where the sum on j runs from 1 to 3. One reason for choosing this<br />
form of U(E) is that it explicitly approaches the identity as E ap-<br />
Let us place primes on the fields in Eq. 22 and substitute Eq. 21 into<br />
it. Then 9 written in terms of the new cp(x) is exactly the same as Eq.<br />
22 if<br />
where tij, are the matrix elements of the 3-by-3 identity matrix. (In<br />
the notation of Eq. 5a, Eq. 23 is U(E)U~(E) = I.) Equation 23 can be<br />
expanded in E, and the linear term then requires that To be an<br />
antisymmetric matrix. Moreover, exp ( ~E~T~) must be a real matrix so<br />
that q(s) remains real after the transformation. This implies that all<br />
elements of the T, are imaginary. These constraints are solved by the<br />
three imaginary antisymmetric 3-by-3 matrices with elements<br />
where ~ 123 = +I and is antisymmetric under the interchange of<br />
any two indices (for example, ~ 321 = -1). (It is a coincidence in this<br />
example that the number of fields is equal to the number of independent<br />
symmetry generators. Also, the parameter E, with one index<br />
should not be confused with the tensor &&with three indices.)<br />
The conditions on U(E) imply that it is an orthogonal matrix; 3-<br />
by-3 orthogonal matrices can also describe rotations in three spatial<br />
dimensions. Thus, the three components of cpI transform in the same<br />
way under isospin rotation as a spatial vector x transforms under a<br />
rotation. Since the rotational symmetry is SU(2), so is the isospin<br />
symmetry. (Thus “isospin” is like spin.) The T, matrices satisfy the<br />
SU(2) commutation relations
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
[Tu, Tb] T,Th - TbTu = iE,hcTc . (25)<br />
Although the explicit matrices of Eq. 24 satisfy this relation, the T,<br />
can be generalized to be quantum-mechanical operators. In the<br />
example of Eqs. 21 and 22, the isospin multiplet has three fields.<br />
Drawing on angular momentum theory, we can learn other<br />
possibilities for isospin multiplets. Spin-J multiplets (or representations)<br />
have 25 + I components, where J can be any nonnegative<br />
integer or half integer. Thus, multiplets with isospin of '12 have two<br />
fields (for example, neutron and proton) and isospin-3/2 multiplets<br />
have four fields (for example, the A++, A+, A', and A- baryons of mass<br />
- 1232 GeV/c2).<br />
The basic structure of all continuous symmetries of the standard<br />
model is completely analogous to the example just developed. In fact,<br />
part of the weak symmetry is called weak isospin, since it also has the<br />
same ma?hematical structure as strong isospin and angular momentum.<br />
Since there are many different applications to particle theory of<br />
given symmetries, it is often useful to know about symmetries and<br />
their multiplets. This mathematical endeavor is called group theory,<br />
and the results of group theory are often helpful in recognizing<br />
patterns in experimental data.<br />
Continuous symmetries are defined by the algebraic properties of<br />
their generators. Group transformations can always be written in the<br />
form of Eq. 21. Thus, if Q, (a = I, .. . , N) are the generators of a<br />
symmetry, then they satisfy commutation relations analogous to Eq.<br />
25:<br />
multiplet of the symmetry.<br />
The general problem of finding all the ways of constructing equations<br />
like Eq. 25 and Eq. 26 is the central problem of Lie-group<br />
theory. First, one must find all sets ofj& This is the problem of<br />
finding all the Lie algebras and was solved many years ago. The<br />
second problem is, given the Lie algebra, to find all the matrices that<br />
represent the generators. This is the problem of finding all the<br />
representations (or multiplets) of a Lie algebra and is also solved in<br />
general, at least when the range of values ofeach E, is finite. Lie group<br />
theory thus offers an orderly approach to the classification of a huge<br />
number of theories.<br />
Once a symmetry of the Lagrangian is identified, then sets of n<br />
fields are assigned to n-dimensional representations of the symmetry<br />
group, and the currents and charges are analyzed just as in Note 2.<br />
For instance, in our example with three real scalar fields and the<br />
Lagrangian of Eq. 22, the currents are<br />
and, if m2 > 0, the global symmetry charge is<br />
where the quantum-mechanical charges Q, satisfy the commutation<br />
relations<br />
where the constants,f,h, are called the structure constants of the Lie<br />
algebra. The structure constants are determined by the multiplication<br />
rules for the symmetry operations, U(E~)U(&~) = U(&j), where ~3<br />
depends on E~ and E ~ Equation . 26 is a basic relation in defining a Lie<br />
algebra, and Eq. 2 I is an example of a Lie group operation. The Q,.<br />
which generate the symmetry, are determined by the "group" structure.<br />
The focus on the generators often simplifies the study of Lie<br />
groups. The generators Q, are quantum-mechanical operators. The<br />
(Tu)', of Eqs. 24 and 25 are matrix elements of Q, for some symmetry<br />
(The derivation of Eq. 29 from Eq. 28 requires the canonical commutation<br />
relations of the quantum cpl(x) fields.)<br />
The three-parameter group SU(2) has just been presented in some<br />
detail. Another group of great importance to the standard model is<br />
SU(3), which is the group of 3-by-3 unitary matrices with unit<br />
determinant. The inverse of a unitary matrix U is Ut, so UtU = I.<br />
There are eight parameters and eight generators that satisfy Eq. 26<br />
with the structure constants ofSU(3). The low-dimensional representations<br />
of SU(3) have I, 3, 6, 8, IO, . .. fields, and the different<br />
representations are referred to as 1,3,3,6,6,8, 10, m, and so on.<br />
61
I<br />
Local Phase<br />
Invariance and<br />
Electrodynamics<br />
The theories that make up the standard model are all based on the<br />
principle of local symmetry. The simplest example of a local symmetry<br />
is the extension of the global phase invariance discussed at the<br />
end of Note 2 to local phase invariance. As we will derive below, the<br />
requirement that a theory be invariant under local phase transformations<br />
irnplies the existence ofa gauge field in the theory that mediates<br />
or carries the “force” between the matter fields. For electrodynamics<br />
the gauge field is the electromagnetic vector potential A,(x) and its<br />
quantum particle is the massless photon. In addition, in the standard<br />
model the gauge fields mediating the strong interactions between the<br />
quarks are the massless gluon fields and the gauge fields mediating<br />
the weak interactions are the fields for the massive Zo and w‘ weak<br />
bosons.<br />
To illustrate these principles we extend the global phase invariance<br />
of the Lagrangian of Eq. I to a theory that has local phase invariance.<br />
Thus, we require 9’ to have the same form for cp’(x) and cp(x), where<br />
the local phase transformation is defined by<br />
cp’(x) = efE(‘)cp(x) . (30)<br />
The potential energy,<br />
already has this symmetry, but the kinetic energy, ?cptd,,cp, clearly<br />
does not. since<br />
Y does not have local phase invariance if the Lagrangian of the<br />
transformed fields depends on E(X) or its derivatives. The way to<br />
eliminate the d , ~ dependence is to add a new field A,(x) called the<br />
gauge field and then require the local symmetry transformation law<br />
for this new field to cancel the d , ~ term in Eq. 32. The gauge field can<br />
be added by generalizing the derivative a, to D,, where<br />
D,, = a,, - ieA,,(x) . (33)<br />
This is just the minimal-coupling procedure of electrodynamics. We<br />
can then make a kinetic energy term of the form (Dpcp)t(Dpcp) if we<br />
require that<br />
When written out with Eq. 33, Eq. 34 becomes an equation for A;(x)<br />
in terms ofA,,(x), which is easily solved to give<br />
I<br />
AG(x) = A,(x) + - e d,E(X) . (35)<br />
Equation 35 prescribes how the gauge field transforms under the local<br />
phase symmetry.<br />
Thus the first step to modifying Eq. 1 to be a theory with local<br />
phase invariance is simply to replace dp by D,, in 9. (A slightly<br />
generalized form of this trick is used in the construction of all the<br />
theories in the standard model.) With this procedure the dominant<br />
interaction of the gauge field AW(x) with the matter field cp is in the<br />
form of a current times the gauge field, ePA,, where J, is the current<br />
defined in Eq. 14.<br />
of the calculation is replacing dpcpta,,cp by (Dpcp)t(D,,cp). However,<br />
Spontaneous<br />
instead of simply substituting Eq. 17 for cp and computing<br />
(Dwcp)t(D,cp) directly, it is convenient to make a local phase trans-<br />
Breaking of Local formation first:<br />
1<br />
Phase Invariance q’(x) = + 90~<br />
exp[lrr(x)/cpo~, (41)<br />
We now show that spontaneous breaking of local symmetry implies<br />
that the associated vector boson has a mass, in spite of the fact<br />
that A”,, by itself is not locally phase invariant. Much of the<br />
calculation in Note 3 can be translated to the Lagrangian of Eq. 38. In<br />
fact, the calculation is identical from Eq. 16 to Eq. 18, so the first new<br />
step is to substitute Eq. 17 into Eq. 38. The only significantly new part<br />
where cp(x) = [p(x) + cpo]/\/Z. (The local phase invariance permits us<br />
to remove the phase of cp(x) at every space-time point.) We<br />
emphasize the difference between Eqs. 17 and 41: Eq. 17 defines the<br />
p(x) and R(X) fields; Eq. 41 is a local phase transformation of cp(x) by<br />
angle n(x). Don’t be fooled by the formal similarity of the two<br />
equations. Thus, we may write Eq. 38 in terms of cp(x) = [~(x) +<br />
cpO]/fi and obtain ,, ’ .,<br />
.’ ,
This leaves a problem. If we simply replace d,cp by D,cp in the<br />
Lagrangian and then derive the equations of motion for A,, we find<br />
that A, is proportional to the current J,. The A, field equation has no<br />
space-time derivatives and therefore A,(x) does not propagate. If we<br />
want A, to correspond to the electromagnetic field potential, we must<br />
add a kinetic energy term for it to 9.<br />
The problem then is to find a locally phase invariant kinetic energy<br />
term for A,(x). Note that the combination of covariant derivatives<br />
D,Dv - D,D,, when acting on any function, contains no derivatives<br />
of the function. We define the electromagnetic field tensor of electrodynamics<br />
as<br />
It contains derivatives of A,. Its transformation law under the local<br />
symmetry is<br />
F;, = F,, . (37)<br />
Thus, it is completely trivial to write down a term that is quadratic in<br />
the derivatives of A,, which would be an appropriate kinetic energy<br />
term. A fully phase invariant generalization of Eq. la is<br />
the key to understanding the electroweak theory.<br />
We now rediscover the Lagrangian of electrodynamics for the<br />
interaction of electrons and photons following the same procedure<br />
that we used for the complex scalar field. We begin with the kinetic<br />
energy term for a Dirac field of the electron y, replace 3, by D,<br />
defined in Eq. 33, and then add - 1/4FpvF,,, where Fpv is defined in<br />
Eq. 36. The Lagrangian for a free Dirac field is<br />
where y, are the four Dirac y matrices and $ = ytyo. Straightening out<br />
the definition of the y, matrices and the components of w is the<br />
problem of describing a spin-% particle in a theory with Lorentz<br />
invariance. We leave the details ofthe Dirac theory to textbooks, but<br />
note that we will use some of these details when we finally write down<br />
the interactions of the quarks and leptons. The interaction of the<br />
electron field w with the electromagnetic field follows by replacing d,<br />
by D,. The electrodynamic Lagrangian is<br />
where the interaction term in i$ypD,y has the form<br />
We should emphasize that 9 has no mass term for A,(x). Thus, when<br />
the fields correspond directly to the particles in Eq. 38, the vector<br />
particles described by A,(x) are massless. In fact, ANA, is not invariant<br />
under the gauge transformation in Eq. 35, so it is not obvious<br />
how the A, field can acquire a mass if the theory does have local<br />
phase invariance. In Note 6 we will show how the gauge field<br />
becomes massive through spontaneous symmetry breaking. This is<br />
where JP= cypy is the electromagnetic current of the electron.<br />
What is amazing about the standard model is that all the electroweak<br />
and strong interactions between fermions and vector bosons are<br />
similar in form to Eq. 40b, and much phenomenology can be<br />
understood in terms of such interaction terms as long as we can<br />
approximate the quantum fields with the classical solutions.<br />
(At the expense of a little algebra, the calculation can be done the<br />
other way. First substitute Eq. 17 for cp in Eq. 38. One then finds an<br />
AV,n term in 9 that can be removed using the local phase transformation<br />
Ab = A, - [ I/(ecpo)]d,n, pf = p, and d = 0. Equation 42<br />
then follows, although this method requires some effort. Thus, a<br />
reason for doing the calculation in the order of Eq. 41 is that the<br />
algebra gets messy rather quickly if the local symmetry is not used<br />
early in the calculation ofthe electroweak case. However, in principle<br />
it makes little difference.)<br />
The Lagrangian in Eq. 42 is an amazing result: the n field has<br />
vanished from 9 altogether (according to Eq. 41, it was simply a<br />
gauge artifact), and there is a term l/2&p; MA, in 9, which is a mass<br />
term for the vector particle. Thus, the massless particle of the global<br />
case has become the longitudinal mode of a massive vector particle,<br />
and there is only one scalar particle p left in the theory. In somewhat<br />
more picturesque language the vector boson has eaten the Goldstone<br />
boson and become heavy from the feast. However, the existence of<br />
the vector boson mass terms should not be understood in isolation:<br />
the phase invariance of Eq. 42 determines the form of the interaction<br />
of the massive A, field with the p field.<br />
This calculation makes it clear that it can be tricky to derive the<br />
spectrum of a theory with local symmetry and spontaneous symmetry<br />
breaking. Theoretical physicists have taken great care to<br />
confirm that this interpretation is correct and that it generalizes to the<br />
full quantum field theory.<br />
63
where is the inverse of the matrix . With these requirements,<br />
it is easily seen that (D!'cp)t(D,cp) is invariant under the group<br />
of local transformations.<br />
The calculation of the field tensor is formally identical to Eq. 36,<br />
except we must take into account that A,(x) is a matrix. Thus, we<br />
define a matrix F,, field tensor as<br />
F,, E<br />
i [D ,D,] = a,A, - &A, - ie [A,,A,] .<br />
(49)<br />
There is a field tensor for each group generator, and some further<br />
matrix manipulation plus Eq. 26 gives the components,<br />
Thus, we can write down a kinetic energy term in analogy to<br />
electrodynamics:<br />
The locally invariant Yang-Mills Lagrangian for spinless fields coupled<br />
to the vector bosons is<br />
Just as in electrodynamics, we can add fermions to the theory in<br />
the form<br />
where D, is defined in Eq. 46 and y~ is a column vector with nfentries<br />
(nf = number of fermions). The matrices To in D, for the fermion<br />
covariant derivative are usually different from the matrices for the<br />
spinless fields, since there is no requirement that cp and I+I need to<br />
belong to the same representation of the group. It is, of course,<br />
necessary for the sets of To matrices to satisfy the commutation<br />
relations of Eq. 26 with the same set of structure constants.<br />
We will not look at the general case of spontaneous symmetry<br />
breaking in a Yang-Mills theory, which is a messy problem<br />
mathematically. There is spontaneous symmetry breaking in the<br />
electroweak sector of the standard model, and we will work out the<br />
steps analogous to Eqs. 41 and 42 for this particular case in the next<br />
Note.
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
The SU(2) X U(1)<br />
@ Et;;weak<br />
The main emphasis in these Notes has been on developing just<br />
those aspects of Lagrangian field theory that are needed for the<br />
standard model. We have now come to the crucial step: finding a<br />
Lagrangian that describes the electroweak interactions. It is rather<br />
difficult to be systematic. The historical approach would be complicated<br />
by the rather late discovery of the weak neutral currents, and<br />
a purely phenomenological development is not yet totally logical<br />
because there are important aspects of the standard model that have<br />
not yet been tested experimentally. (The most important of these are<br />
the details ofthe spontaneous symmetry breaking.) Although we will<br />
write down the answer without excessive explanation, the reader<br />
should not forget the critical role that experimental data played in the<br />
development of the theory.<br />
The first problem is to identify the local symmetry group. Before<br />
the standard model was proposed over twenty years ago, the electromagnetic<br />
and charge-changing weak interactions were known. The<br />
smallest continuous group that can describe these is SU(2), which has<br />
a doublet representation. If the weak interactions can change electrons<br />
to electron neutrinos, which are electrically neutral, it is not<br />
possible to incorporate electrodynamics in SU(2) alone unless a<br />
heavy positively charged electron is added to the electron and its<br />
neutrino to make a triplet, because the sum of charges in an SU(2)<br />
multiplet is zero. Various schemes ofthis sort have been tried but do<br />
not agree with experiment. The only way to leave the electron and<br />
electron neutrino in a doublet and include electrodynamics is to add<br />
an extra U(1) interaction to the theory. The hypothesis of the extra<br />
U( I ) factor was challenged many times until the discovery of the<br />
weak neutral current. That discovery established that the local symmetry<br />
of the electroweak theory had to be at least as large as SU(2) X<br />
U( 1).<br />
Let us now interpret the physical meaning of the four generators of<br />
SU(2) X U(I). The three generators of the SU(2) group are I+, I3,<br />
and I-, and the generator of the U(1) group is called Y, the weak<br />
hypercharge. (The weak SU(2) and U( I) groups are distinguished<br />
from other SU(2) and U( I) groups by the label “W.”) I+ and I- are<br />
associated with the weak charge-changing currents (the general definition<br />
of a current is described in Note 2), and the charge-changing<br />
currents couplc to the W+ and W- charged weak vector bosons in<br />
analogy to Eq. 40b. Both I3 and Yare related to the electromagnetic<br />
current and the weak neutral current. In order to assign the electron<br />
and its neutrino to an SU(2) doublet, the electric charge Qe” is<br />
defined by<br />
Q‘” = I3 + Y/2 , (55)<br />
so the sum ofelectric charges in an n-dimensional multiplet is nY/2.<br />
The charge of the weak neutral current is a different combination of<br />
I3 and Y, as will be described below.<br />
The Lagrangian includes many pieces. The kinetic energies of the<br />
vector bosons are described by Y Y . ~ in , analogy to the first term in<br />
Eq. 38. The three weak bosons have masses acquired through spontaneous<br />
symmetry breaking, so we need to add a scalar piece YSca~,, to<br />
the Lagrangian in order to describe the observed symmetry breaking<br />
(also see Eq. 38). The fermion kinetic energy Yfermion includes the<br />
fermion-boson interactions, analogous to the electromagnetic interactions<br />
derived in Eqs. 39 and 40. Finally, we can add terms that<br />
couple the scalars with the fermions in a term Yyukawa. One physical<br />
significance of the Yukawa terms is that they provide for masses of<br />
the quarks and charged leptons.<br />
The standard model is then a theory with a very long Lagrangian<br />
with many fields. The electroweak Lagrangian has the terms<br />
(The reader may find this construction to be ad hoc and ugly. If so,<br />
the motivation will be clear for searching for a more unified theory<br />
from which this Lagrangian can be derived. However, it is important<br />
to remember that, at present, the standard model is the pinnacle of<br />
success in theoretical physics and describes a broader range of natural<br />
phenomena than any theory ever has.)<br />
The Yang-Mills kinetic energy term has the form given by Eq. 52<br />
for the SU(2) bosons, plus a term for the U( I) field tensor similar to<br />
electrodynamics (Eqs. 36 and 38).<br />
where the U( I) field tensor is<br />
Fpv = a,B, - d,B,<br />
and the SU(2) Yang-Mills field tensor is<br />
(57)<br />
F;:,, = a,, w: - a,, w;: + g&,hc w; WC, , (59)<br />
where the E,/,< are the structure constants for SU(2) defined in Eq. 24<br />
and the W; are the Yang-Mills fields.<br />
65
@ continued<br />
SU(2) X U( 1) has two factors, and there is an independent coupling<br />
constani for each factor. The coupling for the SU(2) factor is called g,<br />
and it has become conventional to call the U( I ) coupling g'/2. The<br />
two couplings can be written in several ways. The U(1) of electrodynamics<br />
is generated by a linear combination of I3 and Y, and the<br />
coupling is, as usual, denoted by e. The other coupling can then be<br />
parameterized by an angle Bw. The relations among g, g', e, and Bw<br />
are<br />
. .- ., :<br />
e = gg'/\/$+g",.and . .. tan Bw = g'/g. (60)<br />
. .<br />
These definitions will be motivated shortly. In the electroweak theory<br />
both couplings must be evaluated experimentally and cannot be<br />
calculated in the standard model.<br />
The scalar Lagrangian requires a choice of representation for the<br />
scalar fields. The choice requires that the field with a nonzero<br />
vacuum value is electrically neutral, so the photon remains massless,<br />
but it must carry nonzero values of 13 and Y so that the weak neutral<br />
boson (the q) acquires a mass from spontaneous symmetry breaking.<br />
The simplest assignment is<br />
assignment that the cp doublet has Y = 1. After the spontaneous<br />
symmetry breaking, three of the four scalar degrees of freedom are<br />
"eaten" by the weak bosons. Thus just one scalar escapes the feast<br />
and should be observable as an independent neutral particle, called<br />
the Higgs particle. It has not yet been observed experimentally, and it<br />
is perhaps the most important particle in the standard model that<br />
does not yet have a firm phenomenological basis. (The minimum<br />
number of scalar fields in the standard model is four. Experimental<br />
data could eventually require more.)<br />
We now carry out the calculation for the spontaneous symmetry<br />
breaking of SU(2) X U( 1) down to the U( I ) of electrodynamics. Just<br />
as in the example worked out in Note 6, spontaneous symmetry<br />
breaking occurs when m2 < 0 in Eq. 62. In contrast to the simpler<br />
case, it is rather important to set up the problem in a clever way to<br />
avoid an inordinate amount ofcomputation. As in Eq. 41, we write<br />
the four degrees of freedom in the complex scalar doublet so that it<br />
looks like a local symmetry transformation times a simple form ofthe<br />
field:<br />
We can then write the scalar fields in a new gauge where the phases of<br />
cp(x) are removed:<br />
cp'(x) = exp [-irca(x)ra/2cpo]cp(x) =<br />
where cp+ has I, = Y2 and Y = 1, and cpo has I3 = --% and Y = I . Since<br />
cp does not have Y = -I fields, it is necessary to make cp a complex<br />
doublet, so (q~+)~ = -9- has I3 = -I12 and Y = -1, and (TO)+ has 13 =<br />
and Y =: -1. Then we can write down the Lagrangian of the scalar<br />
fields as<br />
where we have used the freedom of making local symmetry transformations<br />
to write cp'(x) in a very simple form. This choice, called<br />
the unitary gauge, will make it easy to write out Eq. 63 in explicit<br />
matrix form. Let us drop all primes on the fields in the unitary gauge<br />
and redefine WE by the equation<br />
where<br />
is the covariant derivative. The 2-by-2 matrices ra are the Pauli<br />
matrice:;. The factor of Y2 is required because the doublet representation<br />
of the SU(2) generators is s,/2. The factor of Y2 in the B, term<br />
is due 1.0 the convention that the U(1) coupling is g'/2 and the<br />
where the definition of the Pauli matrices is used in the first step, and<br />
the W' fields are defined in the second step with a numerical factor<br />
that guarantees the correct normalization of the kinetic energy of the<br />
charged weak vector bosons.<br />
Next, we write out the D,cp in explicit matrix form, using Eqs. 63,<br />
65, and 66:<br />
66
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
Finally, we substitute Eqs. 65 and 61 into Eq. 62 and obtain<br />
where p is the, as yet, unobserved Higgs field.<br />
It is clear from Eq. 68 that the Wfields will acquire a mass equal to<br />
g(p0/2 from the term quadratic in the W fields, ($/4)(p6WKWi.<br />
The combination g‘B, - gWi will also have a mass. Thus, we<br />
“rotate” the B, and Wi fields to the fields Z: for the weak neutral<br />
boson and A, for the photon so that the photon is massless.<br />
sin Ow -cos OW) (2:)<br />
sin ow<br />
( ?,’) = ( cos ow<br />
Our purpose here will be to write out Eq. 72 explicitly for the<br />
assignments.<br />
Consider the electron and its neutrino. (The quark and remaining<br />
lepton contributions can be worked out in a similar fashion.) The lefthanded<br />
components are assigned to a doublet and the right-handed<br />
components are singlets. (Since a neutral singlet has no weak charge,<br />
the right-handed component of the neutrino is invisible to weak,<br />
electromagnetic, or strong interactions. Thus, we can neglect it here,<br />
whether or not it actually exists.) We adopt the notation<br />
(73)<br />
where L and R denote left- and right-handed. Then the explicit<br />
statement of Eq. 72 requires constructing D, for the left- and righthanded<br />
leptons.<br />
Upon substituting Eqs. 69 and 70 into Eq. 68, we find that the 8<br />
mass is ‘12 cpo m2, so the ratio of the Wand 2 masses is<br />
Values for Mb, and MZ have recently been measured at the CERN<br />
proton-antiproton collider: MH~<br />
= (80.8 f 2.7) GeV/c2 and Mz =<br />
(92.9 f 1.6) GeV/c2. The ratio Mw/Mz calculated with these values<br />
agrees well with that given by Eq. 71. (The angle Ow is usually<br />
expressed as sin20w and is measured in neutrino-scattering experiments<br />
to be sin2& = 0.224 f 0.015.) The photon field A, does not<br />
appear in YScalar, so it does not become massive from spontaneous<br />
symmetry breaking. Note, also, that the na(x) fields appear nowhere<br />
in the Lagrangian; they have been eaten by three weak vector bosons,<br />
which have become massive from the feast.<br />
The next term in Eq. 56 is 9’fermion. Its form is analogous to Eqs. 39<br />
and 40 for electrodynamics:<br />
The weak hypercharge of the right-handed electron is -2 so the<br />
coefficient of E, in the first term of Eq. 74 is (-g‘/2) X (-2) = g‘. We<br />
leave it to the reader to check the rest of Eq. 74. The absence ofa mass<br />
term is not an error. Mass terms are of the form $w = $LI+JR + $ R~L.<br />
Since wL is a doublet and cR is a singlet, an electron mass term must<br />
violate the SU(2) X U( 1) symmetry. We will see later that the electron<br />
mass will reappear as a result of modification of ,9’yukawa due to<br />
spontaneous symmetry breaking.<br />
The next task is exciting, because it will reveal how the vector<br />
bosons interact with the leptons. The calculation begins with Eq. 74<br />
and requires the substitution of explicit matrices for T~ Wi, WR, and<br />
wL. We use the definitions in Eqs. 66, 69, and 73. The expressions<br />
become quite long, but the calculation is very straightforward. After<br />
simplifying some expressions, we find that 9’lepton for the electron<br />
lepton and its neutrino is<br />
Y~~~~~~ = i$d,e + iiLyV,vL - e $eA,<br />
The physical problem is to assign the left- and right-handed fermions<br />
to multiplets of SU(2); the assignments rely heavily on experimental<br />
data and are listed in “<strong>Particle</strong> <strong>Physics</strong> and the Standard Model.”<br />
(75)
@ continued<br />
The first two terms are the kinetic energies of the electron and the<br />
neutrino. (Note that e = eL + eR.) The third term is the electromagnetic<br />
interaction (cf. Eq. 40) with electrons of charge -e,<br />
where e is defined in Eq. 60. The coupling ofA, to the electron current<br />
does not distinguish left from right, so electrodynamics does not<br />
violate parity. The fourth term is the interaction of the W' bosons<br />
with the weak charged current of the neutrinos and electrons. Note<br />
that these bosons are blind to right-handed electrons. This is the<br />
reason for maximal parity violation in beta decay. The final terms<br />
predict how the weak neutral current of the electron and that of the<br />
neutrino couple to the neutral weak vector boson Zo.<br />
If the left- and right-handed electron spinors are written out<br />
explicitly, with eL = %(I - y&, the interaction of the weak neutral<br />
current of the electron with the Zo is proportional to $'[(I -<br />
4sin2ew) - ys]eZ,. This prediction provided a crucial test of the<br />
standard model. Recall from Eq. 71 that sinZew is very nearly Ih, so<br />
that the weak neutral current of the electron is very nearly a purely<br />
axial current, that is, a current of the form &''Y5e. This crucial<br />
prediction was tested in deep inelastic scattering of polarized electrons<br />
and in atomic parity-violation experiments. The results ofthese<br />
experiments went a long way toward establishing the standard model.<br />
The tests also ruled out models quite similar to the standard model.<br />
We could discuss many more tests and predictions of the model<br />
based on the form of the weak currents, but this would greatly<br />
lengthen our discussion. The electroweak currents of the quarks will<br />
be described in the next section.<br />
We now discuss the last term in Eq. 56, gyukawa. In a locally<br />
symmetric theory with scalars, spinors, and vectors, the interactions<br />
between vectors and scalars, vector and spinors, and vectors and<br />
vectors are determined from the local invariance by replacing 8, by<br />
D,. In contrast, yyukawa, which is the interaction between the scalars<br />
and spinors, has the same form for both local and global symmetries:<br />
This form forpy,kawa is rather schematic; to make it explicit we must<br />
specify the multiplets and then arrange the component fields so that<br />
the form of yyukawa does not change under a local symmetry transformation.<br />
Let us write Eq. 76 explicitly for the part of the standard model we<br />
have examined so far: cp is a complex doublet of scalar fields that has<br />
the form in the unitary gauge given by Eq. 65. The fermions include<br />
the electron and its neutrino. If the neutrino has no right-handed<br />
component, then it is not possible to insert it into Eq. 76. Since the<br />
neutrino has no mass term in Ylepton, the neutrino remains massless<br />
in this theory. (If VR is included, then the neutrino mass is a free<br />
parameter.) The Yukawa terms for the electron are<br />
(77)<br />
where we have used the fact that &eL = &a = 0, and e = eL + 4( is<br />
the electron Dirac spinor. Note that Eq. 77 includes an electron mass<br />
term,<br />
so the electron mass is proportional to the vacuum value of the scalar<br />
field. The Yukawa coupling is a free parameter, but we can use the<br />
measured electron mass to evaluate it. Recall that<br />
m o e o<br />
Mu,= - = ~ - 81 GeV,<br />
2 2sin BW<br />
where e'/47t = 1/137. This implies that cpo = 251 GeV. Since nip.=<br />
0.0005 1 1 GeV, CY = 2.8 X IOp6 for the electron. There are more than<br />
five Yukawa couplings, including those for the p and 'I leptons and<br />
the three quark doublets as well as terms that mix different quarks of<br />
the same electric charge. The standard model in no way determines<br />
the values of these Yukawa coupling constants. Thus, the study of<br />
fermion masses may turn out to have important hints on how to<br />
extend the standard model.
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
@) Quarks<br />
the assumption of local symmetry leads to a Lagrangian whose form<br />
is highly restricted. As far as we know, only the quark and gluon fields<br />
are necessary to describe the strong interactions, and so the most<br />
general Lagrangian is<br />
(79)<br />
Discovery of the fundamental fields of the strong interactions was<br />
not straightforward. It took some years to realize that the hadrons,<br />
such as the nucleons and mesons, are made up of subnuclear constituents,<br />
primarily quarks. Quarks originated from an effort to provide<br />
a simple physical picture of the “Eightfold Way,” which is the SU(3)<br />
symmetry proposed by M. Gell-Mann and Y. Ne’eman to generalize<br />
strong isotopic spin. The hadrons could not be classified by the<br />
fundamental three-dimensional representations of this SU(3) but<br />
instead are assigned to eight- and ten-dimensional representations.<br />
These larger representations can be interpreted as products of the<br />
three-dimensional representations, which suggested to Gell-Mann<br />
and G. Zweig that hadrons are composed of constituents that are<br />
assigned to the three-dimensional representations: the u (up), d<br />
(down), and s (strange) quarks. At the time of their conception, it was<br />
not clear whether quarks were a physical reality or a mathematical<br />
trick for simplifying the analysis of the Eightfold-Way SU(3). The<br />
major breakthrough in the development of the present theory of<br />
strong interactions came with the realization that, in addition to<br />
electroweak and Eightfold-Way quantum numbers, quarks carry a<br />
new quantum number, referred to as color. This quantum number<br />
has yet to be observed experimentally.<br />
We begin this lecture with a description of the Lagrangian of a<br />
strong-interaction theory of quarks formulated in terms of their color<br />
quantum numbers. Called quantum chromodynamics, or QCD, it is<br />
a Yang-Mills theory with local color-SU(3) symmetry in which each<br />
quark belongs to a three-dimensional color multiplet. The eight<br />
color-SU(3) generators commute with the electroweak SU(2) X U( 1)<br />
generators, and they also commute with the generators of the Eightfold<br />
Way, which is a different SU(3). (Like SU(2), SU(3) is a recurring<br />
symmetry in physics, so its various roles need to be distinguished.<br />
Hence we need the label “color.”) We conclude with a discussion of<br />
the weak interactions of the quarks.<br />
The QCD Lagrangian. The interactions among the quarks are<br />
mediated by eight massless vector bosons (called gluons) that are<br />
required to make the SU(3) symmetry local. As we have already seen,<br />
where<br />
The sum on a in the first term is over the eight gluon fields Ai. The<br />
second term represents the coupling of each gluon field to an SU(3)<br />
current of the quark fields, called a color current. This term is<br />
summed over the index i, which labels each quark type and is<br />
independent ofcolor. Since each quark field y~; is a three-dimensional<br />
column vector in color space, D, is defined by<br />
where ha is a generalization of the three 2-by-2 Pauli matrices of<br />
SU(2) to the eight 3-by-3 Gell-Mann matrices of SU(3), and g, is the<br />
QCD coupling. Thus, the color current of each quark has the form<br />
$.aywv. The left-handed quark fields couple to the gluons with<br />
exactly the same strength as the right-handed quark fields, so parity is<br />
conserved in the strong interactions.<br />
The gluons are massless because the QCD Lagrangian has no<br />
spinless fields and therefore no obvious possibility of spontaneous<br />
symmetry breaking. Of course, if motivated for experimental<br />
reasons, one can add scalars to the QCD Lagrangian and spontaneously<br />
break SU(3) to a smaller group. This modification has been<br />
used, for example, to explain the reported observation of fractionally<br />
charged particles. The experimental situation, however, still remains<br />
murky, so it is not (yet) necessary to spontaneously break SU(3) to a<br />
smaller group. For the remainder of the discussion, we assume that<br />
QCD is not spontaneously broken.<br />
The third term in Eq. 79 is a mass term. In contrast to the<br />
electroweak theory, this mass term is now allowed, even in the<br />
absence of spontaneous symmetry breaking, because the left- and<br />
right-handed quarks are assigned to the same multiplet of SU(3). The<br />
numerical coeficients MIJ are the elements of the quark mass matrix;<br />
they can connect quarks of equal electric charge. The PQCD of Eq. 79<br />
permits us to redefine the QCD quark fields so that MIJ = m,6,. The
@ continued<br />
mass matrix is then diagonal and each quark has a definite mass,<br />
which is an eigenvalue of the mass matrix. We will reappraise this<br />
situation below when we describe the weak currents of the quarks.<br />
After successfully extracting detailed predictions of the electroweak<br />
theory from its complicated-looking Lagrangian, we might be<br />
expected to perform a similar feat for the PQ,, of Eq. 79 without too<br />
much difficulty. This is not possible. Analysis of the electroweak<br />
theory was so simple because the couplingsgand g’ are always small,<br />
regardless of the energy scale at which they are measured, so that a<br />
classical analysis is a good first approximation to the theory. The<br />
quantum corrections to the results in Note 8 are, for most processes,<br />
only a few percent.<br />
In QCD processes that probe the short-distance structure of<br />
hadrons, the quarks inside the hadrons interact weakly, and here the<br />
classical analysis is again a good first approximation because the<br />
couplingg, is small. However, for Yang-Mills theories in general, the<br />
renormalization group equations of quantum field theory require<br />
that g, increases as the momentum transfer decreases until the<br />
momentum transfer equals the masses of the vector bosons. Lacking<br />
spontaneous symmetry breaking to give the gluons mass, QCD<br />
contains no mechanism to stop the growth of g,, and the quantum<br />
effects become more and more dominant at larger and larger distances.<br />
Thus, analysis of the long-distance behavior of QCD, which<br />
includes deriving the hadron spectrum, requires solving the full<br />
quantum theory implied by Eq. 79. This analysis is proving to be very<br />
difficult.<br />
Even without the solution of YQCD, we can, however, draw some<br />
conclusions. The quark fields w, in Eq. 79 must be determined by<br />
experiment. The Eightfold Way has already provided three of the<br />
quarks, and phenomenological analyses determine their masses (as<br />
they appear in the QCD Lagrangian). The mass of the u quark is<br />
nearly zero (a few MeV/c*), the dquark is a few MeV/cZ heavier than<br />
the u, and the mass of the s quark is around 300 MeV/cz. If these<br />
results are substituted into Eq. 79, we can derive a beautiful result<br />
from the QCD Lagrangian. In the limit that the quark mass differences<br />
can be ignored, Eq. 79 has a global SU(3) symmetry that is<br />
identical to the Eightfold-Way SU(3) symmetry. Moreover, in the<br />
limit that the u, d, and s masses can be ignored, the left-handed u, d,<br />
and s quarks can be transformed by one SU(3) and the right-handed<br />
u, d, and s quarks by an independent SU(3). Then QCD has the<br />
“chiral” SU(3) X SU(3) symmetry that is the basis of current algebra.<br />
The sums of the corresponding SU(3) generators of chiral SU(3) X<br />
SU(3) generate the Eightfold-Way SU(3). Thus, the QCD Lagrangian<br />
incorporates in a very simple manner the symmetry results of<br />
hadronic physics of the 1960s. The more recently discovered c<br />
(charmed) and b (bottom) quarks and the conjectured t (top) quark<br />
are easily added to the QCD Lagrangian. Their masses are so large<br />
and so different from one another that the SU(3) and SU(3) X SU(3)<br />
symmetries of the Eightfold-Way and current algebra cannot be<br />
extended to larger symmetries. (The predictions of, say, SU(4) and<br />
chiral SU(4) X SU(4) do not agree well with experiment.)<br />
It is important to note that the quark masses are undetermined<br />
parameters in the QCD Lagrangian and therefore must be derived<br />
from some more complete theory or indicated phenomenologically.<br />
The Yukawa couplings in the electroweak Lagrangian are also free<br />
parameters. Thus, we are forced to conclude that the standard model<br />
alone provides no constraints on the quark masses, so they must be<br />
obtained from experimental data.<br />
The mass term in the QCD Lagrangian (Eq. 79) has led to new<br />
insights about the neutron-proton mass difference. Recall that the<br />
quark content of a neutron is uddand that of a proton is uud. If the u<br />
and d quarks had the same mass, then we would expect the proton to<br />
be more massive than the neutron because of the electromagnetic<br />
energy stored in the uu system. (Many researchers have confirmed<br />
this result.) Since the masses of the u and d quarks are arbitrary in<br />
both the QCD and the electroweak Lagrangians, they can be adjusted<br />
phenomenologically to account for the fact that the neutron mass is<br />
1.293 MeV/c2 greater than the proton mass. This experimental<br />
constraint is satisfied if the mass of the d quark is about 3 MeV/c2<br />
greater than that of the u quark. In a way, this is unfortunate, because<br />
we must conclude that the famous puzzle of the n-p mass difference<br />
will not be solved until the standard model is extended enough to<br />
provide a theory of the quark masses.<br />
Weak Currents. We turn now to a discussion of the weak currents of<br />
the quarks, which are determined in the same way as the weak<br />
currents of the leptons in Note 8. Let us begin with just the u and d<br />
quarks. Their electroweak assignments are as follows: the left-handed<br />
components 111. and dl form an SU(2) doublet with Y = %, and the<br />
right-handed components UR and d~ are SU(2) singlets with Y = 4/3
<strong>Particle</strong> <strong>Physics</strong> and the Standard Model<br />
and -%, respectively (recall Eq. 55).<br />
The steps followed in going from Eq. 73 to Eq. 75 will yield the<br />
electroweak Lagrangian of quarks. The contribution to the, Lagrangian<br />
due to interaction of the weak neutral current .IF’ of the u and d<br />
quarks with Zo is<br />
where<br />
The reader will enjoy deriving this result and also deriving the<br />
contribution of the weak charged current of the quarks to the<br />
electroweak Lagrangian. Equation 83 will be modified slightly when<br />
we include the other quarks.<br />
So far we have emphasized in Notes 8 and 9 the construction of the<br />
QCD and electroweak Lagrangians for just one lepton-quark<br />
“family” consisting of the electron and its neutrino together with the<br />
u and d quarks. Two other lepton-quark families are established<br />
experimentally: the muon and its neutrino along with the c and s<br />
quarks and the T lepton and its neutrino along with the f and b quarks.<br />
Just like (v,)~ and eL, (v& and p~ and (v,)~ and TL form weak-SU(2)<br />
doublets; c?R, p~ and TR are each SU(2) singlets with a weak hypercharge<br />
of-2. Similarly, the weak quantum numbers ofcand sand of<br />
t and b echo those of u and d cL and SL form a weak-SU(2) doublet as<br />
do [R and bL. Like UR and dR, the right-handed quarks CR, SR, tR, and<br />
b~ are all weak-SU(2) singlets.<br />
This triplication of families cannot be explained by the standard<br />
model, although it may eventually turn out to be a critical fact in the<br />
development of theories of the standard model. The quantum<br />
numbers of the quarks and leptons are summarized in Tables 2 and 3<br />
in “<strong>Particle</strong> <strong>Physics</strong> and the Standard Model.”<br />
All these quark and lepton fields must be included in a Lagrangian<br />
that incorporates both the electroweak and QCD Lagrangians. It is<br />
quite obvious how to do this: the standard model Lagrangian is<br />
simply the sum of the QCD and electroweak Lagrangians, except that<br />
the terms occurring in both Lagrangians (the quark kinetic energy<br />
terms ic,ypdPyi and the quark mass terms yiMuyj) are included just<br />
once. Only the mass term requires comment.<br />
The quark mass terms appear in the electroweak Lagrangian in the<br />
form 9’yukawa (Eq. 77). In the electroweak theory quarks acquire<br />
masses only because SU(2) X U( I) is spontaneously broken. However,<br />
when there are three quarks of the same electric charge (such as<br />
d, s, and b), the general form of the mass terms is the same as in Eq.<br />
79, ciMijy,, because there can be Yukawa couplings between dand s,<br />
d and b, and s and 6. The problem should already be clear: when we<br />
speak of quarks, we think of fields that have a definite mass, that is,<br />
fields for which Mi is diagonal. Nevertheless, there is no reason for<br />
the fields obtained directly from the electroweak symmetry breaking<br />
to be mass eigenstates.<br />
The final part of the analysis takes some care: the problem is to find<br />
the most general relation between the mass eigenstates and the fields<br />
occurring in the weakcurrents. Wegive theanswer forthecaseoftwo<br />
families ofquarks. Let us denote the quark fields in the weak currents<br />
with primes and the mass eigenstates without primes. There is<br />
freedom in the Lagrangian to set u = u’ and c = c‘. If we do so, then<br />
the most general relationship among d, s, d‘, and s’ is<br />
The parameter €Jc, the Cabibbo angle, is not determined by the<br />
electroweak theory (it is related to ratios of various Yukawa couplings)<br />
and is found experimentally to be about 13”. (When the band<br />
t (=t‘) quarks are included, the matrix in Eq. 84 becomes a 3-by-3<br />
matrix involving four parameters that are evaluated experimentally.)<br />
The correct weak currents are then given by Eq. 83 if all quark<br />
families are included and primes are placed on all the quark fields.<br />
The weak currents can be written in terms of the quark mass<br />
eigenstates by substituting Eq. 84 (or its three-family generalization)<br />
into the primed version of Eq. 83. The ratio of amplitudes for s - u<br />
and d - u is tan 8c; the small ratio of the strangeness-changing to<br />
non-strangeness-changing charged-current amplitudes is due to the<br />
smallness of the Cabibbo angle. It is worth emphasizing again that the<br />
standard model alone provides no understanding of the value of this<br />
angle. 0<br />
71
A<br />
II throughout his history man has<br />
wanted to know the dimensions<br />
of his world and his place in it.<br />
Before the advent of scientific instruments<br />
the universe did not seem very<br />
large or complicated. Anything too small to<br />
detect with the naked eye was not known,<br />
and the few visible stars might almost be<br />
touched if only there were a higher hill<br />
nearby.<br />
Today, with high-energy particle accelerators<br />
the frontier has been pushed down<br />
to distance intervals as small as centimeter<br />
and with super telescopes to cosmological<br />
distances. These explorations<br />
have revealed a multifaceted universe; at<br />
first glance its diversity appears too complicated<br />
to be described in any unified manner.<br />
Nevertheless, it has been possible to<br />
incorporate the immense variety of experimental<br />
data into a small number of<br />
quantum field theories that describe four<br />
basic interactions-weak, strong, electromagnetic,<br />
and gravitational. Their mathematical<br />
formulations are similar in that each<br />
one can be derived from a local symmetry.<br />
This similarity has inspired hope for even<br />
greater progress: perhaps an extension of the<br />
present theoretical framework will provide a<br />
single unified description of all natural<br />
phenomena.<br />
This dream of unification has recurred<br />
again and again, and there have been many<br />
successes: Maxwell’s unification of electricity<br />
and magnetism; Einstein’s unification<br />
of gravitational phenomena with the<br />
geometry of space-time; the quantum-mechanical<br />
unification of Newtonian mechanics<br />
.with the wave-like behavior of matter; the<br />
quantum-mechanical generalization of electrodynamics;<br />
and finally the recent unification<br />
of electromagnetism with the weak<br />
force. Each of these advances is a crucial<br />
component of the present efforts to seek a<br />
more complete physical theory.<br />
Before the successes of the past inspire too<br />
much optimism, it is important to note that a<br />
unified theory will require an unprecedented<br />
extrapolation. The present optimism is generated<br />
by the discovery of theories successful<br />
7 4<br />
at describing phenomena that take place over<br />
distance intervals of order 10-l6 centimeter<br />
or larger. These theories may be valid to<br />
much shorter distances, but that remains to<br />
be tested experimentally. A fully unified theory<br />
will have to include gravity and therefore<br />
will probably have to describe spatial structures<br />
as small as centimeter, the fundamental<br />
length (determined by Newton’s<br />
gravitational constant) in the theory of gravity.<br />
History suggests cause for further<br />
caution: the record shows many failures resulting<br />
from attempts to unify the wrong, too<br />
few, or too many physical phenomena. The<br />
end of the 19th century saw a huge but<br />
unsuccessful effort to unify the description of<br />
all Nature with thermodynamics. Since the<br />
second law of thermodynamics cannot be<br />
derived from Newtonian mechanics, some<br />
physicists felt it must have the most fundamental<br />
significance and sought to derive the<br />
rest of physics from it. Then came a period of<br />
belief in the combined use of Maxwell’s electrodynamics<br />
and Newton’s mechanics to explain<br />
all natural phenomena. This effort was<br />
also doomed to failure: not only did these<br />
theories lack consistency (Newton’s equations<br />
are consistent with particles traveling<br />
faster than the speed of light, whereas the<br />
Lorentz invariant equations of Maxwell are<br />
not), but also new experimental results were<br />
emerging that implied the quantum structure<br />
of matter. Further into this century came the<br />
celebrated effort by Einstein to formulate a<br />
unified field theory of gravity and electromagnetism.<br />
His failure notwithstanding,<br />
the mathematical form of his classical theory<br />
has many remarkable similarities to the<br />
modern efforts to unify all known fundamental<br />
interactions. We must be wary that our<br />
reliance on quantum field theory and local<br />
symmetry may be similarly misdirected, although<br />
we suppose here that it is not.<br />
Two questions will be the central themes<br />
ofthis essay. First, should we believe that the<br />
theories known today are the correct components<br />
of a truly unified theory? The component<br />
theories are now so broadly accepted<br />
that they have become known as the “standard<br />
model.” They include the electroweak<br />
theory, which gives a unified description of<br />
quantum electrodynamics (QED) and the<br />
weak interactions, and quantum chromodynamics<br />
(QCD), which is an attractive candidate<br />
theory for the strong interactions. We<br />
will argue that, although Einstein’s theory of<br />
gravity (also called general relativity) has a<br />
somewhat different status among physical<br />
theories, it should also be included in the<br />
standard model. If it is, then the standard<br />
model incorporates all observed physical<br />
phenomena-from the shortest distance intervals<br />
probed at the highest energy accelerators<br />
to the longest distances seen by<br />
modern telescopes. However, despite its experimental<br />
successes, the standard model remains<br />
unsatisfying; among its shortcomings<br />
is the presence of a large number of arbitrary<br />
constants that require explanations. It remains<br />
to be seen whether the next level of<br />
unification will provide just a few insights<br />
into the standard model or will unify all<br />
natural phenomena.<br />
The second question examined in this essay<br />
is twofold: What are the possible strategies<br />
for generalizing and extending the standard<br />
model, and how nearly do models based<br />
on these strategies describe Nature? A central<br />
problem of theoretical physics is to identify<br />
the features of a theory that should be abstracted,<br />
extended, modified, or generalized.<br />
From among the bewildering array of theories,<br />
speculations, and ideas that have<br />
grown from the standard model, we will<br />
describe several that are currently attracting<br />
much attention.<br />
We focus on two extensions of established<br />
concepts. The first is called supersymmetry;<br />
it enlarges the usual space-time symmetries<br />
of field theory, namely, PoincarC invariance,<br />
to include a symmetry among the bosons<br />
(particles of integer spin) and fermions<br />
(particles of half-odd integer spin). One of<br />
the intriguing features of supersymmetry is<br />
that it can be extended to include internal<br />
symmetries (see Note 2 in “Lecture Notes-<br />
From Simple Field Theories to the Standard<br />
Model). In the standard model internal local<br />
symmetries play a crucial role, both for<br />
classifying elementary particles and for de-
___I__-<br />
Toward a Unified Theory<br />
t<br />
Gravity<br />
I<br />
Classical<br />
Origins<br />
Electra<br />
J<br />
Extended Supergravity<br />
\<br />
Development of<br />
Gravitational Theories<br />
including Other Forces<br />
'1<br />
'---___----_-- -----<br />
Fig. 1. Evolution of fundamental theories of Nature from the<br />
cfassicaffield theories of Newton and Maxwell to thegrandest<br />
theoretical conjectures of today. The relationships among<br />
these theories are discussed in the text. Solid lines indicate a<br />
direct and well-established extension, or theoretical generalization.<br />
The wide arrow symbolizes the goal of present<br />
research, the unification of quantum field theories with<br />
gravity.<br />
75
termining the form of the interactions among<br />
them. The electroweak theory is based on the<br />
internal local symmetry group SU(2) X U( I)<br />
(see Note 8) and quantum chromodynamics<br />
on the internal local symmetry group SU(3).<br />
Gravity is based on space-time symmetries:<br />
general coordinate invariance and local<br />
PoincarC symmetry. It is tempting to try to<br />
unify all these symmetries with supersymmetry.<br />
Other important implications of supersymmetry<br />
are that it enlarges the scope of the<br />
classification schemes of the basic particles<br />
to include fields ofdifferent spins in the same<br />
multiplet, and it helps to solve some technical<br />
problems concerning large mass ratios<br />
that plague certain efforts to derive the standard<br />
model. Most significantly, if supersymmetry<br />
is made to be a local symmetry, then it<br />
automatically implies a theory of gravity,<br />
called supergravity, that is a generalization of<br />
Einsttin’s theory. Supergravity theories require<br />
the unification of gravity with other<br />
kinds of interactions, which may be, in some<br />
future version, the electroweak and strong<br />
interactions. The near successes of this approach<br />
are very encouraging.<br />
The other major idea described here is the<br />
extension of the space-time manifold to<br />
more than four dimensions, the extra<br />
dimensions having, so far, escaped observation.<br />
This revolutionary idea implies that<br />
particles are grouped into larger symmetry<br />
multiplets and the basic interactions have a’<br />
geometrical origin. Although the idea of extending<br />
space-time beyond four dimensions<br />
is not new, it becomes natural in the context<br />
of supergravity theories because these complicated<br />
theories in four dimensions may be<br />
derived from relatively simple-looking theones<br />
in higher dimensions.<br />
We will follow these developments one<br />
step further to a generalization of the field<br />
concept: instead of depending.on space-time,<br />
the fields may depend on paths in spacetime.<br />
When this generalization is combined<br />
with supersymmetry, the resulting theory is<br />
called a superstring theory. (The whimsicality<br />
of the name is more than matched by<br />
the theory’s complexity.) Superstring the-<br />
’76<br />
ories are encouraging because some of them<br />
reduce, in a certain limit, to the only supergravity<br />
theories that are likely to generalize<br />
the standard model. Moreover, whereas<br />
supergravity fails to give the standard model<br />
exactly, a superstring theory might succeed.<br />
It seems that superstring theories can be<br />
formulated only in ten dimensions.<br />
Figure 1 provides a road map for this<br />
essay, which journeys from the origins of the<br />
standard model in classical theory to the<br />
extensions of the standard model in supergravity<br />
and superstrings. These extensions<br />
may provide extremely elegant ways to unify<br />
the standard model and are therefore attracting<br />
enormous theoretical interest. It must be<br />
cautioned, however, that at present no experimental<br />
evidence exists for supersymmetry<br />
or extra dimensions.<br />
Review of the Standard Model<br />
We now review the standard model with<br />
particular emphasis on its potential for being<br />
unified by a larger theory. Over the last<br />
several decades relativistic quantum field<br />
theories with local symmetry have succeeded<br />
in describing all the known interactions<br />
down to the smallest distances that have<br />
been explored experimentally, and they may<br />
be correct to much shorter distances.<br />
Electrodynamics and Local Symmetry. Electrodynamics<br />
was the first theory with local<br />
symmetry. Maxwell’s great unification of<br />
electricity and magnetism can be viewed as<br />
the discovery that electrodynamics is described<br />
by the simplest possible local symmetry,<br />
local phase invariance. Maxwell’s addition<br />
of the displacement current to the field<br />
equations, which was made in order to insure<br />
conservation of the electromagnetic current,<br />
turns out to be equivalent to imposing local<br />
phase invariance on the Lagrangian of electrodynamics,<br />
although this idea did not<br />
emerge until the late 1920s.<br />
A crucial feature of locally symmetric<br />
quantum field theories is this: typically, for<br />
each independent internal local symmetry<br />
there exists a gauge field and its corresponding<br />
particle, which is a vector boson (spin-I<br />
particle) that mediates the interaction between<br />
particles. Quantum electrodynamics<br />
has just one independent local symmetry<br />
transformation, and the photon is the vector<br />
boson (or gauge particle) mediating the interaction<br />
between electrons or other charged<br />
particles. Furthermore, the local symmetry<br />
dictates the exact form of the interaction.<br />
The interaction Lagrangian must be of the<br />
form eJ~(x)A,(x), where P (x) is the current<br />
density of the charged particles and A,(x) is<br />
the field of the vector bosons. The coupling<br />
constant e is defined as the strength with<br />
which the vector boson interacts with the<br />
current. The hypothesis that all interactions<br />
are mediated by vector bosons or, equivalently,<br />
that they originate from local symmetries<br />
has been extended to the weak and<br />
then to the strong interactions.<br />
Weak Interactions. Before the present understanding<br />
of weak interactions in terms of<br />
local symmetry, Fermi’s 1934 phenomenological<br />
theory of the weak interactions had<br />
been used to interpret many data on nuclear<br />
beta decay. After it was modified to include<br />
parity violation, it contained all the crucial<br />
elements necessary to describe the lowenergy<br />
weak interactions. His theory assumed<br />
that beta decay (e.g., n - p + e- + ic)<br />
takes place at a single space-time point. The<br />
form of the interaction amplitude is a prod-<br />
uct of two currents PJ,, where each current<br />
is a product of fermion fields, and J’J, describes<br />
four fermion fields acting at the point<br />
of the beta-decay interaction. This amplitude,<br />
although yielding accurate predictions<br />
at low energies, is expected to fail at centerof-mass<br />
energies above 300 GeV, where it<br />
predicts cross sections that are larger than<br />
allowed by the general principles ofquantum<br />
field theory.<br />
The problem of making a consistent (renormalizable)<br />
quantum field theory to describe<br />
the weak interactions was not solved<br />
until the 1960s, when the electromagnetic<br />
and weak interactions were combined into a<br />
locally symmetric theory. As outlined in Fig.
oson<br />
Toward a Unified Theory<br />
Fermi Theory<br />
Electroweak Theory<br />
Chai<br />
Amp1 itude<br />
Changing Cu<br />
Amplitude<br />
Fig. 2. Comparison of neutrino-quark charged-current scattering in the Fermi<br />
theory and the modern SU(2) X U(1) electroweak theory. (The bar indicates the<br />
Dirac conjugate.) Thepoint interaction of the Fermi theory leads to an inconsistent<br />
quantum theory. The W -+ exchange in the electroweak theory spreads out the<br />
weak interactions, which then leads to a consistent (renormalizable) quantum field<br />
theory. JF) and JL-) are the charge-raising and charge-lowering currents, respectively.<br />
The amplitudes given by the two theories are nearly equal as long as the<br />
square of the momentum transfer, q2 = (pu- pd)', is much less than the square of<br />
the mass of the weak boson, ML).<br />
2, the vector bosons associated with the electroweak<br />
local symmetry serve to spread out<br />
the interaction of the Fermi theory in spacetime<br />
in a way that makes the theory consistent.<br />
Technically, the major problem with<br />
the Fermi theory is that the Fermi coupling<br />
constant, GF, is not dimensionless (C, =<br />
(293 GeV)-*), and therefore the Fermi theory<br />
is not a renormalizable quantum field theory.<br />
This means that removing the infinities<br />
from the theory strips it of all its predictive<br />
power.<br />
In the gauge theory generalization of<br />
Fermi's theory, beta decay and other weak<br />
interactions are mediated by heavy weak<br />
vector bosons, so the basic interaction has<br />
the form gWpJ, and the current-current interaction<br />
looks pointlike only for energies<br />
much less than the rest energy of the weak<br />
bosons. (The coupling g is dimensionless,<br />
whereas GF is a composite number that includes<br />
the masses ofthe weak vector bosons.)<br />
The theory has four independent local symmetries,<br />
including the phase symmetry that<br />
yields electrodynamics. The local symmetry<br />
group of the electroweak theory is SU(2) X<br />
U( I), where U( 1 ) is the group of phase transformations,<br />
and SU(2) has the same structure<br />
as rotations in three dimensions. The<br />
one phase angle and the three independent<br />
angles of rotation in this theory imply the<br />
existence of four vector bosons, the photon<br />
plus three weak vector bosons, W', Zo, and<br />
W-. These four particles couple to the four<br />
SU(2) X U(1) currents and are responsible<br />
for the "electroweak" interactions.<br />
The idea that all interactions must be derived<br />
from local symmetry may seem simple,<br />
but it was not at all obvious how to apply this<br />
idea to the weak (or the strong) interactions.<br />
Nor was it obvious that electrodynamics and<br />
the weak interactions should be part of the<br />
same local symmetry since, experimentally,<br />
the weak bosons and the photon do not share<br />
much in common: the photon has been<br />
known as a physical entity for nearly eighty<br />
years, but the weak vector bosons were not<br />
observed until late 1982 and early 1983 at the<br />
CERN proton-antiproton collider in the<br />
highest energy accelerator experiments ever<br />
77
performed; the mass of the photon is consistent<br />
with zero, whereas the weak vector bosons<br />
have huge masses (a little less than 100<br />
GeV/c*); electromagnetic interactions can<br />
take place over very large distances, whereas<br />
the weak interactions take place on a distance<br />
scale of about centimeter; and<br />
finally, the photon has no electnc charge,<br />
whereas the weak vector bosons carry the<br />
electric and weak charges of the electroweak<br />
interactions. Moreover, in the early days of<br />
gauge theories, it was generally believed, al-<br />
though incorrectly, that local symmetry of a ue<br />
>--<br />
Lagrangian implies masslessness for the vector<br />
bosons.<br />
How can particles as different as the ,<br />
photon and the weak bosons possibly be ~ -<br />
>--<br />
’ Any Charged <strong>Particle</strong><br />
Photon Electromagnetic U(1)<br />
(QED)<br />
Any Charged<br />
w+<br />
Electroweak SU(2) x U( 1)<br />
unified by local symmetry? The answer is<br />
explained in detail in the Lecture Notes; we<br />
mention here merely that if the vacuum of<br />
a locally symmetric theory has a nonzero<br />
symmetry charge density due to the<br />
presence of a spinless field, then the vector<br />
boson associated with that symmetry acquires<br />
a mass. Solutions to the equations of<br />
dwe+<br />
motion in which the vacuum is not invariant -<br />
under symmetry transformations are called<br />
Conjectured<br />
spontaneously broken solutions, and the vec-<br />
Strongtor<br />
boson mass can be arbitrarily large<br />
Electroweak<br />
without upsetting the symmetry of the La- I U<br />
Unification<br />
grangian.<br />
(Proton Decay)<br />
i<br />
In the electroweak theory spontaneous<br />
symmetry breaking separates the weak and<br />
electromagnetic interactions and is the most<br />
important mechanism for generating masses<br />
of the elementary particles. In the theories<br />
dicussed below, spontaneous symmetry<br />
breaking is often used to distinguish interactions<br />
that have been unified by extending<br />
symmetries (see Note 8).<br />
lhe range of validity of the electroweak<br />
theory is an important issue, especially when<br />
considering extensions and generalizations<br />
to a theory of broader applicability. “Range<br />
of validity’’ refers to the energy (or distance)<br />
sale over which the predictions of a theory<br />
arc: valid. The old Fermi theory gives a good<br />
account of the weak interactions for energies<br />
less than 50 GeV, but at higher energies,<br />
where the effect of the weak bosons is to<br />
78
Toward a Unified Theory<br />
Number of<br />
Relative<br />
Vector Range of Strength at Mass<br />
Bosons Force Low Enerav Scale<br />
1 (photon) Infinite 1/137<br />
8 (gluons)<br />
4 (3 weak<br />
bosons, 1<br />
photon)<br />
(Graviton)<br />
24<br />
10-13 cm<br />
10-15 cm<br />
(weak)<br />
Infinite<br />
I oWz9 cm<br />
__<br />
. ...<br />
1<br />
I 0-5<br />
GFIf2 = 290 GeVlc2<br />
10-38 G$/2 = 1.2 X 10'9GeVfc2<br />
10-32<br />
.__ -~ - ~<br />
"_.__._.I." ...... .<br />
I OI5 GeV/cZ<br />
Mass Scale: There is no universal definition of mass scale in particle physics. It is,<br />
however, possible to select a mass scale of physical significance for each of these<br />
theories. For example, in the electroweak and SU(5) theories the mass scale is<br />
associated with the spontaneous symmetry breaking. In both cases the vacuum value<br />
of a scalar field (which has dimensions of mass) has a nonzero value. In the weak<br />
interactions GF is related directly to this vacuum value (see Fig. 2) and, at the same<br />
time, to the masses of the weak bosons. Similarly, the scale of the SU(5) model is<br />
related to the proton-decay rate and to the vacuum value of a different scalar field. In<br />
the Fermi theory GF is the strengt<br />
the strength of the gravitational<br />
massless graviton, the origin of th<br />
be related to a vacuum value but<br />
eak interaction in the Sam<br />
on. However, in gravity<br />
value of GN is not well under<br />
precisely the way that GF is.)<br />
scale is defined in a completely diEerent way. Aside from the quark masses, the<br />
classical QCD Lagrangian has no mass scales and no scalar fields. However, in<br />
quantum field theory the coupling of a gluon to a quark current depends on the<br />
momentum carried by the gluon, and this coupling is found to be large for momentum<br />
transfers below 200 MeVlc. It is thus customary to select p = 200 MeV/c2 (where p is<br />
the parameter governing the scale of asymptotic freedom) as the mass scale for QCD.<br />
spread out the weak interactions in spacetime,<br />
the Fermi theory fails. The electroweak<br />
theory remains a consistent quantum field<br />
theory at energies far above a few hundred<br />
GeV and reduces to the Fermi theory (with<br />
the modification for parity violation) at<br />
lower energies. Moreover, it correctly<br />
predicts the masses of the weak vector bosons.<br />
In fact, until experiment proves otherwise,<br />
there are no logical impediments to<br />
extending the electroweak theory to an<br />
energy scale as large as desired. Recall the<br />
example of electrodynamics and its quantum-mechanical<br />
generalization. As a theory<br />
of light in the mid-19th century, it could be<br />
tested to about IO-' centimeter. How could it<br />
have been known that QED would still be<br />
valid for distance scales ten orders of magnitude<br />
smaller? Even today it is not known<br />
where quantum electrodynamics breaks<br />
down.<br />
Strong Interactions. Quantum chromodynamics<br />
is the candidate theory of the<br />
strong interactions. It, too, is a quantum field<br />
theory based on a local symmetry; the symmetry,<br />
called color SU(3), has eight independent<br />
kinds of transformations, and so the<br />
strong interactions among the quark fields<br />
are mediated by eight vector bosons, called<br />
gluons. Apparently, the local symmetry of<br />
the strong interaction theory is not spontaneously<br />
broken. Although conceptually<br />
simpler, the absence of symmetry breaking<br />
makes it harder to extract experimental<br />
predictions. The exact SU(3) color symmetry<br />
may imply that the quarks and gluons, which<br />
carry the SU(3) color charge, can never be<br />
observed in isolation. There seem to be no<br />
simple relationships between the quark and<br />
gluon fields of the theory and the observed<br />
structure of hadrons (strongly interacting<br />
particles). The quark model of hadrons has<br />
not been rigorously derived from QCD.<br />
One of the main clues that quantum<br />
chromodynamics is correct comes from the<br />
results of "deep" inelastic scattering experiments<br />
in which leptons are used to probe the<br />
structure of protons and neutrons at very<br />
short distance intervals. The theory predicts<br />
79
that at very high momentum transfers or,<br />
equivalently, at very short distances (
Toward a Unified Theory<br />
I<br />
I<br />
I<br />
+.‘<br />
rm<br />
CI<br />
E<br />
0<br />
m<br />
C<br />
.-<br />
n<br />
J<br />
0<br />
0.1<br />
0.0’<br />
I \<br />
1 o2 10’5<br />
Mass (GeV/c2 1<br />
ig. 3. Unification in the SU(5) model. The values of the SU(2), U(l), and SU(3)<br />
mplings in the SU(5) model are shown as functions of mass scale. These values<br />
re calculated using the renormalization group equations of quantum field theory.<br />
t the unification energy scale the proton-decay bosons begin to contribute to the<br />
!normalization group equations; at higher energies, the ratios track together along<br />
te solid curve. If the high-mass bosons were not included in the calculation, the<br />
iuplings would follow the dashed curves.<br />
iodest efforts to unify the fundamental inractions<br />
may be an important step toward<br />
icluding gravity. Moreover, these efforts reuire<br />
the belief that local gauge theories are<br />
irrect to distance intervals around<br />
mtimeter, and so they have made theorists<br />
lore “comfortable” when considering the<br />
ctrapolation to gravity, which is only four<br />
rders of magnitude further. Whether this<br />
utlook has been misleading remains to be<br />
:en. The components of the standard model<br />
-e summarized in Table I.<br />
Iectroweak-Strong Unification<br />
rithout Gravity<br />
The SU(2) X U( I ) X SU(3) local theory is a<br />
:tailed phenomenological framework in<br />
hich to analyze and correlate data on elecoweak<br />
and strong interactions, but the<br />
ioice of symmetry group, the charge assignients<br />
of the scalars and fermions, and the<br />
dues of many masses and couplings must<br />
: deduced from experimental data. The<br />
-oblem is to find the simplest extension of<br />
is part of the standard model that also<br />
iifies (at least partially) the interactions,<br />
assignments, and parameters that must be<br />
put into it “by hand.” Total success at unification<br />
is not required at this stage because<br />
the range of validity will be restricted by<br />
gravitational effects.<br />
One extension is to a local symmetry<br />
group that includes SU(2) X U(1) X SU(3)<br />
and interrelates the transformations of the<br />
standard model by further internal symmetry<br />
transformations. The simplest example<br />
is the group SU(5), although most of the<br />
comments below also apply to other<br />
proposals for electroweak-strong unification.<br />
The SU(5) local symmetry implies new constraints<br />
on the fields and parameters in the<br />
theory. However, the theory also includes<br />
new interactions that mix the electroweak<br />
and strong quantum numbers; in SU(5) there<br />
are vector bosons that transform quarks to<br />
leptons and quarks to antiquarks. These vector<br />
bosons provide a mechanism for proton<br />
decay.<br />
Ifthe SU(5) local symmetry were exact, all<br />
the couplings of the vector bosons to the<br />
symmetry currents would be equal (or related<br />
by known factors), and consequently<br />
the proton decay rate would be near the weak<br />
decay rates. Spontaneous symmetry breaking<br />
of SU(5) is introduced into the theory to<br />
separate the electroweak and strong interactions<br />
from the other SU(5) interactions as<br />
well as to provide a huge mass for the vector<br />
bosons mediating proton decay and thereby<br />
reduce the predicted decay rate. To satisfy<br />
the experimental constraint that the proton<br />
lifetime be at least IO3’ years, the masses of<br />
the heavy vector bosons isn the SU( 5 ) model<br />
must be at least IOl4 GeV/c2. Thus, experimental<br />
facts already determine that the<br />
electroweak-strong unification must introduce<br />
masses into the theory that are<br />
within a factor of IO5 ofthe Planck mass.<br />
It is possible to calculate the proton lifetime<br />
in the SU(5) model and similar unified<br />
models from the values of the couplings and<br />
masses of the particles in the theory. The<br />
couplings of the standard model (the two<br />
electroweak couplings and the strong coupling)<br />
have been measured in low-energy<br />
processes. Although the ratios of the couplings<br />
are predicted by SU(5), the symmetry<br />
values are accurate only at energies where<br />
SU(5) looks exact, which is at energies above<br />
the masses of the vector bosons mediating<br />
proton decay. In general, the strengths of the<br />
couplings depend on the mass scale at which<br />
they are measured. Consequently, the SU(5)<br />
ratios cannot be directly compared with the<br />
values measured at low energy. However, the<br />
renormalization group equations of field theory<br />
prescribe how they change with the mass<br />
scale. Specifically, the change of the coupling<br />
at a given mass scale depends only on all the<br />
elementary particles with masses less than<br />
that mass scale. Thus, as the mass scale is<br />
lowered below the mass of the proton-decay<br />
bosons, the latter must be omitted from the<br />
equations, so the ratios of the couplings<br />
change from the SU(5) values. If we assume<br />
that the only elementary fields contributing<br />
to the equations are the low-mass fields<br />
known experimentally and if the protondecay<br />
bosons have a mass of IOl4 GeVlc’<br />
(see Fig. 3), then the low-energy experimental<br />
ratios of the standard model couplings are<br />
predicted correctly by the renormalization<br />
group equations but the proton lifetime<br />
81
prediction is a little less than the experimental<br />
lower bound. However, adding a few<br />
more “low-mass’’ (say, less than IO’*<br />
GeV/c:’) particles to the equations lengthens<br />
the lifetime predictions, which can thereby<br />
be pushed well beyond the limit attainable in<br />
present -day experiments.<br />
Thus, using the proton-lifetime bound<br />
directly and the standard model couplings at<br />
low mass scale, we have seen that electroweak-strong<br />
unification implies mass<br />
scales close to the scale where gravity must<br />
be included. Even if it turns out that the<br />
electroweak-strong unification is not exactly<br />
correct, it has encouraged the extrapolation<br />
of present theoretical ideas well beyond the<br />
energies available in present accelerators.<br />
Electroweak-strong unified models such as<br />
SU(5) achieve only a partial unification. The<br />
vector bosons are fully unified in the sense<br />
that they and their interactions are determined<br />
by the choice of SU(5) as the local<br />
symmetry. However, this is only a partial<br />
unificaition. The choice of fermion and scalar<br />
multiplets and the choice of symmetrybreaking<br />
patterns are left to the discretion of<br />
the physicist, who makes his selections based<br />
on low-energy phenomenology. Thus, the<br />
“unification” in SU(5) (and related local<br />
symmetries) is far from complete, except for<br />
the vector bosons. (This suggests that theories<br />
in which all particles are more closely<br />
related to the vector bosons might remove<br />
some of the arbitrariness; this will prove to<br />
be the case for supergravity.)<br />
In summary, strong experimental evidence<br />
for electroweak-strong unification,<br />
such as proton decay, would support the<br />
study of quantum field theories at energies<br />
just below the Planck mass. From the vantage<br />
of these theories, the electroweak and<br />
strong. interactions should be the low-energy<br />
limit of the unifying theory, where “low<br />
energy” corresponds to the highest energies<br />
available at accelerators today! Only future<br />
experiments will help decide whether the<br />
standard model is a complete low-energy<br />
theory, or whether we are repeating the ageold<br />
error of omitting some low-energy interactions<br />
that are not yet discovered. Never-<br />
82<br />
theless, the quest for total unification of the<br />
laws of Nature is exciting enough that these<br />
words of caution are not sufficient to delay<br />
the search for theories incorporating gravity.<br />
Toward Unification with Gravity<br />
Let us suppose that the standard model<br />
including gravity is the correct set of theories<br />
to be unified. On the basis of the previous<br />
discussion, we also accept the hypothesis that<br />
quantum field theory with local symmetry is<br />
the correct theoretical framework for extrapolating<br />
physical theory to distances perhaps<br />
as small as the Planck length. Quantum<br />
field theory assumes a mathematical model<br />
of space-time called a manifold. On large<br />
scales a manifold can have many different<br />
topologies, but at short enough distance<br />
scale, a manifold always looks like a flat<br />
(Minkowski) space, with space and time infinitely<br />
divisible. This might not be the structure<br />
of space-time at very small distances,<br />
and the manifold model of space-time might<br />
fail. Nevertheless, all progress at unifying<br />
gravity and the other interactions described<br />
here is based on theories in which space-time<br />
is assumed to be a manifold.<br />
Einstein’s theory of gravity has fascinated<br />
physicists by its beauty, elegance, and correct<br />
predictions. Before examining efforts to extend<br />
the theory to include other interactions,<br />
let us review its structure. Gravity is a<br />
“geometrical” theory in the following sense.<br />
The shape or geometry of the manifold is<br />
determined by two types of tensors, called<br />
curvature and torsion, which can be constructed<br />
from the gravitational field. The<br />
Lagrangian of the gravitational field depends<br />
on the curvature tensor. In particular, Einstein’s<br />
brilliant discovery was that the<br />
curvature scalar, which is obtained from the<br />
curvature tensor, is essentially a unique<br />
choice for the kinetic energy of the gravitational<br />
field. The gravitational field calculated<br />
from the equations of motion then determines<br />
the geometry of the space-time<br />
manifold. <strong>Particle</strong>s travel along “straight<br />
lines” (or geodesics) in this space-time. For<br />
example, the orbits of the planets are<br />
geodesics of the space-time whose geometry<br />
is determined by the sun’s gravitational field.<br />
In Einstein’s gravity all the remaining<br />
fields are called matter fields. The Lagrangian<br />
is a sum of two terms:<br />
the matter fields together with the gravitational<br />
field in something like a curvature<br />
scalar and thereby eliminate 9mat,er In addition,<br />
generalizing the graviton field in this<br />
way might lead to a consistent (renormalizable)<br />
quantum theory of gravity.<br />
There are reasons to hope that the problem<br />
of finding a renormalizable theory of gravity<br />
is solved by superstrings, although the prool<br />
is far from complete. For now, we discuss the<br />
unification of the graviton with othcr field:<br />
without concern for renormalizability.<br />
We will discuss several ways to find mani.<br />
folds for which the curvature scalar depend!<br />
on many fields, not just the gravitationa
Toward a Unified Theory<br />
apply to symmetries such as supersymmetry,<br />
with its anticommuting generators.<br />
These two loopholes in the assumptions of<br />
the theorem have suggested two directions of<br />
research in the attempt to unify gravity with<br />
the other interactions. First, we might suppose<br />
that the dimensionality of space-time is<br />
greater than four, and that spontaneous symmetry<br />
breaking of the PoincarC invariance of<br />
this larger space separates 4-dimensional<br />
space-time from the other dimensions. The<br />
symmetries of the extra dimensions can then<br />
correspond to internal symmetries, and the<br />
symmetries of the states in four dimensions<br />
need not imply an unsatisfactory infinity of<br />
states. A second approach is to extend the<br />
PoincarC symmetry to supersymmetry,<br />
which then requires additional fermionic<br />
fields to accompany the graviton. A combination<br />
of these approaches leads to the<br />
most interesting theories.<br />
Higher Dimensional Space-Time<br />
second dimension, which is wound up in a circle, becomes visible. If space-time has<br />
more than four dimensions, then the extra dimensions could have escaped detection<br />
ifeach is wound into a circle whose radius is less than centimeter.<br />
field. This generally requires extending the 4-<br />
dimensional space-time manifold. The fields<br />
and manifold must satisfy many constraints<br />
before this can be done. All the efforts to<br />
Jnifygravity with the other interactions have<br />
Jeen formulated in this way, but progress<br />
was not made until the role of spontaneous<br />
symmetry breaking was appreciated. As we<br />
low describe, it is crucial for the solutions of<br />
he theory to have less symmetry than the<br />
,agrangian has.<br />
In the standard model the generators of<br />
he space-time PoincarC symmetry commute<br />
vith (are independent of) the generators of<br />
he internal symmetries of the electroweak<br />
nd strong interactions. We might look for a<br />
local symmetry that interrelates the spacetime<br />
and internal symmetries, just as SU(5)<br />
interrelates the electroweak and strong internal<br />
symmetries. Unfortunately, if this<br />
enlarged symmetry changes simultaneously<br />
the internal and space-time quantum<br />
numbers of several states of the same mass,<br />
then a theorem of quantum field theory requires<br />
the existence of an infinite number of<br />
particles of that mass. However, this seemingly<br />
catastrophic result does not prevent the<br />
unification of space-time and internal symmetries<br />
for two reasons: first, all symmetries<br />
of the Lagrangian need not be symmetries of<br />
the states because of spontaneous symmetry<br />
breaking; and second, the theorem does not<br />
If the dimensionality of space-time is<br />
greater than four, then the geometry of spacetime<br />
must satisfy some strong observational<br />
constraints. In a 5-dimensional world the<br />
fourth spatial direction must be invisible to<br />
present experiments. This is possible if at<br />
each 4-dimensional space-time point the additional<br />
direction is a little circle, so that a<br />
tiny person traveling in the new direction<br />
would soon return to the starting point. Theories<br />
with this kind of vacuum geometry are<br />
generically called Kaluza-Klein theories.'<br />
It is easy to visualize this geometry with a<br />
two-dimensional analogue, namely, a long<br />
pipe. The direction around the pipe is<br />
analogous to the extra dimension, and the<br />
location along the pipe is analogous to a<br />
location in 4-dimensional space-time. If the<br />
means for examining the structure of the<br />
pipe are too coarse to see distance intervals<br />
as small as its diameter, then the pipe appears<br />
I-dimensional (Fig. 4). If the probe of<br />
the structure is sensitive to shorter distances,<br />
the pipe is a 2-dimensional structure with<br />
one dimension wound up into a circle.<br />
83
I<br />
The physically interesting solutions of<br />
Einstein’s 4-dimensional gravity are those in<br />
which, if all the matter is removed, spacetime<br />
is flat. The 4-dimensional space-time<br />
we see around us is flat to a good approximation;<br />
it takes an incredibly massive hunk of<br />
high-density (much greater than any density<br />
observed on the earth) matter to curve space.<br />
However, it might also be possible to construct<br />
a higher dimensional theory in which<br />
our 4-dimensional space-time remains flat in<br />
the absense of identifiable matter, and the<br />
extra dimensions are wound up into a “little<br />
ball.” \Ye must study the generalizations of<br />
Einstein’s equations to see whether this can<br />
happen, and if it does, to find the geometry of<br />
the extra dimensions.<br />
The Cosmological Constant Problem. Before<br />
we examine the generalizations of gravity in<br />
more detail, we must raise a problem that<br />
pervades all gravitational theories. Einstein’s<br />
equations state that the Einstein tensor<br />
(which is derived from the curvature scalar<br />
in finding the equations of motion from the<br />
Lagrangian) is proportional to the energymomentum<br />
tensor. If, in the absence of all<br />
matter and radiation, the energy-momentum<br />
tensor is zero, then Einstein’s equations are<br />
solved by flat space-time and zero gravitational<br />
field. In 4-dimensional classical general<br />
relativity, the curvature of space-time<br />
and the gravitational field result from a<br />
nonzero energy-momentum tensor due to<br />
the presence of physical particles.<br />
However, there are many small effects,<br />
such as other interactions and quantum effects,<br />
not included in classical general relativity.<br />
that can radically alter this simple<br />
picture. For example, recall that the electroweak<br />
theory is spontaneously broken,<br />
which means that the scalar field has a<br />
nonzero vacuum value and may contribute<br />
to the vacuum value of the energy-momentum<br />
tensor. If it does, the solution to the<br />
Einstein equations in vacuum is no longer<br />
flat space but a curved space in which the<br />
curvature increases with increasing vacuum<br />
energy. Thus, the constant value of the potential<br />
energy, which had no effect on the<br />
weak interactions, has a profound effect on<br />
gravity.<br />
At first glance, we can solve this difficulty<br />
in a trivial manner: simply add a constant to<br />
the Lagrangian that cancels the vacuum<br />
energy, and the universe is saved. However,<br />
we may then wish to compute the quantummechanical<br />
corrections to the electroweak<br />
theory or add some additional fields to the<br />
theory; both may readjust the vacuum<br />
energy. For example, electroweak-strong unification<br />
and its quantum corrections will<br />
contribute to the vacuum energy. Almost all<br />
the details of the theory must be included in<br />
calculating the vacuum energy. So, we could<br />
repeatedly readjust the vacuum energy as we<br />
learn more about the theory, but it seems<br />
artificial to keep doing so unless we have a<br />
good theoretical reason. Moreover, the scale<br />
of the vacuum energy is set by the mass scale<br />
of the interactions. This is a dilemma. For<br />
example, the quantum corrections to the<br />
electroweak interactions contribute enough<br />
vacuum energy to wind up our 4-dimensional<br />
space-time into a tiny ball about<br />
centimeter across, whereas the scale of the<br />
universe is more like IOzs centimeters. Thus,<br />
the observed value of the cosmological constant<br />
is smaller by a factor of IOs2 than the<br />
value suggested by the standard model.<br />
Other contributions can make the theoretical<br />
value even larger. This problem has the innocuous-sounding<br />
name of “the cosmological<br />
constant problem.” At present<br />
there are no principles from which we can<br />
impose a zero or nearly zero vacuum energy<br />
on the 4-dimensional part of the theory, although<br />
this problem has inspired much research<br />
effort. Without such a principle, we<br />
can safely say that the vacuum-energy<br />
prediction of the standard model is wrong.<br />
At best, the theory is not adequate to confront<br />
this problem.<br />
If we switch now to the context of gravity<br />
theories in higher dimensions, the difficult<br />
question is not why the extra dimensions are<br />
wound up into a little ball, but why our 4-<br />
dimensional space-time is so nearly flat,<br />
since it would appear that a large cosmological<br />
constant is more natural than a<br />
small one. Also, it is remarkable that the<br />
vacuum energy winding the extra<br />
dimensions into a little ball is conceptually<br />
similar to the vacuum charge of a local symmetry<br />
providing a mass for the vector bosons.<br />
However, in the case of the vacuum<br />
geometry, we have no experimental data tha<br />
bear on these speculations other than thc<br />
remarkable flatness of our 4-dimensiona<br />
space-time. The remaining discussion of uni<br />
fication with gravity must be conducted ir<br />
ignorance of the solution to the cosmologica<br />
constant problem.<br />
Internal Symmetries<br />
from Extra Dimensions<br />
The basic scheme for deriving local sym<br />
metries from higher dimensional gravity wa<br />
pioneered by Kaluza and Klein’ in the 1920$<br />
before the weak and strong interactions wer<br />
recognized as fundamental. Their attempt<br />
to unify gravity and electrodynamics in fou<br />
dimensions start with pure gravity in fivi<br />
dimensions. They assumed that the vacuun<br />
geometry is flat 4-dimensional space-timl<br />
with the fifth dimension a little loop of de<br />
finite radius at each space-time point, just a<br />
in the pipe analogy of Fig. 4. The Lagrangiai<br />
consists of the curvature scalar, constructec<br />
from the gravitational field in fivi<br />
dimensions with its five independent com<br />
ponents. The relationship of a higher dimen<br />
sional field to its 4-dimensional fields is sum<br />
marized in Fig. 5 and the sidebar, “Field<br />
and Spin in Higher Dimensions.” The in<br />
finite spectrum in four dimensions include<br />
the massless graviton (two helicity compo<br />
nents of values +2), a massless vector bosoi<br />
(two helicity components of *I), a massles<br />
scalar field (one helicity component of 0:<br />
and an infinite series of massive spin-.<br />
pyrgons of increasing masses. (The tern<br />
“pyrgon” derives from mjpyoo, the Creel<br />
word for tower.) The Fourier expansion fo<br />
each component of the gravitational field i<br />
identical to Eq. 1 of the sidebar. Since thl<br />
extra dimension is a circle, its symmetry is I<br />
phase symmetry just as in electrodynamic<br />
1
.-'<br />
Toward a Unified Theory<br />
D-Dimensional<br />
Of Spin<br />
Field Relabeled<br />
in Terms of<br />
4-Dimensional Spin Ji<br />
Infinite<br />
Towers of<br />
4-Dimensional<br />
Fields<br />
Zero Mode(s) @;"(x)<br />
(Massless) *<br />
4-Dimensional<br />
Space-Time Directions<br />
i<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
I<br />
'ig. 5. A field in D dimensions unifies fields of different<br />
Dins and masses in four dimensions. In step 1 the spin<br />
omponents of a single higher dimensional spin are resolved<br />
rto several spins in four dimensions. (The total number of<br />
omponents remains constant.) Mathematically this is<br />
chieved by finding the spins J,, Jb ... in four dimensions<br />
kat are contained in "spin-$', of D dimensions. Step 2 is<br />
the harmonic expansion of the 4-dimensional spin components<br />
on the extra dimensions, which then resolves a single<br />
massless D-dimensional field into an infinite number of 4-<br />
dimensional fields of varying masses. When the 4-dimensional<br />
mass is zero, the corresponding 4-dimensionalfield is<br />
called a zero mode. The 4-dimensionalfields with 4-dimensional<br />
mass form an infinite sequence ofpyrgons.<br />
he symmetry ofthis vacuum state is not the more realistic theories. The zero modes no low-mass charged particles. (Adding fer-<br />
-dimensional PoincarC symmetry but the (massless particles in four dimensions) are mions to the 5-dimensional theory does not<br />
irect product of the 4-dimensional Poincare electrically neutral. Only the pyrgons carry help, because the resulting 4-dimensional<br />
'oup and a phase symmetry. electric charge. The interaction associated fermions are all pyrgons, which cannot be<br />
This skeletal theory should not be taken with the vector boson in four dimensions low mass either.) Nevertheless, the<br />
:riously, except as a basis for generalizing to cannot be electrodynamics because there are hypothesis that all interactions are conse-<br />
85
i<br />
le in our worl<br />
at the identical points<br />
n set of fields. For exa<br />
-dimensional Ei<br />
“dimensional reduction.” The dimensionally reduced theory should<br />
quences ofthe symmetries of space-time is so<br />
attractive that efforts to generalize the<br />
Kaluza-Klein idea have been vigorously<br />
pursued. These theories require a more complete<br />
discussion of the possible candidate<br />
manifolds of the extra dimensions.<br />
The geometry of the extra dimensions in<br />
the absence of matter is typically a space with<br />
a high degree of symmetry. Symmetry requires<br />
the existence of transformations in<br />
which the starting point looks like the point<br />
reached after the transformation. (For example,<br />
thc environments surrounding each<br />
point on a sphere are identical.) Two of the<br />
most important examples are “group manifolds”<br />
and “coset spaces,” which we briefly<br />
describe.<br />
The tranformations of a continuous group<br />
86<br />
are identified by N parameters, where N is<br />
the number of independent transformations<br />
in the group. For example, N = 3 for SU(2)<br />
and 8 for SU(3). These parameters are the<br />
coordinates of an N-dimensional manifold.<br />
Ifthe vacuum values of fields are constant on<br />
the group manifold, then the vacuum solution<br />
is said to be symmetric.<br />
Coset spaces have the symmetry of a group<br />
too, but the coordinates are labeled by a<br />
subset of the parameters of a group. For<br />
example, consider the space S0(3)/S0(2). In<br />
this example, SO(3) has three parameters,<br />
and SO(2) is the phase symmetry with one<br />
parameter, so the coset space S0(3)/S0(2)<br />
has three minus one, or two, dimensions.<br />
This space is called the 2-sphere, and it has<br />
the geometry of the surface of an ordinary<br />
sphere. Spheres can be generalized to an<br />
number of dimensions: the N-dimensions<br />
sphere is the coset space [SO(N+ I)]/SO(N<br />
Many other cosets, or “ratios” of group!<br />
make spaces with large symmetries. It i<br />
possible to find spaces with the symmetrie<br />
of the electroweak and strong interaction!<br />
One such space is the group manifold SU(2<br />
X U(I) X SU(3). which has twelv<br />
dimensions. More interesting is the lower<br />
dimensional space with those symmetrie:<br />
namely, the coset space [SU(3) X SU(2) :<br />
U(I)]/[SU(2) X U(l) X U(I)], which ha<br />
dimension 8 + 3 + I - 3 - I - 1 = 7. (TI-<br />
SU(2) and the U(l)’s in the denorninatc<br />
differ from those in the numerator, so the<br />
cannot be “canceled.”)Thus, one might hor<br />
that (4 + 7 = I I)-dimensional gravity woul
Toward a Unified Theory<br />
Higher Dimensions<br />
describe the low-energy limit of the theory.<br />
The gravitational field can be generalized to higher (>5) dimensional<br />
manifolds, where the extra dimensions at each 4-dimensional<br />
space-time point form a little ball of finite volume. The mathematics<br />
requires a generalization of Fourier series to “harmonic” expansions<br />
on these spaces. Each field (or field component if it ha5 spin) unifies<br />
an infinite set of pyrgons. and the series may also contain some zero<br />
modes. The terms in the series correspond to fields of increasing 4-<br />
dimensional mass, just as in thc 5-d ensional example. The kinetic<br />
energy in the extra dimensions of each term in the series then<br />
corresponds to a mass in our space-time. The higher dimensional<br />
field quite generally describes mathematically an infinite number of<br />
4-dimensional fields.<br />
Spin in Higher Dimensions. The definition of spin in D dimensions<br />
depends on the D-dimensional Lorentz symmetry; Cdimensional<br />
Lorentz symmetry is naturally embedded in the D-dimensional<br />
symmetry. Consequently a D-dimensional field of a specific spin<br />
unifies 4-dimensional fields with different spins.<br />
Conceptually the description of D-dimensional spin is similar to<br />
that of spin in four dimensions. A massless particle of spin J in four<br />
dimensions has helicities +J and -Jcorresponding to the projections<br />
of spin along the direction of motion. These two helicities are singlet<br />
multiplets of the I -dimensional rotations that leave unchanged the<br />
direction of a particle traveling at the speed of light. The group of 1-<br />
dimensional rotations is the phase symmetry S0(2), and this method<br />
for identifying the physical degrees of freedom is called the “lightcone<br />
classification.” However, the situation is a little more com-<br />
plicated in five dimensions, where there are three directions orthogonal<br />
to the direction of the particle. Then the helicity symmetry<br />
becomes SO(3) (instead of S0(2)), and the spin multiplets in five<br />
dimensions group together sets of 4-dimensional helicity. For example,<br />
the graviton in five dimensions has five components. The SO(2)<br />
of four dimensions is contained in this SO( 3) symmetry, and the 4-<br />
dimensional helicities of the 5-dimensional graviton are 2, I, 0, -1,<br />
and -2.<br />
nerally, the light-cone symmetry that leaves the direction<br />
of motion of a massless particle unchanged in D dimensions is<br />
SO(D - 2), and the D-dimensional helicity corresponds to the multiplets<br />
(or representations) of SO(D - 2). For example, the graviton<br />
has D(D- 3)/2 independent degrees of freedom in D dimensions;<br />
thus the graviton in eleven dimensions belongs to a 44-component<br />
representation of SO(9). The SO(2) of the 4-dimensional helicity is<br />
inside the S0(9), so the forty-four componcnts of the graviton in<br />
eleven dimensions carry labels of 4-dimensional helicity as follows:<br />
one component of helicity 2, seven of helicity I, twenty-eight of<br />
helicity 0, seven of helicity -I and one of helicity -2. (The components<br />
of the graviton in eleven dimensions then correspond to the<br />
graviton, seven massless vector bosons, and twenty-eight scalars in<br />
four dimensions.)<br />
The analysis for massive particles in D dimensions proceeds in<br />
exactly the same way, except the helicity symmetry is the one that<br />
leaves a resting particle at rest. Thus, the massive helicity symmetry<br />
is SO(D- I). (For example, SO(3) describes the spin of a massive<br />
particle in ordinary 4-dimensional space-time.) These results are<br />
summarized in Fig. 5 of the main text.<br />
nify all known interactions.<br />
It turns out that the 4-dimensional fields<br />
nplied by the I I-dimensional gravitational<br />
eld resemble the solution to the 5-dimenonal<br />
Kaluza-Klein case, except that the<br />
-avitational field now corresponds to many<br />
lore 4-dimensional fields. There are methds<br />
of dimensional reduction for group<br />
ianifolds and coset spaces, and the zero<br />
lodes include a vector boson for each symietry<br />
of the extra dimensions. Thus, in the<br />
I + 7)-dimensional example mentioned<br />
>ow. there is a complete set of vector bosns<br />
for the standard model. At first sight this<br />
lode1 appears to provide an attractive uni-<br />
Eation ofall the interactions of the standard<br />
iodel; it explains the origins of the local<br />
mmetries of the standard model as space-<br />
time symmetries of gravity in eleven<br />
dimensions.<br />
Unfortunately, this 1 I-dimensional<br />
Kaluza-Klein theory has some shortcomings.<br />
Even with the complete freedom consistent<br />
with quantum field theory to add fermions, it<br />
cannot account for the parity violation seen<br />
in the weak neutral-current interactions of<br />
the electron. Witten’ has presented very general<br />
arguments that no I I-dimensional<br />
Kaluza-Klein theory will ever give the correct<br />
electroweak theory.<br />
Supersymmetry and Gravity in<br />
Four Dimensions<br />
We return from our excursion into higher<br />
dimensions and discuss extending gravity<br />
not by enlarging the space but rather by<br />
enlarging the symmetry. The local PoincarC<br />
symmetry of Einstein’s gravity implies the<br />
massless spin-2 graviton; our present goal is<br />
to extend the Poincari: symmetry (without<br />
increasing the number ofdirnensions) so that<br />
additional fields are grouped together with<br />
the graviton. However, this cannot be<br />
achieved by an ordinary (Lie group) symmetry:<br />
the graviton is the only known<br />
elementary spin-2 field, and the local symmetries<br />
of the standard model are internal<br />
symmetries that group together particles of<br />
the same spin. Moreover, gravity has an<br />
exceptionally weak interaction, so if the<br />
graviton carries quantum numbers of symmetries<br />
similar to those of the standard<br />
model, it will interact too strongly. We can<br />
87
accommodate these facts if the graviton is a<br />
singlet under the internal symmetry, but then<br />
its multiplet in this new symmetry must<br />
include particles of other spins. Supersymmetry’<br />
is capable of fulfilling this requirement.<br />
Four-Dimensional Supersymmetry. Supersymmetry<br />
is an extension of the Poincart<br />
symmetry, which includes the six Lorentz<br />
generators M,,” and four translations P,,. The<br />
Poincart generators are boson operators, so<br />
they can change the spin components of a<br />
massive field but not the total spin. The<br />
simplest version of supersymmetry adds fermionic<br />
generators Q, to the Poincart generators;<br />
Qa transforms like a spin-% field<br />
under Lorentz transformations. (The index a<br />
is a spinor index.) To satisfy the Pauli exclusion<br />
principle, fermionic operators in<br />
quantum field theory always satisfy anticommutation<br />
relations, and the supersymmetry<br />
generators are no exception. In the algebra<br />
the supersymmetry generators Qo anticommute<br />
to yield a translation<br />
where fp is the energy-momentum 4-vector<br />
and the yg, are matrix elements of the Dirac<br />
y matrices.<br />
The significance of the fermionic generators<br />
is that they change the spin ofa state<br />
or field by t%; that is, supersymmetry unifies<br />
bosons and fermions. A multiplet of<br />
“simple” supersymmetry (a supersymmetry<br />
with one fermionic generator) in four<br />
dimensions is a pair of particles with spins J<br />
and .I- %; the supersymmetry generators<br />
transform bosonic fields into fermionic<br />
fields and vice versa. The boson and fermion<br />
components are equal in number in all supersymmetry<br />
multiplets relevant to particle theories.<br />
We can construct larger supersymmetries<br />
by adding more fermionic generators to the<br />
PoincarC symmetry. “N-extended” supersymmetry<br />
has N fermionic generators. By<br />
applying each generator to the state of spin J,<br />
88<br />
we can lower the helicity up to N times.<br />
Thus, simple supersymmetry, which lowers<br />
the helicity just once, is called N = 1 supersymmetry.<br />
N = 2 supersymmetry can lower<br />
the helicity twice, arid the N = 2 multiplets<br />
have spins J, J- %, and J - 1. There are<br />
twice as many J - lh states as J or J - 1, so<br />
that there are equal numbers of fermionic<br />
and bosonic states. The N = 2 multiplet is<br />
made up of two N == 1 multiplets: one with<br />
spins J and J- I12 and the other with spins<br />
J- %and J- 1.<br />
In principle, this construction can be extended<br />
to any N, but in quantum field theory<br />
there appears to be a limit. There are serious<br />
difficulties in constructing simple field theories<br />
with spin 5/2 or higher. The largest<br />
extension with spin 2 or less has N = 8. In N<br />
= 8 extended supersymmetry, there is one<br />
state with helicity of 2, eight with 3/2,<br />
twenty-eight with 1, fifty-six with 1/2, seventy<br />
with 0, fifty-six with -l/2, twenty-eight<br />
with -1, eight with 3/2 and one with -2.<br />
This multiplet with 256 states will play an<br />
important role in the supersymmetric theories<br />
of gravity or supergravity discussed<br />
below. Table 2 shows the states of N-extended<br />
supersymmetry.<br />
Theories with Supersymmetry. Rather ordinary-looking<br />
Lagrangians can have supersymmetry.<br />
For example, there is a Lagrangian<br />
with simple global supersymmetry<br />
in four dimensions with a single Majorana<br />
fermion, which has one component with<br />
helicity +1/2, one with helicity -l/2, and<br />
two spinless particles. Thus, there are two<br />
bosonic and two fermionic degrees of freedom.<br />
The supersymmetry not only requires<br />
the presence of both fermions and bosons in<br />
the Lagrangian but also restricts the types of<br />
interactions, requires that the mass<br />
parameters in the multiplet be equal, and<br />
relates some other parameters in the Lagrangian<br />
that would otherwise be unconstrained.<br />
The model just described, the Wess-<br />
Zumino model,3 is so simple that it can be<br />
written down easily in conventional field<br />
notation. However, more realistic supersym-<br />
metnc Lagrangians take pages to write down,<br />
We will avoid this enormous complication<br />
and limit our discussion to the spectra 01<br />
particles in the various theories.<br />
Although supersymmetry may be an exaci<br />
symmetry of the Lagrangian, it does not appear<br />
to be a symmetry of the world because<br />
the known elementary particles do not have<br />
supersymmetric partners. (The photon and a<br />
neutrino cannot form a supermultiplet because<br />
their low-energy interactions are different.)<br />
However, like ordinary symmetries,<br />
the supersymmetries of the Lagrangian do<br />
not have to be supersymmetries of the<br />
vacuum: supersymmetry can be spontaneously<br />
broken. The low-energy predictions<br />
of spontaneously broken supersymmetric<br />
models are discussed in “Supersymmetry<br />
at 100 GeV.”<br />
mediates the gravitational interaction; lowmass<br />
spin-% fermions dominate low-energy<br />
phenomenology; and spinless fields provide<br />
the mechanism for spontaneous symmetq<br />
breaking. All these fields are crucial to thc<br />
standard model, although there seems to br<br />
no relation among the fields ofdifferent spin<br />
A spin of 3/2 is not required phenomenologi<br />
cally and is missing from the list. If thl<br />
supersymmetry is made local, the resultin,<br />
theory is supergravity, and the spin-2 gravi<br />
ton is accompanied by a “gravitino” wit1<br />
spin 312.<br />
Local supersymmetry can be imposed on<br />
theory in a fashion formally similar to th<br />
local symmetries of the standard model, ex<br />
cept for the additional complications due t<br />
the fact that supersymmetry is a space-tim<br />
symmetry. Extra gauge fields are required t<br />
compensate for derivatives of the spact<br />
time-dependent parameters, so, just as fc<br />
ordinary symmetries, there is a gauge partic
Toward a Unified Theoty<br />
number of states of each helicity for each possible supermultiplet containing a<br />
graviton but with spin I 2. Simple supergravity (N = 1) has a graviton and<br />
gravitino. N = 4 supergravity is the s<br />
The overlap of the multi<br />
gives rise to large addi<br />
supergravities have the same list o<br />
symmetry implies that the N = 7 theory must have two multiplets (as for N <<br />
7), whereas N = 8 is the first and last case for which particle-antiparticle<br />
symmetry can be satisfied by a single multiplet.<br />
Helicity 1 2 3 4 5 6 701-8<br />
2 1 1 1<br />
312 1 2 3<br />
1 1 3<br />
1 12<br />
0<br />
-112<br />
-I 1 3<br />
-312 1 2 3<br />
-2 1 1 1<br />
Total 4 8 16<br />
ymmetry transformation. However, the<br />
auge particles associated with the supersymietry<br />
generators must be fermions. Just as<br />
he graviton has spin 2 and is associated with<br />
ie local translational symmetry, the gravino<br />
has spin 312 and gauges the local superymmetry.<br />
The graviton and gravitino form<br />
simple (N = 1) supersymmetry multiplet.<br />
‘his theory is called simple supergravity and<br />
i interesting because it succeeds in unifying<br />
le graviton with another field.<br />
The Lagrangian of simple supergravity4 is<br />
n extension of Einstein’s Lagrangian, and<br />
ne recovers Einstein’s theory when the<br />
-avitational interactions ofthe gravitino are<br />
,nored. This model must be generalized to a<br />
lore realistic theory with vector bosons,<br />
1<br />
4<br />
6<br />
4<br />
2<br />
4<br />
6<br />
4<br />
I<br />
32<br />
1 1<br />
5 6<br />
10 16<br />
11 26<br />
10 30<br />
11 26<br />
10 16<br />
5 6<br />
1 1<br />
64 128<br />
1<br />
8<br />
28<br />
56<br />
70<br />
56<br />
28<br />
8<br />
1<br />
256<br />
spin-% fermions, and spinless fields to be of<br />
much use in particle theory.<br />
The generalization is to Lagrangians with<br />
extended local supersymmetry, where the<br />
largest spin is 2. The extension is extremely<br />
complicated. Nevertheless, without much<br />
work we can surmise some features of the<br />
extended theory. Table 2 shows the spectrum<br />
of particles in N-extended supergravity.<br />
We start here with the largx extended<br />
supersymmetry and investigate whether it<br />
includes the electroweak and strong interactions.<br />
In N = 8 extended supergravity the<br />
spectrum is just the N = 8 supersymmetric<br />
multiplet of 256 helicity states discussed<br />
before. The massless particles formed from<br />
these states include one graviton, eight gravi-<br />
tinos, twenty-eight vector bosons, fifty-six<br />
fermions, and seventy spinless fields.<br />
N= 8 supergravity’ is an intriguing theory.<br />
(Actually, several different N = 8 supergravity<br />
Lagrangians can be constructed.) It<br />
has a remarkable set of internal symmetries,<br />
and the choice of theory depends on which of<br />
these symmetries have gauge particles associated<br />
with them. Nevertheless, supergravity<br />
theories are highly constrained and<br />
we can look for the standard model in each.<br />
We single out one of the most promising<br />
versions of the theory, describe its spectrum,<br />
and then indicate how close it comes to<br />
unifying the electroweak, strong, and gravitational<br />
interactions.<br />
In the N = 8 supergravity of de Wit-<br />
Nicolai theory6 the twenty-eight vector bosons<br />
gauge an SO(8) symmetry found by<br />
Cremmer and Julia.’ Since the standard<br />
model needs just twelve vector bosons,<br />
twenty-eight would appear to be plenty. In<br />
the fermion sector, the eight gravitinos must<br />
have fairly large masses in order to have<br />
escaped detection. Thus, the local supersymmetry<br />
must be broken, and the gravitinos<br />
acquire masses by absorbing eight spin-%<br />
fermions. This leaves.56 -’8 = 48 spin-%<br />
fermion fields. For the quarks and leptons in<br />
the standard model, we need forty-five fields,<br />
so this number also is sufficient.<br />
The next question is whether the quantum<br />
numbers of SO(8) correspond to the electroweak<br />
and strong quantum numbers and<br />
the spin4 fermions to quarks and leptons.<br />
This is where the problems start: if we<br />
separate an SU(3) out of the SO(8) for QCD,<br />
then the only other independent ’interactions<br />
are two local phase symmetries of U( I ) X<br />
U( I), which IS not large enough to include<br />
the SU(2) X U( I ) of the electroweak theory.<br />
The rest of the SO@) currents mix the SU(3)<br />
and the two U( 1)’s. Moreover, many of the<br />
fifty-six spin-% fermion states (or forty-eight<br />
if the gravitinos are massive) have the wrong<br />
SU(3) quantum numbers to be quarks and<br />
leptons.’ Finally, even if the quantum<br />
numbers for QCD were right and the electroweak<br />
local symmetry were present, the<br />
weak interactions could still not be ac-<br />
89
counted for. No mechanism in this theory<br />
can guarantee the almost purely axial weak<br />
neutral current of the electron. Thus this<br />
interpretation of N = 8 supergravity cannot<br />
be the ultimate theory. Nevertheless, this is a<br />
model of unification, although it gave the<br />
wrong sets of interactions and particles.<br />
Perhaps the 256 fields do not correspond<br />
directly to the observable particles, but we<br />
need a more sophisticated analysis to find<br />
them. For example, there is a “hidden” local<br />
SU(8) symmetry, independent of the gauged<br />
SO(8) mentioned above, that could easily<br />
contain the electroweak and strong interactions.<br />
It is hidden in the sense that the Lagrangian<br />
does not contain the kinetic energy<br />
terms for the sixty-three vector bosons of<br />
SU(8). These sixty-three vector bosons are<br />
composites of the elementary supergravity<br />
fields, and it is possible that the quantum<br />
corrections will generate kinetic energy<br />
terms. Then the fields in the Lagrangian do<br />
not correspond to physical particles; instead<br />
the photon, electron, quarks, and so on,<br />
which look elementary on a distance scale of<br />
present experiments, are composite. Unfortunately,<br />
it has not been possible to work<br />
out a logical derivation of this kind of result<br />
for N =- 8 supergravity.*<br />
In summary, N = 8 supergravity may be<br />
correct, but we cannot see how the standard<br />
model follows from the Lagrangian. The<br />
basic fields seem rich enough in structure to<br />
account for the known interactions, but in<br />
detail they do not look exactly like the real<br />
world. Whether N = 8 supergravity is the<br />
wrong theory, or is the correct theory and we<br />
simply do not know how to interpret it, is not<br />
yet known.<br />
Supergravity in Eleven<br />
Dimensions<br />
The apparent phenomenological shortcomings<br />
of N = 8 supergravity have been<br />
known for some time, but its basic mathematical<br />
structure is so appealing that many<br />
theorists continue to work on it in hope that<br />
90<br />
some variant will give the electroweak and<br />
strong interactions. One particularly interesting<br />
development is the generalization of N =<br />
8 supergravity in four dimensions to simple<br />
(N = I ) supergravity in eleven dimensions.’<br />
This generalization combines the ideas of<br />
Kaluza-Klein theories with supersymmetry.<br />
The formulation and dimensional reduction<br />
of simple supergravity in eleven<br />
dimensions requires most of the ideas already<br />
described. First we find the fields of l l-<br />
dimensional supergravity that correspond to<br />
the graviton and gravitino fields in four<br />
dimensions. Then we describe the components<br />
of each of the 1 I-dimensional fields.<br />
Finally, we use the harmonic expansion on<br />
the extra seven dimensions to identify the<br />
zero modes and pyrgons. For a certain<br />
geometry of the extra dimensions, the<br />
dimensionally reduced, 1 I-dimensional<br />
supergravity without pyrgons is N = 8 supergravity<br />
in four dimensions; for other<br />
geometries we find new theories. We now<br />
look at each of these steps in more detail.<br />
In constructing the 1 I-dimensional fields,<br />
we begin by recalling that the helicity symmetry<br />
of a massless particle is SO(9) and the<br />
spin components are classified by the multiplets<br />
of SO(9). The multiplets of SO(9) are<br />
either fermionic or bosonic, which means<br />
that all the four-dimensional helicities are<br />
either integers (bosonic) or half-odd integers<br />
(fermionic) for all the components in a single<br />
multiplet. The generators independent of the<br />
SO(2) form an S0(7), which is the Lorentz<br />
group for the extra seven dimensions. Thus,<br />
the SO(9) multiplets can be expressed in<br />
terms of a sum of multiplets of SO(7) X<br />
S0(2), which makes it possible to reduce I I-<br />
dimensional spin to +dimensional spin.<br />
The fields of 1 I-dimensional, N= I supergravity<br />
must contain the graviton and gravitino<br />
in four dimensions. We have already<br />
mentioned in the sidebar that the graviton in<br />
eleven dimensions has forty-four bosonic<br />
components. The smallest SO(9) multiplet of<br />
I I-dimensional spin that yields a helicity of<br />
312 in four dimensions for the gravitinos has<br />
128 components, eight components with<br />
helicity 3/2, fifty-six with 1/2, fifty-six with<br />
ofN= 8 supergravity in four dimen~ions.~ In<br />
this case each of the components is expanded<br />
in a sevenfold Fourier series, one series for<br />
each dimension just as in Eq. 1 in the sidebar,<br />
except that ny is replaced by Cn,v,. The<br />
dimensional reduction consists of keeping<br />
only those fields that do not depend on any<br />
y,, that is, just the 4-dimensional fields corresponding<br />
to n, = n2 =. . . = n, = 0. Thus,<br />
there is one zero mode (massless field in four<br />
dimensions) for each component. The<br />
pyrgons are the 4-dimensional fields witb<br />
any n, # 0, and these are omitted in the<br />
dimensional reduction.<br />
The I I-dimensional theory has a simple<br />
Lagrangian, whereas the 4-dimensiona1, N =<br />
8 Lagrangian takes pages IO write down. Ir<br />
fact the N = 8 Lagrangian was first derived ir<br />
this way.’ It is easy to be impressed by :<br />
formalism in which everything looks simple<br />
This is the first of several reasons to tak<br />
seriously the proposal that the extri<br />
dimensions might be physical, not just I<br />
mathematical trick.<br />
The seven extra dimensions of the 11<br />
dimensional theory must be wound up into
Toward a Unified Theory<br />
Table 3<br />
The relation of si (N= 1) supergr<br />
supergravity in fo ensions. The 256 ents of the mass<br />
11-dimensional, N = 1 supergravity fa<br />
SO(9). The members of these multiplets have definite helicities in four<br />
dimensions. The count of helicity states is given in terms of the size of SO(7)<br />
multiplets, where SO(7) is the Lorentz symmetry of the seven extra dimensions<br />
in the 11-dimensional theory.<br />
4-Dimensional Helicitv<br />
n 2 312 1 112 0 -112 -1 -312 -2<br />
44 1 7 14-27 7 1<br />
84 21 7+35 21<br />
I28 8 8+48 8+48 8<br />
Total 1 8 28 56 70 56 28 8 1<br />
ase described above assumes that the little<br />
all is a 7-torus, which is the group manifold<br />
lade of the product of seven phase symietries.<br />
As a Kaluza-Klein theory, the seven<br />
ector bosons in the graviton (Table 3) gauge<br />
lese seven symmetries. Since the twentyight<br />
vector bosons of N= 8 supergravity can<br />
e the gauge fields for a local S0(8), it is<br />
iteresting to see if we can redo the dimenional<br />
reduction so that I I-dimensional<br />
upergravity is a Kaluza-Klein theory for<br />
0(8), the de Wit-Nicolai theory. Indeed,<br />
lis is possible. If the extra dimensions are<br />
ssumed to be the 7-sphere, which is the<br />
oset space SO(S)/S0(7), the vector bosons<br />
o gauge SO(8).io This is, perhaps, the ulmate<br />
Kaluza-Klein theory, although it does<br />
ot contain the standard model. The main<br />
ifference between the 7-torus and coset<br />
paces is that for coset spaces there is not<br />
ecessarily a one-to-one correspondence beween<br />
components and zero modes. Some<br />
omponents may have several zero modes,<br />
chile others have none (recall Fig. 5).<br />
There are other manifolds that solve the<br />
I-dimensional supergravity equations, aliough<br />
we do not describe them here. The<br />
iternal local symmetries are just those of the<br />
extra dimensions, and the fermions and bosons<br />
are unified by supersymmetry. Thus, 1 1-<br />
dimensional supergravity can be dimensionally<br />
reduced to one of several different 4-<br />
dimensional supergravity theories, and we<br />
can search through these theories for one that<br />
contains the standard model. Unfortunately,<br />
they all suffer phenomenological shortcomings.<br />
Eleven-dimensional supergravity contains<br />
an additional error. In the solution where the<br />
seven extra dimensions are wound up in a<br />
little ball, our 4-dimensional world gets just<br />
as compacted: the cosmological constant is<br />
about 120 orders of magnitude larger than is<br />
observed experimentally.” This is the cosmological<br />
constant problem at its worst. Its<br />
solution may be a major breakthrough in the<br />
search for unification with gravity. Meanwhile,<br />
it would appear that supergravity has<br />
given us the worst prediction in the history of<br />
modern physics!<br />
Superstrings<br />
In view of its shortcomings, supergravity<br />
is apparently not the unified theory of all<br />
elementary particle interactions. In many<br />
ways it is close to solving the problem, but a<br />
theory that is correct in all respects has not<br />
been found. The weak interactions are not<br />
exactly right nor is the list of spin4 fermions.<br />
There seems to be no good reason<br />
that the cosmological constant should be<br />
nearly or exactly zero as observed experimentally.<br />
The issue of the renormalizability<br />
of the quantum theory of gravity<br />
also remains unsolved. Supergravity improves<br />
the quantum structure of the theory<br />
in that the unwanted infinities are not as bad<br />
as in Einstein’s theory with matter, but<br />
troubles still appear. Newton’s constant is a<br />
fundamental parameter in the theory, and 4-<br />
fermion terms similar to those in Fermi’s<br />
weak interaction theory are still present. In N<br />
= 8 supergravity, which is the best case, the<br />
perturbation solution to the quantum field<br />
theory is expected to break down eventually.<br />
In spite of these difficulties we have<br />
reasons to be optimistic that supergravity is<br />
on the right track. It does unify gravity with<br />
some interactions and is almost a consistent<br />
quantum field theory. The line of generalization<br />
followed so far has led to theories that<br />
are enormous improvements, in a mathematical<br />
sense, over Einstein’s gravity. It<br />
would seem reasonable to look for generalizations<br />
beyond supergravity.<br />
Superstring theories may answer some of<br />
these questions. Just as the progress of supergravity<br />
was based on the systematic addition<br />
of fields to Einstein’s gravity, superstring<br />
theory can also be viewed in terms of the<br />
systematic addition of fields to supergravity.<br />
Although the formulation of superstring theory<br />
looks quite different from the formulation<br />
of supergravity, this may be partially<br />
due to its historical origin.<br />
Superstring theories were born from an<br />
early effort to find a theory of the strong<br />
interactions. They began as a very efficient<br />
means of understanding the long list of<br />
hadronic resonances. In particular, hadrons<br />
of high spin have been identified expenmentally.<br />
It is interesting that sets of hadrons of<br />
different spins but the same internal quantum<br />
numbers can be grouped together into<br />
91
“Regge trajectories.” Figure 6 shows examples<br />
of :Regge trajectories (plots of spin versus<br />
mass-squared) for the first few states of the A<br />
and N resonances; these resonances for<br />
hadrons of different spins fall along nearly<br />
straighl. lines. Such sequences appear to be<br />
general phenomena, and so, in the ’60s and<br />
early OS, a great effort was made to incorporate<br />
these results directly into a theory.<br />
The basic idea was to build a set of hadron<br />
amplitudes with rising Regge trajectories<br />
that satisfied several important constraints<br />
of quantum field theory, such as Lorentz<br />
invariance, crossing symmetry, the correct<br />
analytic properties, and factorization of resonance-pole<br />
residues. ’’ Although the theory<br />
was a prescription for calculating the<br />
amplitudes, these constraints are true of<br />
quantum field theory and are necessary for<br />
the theory to make sense.<br />
The constraints of field theory proved to<br />
be too much for this theory of hadrons.<br />
Something always went wrong. Some theories<br />
predicted particles with imaginary mass<br />
(tachyons) or particles produced with<br />
negative probability (ghosts), which could<br />
not be interpreted. Several theories had no<br />
logical difficulties, but they did not look like<br />
hadron theories. First of all, the consistency<br />
requirements forced them to be in ten<br />
dimensions rather than four. Moreover, they<br />
predicted massless particles with a spin of 2;<br />
no hadrons of this sort exist. These original<br />
superstring theories did not succeed in describing<br />
hadrons in any detail, but the solution<br />
of QCD may still be similar to one of<br />
them.<br />
In 1074 Scherk and Schwarzi3 noted that<br />
the quantum amplitudes for the scattering of<br />
the massless spin-2 states in the superstring<br />
are the same as graviton-graviton scattering<br />
in the simplest approximation of Einstein’s<br />
theory. They then boldly proposed throwing<br />
out the hadronic interpretation of the superstring<br />
and reinterpreting it as a fundamental<br />
theory of elementary particle interactions. It<br />
was easily found that superstrings are closely<br />
related to supergravity, since the states fall<br />
into supersymmetry multiplets and massless<br />
spin-2 particles are required.14<br />
.-<br />
E<br />
n<br />
v)<br />
1112<br />
912<br />
712<br />
512<br />
312<br />
1 I2<br />
Fig. 6. Regge trajectories in hadron physics. The neutron and proton (N(938))<br />
on a linearly rising Regge trajectory with other isospin-% states: the N(l680)<br />
spin 5/2, the N(2220) of spin 9/2, and so on. This fact can be interpreted as meani<br />
that the N(1680), for example, looks like a nucleon except that the quarks are in<br />
F wave rather than a P wave. Similarly the isospin-3/2 A resonance at 1232 M<br />
lies on a trajectory with other isospin-3/2 states of spins 7/2, I I ~ 1512, , and so t<br />
The slope of the hadronic Regge trajectories is approximately (1 GeV/cZ)-’. 1<br />
slope of the superstring trajectories must be much smaller<br />
The theoretical development of superstrings<br />
is not yet complete, and it is not<br />
possible to determine whether they will finally<br />
yield the truly unified theory of all<br />
interactions. They are the subject of intense<br />
research today. Our plan here is to present a<br />
qualitative description of superstrings and<br />
then to discuss the types and particle spectra<br />
of superstring theories.<br />
Recent formulations of superstring theories<br />
are generalizations of quantum field<br />
theory.I5 The fields of an ordinary field t<br />
ory, such as supergravity, depend on<br />
space-time point at which the field<br />
evaluated. The fields of superstring the(<br />
depend on paths in space-time. At each n<br />
ment in time, the string traces out a path<br />
space, and as time advances, the str<br />
propagates through space forming a surf<br />
called the “world sheet.” Strings can<br />
closed, like a rubber band, or open, lik<br />
broken rubber band. Theories of both ty,<br />
92
__<br />
-<br />
- , ,<br />
i<br />
Toward a Unified Theory<br />
- ___I-<br />
(a)<br />
- ____ - ~<br />
*2<br />
‘3<br />
___ ,<br />
Fig. 7. Dynamics of closed strings. Thefigures show the string configurations at a<br />
sequence of times (in two dimensions instead of ten). In Fig. 7(a) a string in motion<br />
from times t, to t2 traces out a world sheet. Figure 7(b) shows the three closed string<br />
interaction, where one string at t, undergoes a change of shape until itpinches off at<br />
apoint at time t2 (the interaction time). At time t3 two strings arepropagating away<br />
from the interaction region.<br />
are promising, but the graviton is always<br />
associated with closed strings.<br />
Before analyzing the motion of a superstring,<br />
we must return to a discussion of<br />
space-time. Previously, we described extensions<br />
of space-time to more than four<br />
dimensions. In all those cases coordinates<br />
were numbers that satisfied the rules of ordinary<br />
arithmetic. Yet another extension of<br />
space-time, which is useful in supergravity<br />
and crucial in superstring theory, is the addition<br />
to space-time of “supercoordinates”<br />
that do not satisfy the rules of ordinary arithmetic.<br />
Instead, two supercoordinates ea and<br />
i<br />
Bp satisfy anticommutation relations 0,0, +<br />
= 0, and consequently 0,0, (with no sum<br />
on a) = 0. Spaces with this kind of additional<br />
coordinatc are called superspaces.’6<br />
At first encounter superspaces may appear<br />
to be somewhat silly constructions. Nevertheless,<br />
much of the apparatus of differential<br />
geometry of manifolds can be extended to<br />
superspaces, so applications in physics may<br />
exist. It is possible to define fields that depend<br />
on the coordinates of a superspace.<br />
Rather naturally, such fields are called superfields.<br />
Let us apply this idea to supergravity,<br />
which is a field theory of both fermionic and<br />
bosonic fields. The supergravity fields can be<br />
further unified if they are written as a smaller<br />
number of superfields. Supergravity Lagrangians<br />
can then be written in terms of<br />
superfields; the earlier formulations are recovered<br />
by expanding the superfields in a<br />
powcr series in the supercoordinates. The<br />
anticommutation rule 0,0, = 0 leads to a<br />
finite number of ordinary fields in this expansion.<br />
The motion of a superstring is described<br />
by the motion of each space-time coordinate<br />
and supercoordinate along the string; thus<br />
the motion of the string traces out a “world<br />
sheet” in superspace. The full theory describes<br />
the motions and interactions of<br />
superstrings. In particular, Fig. 7 shows the<br />
basic form of the three closed superstring<br />
interactions. All other interactions of closed<br />
strings can be built up out of this one kind of<br />
intera~tion.’~ Needless to say, the existence<br />
of only one kind of fundamental interaction<br />
would severely restrict theories with only<br />
closed strings.<br />
There is a direct connection between the<br />
quantum-mechanical states of the string and<br />
the elementary particle fields of the theory.<br />
The string, whether it is closed or open, is<br />
under tension. Whatever its source, this tension,<br />
rather than Newton’s constant, defines<br />
the basic energy scale of the theory. To first<br />
approximation each point on the string has a<br />
force on it depending on this tension and the<br />
relative displacement between it and<br />
neighboring points on the string. The prob-<br />
93
lem of unravelling this infinite number of<br />
harmonic oscillators is one of the most<br />
famous problems of physics. The amplitudes<br />
of the Fourier expansion of the string displacement<br />
decouple the infinite set of harmonic<br />
oscillators into independent Fourier<br />
modes. These Fourier modes then correspond<br />
to the elementary-particle fields.<br />
The quantum-mechanical ground state of<br />
this infinite set of oscillators corresponds to<br />
the fields of IO-dimensional supergravity.<br />
Ten space-time dimensions are necessary to<br />
avoid tachyons and ghosts. The excited<br />
modes of the superstring then correspond to<br />
the new fields being added to supergravity.<br />
The harmonic oscillator in three<br />
dimensions can provide insight into the<br />
qualitative features of the superstring. The<br />
maximum value of the spin of a state of the<br />
harmonic oscillator increases with the level<br />
of the excitation. Moreover, the energy<br />
necessary to reach a given level increases as<br />
the spring constant is increased. The superstring<br />
is similar. The higher the excitation of<br />
the string, the higher are the possible spin<br />
values (now in ten dimensions). The larger<br />
94<br />
the string tension, the more massive are the<br />
states ofan excited level.<br />
The consistency requirements restrict<br />
superstring theories to two types. Type I<br />
theories have IO-dimensional N = I supersymmetry<br />
and include both closed and open<br />
strings and five kinds of string interactions.<br />
Nothing more will be said here about Type I<br />
theories, although they are extremely interesting<br />
(see Refs. 14 and 1 s).<br />
Type I1 theories have N = 2 supersymmetry<br />
in ten dimensions and accommodate<br />
closed strings only. There are two N = 2<br />
supersymmetry multiplets in ten dimensions,<br />
and each corresponds to a Type I1<br />
superstring theory. We will now describe<br />
these two superstring theories.<br />
The Type IIA ground-state spectrum is the<br />
one that can be derived by dimensional reduction<br />
of simple supergravity in eleven<br />
dimensions to N = 2 supergravity in ten<br />
dimensions. Thus, if we continue to reduce<br />
from ten to four dimensions with the<br />
hypothesis that the extra six dimensions<br />
form a 6-torus, we will obtain N = 8 supergravity<br />
in four dimensions. The superstring<br />
theory adds both pyrgons and Regge recurrences<br />
to the 256 N = 8 supergravity fields,<br />
but it has been possible (and often simpler)<br />
to investigate several aspects of supergravity<br />
directly from the superstring theory.<br />
The classification ofthe excited IO-dimensional<br />
string states (or elementary fields of<br />
the theory) is complicated by the description<br />
of spin in ten dimensions. However, the<br />
analysis does not differ conceptually from<br />
the analysis of spin for 11-dimensional<br />
supergravity. The massless states, which<br />
form the ground state of the superstring, are<br />
classified by multiplets of SO@), and the<br />
excitations of the string are massive fields in<br />
ten dimensions that belong to multiplets of<br />
SO(9). The ground-state fields of the Type<br />
IIA superstring are found in Table 4.<br />
Thz Type IIB ground-state fields cannot<br />
be derived from 1 I-dimensional supergravity.<br />
Instead the theory has a useful phase<br />
symmetry in ten dimensions. The fields<br />
listed as occumng twice in Table 4 carry<br />
nonzero values of the quantum number associated<br />
with U( I). So far, the main application<br />
of the U(1) symmetry has been the
Toward a Unified Theory<br />
..<br />
Y<br />
1 SO(9) Multiplets 1<br />
I<br />
712 -<br />
( 10-Dimensional Massive Spin States),,,o<br />
I<br />
10-Dimensional Mass2<br />
Fig. 8. The ground state andfirst Regge recurrence of fermionic slates in the 10-<br />
dimensional Type IIB superstring theory. There are a total of 256 fermionic and<br />
bosonic stales in the ground state. (The 56, contains the gravitino.) The first<br />
excited states contain 65,536 component fields. Half of these are fermions. (Each<br />
representation of the fermions shown above appears twice.)<br />
derivation of the equations of motion for the<br />
ground-state fields.” It will certainly have a<br />
crucial role in the future understanding of<br />
Type IIB superstrings.<br />
The quantum-mechanical excitations of<br />
the superstring correspond to the Regge recurrences,<br />
which are massive in ten<br />
dimensions; they belong to multiplets of<br />
SO(9). Thus, it is possible to fill in a diagram<br />
similar to Fig. 6, although the huge number<br />
of states makes the results look complicated.<br />
We give a few results to illustrate the<br />
method.<br />
The sets of Regge recurrences in Type IIA<br />
and IIB are identical. In Figure 8 we show the<br />
first recurrence of the fermion trajectories.<br />
(Note that only one-half of the 32,768 fer-<br />
mionic states of this mode are shown. The<br />
boson states are even messier.) The first excited<br />
level has a total of65,536 states, and the<br />
next two excited levels have 5,308,416 and<br />
235,929,600 states, respectively, counting<br />
both fermions and bosons. (<strong>Particle</strong><br />
physicists seem to show little embarrassment<br />
these days over adding a few fields to a<br />
theory!)<br />
The component fields in ten dimensions<br />
can now be expanded into 4-dimensional<br />
fields as was done in supergravity. Besides<br />
the zero modes and pyrgons associated with<br />
the ground states, there will be infinite ladders<br />
of pyrgon fields associated with each of<br />
the fields of the excited levels of the superstring.<br />
I<br />
The zero modes in four dimensions have<br />
been investigated only for the 6-torus; in this<br />
case all the zero modes come from the<br />
ground states. There is one zero mode for<br />
each component field, since the dimensional<br />
reduction is done as a 6-dimensional Fourier<br />
series on the 6-torus. The answers for other<br />
geometries are not yet known. It may be that<br />
many more fields become zero modes (or<br />
have nearly zero mass) in four dimensions<br />
when the dimensional reduction is studied<br />
for other spaces. An important problem is<br />
the analysis of superstrings on curved spaces,<br />
which has not yet been definitively studied.<br />
Although not much progress has been<br />
made toward understanding the phenomenology<br />
of these superstring theories, there<br />
has been some formal progress. The theory<br />
described here may be a quantum theory of<br />
gravity. (It may take all those new fields to<br />
obtain a renormalizable theory.) Although<br />
local symmetries can be ruined by anomalies,<br />
Type I1 (and several Type I) superstrings<br />
satisfy the constraints. Also, the one-loop<br />
calculation is finite; there are no candidates<br />
for counter terms, so the theory may be<br />
finite. Of course, this promising result needs<br />
support from higher order calculations.<br />
These results give some encouragement<br />
that superstrings may solve some long-standing<br />
problems in particle theory; whether they<br />
will lead to the ultimate unification of all<br />
interactions remains to be seen.<br />
Postscript<br />
The search for a unified theory may be<br />
likened to an old geography problem. Columbus<br />
sailed westward to reach India believing<br />
the world had no edge. By analogy, we<br />
are searching for a unified theory at shorter<br />
and shorter distance scales believing the<br />
microworld has no edge. Perhaps we are<br />
wrong and space-time is not continuous. Or<br />
perhaps we are only partly wrong, like Columbus,<br />
and will discover something new,<br />
but something consistent with what we already<br />
know. Then again, we may finally be<br />
right on course to a theory that unifies all<br />
Nature’s interactions.<br />
95
AUTHORS<br />
Richard C. Slansky has a broad background in physics with more than a<br />
taste ofmetaphysics. He received a B.A. in physics from Harvard in 1962<br />
and then spent the following year as a Rockefeller Fellow at Harvard<br />
Divinity School. Dick then attended the University of California,<br />
Berkeley, where he received his Ph.D. in physics in 1967. A two-year<br />
postdoctoral stint at the California Institute of Technology was followed<br />
by five years as Instructor and Assistant Professor at Yale University<br />
(1969-1974). Dick joined the Laboratory in 1974 as a Staff Member in the<br />
Elementary <strong>Particle</strong>s and Field Theory group of the Theoretical Division,<br />
where his interests encompass phenomenology, high-energy physics, and<br />
the early universe.<br />
References<br />
I. For a modern description of Kaluza-Klein theories, see Edward Witten, Nuclear<br />
<strong>Physics</strong> B186(1981):412 and A. Salam and J. Strathdee, Annals of <strong>Physics</strong><br />
141( 1982):3 16.<br />
2. Two-dimensional supersymmetry was discovered in dual-resonance models by<br />
P. Ramond, Physical Review D 3( 197 1):2415 and by A. Neveu and J. H. Schwarz,<br />
Nuclear <strong>Physics</strong> B3 I( 197 1):86. Its four-dimensional form was discovered by Yu.<br />
A. Gol’fand and E. P. Likhtman, Journal of Experimental and Theoretical<br />
<strong>Physics</strong> Letters I3( 1971):323.<br />
3. J. Wess and B. Zumino, <strong>Physics</strong> Letters 49B(1974):52 and Nuclear <strong>Physics</strong><br />
B70( 1974):39.<br />
4. Daniel Z. Freedman, P. van Nieuwenhuizen, and S. Farrara, Physical Review D<br />
13( 1976):3214; S. Deser and B. Zumino, <strong>Physics</strong> Letters 62B( 1976):335; Daniel<br />
Z. Freedman and P. van Nieuwenhuizen, Physical Review D 14( 1976):912.<br />
5. E. Cremmer and B. Julia, <strong>Physics</strong> Letters 80B( 1982):48 and Nuclear <strong>Physics</strong><br />
BI 59( 1979): 14 I.<br />
96
Toward a Unified Theory<br />
6. B. de Wit and H. Nicolai, <strong>Physics</strong> Letters 108B(1982):285 and Nuclear <strong>Physics</strong><br />
B208( 1982):323.<br />
7. This shortage of appropriate low-mass particles was noted by M. Cell-Mann in a<br />
talk at the 1977 Spring Meeting of the American Physical Society.<br />
8. J. Ellis, M. Gaillard, L. Maiani, and B. Zumino in Unification ofthe Fundamental<br />
<strong>Particle</strong> Interactions, S. Farrara, J. Ellis, and P. van Nieuwenhuizen, editors<br />
(New YorkPlenum Press, 1980), p. 69.<br />
9. E. Cremmer, B. Julia, and J. Scherk, <strong>Physics</strong> Letters 76B( 1978):409. Actually, the<br />
N = 8 supergravity Lagrangian in four dimensions was first derived by<br />
dimensionally reducing the N = 1 supergravity Lagrangian in eleven dimensions<br />
to N = 8 supergravity in four dimensions.<br />
10. M. J. Duff in Supergravity 81, S. Farrara and J. G. Taylor, editors (London:<br />
Cambridge University Press, 1982), p. 257.<br />
11. Peter (3.0. Freund and Mark A. Rubin, <strong>Physics</strong> Letters 97B( 1983):233.<br />
12. “Dual Models,” <strong>Physics</strong> Reports Reprint, Vol. I, M. Jacob, editor (Amsterdam:<br />
North-Holland, 1974).<br />
13. J. Scherk and John H. Schwarz, Nuclear <strong>Physics</strong> B8 I( 1974): I 18.<br />
14. For a history of this development and a list of references, see John H. Schwarz,<br />
<strong>Physics</strong> Reports 89(1982):223 and Michael B. Green, Surveys in High Energy<br />
<strong>Physics</strong> 3( 1983): 127.<br />
15. M. B. Green and J. H. Schwarz, Caltech preprint CALT-68-1090, 1984.<br />
16. For detailed textbook explanations of superspace, superfields, supersymmetry,<br />
and supergravity see S. James Gates, Jr., Marcus T. Grisaru, Martin Rocek, and<br />
Warren Siegel, Superspace: One Thousand and One Lessons in Supersymmetry<br />
(Reading, Massachusetts: Benjamin/Cummings Publishing Co., Inc., 1983) and<br />
Julius Wess and Jonathan Bagger, Supersymmetry and Supergravity (Princeton,<br />
New Jersey:Princeton University Press, 1983).<br />
17. John H. Schwarz, Nuclear <strong>Physics</strong> B226( 1983):269; P. S. Howe and P. C. West,<br />
1 Nuclear <strong>Physics</strong> B238( 1984): 18 1.<br />
97
Supersymmetrj<br />
S upersymmetry is a symmetry that connects particles of integral and half-integral spix<br />
Invented about ten years ago by physicists in Europe and the Soviet Union, supersymmetr<br />
was immediately recognizled as having amazing dynamical properties. In particula<br />
this symmetry provides a rational framework for unifying all the known forces betwee<br />
elementary particles-tlhe strong, weak, electromagnetic, and gravitational. Indeed,<br />
may also unify the separate concepts of matter and force into one comprehensiv<br />
framework.<br />
In the supersymmetric world depicted here, each boson pairs with a fermion partne,<br />
98<br />
\
There are two types of symmetries in<br />
nature: external (or space-time) symmetries<br />
and internal symmetries. Examples of internal<br />
symmetries are the symmetry of isotopic<br />
spin that identifies related energy levels of<br />
the nucleons (protons and neutrons) and the<br />
more encompassing SU(3) X SU(2) X U( 1)<br />
symmetry of the standard model (see “<strong>Particle</strong><br />
<strong>Physics</strong> and the Standard Model”).<br />
Operations with these symmetries do not<br />
change the space-time properties of a particle.<br />
External symmetries include translation<br />
invariance and invariance under the Lorentz<br />
transformations. Lorentz transformations,<br />
in turn, include rotations as well as the<br />
special Lorentz transformations, that is, a<br />
“boost” or a change in the velocity of the<br />
frame of reference.<br />
Each symmetry defines a particular operation<br />
that does not affect the result of any<br />
experiment. An example of a spatial translation<br />
is to, say, move our laboratory (accelerators<br />
and all) from Chicago to New<br />
Mexico. ’We are, of zourse, not surprised that<br />
the resull of any experiment is unaffected by<br />
the move, and we say that our system is<br />
translationally invariant. Rotational invariance<br />
is similarly defined with respect to<br />
rotating our apparatus about any axis. Invariance<br />
under a special Lorentz transformation<br />
corresponds to finding our results unchanged<br />
when our laboratory, at rest in our<br />
reference frame, is replaced by one moving at<br />
a constant velocity.<br />
Corresponding to each symmetry operation<br />
is a quantity that is conserved. Energy<br />
and momentum are conserved because of<br />
time and space-translational invariance, respectively.<br />
The energy of a particle at rest is<br />
its mass (E = mc2). Mass is thus an intrinsic<br />
property of a particle that is conserved because<br />
of invariance of our system under<br />
space-time translations.<br />
Spin. Angular momentum conservation is a<br />
result of ILorentz invariance (both rotational<br />
and special). Orbital angular momentum refers<br />
to the angular momentum ofa particle in<br />
motion, whereas the intrinsic angular<br />
100<br />
momentum of a particle (remaining even at<br />
rest) is called spin. (<strong>Particle</strong> spin is an external<br />
symmetry, whereas isotopic spin,<br />
which is not based on Lorentz invariance, is<br />
not.)<br />
In quantum mechanics spin comes in integral<br />
or half-integral multiples of a fundamental<br />
unit h (h = h/2n: where h is Planck‘s<br />
constant). (Orbital angular momentum only<br />
comes in integral multiples of h.) <strong>Particle</strong>s<br />
with integral values of spin (0, h, 2h, . . .)are<br />
called bosons, and those with half-integral<br />
spins (h/2, 3h/2, 5h,’2,. . .) are called fer-,<br />
mions. Photons (spin I), gravitons (spin 2),<br />
and pions (spin 0) are examples of bosons.<br />
Electrons, neutrinos, quarks, protons, and<br />
neutrons-the particles that make up ordinary<br />
matter-are all spin-% fermions.<br />
The conservation laws, such as those of<br />
energy, momentum, or angular momentum,<br />
are very useful concepts in physics. The following<br />
example dealing with spin and the<br />
conservation of angular momentum<br />
provides one small bit of insight into their<br />
utility.<br />
In the process of beta decay, a neutron<br />
decays into a proton, an electron, and an<br />
antineutrino. The antineutrino is massless<br />
(or very close to being massless), has no<br />
charge, and interacts only very weakly with<br />
other particles. In short, it is practically invisible,<br />
and for many years beta decay was<br />
thought to be simply<br />
n-p+ee-<br />
However, angular momentum is not conserved<br />
in this process since it is not possible<br />
for the initial angular momentum (spin 1/2<br />
for the neutron) to equal the final total<br />
angular momentum (spin 1/2 for the proton<br />
k spin 1/2 for the electron k an integral value<br />
for the orbital angular momentum). As a<br />
result, W. Pauli predicted that the neutrino<br />
must exist because its half-integral spin<br />
restores conservation of angular momentum<br />
to beta decay.<br />
There is a dramatic difference between the<br />
behavior of the two groups of spin-classified<br />
particles, the bosons and the fermions. This<br />
difference is clarified in the so-called spinstatistics<br />
theorem that states that bosons<br />
must satisfy commutation relations (the<br />
quantum mechanical wave function is symmetric<br />
under the interchange of identical<br />
bosons) and that fermions must satisfy anticommutation<br />
relations (antisymmetric wave<br />
functions). The ramification of this simple<br />
statement is that an indefinite number of<br />
bosons can exist in thp same place at the<br />
same time, whereas only one fermion can be<br />
in any given place at a given time (Fig. 1).<br />
Hence “matter” (for example, atoms) is<br />
made of fermions. Clearly, if you can’t put<br />
more than one in any given place at a time,<br />
then they must take up space. If they are also<br />
observable in some way, then this is exactly<br />
our concept of matter. Bosons, on the other<br />
hand, are associated with “forces.” For example,<br />
a large number of photons in the<br />
same place form a macroscopically observable<br />
electromagnetic field that affects<br />
charged particles.<br />
Supersymmetry. The fundamental property<br />
of supersymmetry is that it is a spacetime<br />
symmetry. A supersymmetry operation<br />
alters particle spin in half-integral jumps,<br />
changing bosons into fermions and vice<br />
versa. Thus supersymmetry is the first symmetry<br />
that can unify matter and force, the<br />
basic attributes of nature.<br />
If supersymmetry is an exact symmetry in<br />
nature, then for every boson of a given mass<br />
there exists a fermion of the same mass and<br />
vice versa; for example, for the electron there<br />
should be a scalar electron (selectron), for the<br />
neutrino, a scalar neutrino (sneutrino), for<br />
quarks, scalar quarks (squarks), and so forth.<br />
Since no such degeneracies have been observed,<br />
supersymmetry cannot be an exact<br />
symmetry of nature. However, it might be a<br />
symmetry that is inexact or broken. If so, it<br />
can be broken in either of two inequivalent<br />
ways: explicit supersymmetry breaking in<br />
which the Lagrangian contains explicit terms<br />
that are not supersymmetric, or spontaneous<br />
supersymmetry breaking in which the Lagrangian<br />
is supersymmetric but the vacuum<br />
is not (spontaneous symmetry breaking is<br />
l
Supersymmetry ai 100 Ge V<br />
t lclB<br />
Fig. 1. (a) An example of a symmetric wave function for apair of bosons and (b) an<br />
antisymmetric wave function for a pair of fermions, where the vector r represents<br />
the distance between each pair of identical particles. Because the boson wave<br />
function is symmetric with respect to exchange (yB (r) = yB(-r)), there can be a<br />
nonzero probablity (vi) for two bosons to occupy the same position in space (r =<br />
0), whereas for the asymmetric fermion wave function (yF (r) = -yF (-r)) the<br />
probability (vi) of two fermions occupying the same position in space must be<br />
zero.<br />
explained in Notes 3 and 6 of “Lecture<br />
Notes-From Simple Field Theories to the<br />
Standard Model”). Either way will lift the<br />
boson-fermion degeneracy, but the latter way<br />
will introduce (in a somewhat analogous way<br />
to the Higgs boson of weak-interaction symmetry<br />
breaking) a new particle, the Goldstone<br />
fermion. (We develop mathematically<br />
some of the ideas of this paragraph in<br />
“Supersymmetry and Quantum Mechanics”.)<br />
A question of extreme importance is the<br />
scale of supersymmetry breaking. This scale<br />
can be characterized in terms of the so-called<br />
supergup, the mass splitting between fermions<br />
and their bosonic partners (8’ = Mi -<br />
M;). Does one expect this scale to be of the<br />
order of the weak scale (- 100 GeV), or is it<br />
much larger? We will discuss the first<br />
possiblity at length because if supersymmetry<br />
is broken on a scale of order 100 GeV<br />
W<br />
+F<br />
there are many predictions that can be verified<br />
in the next generation of high-energy<br />
accelerators. The second possibility would<br />
not necessarily lead to any new low-energy<br />
consequences.<br />
We will also discuss the role gravity has<br />
played in the description of low-energy<br />
supersymmetry. This connection betweeen<br />
physics at the largest mass scale in nature<br />
(the Planck scale: Mpl = ( ~c/C~)’/~ = 1.2 X<br />
IOl9 GeV/c*, where CN is Newton’s gravitational<br />
constant) and physics at the low<br />
energies of the weak scale (Mw 83 GeV/c’<br />
where Mw is the mass of the W boson responsible<br />
for weak interactions) is both<br />
novel and exciting.<br />
Motivations. Why would one consider<br />
supersymmetry to start with?<br />
First, supersymmetry is the largest<br />
possible symmetry of nature that can com-<br />
bine internal symmetries and space-time<br />
symmetries in a nontrivial way. This combination<br />
is not a necessary feature of supersymmetry<br />
(in fact, it is accomplished by extending<br />
the algebra of Eqs. 2 and 3 in “Supersymmetry<br />
and Quantum Mechanics” to include<br />
more supersymmetry generators and<br />
internal symmetry generators). However, an<br />
important consequence of such an extension<br />
might be that bosons and fermions in different<br />
representations of an internal symmetry<br />
group are related. For example, quarks<br />
(fermions) are in triplets in the strong-interaction<br />
group SU(3), whereas the gluons (bosons)<br />
are in octets. Perhaps they are all related<br />
in an extended supersymmetry, thus providing<br />
a unified description of quarks and their<br />
forces.<br />
Second, supersymmetry can provide a theory<br />
of gravity. If supersymmetry is global,<br />
then a given supersymmetry rotation must<br />
be the same over all space-time. However, if<br />
supersymmetry is local, the system is invariant<br />
under a supersymmetry rotation that<br />
may be arbitrarily different at every point.<br />
Because the various generators (supersymmetry<br />
charges, four-momentum translational<br />
generators, and Lorentz generators for<br />
both rotations and boosts) satisfy a common<br />
dgebra of commutation and anticommutation<br />
relations, consistency requires that all<br />
the symmetries are local. (In fact, the anticommutator<br />
of two supersymmetry generators<br />
is a translation generator.) Thus different<br />
points in space-time can transform in<br />
different ways; put simply, this can amount<br />
to acceleration between points, which, in<br />
turn, is equivalent to gravity. In fact, the<br />
theory of local translations and Lorentz<br />
transformations is just general relativity, that<br />
is, Einstein’s theory of gravity, and a supersymmetric<br />
theory of gravity is called supergravity.<br />
It is just the theory invariant under<br />
local supersymmetry. Thus, supersymmetry<br />
allows for a possible unification of all of<br />
nature’s particles and their interactions.<br />
These two motivations were realized quite<br />
soon after the advent of supersymmetry.<br />
They are possibilities that unfortunately<br />
have not yet led to any reasonable prediccontinued<br />
on page I06<br />
101
Supers *<br />
in<br />
Quantum<br />
nlcs<br />
I<br />
intend to develop here Some of the algebra pertinent to the<br />
basic concepts OfsupersYmmetry. f will do this by showing an<br />
analogy between the quantum-mechanical harmonic oscillator<br />
and a bosonic field and a further analogy between the<br />
quantum-mechanical spin-% particle and a fermionic field. One<br />
result of combining the two resulting fields will be to show that a<br />
“tower” ofdegeneracies between the states for bosons and fermions is<br />
a natural feature ofeven the simplest of supersymmetry theories.<br />
A supersymmetry operation changes bosons into fermions and<br />
vice versa, which can be represented schematically with the operators<br />
QL and Q, and the equations<br />
Qilboson) = Ifemion),<br />
and<br />
Qalfermion) = Iboson), .<br />
example, the operation of changing a fermion to a boson and back<br />
again results in changing the position of the fermion.<br />
If supersymmetry is an invariance of nature, then<br />
[H, e,] = 0 ,<br />
that is, (2, commutes with the Hamiltonian H of the universe. Also,<br />
in this case, the vacuum is a supersymmetric singlet (Q,Ivac) = 0).<br />
Equations 1 through 3 are the basic defining equations of supersymmetry.<br />
In the form given, however, the supersymmetry is solely<br />
an external or space-time symmetry (a supersymmetry operation<br />
changes particle spin without altering any of the particle’s internal<br />
symmetries). An extended supersymmetry that connects external and<br />
internal symmetries can be constructed by . expanding . the number of<br />
operators of Eq. 2. However, for our purposes, we need not consider<br />
that complication.<br />
’L<br />
-<br />
1.1<br />
(3)<br />
In the simplest version of SuPersYmmetV, there are four such<br />
operators Or generators of supersymmetry (Qc~ and the krmitian<br />
conjugate QI with a = 1, 2). Mathematically, the generators are<br />
Lorentz spinors satisfying fermionic anticommutation relations<br />
The Harmonic Oscillator. In order to illustrate the consequences<br />
of Eqs. I through 3, we first need to review the quantum-mechanical<br />
treatment ofthe harmonic oscillator,<br />
The Hamiltonian for this system is<br />
where pJ’ is the energy-momentum four-vector bo = H, pi = threemomentum)<br />
and the o,, are two-by-two matrices that include the<br />
Pauli spin matrices o‘ (o,, = (1, 6’) where I = 1. 2, 3). Equation 2<br />
represents the unusual feature of this symmetry: the supersymmetry<br />
operators combine to generate translation in space and time. For<br />
where p and q are, respectively, the momentum and position<br />
coordinates of a nonrelativistic particle with unit mass and a 2n/m<br />
period of oscillation. The coordinates satisfy the quantum-mechanical<br />
commutation relation<br />
102
Supersymmetry at 100 Ge V<br />
he well-known solution to the harmonic oscillator (the set of<br />
nstates and eigenvalues of HOsJ is most conveniently expressed<br />
terms of the so-called raising and lowering operators, ut and a,<br />
respectively, which are defined as<br />
Finally, we find that<br />
that is, the states In) have ene<br />
perator for ut and lowering operator for a.<br />
counting operator since ut u In) = n I n).<br />
The Bosonic Field. There is a simple analogy between the quantum<br />
oscillator and th<br />
written as<br />
scillators ( ~ f up], , where p is<br />
d which satisfy the commutation relation<br />
terms of these operators, the Hamiltonian becomes<br />
with eigenstates<br />
t 7)<br />
(8)<br />
Ifscalar= 9 Iw,(ufup (13)<br />
with the summation taken over the individual oscillators p.<br />
round state of the free scalar quantum field is called the<br />
(it contains no scalar particles) and is described mathematically<br />
by the conditions<br />
up Ivac) = 0<br />
and (14)<br />
(vaclvac) = 1 .<br />
where N,, is a normalization factor and 10) is the ground state<br />
satisfying<br />
UlO) = 0<br />
and<br />
IO)= 1 .<br />
is easy to show that<br />
din) = &TI ~n + I)<br />
and<br />
)= fi In- I),<br />
The uf and up operators create or annihilate, respectively, a single<br />
scalar particle with energy ha, (ha,= s2,<br />
where p is the<br />
momentum carried by the created particle and m is the mass). A<br />
scalar particle is thus an excitation of one particular oscillator mode.<br />
The Fermionic Field. The simple quantum-mechanical analogue of<br />
a spin-% field needed to represent fermions is just a quantum particle<br />
with spin ‘12. This is necessary because, whereas bosons can be<br />
represented by scalar particles satisfying commutation relations,<br />
fermions must be represented by spin-% particles satisfying anticommutation<br />
relations.<br />
A spin-% particle has two spin states: 10) for spin down and 11) for<br />
spin up. Once again we define raising and lowering operators, here bt<br />
and 6, respectively. These operators satisfy the anticommutation<br />
relations<br />
{b, bt) = (bbt + b’b) = 1<br />
103
tates illustrates<br />
same energy as their fermionic part<br />
Moreover, it is easy to see that<br />
relations<br />
104
Supersymmetry at 100 GeV<br />
First we may add a small symmetry breaking term to the Hamiltonian,<br />
that is, H - H + EH’. where E is a small parameter and<br />
Energy<br />
States<br />
[ W, Q] # 0 . (25)<br />
Boson Fermion<br />
0 10,0><br />
hw I 1,0> 10,1><br />
2hw I2,0> 11,1><br />
3hw I3,0> 12,1><br />
This mechanism is called euplicit symmetry breaking. Using it we can<br />
give scalars a mass that is larger than that of their fermionic partners,<br />
as is observed in nature. Although this breaking mechanism may be<br />
perfectly self-consistent (even this is in doubt when one includes<br />
gravity), it is totally ad hoc and lacks predictive power.<br />
The second symmetry breaking mechanism is termed spontaneous<br />
symmetry breaking. This mechanism is characterized by the fact that<br />
the Hamiltonian remains supcrsymmetnc,<br />
but the ground state does not,<br />
The boson- fermion degeneracy for exact supersymmetry in<br />
which thefirst number in In,m) corresponds to the state for<br />
the oscillator degree of freedom (the scalar, or bosonic,<br />
field) and the second number to that for the spin-% degree of<br />
freedom (the fermionic field).<br />
which areanalogous to Eq. 1 because they represent the conversion of<br />
a fermionic state to a bosonic state and vice versa.<br />
The above example is a simple representation ofsupersymmetry in<br />
quantum mechanics. It is, however, trivial since it describes noninteracting<br />
bosons (oscillators) and fermions (spin-% particles). Nontrivial<br />
interacting representations of supersymmetry may also be<br />
obtained. In some of these representations it it possible to show that<br />
the ground state is not supersymmetric even though the Hamiltonian<br />
is. This is an example of spontaneous supersymmetry breaking.<br />
Symmetry Breaking. If supersymmetry were an exact symmetry of<br />
nature, then bosons and fermions would come in degenerate pairs.<br />
Since this is not the case, the symmetry must be broken. There are<br />
two inequivalent ways in which to do this and thus to have the<br />
degeneracy removed.<br />
Supersymmetry can either be a global symmetry, such as the<br />
rotational invariance ofa ferromagnet, or a local symmetry, such as a<br />
phase rotation in electrodynamics. Spontaneous breaking of a<br />
global symmetry leads to a massless Nambu-Goldstone particle. In<br />
supersymmetry we obtain a massless fermion c, the goldstino.<br />
Spontaneous breaking of a local symmetry, however, results in the<br />
gauge particle becoming massive. (In the standard model, the W<br />
bosons obtain a mass Me = gV by “eating” the massless Higgs<br />
bosons, where g is the SU(2) coupling constant and Vis the vacuum<br />
expectation value of the neutral Higgs boson.) The gauge particle of<br />
local supersymmetry is called a gravitino. It is the spin-3/2 partner of<br />
the graviton; that is, local supersymmetry incorporates Einstein’s<br />
theory ofgravity. When supersymmetry is spontaneously broken, the<br />
gravitino obtains a mass<br />
by “eating” the goldstino (here GN is Newton’s gravitational constant<br />
and A,, is the vacuum expectation of some field that spontaneously<br />
breaks supersymmetry).<br />
Thus, if the ideas of supersymmetry are correct, there is an<br />
underlying symmetry connecting bosons and fermions that is “hidden”<br />
in nature by spontaneous symmetry breaking. W<br />
105
coniinuedfiorn page 101<br />
tions. Many workers in the field are, however,<br />
still pursuing these elegant notions.<br />
Recently a third motivation for supersymmetry<br />
has been suggested. I shall describe the<br />
motivation and then discuss its expected<br />
consequences.<br />
For many years Dirac focused attention on<br />
the “problem of large numbers” or, more<br />
recently, the “hierarchy problem.” There are<br />
many extremely large numbers that appear<br />
in physics and for which we currently have<br />
no good understanding of their origin. One<br />
such large number is the ratio of the gravitational<br />
and weak-interaction mass scales<br />
mentioned earlier (Mpl/Mw- IO”).<br />
The gravitational force between two particles<br />
is proportional to the product of the<br />
energy (or mass if the particles are at rest) of<br />
the two particles times GN. Thus, since GN cc<br />
l/M& the force between two W bosons at<br />
rest is proportional to M&/M~,I - This<br />
is to be compared to the electric force between<br />
W bosons, which is proportional to a<br />
= e2/(4nhc) - lop2, where e is the electromagnetic<br />
coupling constant. Hence gravitational<br />
interactions between all known<br />
elementary particles are, at observable<br />
energies, at least IO” times weaker than their<br />
electromagnetic interactions.<br />
The key word is observable, for if we could<br />
imagine reaching an energy of order Mplc2,<br />
then the gravitational interactions would become<br />
quite strong. In other words, gravitationally<br />
bound states can be formed, in pnnciple,<br />
with mass of order Mpl - IOl9 GeV.<br />
The Planck scale might thus be associated<br />
with particles, as yet unobserved, that have<br />
strong gravitational interactions.<br />
At a somewhat lower energy, we also have<br />
the grand unification scale (MG - IOi5 GeV<br />
or greater), another very large scale with<br />
similar theoretical significance. New particles<br />
and interactions are expected to become<br />
important at MG.<br />
In either case, should these new<br />
phenomena exist, we are faced with the ques-<br />
tion ofwhy there are two such diverse scales,<br />
Mw and Mpl (or MG), in nature.<br />
The problem is exacerbated in the context<br />
of the standard model. In this mathematical<br />
H<br />
Y<br />
Perturbation Mass<br />
Fig. 2. If A. (lefi) represents a perturbative mass correction for an ordinary particle<br />
H due to the creation of a virtual photon y, then a supersymmetry rorarion of the<br />
central region of the diagram will generate a second mass correction A, (right)<br />
involving the supersymmetric partners H and thephotino 7. If supersymmetry is an<br />
exact symmetry, then the total mass correction is zero.<br />
framework, the W boson has a nonzero mass<br />
Mw because of spontaneous symmetry<br />
breaking and the existence of the scalar particle<br />
called the Higgs boson. Moreover, the<br />
mass of the W and the mass of the Higgs<br />
particle must be approximately equal. Unfortunately<br />
scalar masses are typically extremely<br />
sensitive to the details of the theory<br />
at very high energies. In particular, when one<br />
calculates quantum mechanical corrections<br />
to the Higgs mass p~ in perturbation theory,<br />
one finds<br />
where<br />
zero, and 6p2 is the perturbative correction.<br />
The parameter a is a generic coupling constant<br />
connecting the low mass states of order<br />
Mw and the heavy states of order Miarge, that<br />
is, the largest mass scale in the theory. For<br />
example, some of the theorized particles with<br />
mass Mpl or MG will have electric charge and<br />
interact with known particles. In this case, a<br />
= e2/4nAc, a measure of the electromagnetic<br />
coupling. Clearly p~ is naturally very large<br />
here and not approximately equal to the<br />
mass of the W.<br />
Supersymmetry can ameliorate the problem<br />
because, in such theories, scalar particles<br />
are no longer sensitive to the details at high<br />
energies. As a result of miraculous cancellations,<br />
one finds<br />
In these equations pb is the zeroth order<br />
value of the Higgs boson mass, which can be This happens in the following way (Fig. 2).<br />
106
Supersymmetry af 100 GeV<br />
Table 1<br />
The Supersymmetry Doubling of <strong>Particle</strong>s<br />
spi n-lh<br />
qu’arks<br />
spin-%<br />
leptons<br />
Standard Model<br />
-<br />
(i) ;<br />
(There are two other quark-lepton families similar to this one.)<br />
spin- 1<br />
gauge bosons<br />
y, W*, Zo, g<br />
spin-0 H+ A-<br />
Higgs bosons ( H o) ( H o)<br />
spin-0<br />
scalar partner<br />
spin-0<br />
scalar partner<br />
spin-2<br />
graviton 8<br />
__- -- ___ - ... -..<br />
G<br />
G<br />
For each ordinary mass correction, there will<br />
be a second mass correction related to the<br />
first by a supersymmetry rotation (the symmetry<br />
operation changes the virtual particles<br />
of the ordinary correction into their corresponding<br />
supersymmetric partners). Although<br />
each correction separately is proportional<br />
to a Mirge, the sum of the two corrections<br />
is given by Eq. 3. In this case, if & = 0,<br />
then pH = 0 and will remain zero to all orders<br />
in perturbation theory as long as supersymmetry<br />
remains unbroken. Hence supersymmetry<br />
is a symmetry that prevents scalars<br />
from getting “large” masses, and one can<br />
even imagine a limit in which scalar masses<br />
vanish. Under these conditions we say<br />
scalars are “naturally” light.<br />
How then do we obtain the spontaneous<br />
Global Supersymmetry<br />
Local Supersymmetry<br />
(5) ;<br />
IC<br />
Supersymmetric Partners<br />
spin-0<br />
squarks<br />
spin-0<br />
sleptons<br />
breaking of the weak interactions and a W<br />
boson mass? We remarked that supersymmetry<br />
cannot be an exact symmetry of<br />
nature; it must be broken. Once supersymmetry<br />
is broken, the perturbative correction<br />
(Eq. 3) is replaced by<br />
where A,, is the scale of supersymmetry<br />
breaking. If supersymmetry is broken spontaneously,<br />
then A,, is not sensitive to Mlarge<br />
and could thus have a value that is much less<br />
than Mlarge. This correction to the Higgs<br />
boson mass can then result in a spontaneous<br />
breaking of the weak interactions, with the<br />
standard mechanism, at a scale of order Ass<br />
Mlarge ‘<br />
The <strong>Particle</strong>s. We’ve discussed a bit of the<br />
motivation for supersymmetry. Now let’s<br />
describe the consequences of the minimal<br />
supersymmetric extension of the standard<br />
model, that is, the particles, their masses, and<br />
their interactions.<br />
The particle spectrum is literally doubled<br />
(Table I). For every spin-% quark or lepton<br />
there is a spin-0 scalar partner (squark or<br />
slepton) with the same quantum numbers<br />
under the SU(3) X SU(2) X U( 1) gauge interactions.<br />
(We show only the first family of<br />
quarks and leptons in Table I; the other two<br />
families include the s, e, b, and I quarks, and,<br />
for leptons, the muon and tau and their<br />
associated neutrinos.)<br />
The spin-l gauge bosons (the photon y, the<br />
weak interaction bosons W’ and Zo, and<br />
the gluons g) have spin-% fermionic partners,<br />
called gauginos.<br />
Likewise, the spin-0 Higgs boson, responsible<br />
for, the spontaneous symmetry breaking<br />
ofthe weak interaction, should have a spin-%<br />
fermionic partner, called a Higgsino. However,<br />
we have included two sets of weak<br />
doublet Higgs bosons, denoted H and H,<br />
giving a total of four Higgs bosons and four<br />
Higgsinos. Although only one weak doublet<br />
of Higgs bosons is required for the weak<br />
breaking of the standard model, a consistent<br />
supersymmetry theory requires the two sets.<br />
As a result (unlike the standard model, which<br />
predicts one neutral Higgs boson), supersymmetry<br />
predicts that we should observe two<br />
charged and three neutral Higgs bosons.<br />
Finally, other particles, related to symmetry<br />
breaking and to gravity, should be<br />
introduced. For a global supersymmetry,<br />
these particles will be a massless spin-%<br />
Goldstino and its spin-0 partner. However,<br />
in the local supersymmetry theory needed<br />
for gravity, there will also be a graviton and<br />
its supersymmetric partner, the gravitino.<br />
We will discuss this point in greater detail<br />
later, but local symmetry breaking combines<br />
the Goldstino with the gravitino to form a<br />
massive, rather than a massless, gravitino.<br />
In many cases the doubling of particles<br />
just outlined creates a supersymmetric partner<br />
that is absolutely stable. Such a particle<br />
107
_-----<br />
Standard Model<br />
1<br />
I<br />
i<br />
I<br />
I<br />
L .___- _ _ _ ~ _ _ _ ~ _ _<br />
_^__--_- ~-<br />
-<br />
Fig. 3. Examples of interactions between ordinary particles<br />
(lefl) and the corresponding interactions between an ordinav<br />
particle and two supersymmetric particles (right)<br />
_ _ _ _-_____ ~ ____<br />
______<br />
obtained by performing a supersymmetry rotation on the<br />
first interaction.<br />
could, in fact, be the dominant form of matter<br />
in our universe.<br />
following manner.<br />
Although an unbroken supersymmetry<br />
can keep scalars massless, once supersym-<br />
The Masses. What is the expected mass for metry is broken, all scalars obtain quantum<br />
the supersymmetric partners of the ordinary corrections to their masses proportional to<br />
particles? The theory, to date, does not make the supersymmetry breaking scale Ass, that is<br />
any firm predictions; we can nevertheless<br />
obtain an order-of-magnitude estimate in the 6p2 - a A:s , (5)<br />
which is Eq. 4 with the first negligible term<br />
dropped. If we demand the Higgs mass &-<br />
Zip2 to be of order Mb, then Ais - MZw/a is at<br />
most oforder 1000 GeV. Moreover, the mass<br />
splitting between all ordinary particles and<br />
their supersymmetric partners is again of<br />
order Mw,. We thus conclude that if supersymmetry<br />
is responsible for the large ratio<br />
108
Supersymmetry ut 100 Ge V<br />
I<br />
and e- and the photino 7) that experimentally would be easily recognizable.<br />
\<br />
\ Jet<br />
Fig. 5. A process involving supersymmetric particles (a gluino and squarks 3 that<br />
generates two hadronic jets.<br />
I<br />
M,,/M,,., then the new particles associated<br />
with supersymmetry will be seen in the next<br />
generation of high-energy accelerators.<br />
The Interactions. As a result of supersymmetry,<br />
the entire low-energy spectrum of<br />
particles has been doubled, the masses of the<br />
new particles are of order Mw, but these<br />
masses cannot be predicted with any better<br />
accuracy. A reasonable person might therefore<br />
ask what properties, if any, can we<br />
predict. The answer is that we know all the<br />
interactions of the new particles with the<br />
ordinary ones, of which several examples are<br />
shown in Fig. 3. To get an interaction between<br />
ordinary and new particles, we can<br />
start with an interaction between three or-<br />
Fig. 4. A possible interaction involving supersymmetric particles (the selectrons ;+ dinary particles and rotate two of these (with<br />
a supersymmetry operation) into their supersymmetric<br />
partners. The important point is<br />
that as a result of supersymmetry the coupling<br />
constants remain unchanged.<br />
Since we understand the interactions of<br />
the new particles with the ordinary ones, we<br />
know how to find these new objects. For<br />
example, an electron and a positron can annihilate<br />
and produce a pair of selectrons that<br />
subsequently decay into an electron-positron<br />
pair and two photinos (Fig. 4). This process<br />
is easily recognizable and would be a good<br />
signal of supersymmetry in high-energy electron-positron<br />
colliders.<br />
Supersymmetry is also evident in the process<br />
illustrated in Fig. 5. Here one of the three<br />
quarks in a proton interacts with one of the<br />
quarks in an antiproton; the interaction is<br />
mediated by a gluino. The result is the generation<br />
of two squarks that decay into quarks<br />
and photinos. Because quarks do not exist as<br />
free particles, the experimenter should observe<br />
two hadronic jets (each jet is a collection<br />
of hadrons moving in the same direction<br />
as, and as a consequence of, the initial motion<br />
of a single quark). The two photinos will<br />
generally not interact in the detector, and<br />
thus some of the total energy of the process<br />
will be “missing”.<br />
The theories we have been discussing until<br />
now have been a minimal supersymmetric<br />
extension of the standard model. There are,<br />
109
however, two further extrapolations that are<br />
interestihg both theoretically and phenomenologically.<br />
The first concerns gravity and<br />
the second, grand unified supersymmetry<br />
models.<br />
Gravity. We have already remarked that<br />
supersymmetry may be either a global or a<br />
local symmetry. If it is a global symmetry,<br />
the Goldstino is massless and the lightest<br />
supersymmetric partner. However, if supersymmetry<br />
is a local symmetry, it necessarily<br />
includes the gravity of general relativity and<br />
the Goldstino becomes part of a massive<br />
gravitino (the spin-3/2 partner of the graviton)<br />
with mass<br />
P<br />
With Ass of order Mw/& or 1000 GeV, mG<br />
is extremely small (- lo-’’ times the mass of<br />
the electron).<br />
Recently it was realized that under certain<br />
circumstances A,, can be much larger than<br />
Mw, but, at the same time, the perturbative<br />
corrections 6p2 can still satisfy the constraint<br />
that they be of order Mb. In these special<br />
cases, supersymmetry breaking effects vanish<br />
in the limit as some very large mass<br />
diverges, that is, we obtain<br />
Fig. 6. The decay mode of the proton predicted by the minimal unification<br />
symmetry SU(5). The expected decay products are a neutral pion no and a positron<br />
e+.<br />
(7)<br />
instead of Eq. 5. An example is already<br />
provided by the gravitino mass tnG (where<br />
Miarge = Mp,). In fact, models have now been<br />
constructed in which the gravitino mass is of<br />
order .44wand sets the scale ofthe low-energy<br />
supegap 62 between bosons and fermions.<br />
In either case (an extremely small or a very<br />
large gravitino mass), the observation of a<br />
massive gravitino is a clear signal of local<br />
supersymmetry in nature, that is, the nontrivial<br />
extension of Einstein’s gravity or<br />
supegravi ty.<br />
Grand Unification. Our second extrapolation<br />
of supersymmetry has to do with grand<br />
110<br />
unified theories, which provide a theoretically<br />
appealing unification of quarks and<br />
leptons and their strong, weak, and electromagnetic<br />
interactions. So far there has<br />
been one major experimental success for<br />
grand unification and two unconfirmed<br />
predictions.<br />
The success has to do with the relationship<br />
between various coupling constants. In the<br />
minimal unification symmetry SU(5), two<br />
independent parameters (the coupling constant<br />
gG and the value ofthe unification mass<br />
MG) determine the three independent coupling<br />
constants (g,, g, and g‘) of the standardmodel<br />
SU(3) X SU(2) X U( I) symmetry. As a<br />
result, we obtain one prediction, which is<br />
typically expressed in terms of the weakinteraction<br />
parameter:<br />
d2<br />
sin2Qw = -<br />
g2+gz’<br />
The theory of minimal SU( 5) predicts sin29w I<br />
= 0.2 1, whereas the experimentally observed<br />
value is 0.22 k 0.01, in excellent agreement.<br />
The two predictions of SU(5) that have<br />
not been verified experimentally are the existence<br />
of magnetic monopoles and proton<br />
decay. The expected abundance of magnetic<br />
monopoles today is crucially dependent on<br />
poorly understood processes occumng in the<br />
first second of the history of the universe.<br />
As a result, if they are not seen, we may<br />
ascribe the problem to our poor understand- 1<br />
ing of the early universe. On the other hand,<br />
if proton decay is not observed at the ex-<br />
I
Supersymmetry ut 100 Ge V<br />
p [Supersymmetry Proton Decay ) K" +<br />
N<br />
n pupersymmetry Neutron D eea3 KO + 7<br />
i<br />
1<br />
I<br />
where rnp is the proton mass.<br />
Recent experiments, especially sensitive<br />
to the decay modes of Eq. 9, have found 7p 2<br />
years, in contradiction with the prediction.<br />
Hence minimal SU(5) appears to be in<br />
trouble. There are, of course, ways to complicate<br />
minimal SU(5) so as to be consistent<br />
with the experimental values for both sin20w<br />
and proton decay. Instead of considering<br />
such ad hoc changes, we will discuss the<br />
unexpected consequences of making minimal<br />
SU(5) globally supersymmetric. The parameter<br />
sin2& does not change considerably,<br />
whereas MG increases by an order of<br />
magnitude. Hence, the good prediction for<br />
sin20w remains intact while the proton lifetime,<br />
via the gauge boson exchange process<br />
of Fig. 6, naturally increases and becomes<br />
unobservable.<br />
It was quickly realized, however, that<br />
other processes in supersymmetric SU(5)<br />
give the dominant contribution towards<br />
proton decay (Fig. 7). The decay products<br />
resulting from these processes would consist<br />
ofKmesons and neutrinos or muons, that is,<br />
--<br />
fig. 7. The dominant proton-decay and neutron-decay modes predicted by supersymmetry.<br />
The expected decay products are K mesons (K' and KO) and neutrinos<br />
6).<br />
and so would differ from the expected decay<br />
products of n mesons and positrons. This is<br />
very exciting because detection of the<br />
products of Eq. 11 not only may signal<br />
nucleon decay but also may provide the first<br />
signal of supersymmetry in nature. Experiments<br />
now running have all seen candidate<br />
events of this type. These events are, however,<br />
consistent with background. It may<br />
take several more years before a signal rises<br />
up above the background.<br />
pected rate, then minimal SU(5) is in serious<br />
trouble.<br />
The dominant decay modes predicted by<br />
minimal SU(5) for the nucleons are<br />
p - 7COef<br />
and<br />
n - n-e+<br />
(9)<br />
These processes involve the exchange ofa socalled<br />
X or Y boson with mass of order Mc<br />
(Fig. 6), so that the predicted proton lifetime<br />
T, is<br />
Experiments. An encouraging feature of the<br />
theory is that low-energy supersymmetry can<br />
be verified in the next ten years, possibly as<br />
early as next year with experiments now in<br />
progress at the CERN proton-antiproton collider.<br />
Experimenters at CERN recently dis-<br />
111
covered the W' and Z'bosons, mediators of<br />
the weak interactions, and produced many of<br />
these bosons in high-energy collisions between<br />
protons and antiprotons (each with<br />
momentum - 270 GeV/c). For example,<br />
Fig. 8 shows the process for the generation of<br />
a W- boson, which then decays to a highenergy<br />
electron (detectable) and a highenergy<br />
neutrino (not detectable). A single<br />
electron with the characteristic energy of<br />
about 42 GeV was a clear signature for this<br />
process.<br />
Let us now consider some of the signatures<br />
of supersymmetry for pi or pp colliders. A<br />
clear signal for supersymmetry are multi-jet<br />
events with missing energy. For example,<br />
events containing one, two, three, or four<br />
hadronic jets and nothing more can be interpreted<br />
as a signal for either quark or gluino<br />
production (Figs. 5 and 9). A two- or four-jet<br />
signal is canonical, but these events can look<br />
like one- or three-jet events some fraction of<br />
the time.<br />
There may also be events with two jets, a<br />
highenergy electron, and some missing<br />
energy. This is the characteristic signature of<br />
top quark production via W decay (Fig. lo),<br />
and thus such events may be evidence for top<br />
quarks. But there is also an event predicted<br />
by supersymmetry with the same signature,<br />
namely, the production of a squark pair (Fig.<br />
11). It would require many such events to<br />
disentangle these two possibilites.<br />
The CERN proton-antiproton collider<br />
began taking more data in September 1984<br />
with momentum increased to 320 GeV/c per<br />
beam and with increased luminosity. No<br />
clear evidence for supersymmetric partners<br />
has been observed. As a result, the so-called<br />
UA-1 Collaboration at CERN has put lower<br />
limits on gluino and squark masses of ap<br />
proximately 60 and 80 GeV, respectively. As<br />
of this writing it is apparent that the discovery<br />
of supersymmetric partners, and perhaps<br />
idso the top quark, must wait for the<br />
next generation of high-energy accelerators.<br />
Hopefully, it will not be too long before we<br />
learn whether or not the underlying structure<br />
of the universe possesses this elegant, highly<br />
unifying type of symmetry. 1<br />
Fig. 8. The generation, in a high-energy proton-antiproton collision, of a W-<br />
particle, which then decays into an electron (e-) and an antineutrino (G).<br />
Fig. 9. A proton-antiproton collision involving supersymmetric particles (gluinos<br />
g, squarks antisquarks and photinos T) that generates four hadronic jets.<br />
Jet<br />
Fig. 10. Two-jet events observed by the UA-1 Collaboration at CERN can be<br />
interpreted, as shown here, as a process involving top quark t production.<br />
112
Supersymmetry ut 100 Ge Y<br />
- - Y w+/A<br />
w+ \<br />
U<br />
P<br />
-<br />
c 1<br />
9 q<br />
Jet<br />
\<br />
Fig. 11. The same event discussed in Fig. 10, only here interpreted as a supersymmetric<br />
process involving squarks and antisquarks.<br />
Further Reading<br />
Daniel Z. Freedman and Peter van Nieuwenhuizen, “Supergravity and the Unification of the Laws of<br />
<strong>Physics</strong>.” Scienirfic American (February 1978): 126-143.<br />
Stuart A. Raby did his undergraduate work at the University of Rochester, receiving his B.Sc. in<br />
physics in 1969. Stuart spent six years in Israel as a student/teacher, receiving a M.Sc. in physics from<br />
Tel Aviv University in 1973 and a Ph.D. in physics from the same institution in 1976. Upon<br />
graduating, he took a Research Associate position at Cornell University. From 1978 to 1980, Stuart<br />
was Acting Assistant Professor of physics at Stanford University and then moved over to a three-year<br />
assignment as Research Associate at the Stanford Linear Accelerator Center. He came to the<br />
Laboratory as a Temporary Staff Member in 1981, cutting short his SLAC position, and became a<br />
Staff Member of the Elementary <strong>Particle</strong>s and Field Theory Group of Theoretical Division in 1982.<br />
He has recently served as Visiting Associate Research Scientist for the University of Michigan. He<br />
and his wife Michele have two children, Eric and Liat.<br />
113
:. . . .<br />
. . . . . .<br />
*<br />
. . . . . . . . . ,<br />
. . ,<br />
. . . . - . .<br />
.<br />
. . . .<br />
. I<br />
.. . .. . ... . . ....<br />
* '<br />
. . . . . . . . . . ... .. . ..<br />
. . .<br />
. .' ... . . . . . . . . . - ..<br />
. . . . . .<br />
. . . . . . ..... .<br />
.'. . . .<br />
#<br />
. ...:<br />
....<br />
.. ,.<br />
. . .<br />
.......<br />
...<br />
e . .<br />
....<br />
. .<br />
.* .<br />
. ..<br />
. .<br />
. .<br />
.. . .<br />
. -<br />
. .
...<br />
:..The Family Problem<br />
by T. Goldman and Michael Martin Nieto<br />
The roster of elementary particles includes replicas, exact in every detail but mass,<br />
of those that make up ordinary matter. More facts are needed to explain this<br />
seemingly unnecessary extravagance.<br />
T<br />
he currently “standard” model of particle physics phenomenologically<br />
describes virtually all of our observations of<br />
the world at the level of elementary particles (see “<strong>Particle</strong><br />
<strong>Physics</strong> and the Standard Model”). However, it does not<br />
explain them with any depth. Why is SU(3)c the gauge group of the<br />
strong force? Why is the symmetry of the electroweak force broken?<br />
Where does gravity fit in? How can all ofthese forces be unified? That<br />
is, from what viewpoint will they appear as aspects of a common,<br />
underlying principle? These questions lead us in the directions of<br />
supersymmetry and of grand unification, topics discussed in<br />
“Toward a Unified Theory.”<br />
Yet another feature of the standard model leaves particle physicists<br />
dissatisfied: the multiple repetitions of the representations* of the<br />
particles involved in the gauge interactions. By definition the adjoint<br />
representationt of the gauge fields must occur precisely once in a<br />
gauge theory. However, quantum chromodynamics includes no less<br />
than six occurrences of the color triplet representation of quarks: one<br />
for each of the u, c, t, d, s, and b quarks. The u, c, and t quarks have a<br />
common electric charge of */3 and so are distinguished from the d, s,<br />
and b quarks, which have a common electric charge of -%. But the<br />
quarks with a common charge are distinguished only by their dif-<br />
ferent masses, as far as is now known. The electroweak theory<br />
presents an even worse situation, being burdened with nine leftchiralS<br />
quark doublets, three left-chiral lepton doublets, eighteen<br />
right-chiral quark singlets, and three right-chiral lepton singlets (Fig.<br />
1).<br />
Nonetheless, some organization can be discerned. The exact symmetry<br />
of the strong and electromagnetic gauge interactions, together<br />
with the nonzero masses of the quarks and charged leptons, implies<br />
that the right-chiral quarks and charged leptons and their left-chiral<br />
partners can be treated as single objects under these interactions. In<br />
addition, each neutral lepton is associated with a particular charged<br />
lepton, courtesy of the transformations induced by the weak interaction.<br />
Thus, it is natural to think in terms ofthree quark sets (u and d, c<br />
and s, and t and b) and three lepton sets (e- and v,, p- and v~, and 7-<br />
and v,) rather than thirty-three quite repetitive representations.<br />
Furthermore, the relative lightness of the u and dquark set and of the<br />
e- and v, lepton set long ago suggested to some that the quarks and<br />
leptons are also related (quark-lepton symmetry). Subtle mathematical<br />
properties of modern gauge field theories have provided new<br />
backing for this notion of three “quark-lepton families,” each consisting<br />
of successively heavier quark and lepton sets (Table 1).<br />
*We give a geometric definition of “representation,” using as an example the<br />
SU(3)c triplet representation of; say, the up quark. (This triplet, the smallest<br />
non-singlet representation of Scl(3)c, is called the fundamental representation.)<br />
The members of this representation (U,d, Ublao and usreed correspond to the set<br />
of three vectors directed from the origin of a two-dimensional coordinate system<br />
to the vertices of an equilateral triangle centered at the origin. (The triangle is<br />
usually depicted as standing on a vertex.) The “conjugate” of the triplet<br />
representation, which contains the three anticolor varieties of the up quark with<br />
charge -%, can be defined similarly: it corresponds to the set of three vectors<br />
obtained by reflecting the vectors of the triplet representation through the origin.<br />
(The vectors of the conjugate representation are directed toward the vertices of<br />
an equilateral triangle standing on its side, like a pyramid.) The “group<br />
transformations” correspond to the set of operations by which any one of the<br />
quark or antiquark vectors is transformed into any other.<br />
t The “adjoint” representation of SU(3)c, which contains the ei&ht vector bosons<br />
(the gluons), is found in the ‘(product” of the triplet representation and its<br />
conjugate. This product corresponds to the set of nine vectors obtained by<br />
forming the vector sums of each member of the triplet representation with each<br />
member of its conjugate. This set can be decomposed into a singlet containing a<br />
null vector (a point at the origin) and an octet, the adjoint representation,<br />
containing two null vectors and six vectors directed from the origin to the<br />
vertices of a regular hexagon centered at the origin. Note that the adjoint<br />
representation is symmetric under reflection through the origin.<br />
#A massless particle is said to be le/f-handed (right-handed) if the direction of<br />
its spin vector is opposite (the same as) that of its momentum. Chiraliw is the<br />
Lorentz-invariant generalization of this handedness to massive particles and is<br />
equivalent to handedness for massless particles.
If the underlying significance of this<br />
grouping by mass is not apparent to the<br />
reader, neither is it to particle physicists. NO<br />
one has put forth any compelling reason for<br />
deciding which charge Y3 quark and which<br />
charge --%quark to combine into a quark set<br />
or for deciding which quark set and which<br />
charged and neutral lepton set should be<br />
combined in a quark-lepton family. Like<br />
Mendeleev, we are in possession of what<br />
appears to be an orderly grouping but<br />
without a clue as to its dynamical basis. This<br />
is one theme of “the family problem.”<br />
Still, we do refer to each quark and lepton<br />
set together as a family and thus reduce the<br />
problem to that of understanding only three<br />
families-unless, of course, there are more<br />
families as yet unobserved. This last is another<br />
question that a successful “theory of<br />
families” must answer. Grand unified theories,<br />
supersymmetry theories, and theories<br />
wherein quarks and leptons have a common<br />
substructure can all accommodate quark-lepton<br />
symmetry but as yet have not provided<br />
convincing predictions as to the number of<br />
families. (These predictions range from any<br />
even number to an infinite spectrum.)<br />
Such concatenations of wild ideas (however<br />
intriguing) may not be the best approach<br />
to solving the family problem. A more conservative<br />
approach, emulating that leading<br />
to the standard model, is to attack the family<br />
problem as a separate question and to ask<br />
directly if the different families are<br />
dynamically related.<br />
Here we face a formidable obstacle-a<br />
paucity of information. A fermion from one<br />
family has never been observed to change<br />
into a fermion from another family. Table 2<br />
lists some family-changing decays that have<br />
been sought and the experimental limits on<br />
their occurrence. True, a p- may appear to<br />
decay into an e-, but, as has been experimentally<br />
confirmed, it actually is transformed<br />
into a v,,, and simultaneously the e- and a<br />
appear. Being an antiparticle, the ie carries<br />
the opposite of whatever family quantum<br />
numbers distinguish an e- from any other<br />
charged lepton. Thus, no net “first-familines”<br />
is created, and the “second-familiness”<br />
Fig. I. The electroweak representations of the fermions of the standard model,<br />
which comprise nine left-chiral quark doublets, eighteen right-chiral quark<br />
singlets, three left-chiral lepton doublets, and three right-chiral lepton singlets.<br />
The subscripts r, b, and g denote the three color charges of the quarks, and the<br />
subscripts R and L denote right- and left-chiralprojections. The symbols d’, s’, and<br />
b‘ indicate weak-interaction mass eigenstates, which, as discussed in the text, are<br />
mixtures of the strong-interaction mass eigenstates d, s, and b. Since quantum<br />
chromodynamics does not include the weak interaction, and hence is not concerned<br />
with chirality, the SU(3), representations of the fermions are fewer in number: six<br />
triplets, each containing the three color-charge varieties of one of the quarks, and<br />
three singlets, each containing a charged lepton and its associated neutral lepton.<br />
of the original p- is preserved in the v,,.<br />
In spite of the lack of positive experimental<br />
results, current fashions (which are based<br />
on the successes ofthe standard model) make<br />
irresistible the temptation to assign a family<br />
symmetry group to the three known families.<br />
Some that have been considered include<br />
SU(2), SU(2) X U( I), SU(3), and U( 1) X U( 1)<br />
X U(1). The impoverished level of our understanding<br />
is apparent from the SU(2) case,<br />
in which we cannot even determine whether<br />
the three families fall into a doublet and a<br />
singlet or simply form a triplet.<br />
The clearest possible prediction from a<br />
family symmetry group, analogous to<br />
Mendeleev’s prediction of new elements and<br />
their properties, would be the existence of<br />
one or more additional families necessary to<br />
complete a representation. Such a prediction<br />
can be obtained most naturally from either of<br />
two possibilities for the family symmetry: a<br />
spontaneously broken local gauge symmetry<br />
1161
p-,<br />
The Family Problem<br />
Table 1<br />
Members of the three known quark-lepton families and their masses, Each<br />
family contains one particle from each of the four types of fermions: leptons<br />
with an electric charge of -1 (the electron, the muon, and the tau); neutral<br />
leptons (the electron neutrino, the muon neutrino, and the tau neutrino);<br />
quarks with an electric charge of 2/3 (the up, charmed, and top quarks); and<br />
quarks with an electric charge of -V3 (the down, strange, and bottom<br />
quarks). Each family also contains the antiparticles of its members. (The<br />
antiparticles of the charged leptons are distinguished by opposite electric<br />
charge, those of the neutral leptons by opposite chirality, and those of the<br />
quarks by opposite electric and color charges. For historical reasons only<br />
the antielectron has a distinctive appellation, the positron.) Family membership<br />
is determined by mass, with the first family containing the least<br />
massive example of each type of fermion, the second containing the next<br />
most massive, and so on. What, if any, dynamical basis underlies this<br />
grouping by mass is not known, nor is it known whether other heavier<br />
families exist. The members of the first family dominate the ordinary world,<br />
whereas those of the second and third families are unstable and are found<br />
only among the debris of collisions between members of the first family.<br />
First Family Second Family Third Family<br />
electron, e-<br />
0.511 MeV/c2<br />
electron neutrino, vr<br />
0.00002 MeV/c2 (?)<br />
up quark, u<br />
=5 MeV/c2<br />
down quark, d<br />
10 MeV/c2<br />
Table 2<br />
muon, p-<br />
105.6 MeVlc’<br />
muon neutrino, vp<br />
50.5 MeVJc2<br />
charmed quark, c<br />
=1500MeV/c2<br />
strange quark, s<br />
70 MeV/c2<br />
tau, ‘E-<br />
1782 MeVJc’<br />
tau neutrino, vT<br />
5 147 MeV/c2<br />
top quark, t<br />
240,000 MeVJc2 (?)<br />
bottom quark, b<br />
-4500 MeVJc2<br />
Experimental limits on the branching ratios for some family-changing<br />
decays. The branching ratio for a particular decay mode is defined as the<br />
ratio of the number of decays by that mode to the total number of decays by<br />
all modes. An experiment capable of determining a branching ratio for p+ -<br />
e+y as low as lo”* is currently in progress at <strong>Los</strong> <strong>Alamos</strong> (see “Experiments<br />
To Test Unification Schemes”).<br />
Branching Ratio<br />
Dominant<br />
Decay Mode (upper bound) Decay Mode($<br />
10-10 pi e+v&<br />
10-12 p+ e+v,v,,<br />
I 0-7<br />
no-+ w<br />
10-8<br />
K+ n+no or p+v<br />
10-8<br />
K~ - x+x-no or n 8 nono<br />
10-5 c+ - pno<br />
i<br />
I<br />
or a spontaneously broken global symmetry.*<br />
What follows is a brief ramble<br />
(whose course depends little on detailed assumptions)<br />
through the salient features and<br />
implications of these two possibilities.<br />
Family Gauge Symmetry<br />
All of the unseen decays listed in Table 2<br />
would be strictly forbidden if the family<br />
gauge symmetry were an exact gauge symmetry<br />
as those of quantum electrodynamics<br />
and quantum chromodynamics are widely<br />
believed to be. Here, however, we do not<br />
expect exactness because that would imply<br />
the existence, contrary to experience, of an<br />
additional fundamental force mediated by a<br />
massless vector boson (such as a long-range<br />
force like that of the photon or a strong force<br />
like that of the gluons but extending to leptons<br />
as well as quarks). But we can, as in the<br />
standard model, assume a broken gauge symmetry.<br />
We begin by placing one or more families<br />
in a representation of some family gauge<br />
symmetry group. (The correct group might<br />
be inferred from ideas such as grand unification<br />
or compositeness of fermions. However,<br />
it is much more likely that, as in the case of<br />
the standard model, this decision will best be<br />
guided by hints from experimental observations.)<br />
Together, the group and the representation<br />
determine currents that describe interactions<br />
between members of the representation.<br />
(These currents would be conserved if<br />
the family symmetry were exact.) For example,<br />
if the first and the second families are<br />
placed in the representation, an electrically<br />
neutral current describes the transformation<br />
-<br />
e ++ just as the charged weak current of<br />
the electroweak theory describes the transformation<br />
e- ++ v,. Since the other family<br />
*In principle, we should also consider the<br />
possibilities of a discrete symmetry or an explicit<br />
breaking of family symmetry (probably caused by<br />
some dynamics of a fermion substructure). However,<br />
these ideas would be radical departures from<br />
the gauge symmetries that have proved so successful<br />
to date. We will not pursue them here.<br />
117
members necessarily fall into the same representation,<br />
the e- - p- current includes<br />
contributions from interactions between<br />
these other members (d - s, for example),<br />
just as the charged weak current for<br />
e- - v, includes contributions from p- - vp<br />
and 5- e+ v,.<br />
If we now allow the family symmetry to be<br />
a local gauge symmetry, we find a “family<br />
vector boson,” F, that couples to these currents<br />
(Fig. 2) and mediates the family-changing<br />
interactions. As in the standard model,<br />
the coupled currents can be combined to<br />
yield dynamical predictions such as scattering<br />
amplitudes, decay rates, and relations<br />
between different processes.<br />
Scale of Family Gauge Symmetry<br />
Breaking. Weak interactions occur relatively<br />
infrequently compared to electromagnetic<br />
and strong interactions because<br />
of the large dynamical scale (approximately<br />
100 GeV) set by the masses of the W’ and<br />
Zo bosons that break the electroweak symmetry.<br />
We can interpret the extremely low<br />
rate of family-changing interactions as being<br />
due to an analogous but even larger<br />
dynamical scale associated with the breaking<br />
of a local family gauge symmetry, that is, to a<br />
large value for the mass MF of the family<br />
vector boson. The branching-ratio limit<br />
listed in Table 2 for the reaction KL - p’ +<br />
e’ allows us to estimate a lower bound for<br />
MF as follows.<br />
Like the weak decay of muons, the KL -<br />
pe decay proceeds through formation of a<br />
virtual family vector boson (Fig. 3). The rate<br />
for the decay, r, is given by<br />
Note that the fourth power of MF appears in<br />
Eq. I just as the fourth power of Mw does<br />
(hiding in the square of the Fermi constant)<br />
in the rate equation for muon decay. (Certain<br />
chirality properties of the family interaction<br />
could require that two of the five powers of<br />
the k.aon mass rnK in Eq. 1 be replaced by the<br />
muon mass. However, since the inferred<br />
value of MF varies as the fourth root of this<br />
term, the change would make little numerical<br />
difference.) It is usual to assume that gfamily,<br />
the family coupling constant, is comparable<br />
in magnitude to those for the weak and electromagnetic<br />
interactions. This assumption<br />
reflects our prejudice that family-changing<br />
interactions may eventually be unified with<br />
those interactions. Using Eq. 1 and the<br />
branching-ratio limit from Table 2, we obtain<br />
MF 2 lo5 GeV/c2 .<br />
Such a large lower bound on MF implies that<br />
the breaking of a local family gauge symmetry<br />
produces interactions much weaker<br />
than the weak interactions.<br />
Alternatively, processes like KL - pe may<br />
be the result of family-conserving grand unified<br />
interactions in which quarks are turned<br />
into leptons. However, the experimental<br />
limit on the rate of proton decay implies that<br />
such interactions occur far less frequently<br />
than the family-vio’lating interactions considered<br />
here.<br />
Experiments with neutrinos, also, indicate<br />
a similarly large dynamical scale for the<br />
breaking of a local family gauge symmetry. A<br />
search for the radiative decay vp - v,+ y has<br />
yielded a lower bound on the v,, lifetime of<br />
lo5 (m,/MeV) seconds. If the mass of the<br />
muon neutrino is near its experimentally<br />
observed upper bound of 0.5 MeVlc’, this<br />
lower bound on the lifetime is greater than<br />
the standard-model prediction of approximately<br />
lo’ (MeV/rn,)5 seconds. Thus, some<br />
family-conservation principle may be suppressing<br />
the decay.<br />
More definitive information is available<br />
from neutrino-scattering experiments.<br />
Positive pions decay overwhelmingly ( IO4 to<br />
1) into positive muons and muon neutrinos.<br />
In the absence of family-changing interactions,<br />
scattering of these neutrinos on nuclear<br />
targets should produce only negative<br />
muons. This has been accurately confirmed:<br />
neither positrons nor electrons appear more<br />
frequently than permitted by the present systematic<br />
experimental uncertainty of 0.1 per-<br />
Fig. 2. Examples of neutral familychanging<br />
currents coupled to a family<br />
vector boson (F). Such couplings follow<br />
from the assumption of a local gauge<br />
symmetry for the family symmetry.<br />
cent. An investigation of the neutrinos from<br />
muon decay has yielded similar results. The<br />
decay of a positive muon produces, in addition<br />
to a positron, an electron neutrino and a<br />
muon antineutrino. Again, in the absence of<br />
family-changing interactions, scattering of<br />
these neutrinos should produce only electrons<br />
and positive muons, respectively. A<br />
LAMPF experiment (E-31) has shown, with<br />
an uncertainty of about 5 percent, that no<br />
negative muons or positrons are produced.<br />
The energy scale of Eq. 2 will not be<br />
directly accessible with accelerators in the<br />
I 118
The Family Problem<br />
Fig. 3. Feynman diagram for the family-changing decay K, - p- + e', which is<br />
assumed to occur through formation of a virtual family vector boson (F). The K,<br />
meson is the longer lived of two possible mixtures of the neutral kaon (KO) and its<br />
antiparticle (EO). Neither this decay nor the equallyprobable decay K, - p' + e-<br />
has been observed experimentally; the current upper bound on the branching ratio<br />
is IO-*.<br />
_-<br />
foreseeable future. The Superconducting<br />
Super Collider, which is currently being considered<br />
for construction next decade, is conceived<br />
of as reaching 40,000 GeV but is<br />
estimated to cost several billion dollars. We<br />
cannot expect something yet an order of<br />
magnitude more ambitious for a very long<br />
time. Thus, further information about the<br />
breaking of a local family gauge symmetry<br />
will not arise from a brute force approach but<br />
+<br />
e<br />
rather, as it has till now, from discriminating<br />
searches for the needle ofa rare event among<br />
a haystack of ordinary ones. Clearly, the<br />
larger the total number of events examined,<br />
the more definitive is the information obtained<br />
about the rate of the rare ones. For<br />
this reason the availability of high-intensity<br />
beams of the reacting particles is a very<br />
important factor in the experiments that<br />
need to be undertaken or refined, given that<br />
they are to be carried out by creatures with<br />
finite lifetimes!<br />
For example, consider again the decay KL<br />
- pe. Since the rate of this decay vanes<br />
inversely as the fourth power of the mass of<br />
the family vector boson, a value of MI- in the<br />
million-GeV range implies a branching ratio<br />
lower by four orders of magnitude than the<br />
present limit. A search for so rare a decay<br />
would be quite feasible at a high-intensity,<br />
mediumenergy accelerator (such as the<br />
proposed LAMPF 11) that is designed to<br />
produce kaon fluxes on the order of 10' per<br />
second. (Currently available kaon fluxes are<br />
on the order of IO6 per second.) A typical<br />
solid angle times efficiency factor for an inflight<br />
decay experiment is on the order of IO<br />
percent. Thus, IO' kaons per second could be<br />
examined for the decay mode of interest. A<br />
branching ratio larger than lo-'* could be<br />
found in a one-day search, and a year-long<br />
experiment would be sensitive down to the<br />
lopJ4 level. Ofcourse, we do not know with<br />
absolute certainty whether a positive signal<br />
will be found at any level. Nonetheless, the<br />
need for such an observation to elucidate<br />
family dynamics impels us to make the attempt.<br />
Positive Evidence for Family<br />
Symmetry Breaking<br />
Thus, despite expectations to the contrary,<br />
we have at present no positive evidence in<br />
any neutral process for nonconservation of a<br />
family quantum number, that is, for familychanging<br />
interactions mediated by exchange<br />
ofan electrically neutral vector boson such as<br />
the Fof Figs. 2 and 3. Is it possible that our<br />
expectations are wrong-that this quantum<br />
number is exactly conserved as are electric<br />
charge and angular momentum? The answer<br />
is an unequivocal NO! We have-for<br />
quarks-positive evidence that family is a<br />
broken symmetry. To see this, we must<br />
examine the effect of the electroweak interaction<br />
on the quark mass eigenstates defined by<br />
the strong interaction.<br />
We know, for instance, that a Kf (= u + s)<br />
decays by the weak interaction into a p' and<br />
119
a vp and also decays into a n+ and a no (Fig.<br />
4). In quark terms this means that the u<br />
quark and the s quark in the kaon are coupled<br />
through a W+ boson. The two families<br />
(up-down and charmed-strange) defined by<br />
the quark mass eigenstates under the strong<br />
interaction are mixed by the weak interaction.<br />
Since the kaon decays occur in both<br />
purely leptonic and purely hadronic channels,<br />
they are not likely to be due to peculiar<br />
quark-lepton couplings. Similar evidence for<br />
family violation is found in the decays of D<br />
mesons, which contain charmed quarks.<br />
Weak-interaction eigenstates d' and s'<br />
may be defined in terms of the strong-interaction<br />
mass eigenstates d and s by<br />
coset sinec<br />
(.") = (-sin ec cos ec) ( P) (3)<br />
where e,, the Cabibbo mixing angle, is experimentally<br />
found to be the angle whose<br />
sine is 0.23 & 0.01. (The usual convention,<br />
which entails no loss of generality, is to assign<br />
all the mixing effects of the weak interaction<br />
to the down and strange quarks, leaving<br />
unchanged the up and charmed quarks.) The<br />
fact that the mass and weak-interaction<br />
eigenstates are different implies that a conserved<br />
family quantum number cannot be<br />
defined in the presence of both the strong<br />
and the weak interactions. We can easily<br />
show, however, that this conclusion does not<br />
contradict the observed absence of neutral<br />
family-violating interactions.<br />
lhe weak charged-current interaction describing,<br />
say, the transformation of a d'<br />
quark into a u quark by absorption of a W+<br />
boson has the form<br />
(id' + CS')W+ = (ii, C) ( w+ w+) 0 (.") ,<br />
which, after substitution of Eq. 3, becomes<br />
(4)<br />
(Ud' + &w+ = ii(dc0~ ec + ssin Bc)W+<br />
+ ;(-&in e, + scos e,) w+ .<br />
(5)<br />
(Here we suppress details of the Lorentz<br />
algebra.)<br />
Fig. 4. Feynman diagrams for the decays of a positive kaon into (a) a positive muoi<br />
and a muon neutrino and (b) a positive and a neutral pion. The ellkse witr<br />
diagonal lines represents any one of several possible pathways for production of 4<br />
positive and a neutral pion from an up quark and an antidown quark. These decays<br />
in which the up-down and charmed-strange quark families are mixed by the weai<br />
interaction (as: indicated by sin 0, and cos %), are evidence that the family sym<br />
metry of quarks is a broken symmmetry.<br />
Because of the mixing given by Eq. 3, the<br />
statement we made near the beginning of this<br />
article, that no family-changing decays have<br />
been observed, must be sharpened. True, no<br />
s' - u decay has been seen, but, of course,<br />
the s - u decay implied by Eq. 5 does occur.<br />
Thus, "No family-changing decays of weal<br />
interaction family eigenstates have been 01<br />
served" is the more precise statement.<br />
The weak neutral-current interaction dl<br />
scribing the scattering of a d' quark when<br />
absorbs a Zo has a form like that of Eq. 4:<br />
120
The Family Problem<br />
KO -+<br />
-<br />
d W+ d<br />
-<br />
S W- S<br />
-<br />
-C KO<br />
Gg. 5. Feynman diagram for a CP-violating reaction that transforms the neutral<br />
:aon into its antiparticle. This second-order weak interaction occurs through<br />
ormation of virtual intermediate states including either a u, c, or t quark.<br />
dtdt + ?st)Zo = (511, it)( zo zo) 0 ($)<br />
hce the Cabibbo matrix in Eq. 3 is unitary,<br />
:q. 6 is unchanged (except for the disapiearance<br />
of primes on the quarks) by subtitution<br />
of Eq. 3:<br />
(ad’ + ?s’)Zo = (ad+ Ss)Zo. (7)<br />
-bus, the weak neutral-current interaction<br />
loes not change d quarks into s quarks anynore<br />
than it changes d’ quarks into s’ quarks.<br />
t is only the presumed family vector boson<br />
if mass greater than IO5 GeV that may effect<br />
uch a change.<br />
Tamily Symmetry Violation and<br />
:P Violation<br />
The combined operation of charge con-<br />
Jgation and parity reversal (CP) is, like<br />
sarity reversal alone, now known not to be<br />
n exact symmetry of the world. An under-<br />
:anding of CP violation and proton decay<br />
rould be of universal importance to explain<br />
big-bang’’ cosmology and the observed ex-<br />
:ss of matter over antimatter.<br />
The generalization by Kobayashi and<br />
Maskawa of Eq. 3 to the three-family case is<br />
introduced in “<strong>Particle</strong> <strong>Physics</strong> and the Standard<br />
Model”; it yields a relation between<br />
family symmetry violation and CP violation.<br />
Although other sources of CP violation may<br />
exist outside the standard model, this relation<br />
permits extraction of information about<br />
violation of family symmetry from studies of<br />
CP violation.<br />
The phenomenon of CP violation has, so<br />
far, been observed only in the KO- Eo system.<br />
The CP eigenstates of this system are<br />
the sum and the difference of the KO and Eo<br />
states. The violation is exhibited as a small<br />
tendency for the long-lived state, KL , which<br />
normally decays into three pions, to decay<br />
into two pions (the normal decay mode of<br />
the short-lived state, Ks) with a branching<br />
ratio of approximately This tendency<br />
can be described by saying that the Ks and<br />
KL states differ from the sum and difference<br />
states by a mixing of order E:<br />
[Ks) z [KO) + (1 - E) [EO)<br />
and (8)<br />
IKL) = [KO)- (1- E) [KO).<br />
The quark-model analysis based on the work<br />
of Kobayashi and Maskawa and the secondorder<br />
weak interaction shown in Fig. 5<br />
predict an additional CP-violating effect not<br />
describable in terms of the mixing in Eq. 8;<br />
that is, it would occur even if E were zero.<br />
The effect, which is predicted to be of order<br />
E’, where E‘IE is about IO-*, has not yet been<br />
observed, but experiments sufficiently sensitive<br />
are being mounted.<br />
Both E and &’are related to the Kobayashi-<br />
Maskawa parameters that describe family<br />
symmetry violation. This guarantees that if<br />
the value of E’ is found to be in the expected<br />
range, higher precision experiments will be<br />
needed to determine its exact value . If no<br />
positive result is obtained in the present<br />
round of experiments, it will be even more<br />
important to search for still smaller values.<br />
In either case intense kaon beams are highly<br />
desirable since the durations of such experiments<br />
are approaching the upper limit of<br />
reasonability.<br />
Of course, in principle, CP violation can<br />
be studied in other quark systems involving<br />
the heavier c, b, and t quarks. However, these<br />
are produced roughly 10’ times less<br />
copiously than are kaons, and the CP-violating<br />
effects are not expected to be as large as in<br />
the case of kaons.<br />
Global Family Symmetry<br />
In our discussion of family-violating<br />
processes like K - pe, we have, so far,<br />
assumed the existence of a massive gauge<br />
vector boson reflecting family dynamics. The<br />
general theorem, due to Goldstone, offers<br />
two mutually exclusive possibilities for the<br />
realization of a broken symmetry in field<br />
theory. One is the development ofjust such a<br />
massive vector boson from a massless one;<br />
the other is the absence of any vector boson<br />
and the appearance of a massless scalar<br />
boson, or Goldstone boson. The possible<br />
Goldstone boson associated with family<br />
symmetry has been called the familon and is<br />
denoted byj As is generally true for such<br />
scalar bosons, the strength of its coupling<br />
falls inversely with the mass scale of the<br />
symmetry breaking. Cosmological argu-<br />
121
ments suggest a lower bound on the coupling<br />
ofapproximately IO-’* GeV-’ , a value very<br />
near (within three orders of magnitude) the<br />
upper bound determined from particle-physics<br />
experiments.<br />
The familon would appear in the two-<br />
body decays p - e + f and s - d +f:<br />
The<br />
latter can be observed in the decay K f (= u +<br />
S) -. xt (= u + 71, + nothing else seen. The<br />
familon would not be seen because it is about<br />
as weakly interacting as a neutrino. The only<br />
signal lhat the decay had occurred would be<br />
the appearance of a positive pion at the<br />
kinematically determined momentum of 227<br />
MeV/c.<br />
Such a search for evidence of the familon<br />
would encounter an unavoidable back-<br />
ground of positive pions from the reaction<br />
Kf - xf + v, + if, where the index i covers<br />
all neutrino types light enough to appear in<br />
the reaction. This decay mode occurs<br />
through a one-loop quantum-field correction<br />
to the electroweak theory (Fig. 6) and is<br />
interesting in itself for two reasons. First, it<br />
depends on a different combination of the<br />
parameters involved in CP violation and on<br />
the number N, of light neutrino types. Since<br />
N, is expected to be determined in studies of<br />
Zo decay, an uncertainty in the value of a<br />
matrix element in the standard-model<br />
prediction of the K+ - x+v,;, branching<br />
ratio can be eliminated. Present estimates<br />
place the branching ratio in the range between<br />
and IO-” times N,. Second, a<br />
discrepancy with the N, value determined<br />
from decay of the Zo , which is heavier than<br />
the kaon, would be evidence for the existence<br />
of at least one neutrino with a mass greater<br />
than about 200 MeVlc’.<br />
Fermion Masses and Family Symmetry<br />
Breaking<br />
The mass spectrum of the fermions is itself<br />
unequivocal evidence that family symmetry<br />
is broken. These masses, which are listed in<br />
Table I, should be compared to the W* and<br />
Zo masses of 83 and 92 GeVlc’, respectively,<br />
which set the dynamical scale of electroweak<br />
Fig. 6. Feynman diagram for the decay K + - n+ + vi + ii, where the index i covel<br />
all neutrino types light enough to appear in the reaction. The symbol Qi standsfc<br />
the charged lepton associated with vi and ii.<br />
interactions. (The masses quoted are the theoretical<br />
values, which agree well with the<br />
recently measured experimental values.) The<br />
very existence of the fermion masses violates<br />
electroweak symmetry by connecting doublet<br />
and singlet representations, and the<br />
variations in the pattern of mass splittings<br />
within each family show that family symmetry<br />
is broken. But since we neither know<br />
the mass scale nor understand the pattern of<br />
the family symmetiy breaking, we do not<br />
really know the relation between the mass<br />
scale of electroweak symmetry breaking and<br />
the fermion mass spectrum. It is possible to<br />
devise models in which the first family is<br />
light because the family symmetry breaking<br />
suppresses the electroweak symmetry breaking.<br />
Thus, the “natural” scale of electroweak<br />
symmetry breaking among the fermions<br />
could remain approximately 100 GeV/cZ,<br />
despite the small masses (a few MeV/c2) of<br />
some fermions.<br />
Experiments to establish the masses of the<br />
neutrinos are of great interest to the family<br />
problem and to particle physics in general.<br />
Being electrically neutral, neutrinos are<br />
unique among the fermions in possibly being<br />
endowed with a so-called Majorana mass* in<br />
addition to the usual Dirac mass. One approach<br />
to determining these masses is by<br />
applying kinematics to suitable reactions.<br />
For example, one can measure the end-point<br />
energy of the electron in the beta decay ’H-<br />
or of the muon in the decay nf<br />
’He + e- + ic<br />
- p+ + vp.<br />
Another quite different approach is to<br />
search for “neutrino oscillations.” If the ne1<br />
trino masses are nonzero, weak interactioi<br />
can be expected to mix neutrinos from di<br />
ferent families just as they do the quark<br />
This mixing would cause a beam of, sa<br />
essentially muon neutrinos to be tran<br />
formed into a mixture (varying in space ar<br />
in time) of electron, muon, and tau nei<br />
trinos. Detection of these oscillations wou<br />
not only settle the question ofwhether or ni<br />
neutrinos have nonzero masses but wou’<br />
also provide information about the di<br />
ferences between the masses of neutrinc<br />
from different families. Experiments are i<br />
progress, but, since neutrino interactions a<br />
infamously rare, high-intensity beams a<br />
required to detect any neutrinos at all, I<br />
alone possible small oscillations in the<br />
family identity. (For details about the tritiui<br />
beta decay and neutrino oscillation expel<br />
ments in progress at <strong>Los</strong> <strong>Alamos</strong>, see “E<br />
periments To Test Unification Schemes.”)<br />
Conclusion<br />
The family symmetry problem is a fund<br />
mcntal one in particle physics, apparent<br />
without sufficient information available<br />
present to resolve it. Yet it is as crucial ar<br />
important a problem as grand unificatio<br />
*Majorana mass terms are not allowed for ele<br />
trically charged particles. Such terms induce tran<br />
formations of particles into antiparticles and<br />
would be inconsistent with conservation of elect)<br />
charge.<br />
122
The Family Problem<br />
and it may well be a completely independent may, however, be accessible in studies of rare intensity, mediumenergy accelerator could<br />
me. The known bound of IO5 GeV on the decays of kaons and other mesons, of CP beahighlyeffectivemeansofachievingthese<br />
scale of family dynamics is an order of mag- violation, and of neutrino oscillations. To needs. Unlike many other experimental<br />
nitude beyond the direct reach of any present undertake these experiments at the necessary questions in particle physics, those on the<br />
3r proposed accelerator, including the Super- sensitivity requires intense fluxes of particles high-intensity frontier are clearly defined.<br />
:onducting Super Collider. These dynamics from the second or later families. A high- We await the answers expectantly. W<br />
Further Reading<br />
Howard Georgi. “A Unified Theory of Elementary <strong>Particle</strong>s and Forces.” Scientific American, April<br />
1981, p. 48.<br />
T. (Terry) Goldman received a B.Sc. in physics and mathematics in 1968<br />
from the University of Manitoba and an A.M. and a Ph.D. in physics<br />
from Harvard in 1969 and 1973, respectively. He was a Woodrow Wilson<br />
Fellow from 1968 to 1969 and a National Research Council of Canada<br />
postdoctoral fellow at the Stanford Linear Accelerator Center from 1973<br />
to 1975, when he joined the Laboratory’s Theoretical Division as a postdoctoral<br />
fellow. In 1978 he became a staff member in the same division.<br />
From 1978 to 1980 he was on leave from <strong>Los</strong> <strong>Alamos</strong> as a Senior<br />
Research Fellow at California Institute of Technology, and during the<br />
academic year 1982-83 he was a Visiting Associate Professor at the<br />
University of California, Santa Cruz. His professional work has centered<br />
around weak interactions and grand unified theories. He is a member of<br />
the American Physical Society.<br />
Michael Martin Nieto received a B.A. in physics from the University of<br />
California, Riverside, in 1961 and a Ph.D. in physics, with minors in<br />
mathematics and astrophysics, from Cornell University in 1966. He<br />
joined the Laboratory in 1972 after occupying research positions at the<br />
State University of New York at Stony Brook; the Niels Bohr Institute in<br />
Copenhagen; the University of California, Santa Barbara; Kyoto University;<br />
and Purdue University. His main interests are quantum mechanics,<br />
coherence phenomena, elementary particle physics, and astrophysics. He<br />
is a member of Phi Beta Kappa and the International Association of<br />
Mathematical Physicists and a Fellow of the American Physical Society.<br />
e. For example, in I972 Nieto authored a survey of important experiments in particle physics that could be done at the then<br />
I<br />
ciety.<br />
123<br />
1
Addendum<br />
CP Violation<br />
in Heavy-Quark<br />
Systems<br />
H<br />
ere we extend the discussion of CP<br />
violation in “The Family Problem”<br />
to<br />
heavier quark systems. This requires<br />
generalizing the Cabibbo mixing matrix<br />
(&. 3 in the main text) to more than two<br />
families. The Cabibbo matrix relates the<br />
weak-interaction eigenstates of the ud and cs<br />
quark families to their strong-interaction<br />
mass eigenstates. Now, in general, the unitary<br />
transformation relating the weak and<br />
strong eigenstates among n families will have<br />
%n(n - 1) rotations and Y2(n - l)(n - 2)<br />
physical phases.<br />
We are interested in the generalization to<br />
three families since the third family, containing<br />
the t and b qua.rks, is known to exist. This<br />
extension of the Cabibbo mixing matrix i<br />
called the Kobayashi-Maskawa (K-M) ma<br />
trix after the two physicists who elucidate(<br />
the problem. They realized that the mixin]<br />
matrix for three families would naturall:<br />
encompass a parameterization of CP viola<br />
tion. The K-M matrix can be written as i<br />
product of three rotations (which can bc<br />
thought of as the Euler rotation angles o<br />
classical physics even though the conventiox<br />
is not the standard one) and a singlt<br />
physically meaningful phase (which can bc<br />
identified as the CP-violating parameter). Ir<br />
particular, we define the K-M matrix V foi<br />
the three quark families (ud, cs, and tb) ai<br />
follows:<br />
(;)=V(t) 9<br />
where<br />
1 0 0 1 0 0<br />
0 0 1 0 -s3 c3<br />
12,4
Addendum<br />
Note the form of V in Eq. A2. The first,<br />
third, and fourth matrices are rotations<br />
about particular axes. Except for the unusual<br />
convention, this is just a general orthogonal<br />
rotation in a three-dimensional Cartesian<br />
system. The si and the ci are the sines and<br />
cosines of the three rotation angles 8,. Note<br />
that the i = 1 rotation is the Cabibbo rotation<br />
Oc described in the text.<br />
What is new is the second matrix factor in<br />
Eq. A2, which contains the complex<br />
amplitude with phase 6 that parameterizes<br />
CP violation. Indeed, this is the factor that<br />
makes V not an orthogonal transformation<br />
but a unitary transformation. Vis still normpreserving,<br />
but contains phase information,<br />
something that quantum mechanics allows.<br />
In principle, another matrix U relates the<br />
weak and strong eigenstates of the u, c, and t<br />
quarks, and the product UtV describes the<br />
mixing of weak charged currents. However,<br />
we follow the standard convention and take<br />
U = I, thereby putting all of the physics of<br />
UtVinto Vitself. (Note that the unitarity of<br />
V produces a result equivalent to that given<br />
by Eq. 7: there are still no family-changing<br />
neutral currents.) Because Vis “really” Ut V,<br />
the rows of V can be labeled by the u, c, and t<br />
quarks. Thus, we can write Vas<br />
Physically, this means that the matrix elements<br />
Vu can be considered coupling constants<br />
or decay amplitudes between the<br />
quarks and the weak charged bosons W*.<br />
For example, V, = sin 81 = sin 8c is the left<br />
vertex in Fig. 4a of the main text, which can<br />
be considered a u quark “decaying” into an s<br />
quark.<br />
We know from experiment that sin 8, =<br />
0.23 f 0.01. But further, from recent<br />
measurements of the lifetime of the b quark<br />
and the branching ratio l-b JrbC, we know<br />
that €Iz and O3 are both small. That is, we<br />
have the information<br />
and<br />
These results imply that we can take c2 and c3<br />
to be unity and obtain the approximation<br />
(‘47)<br />
In terms of quark mixing, CP violation in<br />
the Ko-Ro system is described by a secondorder<br />
imaginary amplitude proportional to<br />
s233 sin 6. In other words, the upper 2 by 2<br />
piece of the matrix in Eq. A7 has this new<br />
imaginary contribution when compared with<br />
the Cabibbo matrix of Eq. 3. By using the<br />
Feynman diagram of Fig. 5 in the main text,<br />
the Ko-Ro transition-matrix element (traditionally<br />
called Mlz) can be calculated in<br />
terms of the weak-interaction Hamiltonian<br />
and the entries of the mixing matrix V.<br />
The older parameterization of CP violation,<br />
which involves the parameter E, is<br />
model-independent. It focuses only on the<br />
properties of CP symmetry and the kaons<br />
themselves. It does not even need quarks.<br />
The value of E is determined by experiments<br />
(see below) and is directly related to Mk2. It<br />
remains for a particular formalism (such as<br />
that described here) to successfully predict<br />
M12 in a consistent manner. In particular,<br />
within the K-M formalism it is hard to obtain<br />
a large enough value for the CP-violating<br />
amplitude E even if one assumes 6 = x/2,<br />
because sz and s3 are so small. In fact, agree-<br />
ment with the measured value of E cannot be<br />
obtained unless the mass of the t quark is<br />
equal to or greater than 60 GeV/c2. Because<br />
the t quark has not yet been found, this<br />
possibility remains open.<br />
One way in which CP violation is observed<br />
in the Ko-KO system was described in<br />
the main text. Another way is to detect an<br />
anomalous number of decays to leptons of<br />
the “wrong” sign. In the absence of mixing<br />
one ordinarily expects positively charged<br />
leptons from the KO parent and negatively<br />
charged leptons from the Io parent; that is,<br />
KO = ds decays into d(uau) or d(uQ+v), and<br />
Io = as decays into a(udi) or a(uQ-C), as<br />
shown in Fig. Al. However, to describe the<br />
propagation of a KO (or a Ro), it must be<br />
decomposed into KL and Ks states each of<br />
which is an approximate CP eigenstate containing<br />
approximately equal amplitudes of<br />
KO and KO. Since the KS lifetime is negligibly<br />
short, it is easy to design experiments to<br />
measure decays of the KL only. If CP were an<br />
exact symmetry, then the KO and Io components<br />
of the KL would have equal amplitudes<br />
and would each provide exactly the same<br />
number of leptonic decays; that is, just as<br />
many “wrong”-sign leptons would come<br />
From decays of the Ko component (the antiparticles<br />
of Fig. Al) as “right”-sign leptons<br />
come from decays of the KO component (Fig.<br />
AI). The deviation from exact equality is<br />
another measure of CP violation.<br />
What about CP violation in other neutralboson<br />
systems? If one does the same type of<br />
anaylsis as is often done for the kaon system,<br />
one can phenomenologically describe CP<br />
violation by<br />
where cpo is a neutral boson, Go is its conjugate<br />
under C, E? is the CP-violating parameter<br />
specific to that boson, and<br />
L(t) = %(exp[-(irnl + rl/2)t]<br />
k exp[-(imZ + Tz/2)t]] .<br />
(A9)<br />
125
Addendum<br />
(Here the labels “I” and “2” refer to the<br />
approximate CP eigenstates.) The value of<br />
IcI
Addendum<br />
W-<br />
-- - -.__<br />
- -__ - -__.<br />
- __ -. _- --____<br />
- ___--- - I<br />
Fig. A2. Feynman diagram for the mixing between B: and B: mesons induced by<br />
second-order weak interactions. This diagram is analogous to that presented in the<br />
main text (Fig. 5) for mixing in the KO-Ko system.<br />
momentum are the decay products of a comnon<br />
b6 quark pair. Such parent quark pairs<br />
host always appear as BBmeson pairs.<br />
Suppose there was little or no mixing beween<br />
B:and E:. Then one would expect the<br />
Ibserved ratio of the decays of B,B: pairs<br />
nto back-to-back muon pairs with the same<br />
:harge [(+ +) or (- -)] to the decays of B:B:<br />
to be<br />
,airs with opposite charge [(+-)I<br />
tbout 25 percent. This ratio is deduced by<br />
he following argument. Without mixing (a)<br />
he main contribution to unlike pairs comes<br />
?om the direct decay of both quarks (b -<br />
y-i and 6 - $+v), and (b) the main conribution<br />
to like pairs comes from one prinary<br />
decay and one secondary decay (for<br />
:xample, b - cp-i and 6 - - 3.p-i). The<br />
,elathe rates can be calculated from the<br />
:nown weakdecay parameters, and one obaim<br />
the value 0.24 for the ratio of like- to<br />
mlike-sign pairs.<br />
However, with mixing (such as that shown<br />
in Fig. A2) one can sometimes have<br />
processes like sb - scp-i and s6 - 3.b -<br />
kp-i. This transforms some of the expected<br />
unlike-sign events into like-sign events. In<br />
fact, for a mixing of 10 percent, this changes<br />
the ratio of like- to unlike-sign events from<br />
about Y4 to about 112.<br />
Indeed, the UAl experiment at CERN<br />
sees a ratio of 50 percent. This result can be<br />
explained only by a large mixing between<br />
B: and Bf, which overwhelms the tendency<br />
for the band 6 quarks to decay into oppositesign<br />
pairs. Since one needs significant mixing<br />
to observe CP violation, there is hope of<br />
learning more about CP, depending on the<br />
(as yet undetermined) values of the mass-<br />
matrix parameters for B!and B:(that is, m1,<br />
m2, r~, and r2).<br />
For further details of this fascinating subject,<br />
we recommend the review “Quark Mixing<br />
in Weak Interactions” by Ling-Lie Chau<br />
(<strong>Physics</strong> Reports 95: 1( 1983)). W<br />
127
Experiments to Test<br />
Flash chambers discharging like neon lights, giant spectrometers, stacks of<br />
crystals, tons ofplastic scintillators, thousands ofprecisely strung<br />
wires-all employed to test the ideas of unifiedfield theories.<br />
I<br />
t has long been a dream of physicists to produce a unified field theory of the<br />
forces in nature. Much of the current experimental work designed to test such<br />
theories occurs at the highest energies capable of being produced by the latest<br />
accelerators. However, elegant experiments can be designed at lower energies<br />
that probe the details of the electroweak theory (in which the electromagnetic and<br />
weak interactions have been partially unified) and address key questions about the<br />
further unification of the electroweak and the strong interactions. (See “An Experimentalist’s<br />
View of the Standard Model” for a brief look at the current status of<br />
the quest for a unified field theory.)<br />
In this article we will describe four such experiments being conducted at <strong>Los</strong><br />
<strong>Alamos</strong>, often with outside collaborators. The first, a careful study of the beta decay<br />
of tritium, is an attempt to determine whether or not the neutrino has a mass and<br />
thus whether or not there can be mixing between the three known lepton families<br />
(the electron, muon, and tau and their associated neutrinos).<br />
Two other experiments examine the decay of the muon. The first is a search for<br />
rare decays that do not involve neutrinos, that is, the direct conversion across<br />
lepton families of the muon to an electron. The muon is a duplicate, except for a<br />
greater mass, of the electron, making such a decay seem almost mandatory.<br />
Detection ofa rare decay, or even the lowering of the limits for its occurrence, would<br />
tell us once again more about the mixing between lepton families and about possible<br />
violation of lepton conservation laws. At the same time, precision studies of<br />
ordinary muon decay, in which neutrinos are generated (the muon is accompanied<br />
by its own neutrino and thereby preserves muon number), will help test the stucture<br />
of the present theory describing the weak interaction, for example, by setting limits<br />
on whether or not parity conservation is restored as a symmetry at high energies.<br />
The electron spectrometer for the tritium beta decay experiment under<br />
construction. The thin copper strips evident in the entrance cone region to<br />
the right and at thefirst narrow region toward the center are responsible<br />
for the greatly improved transmission of this spectrometer.<br />
128
Unification Schemes<br />
by Gary H. Sanders<br />
I<br />
129
The intent of the fourth experiment is to<br />
measure interference effects between the<br />
neutral and charged weak currents via scattering<br />
experiments with neutrinos and electrons.<br />
If destructive interference is detected,<br />
then the present electroweak theory should<br />
be applicable even at higher energies; if constructive<br />
interference is detected, then the<br />
theory will need to be expanded, say by<br />
including vector bosons beyond those (the<br />
Zo and the L@) already in the standard<br />
model.<br />
Tritium Beta Decay<br />
In 1930 Pauli argued that the continuous<br />
kinetic energy spectrum of electrons emitted<br />
in beta decay would be explained by a light,<br />
neutral particle. This particle, the neutrino,<br />
was used by Fermi in 1934 to account quantitatively<br />
for the kinematics of beta decay. In<br />
1953, the elusive neutrino was observed<br />
directly by a <strong>Los</strong> <strong>Alamos</strong> team, Fred Reines<br />
and Clyde L. Cowan, using a reactor at Hanford.<br />
Though the neutrino has generally been<br />
taken to be massless, no theory requires neutrinos<br />
to have zero mass. The current experimental<br />
upper limit on the electron neutrino<br />
mass is 55 electron volts (ev), and the<br />
Russian team responsible for this limit<br />
claims a lower limit of 20 eV. The mass of the<br />
neutrino is still generally taken to be zero, for<br />
historical reasons, because the experiments<br />
done by the Russian team are extremely<br />
complex, and because masslessness leads to a<br />
pleasing simplification of the theory.<br />
A more careful look, however, shows that<br />
no respectable theory requires a mass that is<br />
identically zero. Since we have many neutrino<br />
flavors (electron, muon and tau neutrinos,<br />
at least), a nonzero mass would immediately<br />
open possibilities for mixing between<br />
these three known lepton families.<br />
Without regard to the minimal standard<br />
model or any unification schemes, the<br />
possible existence of massive neutrinos<br />
points out our basic ignorance ofthe origin of<br />
the known particle masses and the family<br />
structure of particles.<br />
An Experimentalist’s<br />
View of the<br />
Standard Model<br />
T<br />
he dream of physicists to produce a<br />
unified field theory has, at different<br />
times in the history of physics, appeared<br />
in a different light. For example, one<br />
of the most astounding intellectual achievements<br />
in nineteenth century physics was the<br />
realization that electric forces and magnetic<br />
forces (and their corresponding fields) are<br />
different manifestations of a single electromagnetic<br />
field. Maxwell’s construction of<br />
the differential equations relating these two<br />
fields paved the way for their later relation to<br />
special relativity.<br />
QED. The most wccessful field theory to<br />
date, quantum electrodynamics (QED), appears<br />
to have provided us with a complete<br />
description of the electromagnetic force.<br />
This theory has withstood an extraordinary<br />
array of precision tests in atomic, nuclear,<br />
and particle physics, and at low and high<br />
energies. A generation of physicists has<br />
yearned for comparable field theories describing<br />
the remaining forces: the weak interaction,<br />
the strong interaction, and gravity.<br />
An even more romantic goal has been the<br />
notion that a single field theory might describe<br />
all the known physical interactions.<br />
Electroweak Theory. In the last two decades<br />
we have comt: a long way towards realizing<br />
this goal. The electromagnetic and weak<br />
interactions appear to be well described by<br />
the Weinberg-Salam-Glashow model that<br />
unifies the two fields in a gauge theory. (See<br />
“<strong>Particle</strong> <strong>Physics</strong> and the Standard Model”<br />
for a discussion of gauge theories and other<br />
details just briefly mentioned here.) This<br />
electroweak theory appears to account for<br />
the apparent difference, at low energies. between<br />
the weak interaction and the electromagnetic<br />
interaction. As the energy of an<br />
interaction increases, a unification is<br />
achieved.<br />
So far, at energies accessible to modern<br />
high-energy accelerators, the theory is supported<br />
by experiment. In fact, the discovery<br />
at CERN in 1983 of the heavy vector bosons<br />
UT+, W-, and Zo, whose large mass (compared<br />
to the photon) accounts for the relatively<br />
“weak” nature of the weak force,<br />
beautifully confirms and reinforces the new<br />
theory.<br />
The electroweak theory has many experimental<br />
triumphs, but experimental<br />
physicists have been encouraged to press<br />
ever harder to test the theory, to explore its<br />
range of validity, and to search for new fundamental<br />
interactions and particles. The experience<br />
with QED, which has survived<br />
decades of precision tests, is the standard by<br />
which to judge tests of the newest field theories.<br />
QcD. A recent, successful field theory that<br />
describes the strong force is quantum<br />
chromodynamics (QCD). In this theory the<br />
strong force is mediated by the exchange of<br />
color gluons and a coupling constant is determined<br />
analogous to the fine structure constant<br />
of the electroweak theory.<br />
Standard Model. QCD and the electroweak<br />
theory are now embedded and<br />
united in the minimal standard model. This<br />
model organizes all three fields in a gauge<br />
130
~<br />
Experiments To Test Unification Schemes<br />
I<br />
Table<br />
The first three generations of elementary particles.<br />
Family:<br />
Doublets<br />
Singlets:<br />
Quarks:<br />
Leptons:<br />
I<br />
( ;IL<br />
( 3),<br />
theory of electroweak and strong interactions.<br />
There are two classes of particles: spin-<br />
I/z particles called fermions (quarks and leptons)<br />
that make up the particles of ordinary<br />
matter, and spin- I particles called bosons<br />
that account for the interactions between the<br />
fermions<br />
In this theory the fermions are grouped<br />
asymmetrically according to the "handedness''<br />
of their spin to account for the experimentally<br />
observed violation of CP symmetry<br />
<strong>Particle</strong>s with right-handed spin are<br />
grouped in pairs or doublets; particles with<br />
left-handed spin are placed in singlets. The<br />
exchange of a charged vector boson can convert<br />
one particle in a given doublet to the<br />
other, whereas the singlet particles have no<br />
weak charge and so do not undergo such<br />
transitions.<br />
The Table shows how the model, using<br />
this scheme, builds the first three generations<br />
of leptons and quarks. Since each quark (u, d,<br />
c, s, t, and b) comes in three colors and all<br />
fermions have antiparticles, the model includes<br />
90 fundamental fermions<br />
The spin-l boson mediating the electromagnetic<br />
force is a massless gauge boson,<br />
that is, the photon y. For the weak force,<br />
there are both neutral and charged currents<br />
that involve, respectively, the exchange of<br />
the neutral vector boson Zo and the charged<br />
vector bosons W+ and W-. The color force<br />
of QCD involves eight bosons called gluons<br />
that carry the color charge.<br />
The coupling constants for the weak and<br />
electromagnetic interactions, gwk and gem,<br />
arc related by the Weinberg angle Ow. a mixing<br />
angle used in the theory to parametrize<br />
the combination of the weak and electromagnetic<br />
gauge fields. Specifically,<br />
Only objects required by experimental results<br />
are in the standard model, hence the<br />
term minimal. For example, no right-handed<br />
neutrinos are included. Other minimal assumptions<br />
are massless neutrinos and no<br />
requirement for conservation of total lepton<br />
number or of individual lepton flavor (that<br />
is, electron, muon, or tau number).<br />
The theory, in fact, includes no mass for<br />
any of the elementary particles. Since the<br />
vector bosons for the weak force and all the<br />
fermions (except perhaps the neutrinos) are<br />
known to be massive, the symmetry of the<br />
theory has to be broken. Such symmetrybreaking<br />
is accomplished by the Higgs mechanism<br />
in which another gauge field with its<br />
yet unseen Higgs particle is built into the<br />
theory. However. no other Higgs-type particles<br />
are included.<br />
Many important features are built into the<br />
minimal standard model. For example, lowenergy,<br />
charged-current weak interactions<br />
are dominated by V- :I (vector minus axial<br />
vector) currents; thus, only left-handed W'<br />
bosons have been included. Also, since neutrinos<br />
are taken to be massless, there are<br />
supposed to be no oscillations between neutrino<br />
flavors.<br />
There are many possibilities for extensions<br />
to the standard model. New bosons,<br />
families of particles, or fundamental interactions<br />
may be discovered, or new substructures<br />
or symmetries may be required. The<br />
standard model, at this moment, has no<br />
demonstrated flaws, but there are many potential<br />
sources of trouble (or enlightenment).<br />
GUT. One of the most dramatic notions<br />
that goes beyond the standard model is the<br />
grand unified theory (GUT). In such a theory,<br />
the coupling constants in the electroweak<br />
and strong sectors run together at<br />
extremely high energies (IOi5 to IOi9 gigaelectron<br />
volts (GcV)). All the fields are unified<br />
under a single group structure, and a new<br />
object, the X, appears to generate this grand<br />
symmetry group. This very high-energy mass<br />
scale is not directly accessiblc at any conceivable<br />
accelerator. To explore the wilderness<br />
between present mass scales and the<br />
GUT scale. alas, all high-energy physicists<br />
will have to be content to work as low-energy<br />
physicists. Some seers believe the wilderness<br />
will be a desert, devoid of striking new physics.<br />
In the likely event that the desert is found<br />
blooming with unexplored phenomena, the<br />
journey through this terra incognita will be a<br />
long and fruitful one, even ifwe ure restricted<br />
to feasible tools. 8<br />
131
The reaction studied by all of the experiments<br />
mentioned is<br />
3H +<br />
3He+ + e- + ;e.<br />
This simple decay produces a spectrum of<br />
electrons with a definite end point energy<br />
(that is, conservation of energy in the reaction<br />
does not allow electrons to be emitted<br />
with energies higher than the end point<br />
energy). In the absence of neutrino mass, the<br />
spectrum, including this end point energy,<br />
can be calculated with considerable<br />
precision. Any experiment searching for a<br />
nonzero mass must measure the spectrum<br />
with sufficient resolution and control of systematic<br />
effects to determine if there is a<br />
deviation from the expected behavior.<br />
Specifically, an end point energy lower<br />
than expected would be indicative of energy<br />
carried away as mass by the neutrino.<br />
In 1972 Karl-Erik Bergkvist of the University<br />
of Stockholm reported that the mass of<br />
the electron antineutrino cr was less than 55<br />
eV. This experiment used tritium embedded<br />
in an aluminum oxide base and had a resolution<br />
of 50 eV. The Russian team set out to<br />
improve upon this result using a better spectrometer<br />
and tritium bound in valine<br />
molecules.<br />
Valine is an organic compound, an amino<br />
acid. A molecular biologist in the Russian<br />
collaboration provided the expertise<br />
necessary to tag several of the hydrogen sites<br />
on the molecule with tritium. This knowledge<br />
is important since one of the effects<br />
limiting the accuracy of the result is the<br />
knowledge of the final molecular states after<br />
the decay.<br />
Also important was the accurate determination<br />
of the spectrometer resolution<br />
funcfion, which involved a measurement of<br />
the energy loss of the beta electrons in the<br />
valine. This was accomplished by placing an<br />
ytter-bium-169 beta source in an identical<br />
source assembly and measuring the energy<br />
loss of these electrons as they passed through<br />
the valine.<br />
The beta particles emitted from the source<br />
were analyzed magnetically in a toroidal beta<br />
spectrometer. This. kind of. spectrometer<br />
provides the largest acceptance for ,a given<br />
resolution of any known design, and the<br />
Russians made very significant advances. ,<br />
The <strong>Los</strong> <strong>Alamos</strong> research group, as we shall<br />
see, has improved the spectrometer design<br />
even further.<br />
In 1980 the Russian group published a<br />
positive result for the electron antineutrino<br />
mass. After including corrections for the uncertainties<br />
in resolution and the final state:<br />
spectrum, they quoted a 99 per cent confidence<br />
level value of<br />
. .<br />
I4 < mie< 46 eV .<br />
The result was received with.great excitement,<br />
but two specific criticisms emerged.<br />
John J. Simpson of the University of Guelph<br />
pointed out that the spectrometer resolution<br />
was estimated neglecting the intrinsic<br />
linewidth of the spectrum .of the ytterbium-I69<br />
calibration source. The ex-,<br />
perimenters then measured the source<br />
linewidth to be 6.3 eV; their revised analysis<br />
lowered the best value of the neutrino mass<br />
from 34.3 to 28. eV. The basic result of a<br />
finite mass survives this reanalysis, according<br />
to the authors, but it should,be noted that<br />
the result is very sensitive to the calibration<br />
linewidth. Felix Boehm of the California Institute<br />
of Technology has observed that with<br />
an intrinsic linewidth ofonly 9 eV, the 99 per<br />
cent confidence level result would become<br />
consistent with zero.<br />
The second criticism related to .th.e assumption<br />
made about the energy of the final<br />
atomic states of helium-3. The valine<br />
molecule provides a complex environment,<br />
and the branching ratios into the 2s and<br />
Is states of heliumi3 are difficult to estimate.<br />
Thus the published result may prove to be<br />
false.<br />
This discussion illustrates the.difficulty of<br />
experiments of this kind. Each effort<br />
produces, in addition to the published measurement,<br />
a roadniap to the next generation<br />
experiment. The Russian team built upon its<br />
1980 result and produced a substantially improved<br />
apparatus that yielded a new meas-<br />
urement in 1983.<br />
The spectrometer was improved by adding<br />
an electrostatic field between the source and<br />
the magnetic spectrometer that could be used<br />
to accelerate the incoming electrons. The<br />
beta spectrum could then be measured,<br />
under conditions of constant magnetic field,<br />
by sweeping the electrostatic field to select<br />
different portions of the spectrum. This technique<br />
(originally suggested by the <strong>Los</strong> <strong>Alamos</strong><br />
group) provides a number of advantages.<br />
The magnetic spectrometer always<br />
sees electrons in the same energy range,<br />
providing constant detection efficiency<br />
throughout the measured spectrum. The<br />
magnetic field can also be set above the beta<br />
spectrum end point with the electrostatic<br />
field accelerating electrons from decays in<br />
the source into the spectrometer acceptance.<br />
This reduces the background by a large factor<br />
by making the spectrometer insensitive to<br />
electrons from decays of tritium contamination<br />
in the spectrometer volume.<br />
Also, finite source size, which produces a<br />
larger image at the spectrometer focal plane,<br />
was optically reduced by improved focusing<br />
at the source, yielding a higher count rate<br />
with better resolution.<br />
The improved spectrometer had a resolution<br />
of25 eV, compared to 45 eV in the 1980<br />
experiment. Background was reduced by a<br />
factor of 20, and the region of the spectrum<br />
scanned was increased from 700 eV to 1750<br />
eV.<br />
The controversial spectrometer resolution<br />
function was determined using a different<br />
line of the ytterbium-I69 source, and the<br />
Russians measured its intrinsic linewidth to<br />
be 14.7 eV. They also studied ionization<br />
losses by measuring the ytterbium- I69 spectrum<br />
through varying thicknesses of valine,<br />
yielding a considerably more accurate resolution<br />
function.<br />
The data were taken in 35 separate runs<br />
and the beta spectrum (Fig. 1) was fit by an<br />
expression that included the ideal spectral<br />
shape and the experimental corrections. The<br />
best fit gave<br />
I<br />
132.
Experiments To Test Unification Schemes<br />
0.04 1<br />
18.50 18.58 18.66<br />
Kinetic Energy (keV)<br />
Fig. I. Electron energy spectrum for<br />
tritium decay. This figure shows the<br />
1983 Russian data as the spectrum<br />
drops toward an end point energy of<br />
about 18.58 keV. The difference in the<br />
best fit to the data (solid line) and the<br />
‘t for a zero neutrino mass (dashed<br />
ne) is a shift to lower energies that<br />
wresponds to a mass of about 33.0 eV.<br />
Figure adapted from Michael H.<br />
aaevitz, “Experimental Results on<br />
leutrino Masses and Neutrino Osillations,<br />
’’ page 140, in Proceedings<br />
f the 1983 International Symposium<br />
n Lepton and Photon Interactions at<br />
[igh Energies, edited by David G.<br />
:assel and David L. Kreinick (Ithaca,<br />
lew York:F;R. Newman Laboratory of<br />
ruclear Studies, Cornell University,<br />
983).)<br />
ith a 99 per cent confidence limit range of<br />
20 < mCc< 55 eV .<br />
hese results were derived by making<br />
trticular choices for the final state spectra.<br />
ifferent assumptions for the valine molecu-<br />
lar final states and the helium-3 molecular,<br />
atomic, and nuclear final states can produce<br />
widely varying results.<br />
The physics community has been rantalized<br />
by the prospect that neutrinos have<br />
significant masses. Lepton flavor transitions,<br />
neutrino oscillations, and many other<br />
phenomena would be expected if the result is<br />
confirmed. The range of systematic effects,<br />
however, urges caution and enhanced efforts<br />
by experimenters to attack this problem in an<br />
independent manner. There are currently<br />
more than a dozen groups around the world<br />
engaged in improved experiments on tritium<br />
beta decay. A wide range of tritium sources,<br />
beta spectrometers, and analysis techniques<br />
are being employed.<br />
The Tritium Source. In an ambitious attempt<br />
to use the simplest possible tritium<br />
source, a team from a broad array of technical<br />
fields at <strong>Los</strong> <strong>Alamos</strong> is attempting to<br />
develop a source that consists of a gas of free<br />
(unbound) tritium atoms. Combining diverse<br />
capabilities in experimental particle<br />
physics, nuclear physics, spectrometer design,<br />
cryogenics, tritium handling, ultraviolet<br />
laser technology, and materials science, this<br />
team has developed a nearly ideal source and<br />
has made numerous improvements in electrostatic-magnetic<br />
beta spectrometers.<br />
The two most significant problems come<br />
from the scattering and energy loss of the<br />
electrons in the source and from the atomic<br />
and molecular final states of the helium-3<br />
daughter. These effects are associated with<br />
any solid source. Thus the ideal source would<br />
appear to be free tritium nuclei, but this is<br />
ruled impractical by the repulsive effects of<br />
their charge.<br />
The next best source is a gas of free tritium<br />
atoms. Detailed and accurate calculations of<br />
the atomic final states and electron energy<br />
losses can be performed. Molecular effects,<br />
including final state interactions, breakup,<br />
and energy loss in the substrate, are<br />
eliminated. Since the gas contains no inert<br />
atoms, the effect of energy loss and scattering<br />
in the source are reduced accordingly. Even<br />
the measurement of the beta spectrometer<br />
resolution function is simplified.<br />
The forbidding technical problem of such<br />
a design is building a source rich enough and<br />
compact enough to yield a useful count rate.<br />
Only one decay in IO’ produces an electron<br />
with energy in the interesting region near the<br />
end point where the spectrum is sensitive to<br />
neutrino mass.<br />
The <strong>Los</strong> <strong>Alamos</strong> group was motivated by a<br />
1979 talk given by Gerard Stephenson, of the<br />
<strong>Physics</strong> and Theoretical Divisions, on neutrino<br />
masses. They recognized quite early, in<br />
fact before the 1980 Russian result, that<br />
atomic tritium would be a nearly ideal<br />
source. In their first design, molecular<br />
tritium was to be passed through an extensive<br />
gas handling and purification system<br />
and atomic tritium prepared using a discharge<br />
in a radio-frequency dissociator. The<br />
pure jet of atomic tritium was then to be<br />
monitored for beta decays. It was clear, however,<br />
that the tritium atoms needed to be<br />
used more efficiently.<br />
Key suggestions were made at this point<br />
by John Browne of the <strong>Physics</strong> Division and<br />
Daniel Kleppner of the Massachusetts Institute<br />
of Technology. Advances had been<br />
made in the production of dense gases of<br />
spin-polarized hydrogen. The new techniques-in<br />
which the atomic beam was<br />
cooled and then contained in a bottle made<br />
of carefully chosen materials observed to<br />
have a low probability for promoting recombination<br />
of the atoms-promised a possible<br />
intense source of free atomic tritium. The<br />
collaboration set out to develop and demonstrate<br />
this idea. Crucial to the effort was the<br />
participation of Laboratory cryogenics<br />
specialists.<br />
The resulting tritium source (Fig. 2)<br />
circulates molecular tritium through a radiofrequency<br />
dissociator into a special tube of<br />
aluminum and aluminum oxide. Because the<br />
recombination rate for this material near 120<br />
kelvins is very low, the system achieves 80 to<br />
90 per cent purity of atomic tritium. The<br />
electrons from the beta decay of the atomic<br />
tritium are captured by a magnetic field, and<br />
then electrostatic acceleration, similar to that<br />
employed by the Russians, is used to trans-<br />
133
I<br />
!<br />
Atomic Tritium Source Region<br />
Transport and Focusing<br />
Region<br />
I<br />
I<br />
!<br />
Electron Gun<br />
Superconducting<br />
!<br />
,<br />
I<br />
!<br />
!<br />
I<br />
I<br />
I<br />
!<br />
I<br />
I<br />
Fig. 2. The tritium source. Molecular tritium passes through<br />
the radio- frequency dissociator and then into a I-meterlong<br />
tube as a gas of free atoms. The tube-aluminum with a<br />
surface layer of aluminum oxide-has a narrow range<br />
around a temperature of 120 kelvins at which the molecular<br />
recombination rate is very low, permitting an atom to<br />
experience approximately 50,000 collisions before a<br />
molecule is formed. The resulting diffuse atomic gas fills the<br />
tube, and mercury-diffusion pumps at the ends recirculate it<br />
through the dissociator. Typically, the system achieves 80 fo<br />
90 per cent purity of atomic tritium. By measuring the<br />
spectrum when the dissociator is off; the contribution from<br />
the 10 to 20per cent contamination of molecular tritium can<br />
be determined and subtracted, resulting in a pure atomic<br />
tritium electron spectrum.<br />
A superconducting coil surrounds the tube with afield oj<br />
1.5 kilogauss. At one end the winding has a reflectingfield<br />
provided by a magnetic pinch. These fields capture electrons<br />
from beta decays with 95per cent efficiency.<br />
The other end of the tube connects to a vacuum region and<br />
has coils that transport and, importantly, focus an image oj<br />
the electrons into the spectrometer (Fig 3). The tube is held<br />
at a selecfed voltage between -4 and -20 kilovolfs, and<br />
electrons exit the source to ground potential. Thus, electrons<br />
from decays in the source tube are accelerated by a known<br />
amount to an energy above that of electrons from decays in<br />
port the electrons toward the spectrometer.<br />
During this transport, focusing coils and a<br />
collimator are used to form a small image of<br />
the electron source in the spectrometer.<br />
Development of this tritium source required<br />
solving an array of problems associated<br />
with a system that was to recirculate<br />
atomic tritium. Everything had to be extremely<br />
clean, and no organic materials were<br />
allowed; all surfaces are glass or metal. Conducting<br />
materials had to be used wherever<br />
insulators could collect charge and introduce<br />
a bias. The aluminum oxide coating in the<br />
tube is so thin that electrons simply tunnel<br />
through it, thus providing a conducting surface<br />
that does not encourage recombination.<br />
Special mercury-diffusion pumps and custom<br />
cryopumps, free of oil or other organic<br />
materials, had to be fabricated. Every part of<br />
the tritium source was an exercise in<br />
materials science.<br />
134<br />
The idea of using electrostatic acceleration<br />
at the output of the source was first proposed<br />
by the group at <strong>Los</strong> <strong>Alamos</strong> in 1980 and<br />
subsequently used in the measurement described<br />
in the 1983 Russian publication. Accelerating<br />
the electrons to an energy above<br />
that of electrons from tritium that decays in<br />
the spectrometer both strongly reduces the<br />
background and also improves the acceptance<br />
of electrons into the spectrometer.<br />
However, this technique necessitates a larger<br />
spectrometer.<br />
There are two other important systematic<br />
effects that need to be dealt with: the source<br />
image seen by the spectrometer should be<br />
small, and electrons produced by decays in<br />
the tube that suffer scattering off the walls<br />
have an energy loss that distorts the<br />
measured spectrum. The focusing coil and<br />
the final collimator address both effects,<br />
providing a small image. The only energy<br />
loss mechanism remaining is in the tritium<br />
gas itself, where losses are less than 2 eV.<br />
development of the Russian design. Electrons<br />
from the source pass through the entrance<br />
cone and are focused onto the spectrometer<br />
axis. One very significant improvement<br />
in the spectrometer is the design of the<br />
conductors running parallel to the spec.<br />
trometer axis that do this focusing. In thc<br />
Russian apparatus, the conductors were<br />
thick water-cooled tubes. Most electron:<br />
strike the tubes and, as a result of this loss
Experiments To Test Unification Schemes<br />
Thin Curved<br />
Conductors<br />
Image<br />
Points<br />
or<br />
Decay<br />
Electrons<br />
Cone<br />
Flow of Thick Thin<br />
Electrons Conductors Conductors<br />
the spectrometer. Additional pumps also sharply reduce the<br />
amount of tritium escaping into the spectrometer.<br />
Several sophisticated diagnostic systems monitor source<br />
output and stability. Beta detectors mounted in the focus<br />
region in front of the collimator measure the total decay rate<br />
from molecular and atomic tritium, whereas the fraction of<br />
tritium in molecular form is monitored by an ultraviolet<br />
(1 027 angstroms wavelength) laser system developed by<br />
members of Chemistry Division that uses absorption lines of<br />
molecular tritium. A high-resolution electron gun is used to<br />
monitor energy loss in both the gas and the spectrometer.<br />
This gun is also used to measure the important spectrometer<br />
esolution function directly.<br />
Fig. 3. The spectrometer. Electrons from the source (Fig. 2)<br />
that pass through the collimator (with an approximate<br />
aperture of 1 centimeter) open into a cone shaped region in<br />
the spectrometer with a maximum half angle of 30 degrees.<br />
Electrons between 20 and 30 degrees pass between thin<br />
conducting strips into the spectrometer and are focused onto<br />
the spectrometer axis. This focus serves as a virtual image of<br />
the source. Transmission has been greatly improved over the<br />
Russian design through the use of thin conductors in all<br />
regions of electron flow (see opening photograph for a view<br />
of these conductors). The final focal plane detector is a<br />
position-sensitive, multi- wire proportional gas counter, also<br />
an improvement over previous detectors.<br />
heir spectrometer has low transmission.<br />
The <strong>Los</strong> <strong>Alamos</strong> spectrometer uses thin<br />
!O-mil strips for each ofthe conductors in the<br />
.egion within the transport aperture. This<br />
ichieves an order of magnitude higher transnission,<br />
essential in yielding a useful count<br />
‘ate in an experiment with a dilute gas<br />
ource.<br />
Another benefit of the thin strips is that<br />
hey can be formed easily. In fact, optical<br />
~alculations accurate to third order dictate<br />
he curvature of the entrance and exit strips.<br />
The improved focusing properties of this<br />
irrangement yield an acceptance three times<br />
iigher than the Russian device with no comwomise<br />
in resolution.<br />
The experimenters expect to be taking<br />
lata throughout the latter part of 1984. They<br />
xpect an order of magnitude less backround<br />
and an order of magnitude larger<br />
eometric acceptance than the Russian ex-<br />
periment. The design calls for a resolution<br />
between 20 and 30 eV, with a sensitivity to<br />
neutrino masses less than IO eV. Even with<br />
their dilute gas source, they estimate a data<br />
rate in the region within 100 eV of the spectrum<br />
end point of about 1 hertz, fully competitive<br />
with rates obtained using solid<br />
sources.<br />
Many groups around the world are<br />
vigorously pursuing this measurement. No<br />
other effort. however, will produce a result as<br />
free of systematic problems as the <strong>Los</strong> <strong>Alamos</strong><br />
project. Other experiments are employing<br />
solid sources or, at best, molecular<br />
sources. Many have adopted an electrostatic<br />
grid system that introduces its own problems.<br />
To date, no design promises as clean a<br />
measurement. This year may well be the year<br />
in which the problem of neutrino mass is<br />
settled. The quantitative answer willbe an<br />
important tool in uncovering the very poorly<br />
understood relations between lepton<br />
families. No deep understanding of the models<br />
that unify the forces in nature can be<br />
expected without precise knowledge of the<br />
masses of neutrinos.<br />
Rare Decays of the Muon<br />
The muon has been the source of one<br />
puzzle after another. It was discovered in<br />
1937 in cosmic radiation by Anderson and<br />
Neddermeyer and by Street and Stevenson<br />
and was assumed to be the meson of<br />
Yukawa’s theory of the nuclear force.<br />
Yukawa postulated that the nuclear force,<br />
with its short range, should be mediated by<br />
the exchange of a massive particle, a meson.<br />
This differs from the massless photon of the<br />
infinite-range electromagnetic force. The<br />
muon mass, about 200 times the electron<br />
mass, fit Yukawa’s theory well.<br />
135
It was only after World War I1 ended that<br />
measurements of the muon’s range in<br />
materials were found to be inconsistent with<br />
a particle interacting via a strong nuclear<br />
force. Discovery of the pion, or pi meson,<br />
settled the controversy. To this day, however,<br />
casual usage sometimes includes the<br />
erroneous phrase “mu meson”.<br />
With the resolution of the meson problem,<br />
however, the muon had no reason to be. It<br />
was simply not necessary. The muon appeared<br />
to be, in all known ways, a massive<br />
electron with no other distinguishing attributes.<br />
A famous quotation of I. I. Rabi<br />
summarized the mystery: “The muon, who<br />
ordered that?’<br />
This question is none other than the<br />
family problem described earlier. Today, the<br />
mystery remains, but its complexity has<br />
grown. Three generations of fermions exist,<br />
and the mysterious relation of the muon to<br />
the electron is replicated in the existence of<br />
the tau, discovered in 1976 by Martin Per1<br />
and collaborators. The three generation<br />
scheme is built into the minimal standard<br />
model, but there is little insight to guide us to<br />
the ultimate number of generations.<br />
Is there a conservation number associated<br />
with each family or generation? Are there<br />
selection rules or fundamental symmetries<br />
that account for the apparent absence of<br />
some transitions between these multiplets?<br />
Vertical and horizontal transitions between<br />
quark states do occur. Processes involving<br />
neutrinos connect the lepton generations.<br />
Can the pattern of these observed transitions<br />
give us a clue as to why we are blessed with<br />
this peculiar zoology? Should we look harder<br />
for the processes we have not observed?<br />
Rabi’s question, in its most modern form, is<br />
a rich and bewildering one, and many experimental<br />
groups have taken up its<br />
challenge by pursuing high sensitivity studies<br />
of the rare and unobserved reactions that<br />
may connect the generations.<br />
With the muon and electron virtual<br />
duplicates of each other, it was expected that<br />
the heavier muon would decay by simple,<br />
neutrinoless processes to the electron. Transitions<br />
such as p+ - e+ e+ e-, p+ - e+ y, or<br />
’<br />
Conservation ~aws: CL~ = Constant, ZL, = Constant, XL<br />
Allowed Decay: p+ - e+ v, c,<br />
p- Z - e- Z (where Z signifies that the<br />
interaction is with a nucleus) were expected.<br />
Estimates of the rates for these processes<br />
using second-order, current-current weak interactions<br />
gave results too small to observe.<br />
In fact, the results were much smaller than<br />
the 1957 limit for the branching ratio for p+<br />
- e+ y, which was < 2 X IO-’ (a branching<br />
ratio is the ratio of the probability a decay<br />
will occur to the probability of the most<br />
common decay).<br />
A better early model appeared in 1957<br />
when Schwinger proposed the intermediate<br />
vector boson (now called W and observed<br />
directly in 1983) as the mediator of the<br />
charged-current weak interaction. With this<br />
model and under most assumptions, rates<br />
larger than the experimental limits were<br />
predicted for the three reactions. The failure<br />
to observe these decays required a dynamical<br />
suppression or a new conservation law. Despite<br />
the discussion to follow, the situation<br />
today has changed very little. The measured<br />
limits are more swingent, though, by many<br />
orders of magnitude.<br />
The first proposal for lepton number con-<br />
Forbidden Decays:<br />
p+ - e+ e+ e- I<br />
p-Z-e-Z<br />
p- 2 - e- (2-2)<br />
p+ - e+Vevu<br />
servation came in 1953. In fact, there have<br />
been three different schemes for conserving<br />
lepton number. The 1953 Konopinski-<br />
Mahmoud scheme cannot accommodate<br />
three lepton generations and has not<br />
survived. A scheme in which lepton number<br />
is conserved by a multiplicative law was<br />
proposed in 196 I by Feinberg and Weinberg,<br />
but this method is not the favored conservation<br />
law. An early experiment with a neutrino<br />
detector at the Clinton P. Anderson<br />
Meson <strong>Physics</strong> Facility in <strong>Los</strong> <strong>Alamos</strong><br />
(LAMPF) has removed the multiplicative<br />
law from favor, and the current experiment<br />
to study neutrino-electron scattering, described<br />
later in this article, has set even more<br />
stringent limits on such a law.<br />
The most favored scheme is additive lepton<br />
number conservation, proposed in 1957<br />
by Schwinger, Nishijima, and Bludman. In<br />
this scheme, any process must separately<br />
conserve the sum of muon number and the<br />
sum of electron number. Table 1 shows the<br />
assignment of lepton numbers used. The extension<br />
to the third lepton flavor, tau, is<br />
obvious and natural.<br />
136
Experiments To Test Unification Schemes<br />
10-1<br />
e o e p+ +e+?<br />
e 5 p+ -+ e+e+e-<br />
.- 0<br />
0<br />
+<br />
o p-Z -+ e-Z<br />
a m10-4 - a p+ 4 e+yy<br />
OI<br />
.- C<br />
Qlp<br />
0<br />
-6<br />
-<br />
m 10-7<br />
10-10 i-<br />
..$I *e 0 0<br />
A<br />
4<br />
0 e<br />
<<br />
1950 1960 1970 l!<br />
Year<br />
Fig. 4. The progressive drop in the experimentally<br />
determined upper limit<br />
for the branching ratio of several<br />
muon-number violating processes<br />
shows a gap in the late 1960s. Essentially,<br />
this gap was the result of a belief<br />
by particle physicists in lepton number<br />
conservation.<br />
These schemes require, as the table hints, a<br />
distinct neutrino associated with each lepton.<br />
In a 1962 experiment the existence of<br />
separate muon and electron neutrinos was<br />
confirmed.<br />
With a conservation law firmly entrenched<br />
in the minds of physicists, searches<br />
for decays that did not conserve lepton number<br />
seemed pointless. In a 1963 paper<br />
Sherman Frankel observed “Since it now<br />
appears that this decay is not lurking just<br />
beyond present experimental resolution, any<br />
further search . . . seems futile.”<br />
In retrospect it can be said that the particle<br />
physics community erred. The conclusion<br />
stated in the previous paragraph resulted in a<br />
nearly complete halt to efforts to detect<br />
processes that did not conserve lepton number-and<br />
this on the basis ofa law postulated<br />
without any rigorous or fundamental basis!<br />
It is easy to justify these assertions. Figure<br />
4 shows that the experimental limits on rare<br />
decays were not aggressively addressed between<br />
1964 and the late 1970s. This era of<br />
inattention ended abruptly when an experimental<br />
rumor circulated in 1977-an er-<br />
roneous report terminated a decade of theoretical<br />
prejudice almost overnight! This<br />
could not have been the case if lepton conservation<br />
was required by fundamental ideas.<br />
In 1977 a group searching for the process<br />
p+ - e’ y at the Swiss Institute for Nuclear<br />
Research (SIN) became the inadvertent<br />
source of a report that the decay had been<br />
seen. The experiment, sometimes referred to<br />
as the “original SIN” experiment, was an<br />
order of magnitude more sensitive than any<br />
prior search for this decay and eventually set<br />
a limit on the branching ratio of 1.0 X low9 .<br />
A similar effort at the Canadian meson factory,<br />
TRIUMF, produced a limit of 3.6 X<br />
at about the same time.<br />
The Crystal Box. The extraordinary controversy<br />
generated by the “original SIN” report<br />
motivated a <strong>Los</strong> <strong>Alamos</strong> group to attempt<br />
a search for p’ - e+ y with a sensitivity<br />
to branching ratios of about IO-”. This<br />
experiment was carried out in 1978 and<br />
1979, using several new technologies and a<br />
new type of muon beam at LAMPF, and<br />
yielded an upper limit of 1.7 X 10-”(90 per<br />
cent confidence level). That result stands as<br />
the most sensitive limit on the decay to date<br />
but should be surpassed this year by an experiment<br />
at LAMPF called the Crystal Box<br />
experiment.<br />
This experiment was conceived as the<br />
earlier experiment came to an end. By<br />
searching for three rare muon decays simultaneously,<br />
the experiment would be a major<br />
advance in sensitivity and breadth. Several<br />
new technologies would be exploited as well<br />
as the capabilities of the LAMPF secondary<br />
beams.<br />
In any search for a very rare decay, sensitivity<br />
is limited by two factors: the total<br />
number of candidate decays observed, and<br />
any other process that mimics the decay<br />
being searched for. The design of an experi-<br />
ment must allow the reliable estimate of the<br />
contribution of other processes to a false<br />
signal. This is generally done by a Monte-<br />
Carlo simulation of these decays that includes<br />
taking into account the detector<br />
properties.<br />
In the absence of background or a positive<br />
signal for the process being studied, the number<br />
of seconds the experiment is run translates<br />
linearly into experimental sensitivity.<br />
However, when a background process is detected,<br />
sensitivity is gained only as the square<br />
root of the running time. This happens because<br />
one must subtract the number ofbackground<br />
events from the number of observed<br />
events, and the statistical uncertainties in<br />
these numbers determine the limit. Generally,<br />
when an experiment reaches a level<br />
limited by background, it is time to think of<br />
an improved detector.<br />
The Crystal Box detector is shown in Fig 5.<br />
A beam of muons from the LAMPF accelerator<br />
enters on the axis and is stopped in<br />
a thin polystyrene target. This beam consists<br />
of surface muons-a relatively new innovation<br />
developed during the 1970s and employed<br />
almost immediately at LAMPF and<br />
other meson factories.<br />
Normal beams of muons are prepared in a<br />
three-step process: a proton beam from the<br />
accelerator strikes a target, generating pions;<br />
the pions decay in flight, producing muons;<br />
finally, the optics in the beam line are adjusted<br />
to transport the daughter muons to the<br />
experiment while rejecting any remaining<br />
pions. A more efficient way to collect lowmomentum<br />
positive muons involves the use<br />
of a beam channel that collects muons from<br />
decays of positive pions generated in the<br />
target, but the muons collected are from<br />
pions that have only just enough momentum<br />
to travel from their production point in the<br />
target to its surface. Stopped in the surface,<br />
their decay produces positive muons of low<br />
momentum, near 29 MeV/c (where c is the<br />
speed of light). This technique enables experimenters<br />
to produce beams of surface<br />
muons that can be stopped in a thin experimental<br />
target with rates up to a hundred<br />
times more than conventional decay beams.<br />
137
The muons stopped in the target decay<br />
virtually 100 per cent of the time by the<br />
mode<br />
p+ - e+ v, Gp ,<br />
with a characteristic muon lifetime of 2.2<br />
microseconds. The Crystal Box detector accepts<br />
about 50 per cent of these decays and,<br />
therefore, must reject the positrons from several<br />
hu.ndred thousand ordinary decays OCcurring<br />
each second. At the same time the<br />
detector must select those decays that appear<br />
to be generated by the processes of interest.<br />
The Crystal Box was designed to simultaneously<br />
search for the decay modes<br />
p+ e+ e+ e-<br />
- e+y<br />
-e+yy.<br />
(Since the Crystal Box does not measure the<br />
charge of the particles, we shall not generally<br />
distinquish between positrons and electrons<br />
in our discussion.)<br />
The detector properties necessary for<br />
selecting final states from these reactions and<br />
rejecting events from ordinary muon decay<br />
are:<br />
1. Energy resolution-The candidate<br />
decays produce two or three particles whose<br />
energies sum to the energy of a muon at rest.<br />
The ordinary muon decay and most background<br />
processes include particles from several<br />
decays or neutrinos that remain undetected<br />
but carry away some of the energy.<br />
These processes are extremely unlikely to<br />
yield the correct energy sum.<br />
2. Momentum resolution- Given energy<br />
resolution adequate to accomplish the first<br />
requirement, vector momentum resolution<br />
requires a measurement of the directions of<br />
the particle trajectories. Since muons are<br />
stopped in the target, the decays being sought<br />
for will have vector momentum sums<br />
clustered, within experimental resolution,<br />
about zero. <strong>Particle</strong>s from the leading background<br />
processes (p+ - e+ e+ e- v, ;, p+ -<br />
e+ y v, Gw, or coincidences of different ordinary<br />
muon decays) will tend to have non-<br />
138<br />
Fig. 5. The Crystal Box detector. (a) A beam of muons enters the detector on axis.<br />
Because these are low-momenta surface muons, a thin polystyrene target is able to<br />
stop them at rates up to 100 times more than conventional muon beams. The beam<br />
intensity is generally chosen to be between 300,000 and 600,000 muons per second<br />
with pulses produced at a frequency of 120 hertz and a net duty factor between 6<br />
and 10 per cent. Three kinds of detectors (drvt chamber, plastic-scintillation<br />
counters, and NaI(T1) crystals) surround the target. The detector elements are<br />
divided into four quadrants, each containing nine rows of crystals with a plastic<br />
scintillator in .front of each row. This combination of detectors provides information<br />
on the energies, times of passage, and directions of the photons and electrons<br />
that result from muon decay in the target. The information is used to filter from<br />
several hundred thousand ordinary decays per second the perhaps several per<br />
second that may be of interest.<br />
A sophisticated calibration and stabilization system was developed to achieve
Experiments To Test Unification Schemes<br />
and maintain the desired energy and time resolution for 4 X 106 seconds of data<br />
taking. Before a run starts, a plutonium-beryllium radioactive source is used for<br />
electron energy calibration. Also, a liquid hydrogen target is substituted<br />
periodically for the experimental target, and the photons emitted in the subsequent<br />
pion charge exchange are used for photon energy calibration. During data taking,<br />
energy calibration is monitored by a fiber optic flasher system that exposes each<br />
photomultiplier channel to a known light pulse. A small number of positrons are<br />
accepted from ordinary p+ - e+ v, Vll decays, and the muon decay spectrum cutoff<br />
at 52.8 MeV is used as a reference.<br />
(b) The inner dectector, the drift chamber, consists of 728 cells in 8 annular<br />
rings with about 5000 wires strung to provide the drift cell electrostatic geometry.<br />
A 5-axis, computer-controlled milling machine was used to accurately drill the<br />
array of 5000 holes in each end plate. These holes, many drilled at angles up to<br />
about 10 degrees, had to be located within 0.5 mil so that the chamber wires could<br />
be placed accurately enough to achieve a final resolution of about 1 millimeter in<br />
measuring the position of a muon decay in the target. The area of the stopping<br />
muon spot is about 100 cm2. (Photo courtesy Richard Bolton.)<br />
(c) The outer layer of the detector (here shown under construction) contains 396<br />
thallium-doped sodium iodide crystals and achieves an electron and photon energy<br />
resolution of 5 to 6 per cent. This layer is highly segmented so that the electromagnetic<br />
shower produced by an event is spread among a cluster of crystals. A<br />
weighted average of the energy deposition can then be used to localize the<br />
interaction point of the photons with a position resolution of about 2 cm.<br />
zero vector sums.<br />
3. Time resolution-<strong>Particle</strong>s from the<br />
decay ofa single muon are produced simultaneously.<br />
A leading source of background for,<br />
say p+ - e+ e+ e-, is three electrons from the<br />
decay of three different muons. Such threebody<br />
final states are unlikely to occur simultaneously.<br />
Precision resolution in the time<br />
measurement, significantly better than I<br />
nanosecond, provides a powerful rejection of<br />
those random backgrounds.<br />
4. Position resolution-Decays from a<br />
single muon will originate from a single point<br />
in the stopping target. Sometimes other<br />
processes will add extra particles to an event.<br />
The ability to accurately measure the trajectory<br />
of each particle in an event is crucial if<br />
experimental triggers that have extra tracks<br />
or that originate in separate vertices are to be<br />
rejected.<br />
These parameters are used to filter<br />
measured events. In a sample of lo’*<br />
muons-the number required to reach<br />
sensitivities below the IO-” level-most of<br />
this filtering must be done immediately, as<br />
the data is recorded. The Crystal Box experiment<br />
is exposed to approximately 500,000<br />
muons stopping per second. The experimental<br />
“trigger” rate, the rate of decays that<br />
satisfy crude requirements, is about 1000<br />
hertz. The detector has been designed with<br />
enough intelligence in its hardwired logic<br />
circuits to pass events to the data acquisition<br />
computer at a rate of less than 10 hertz. In<br />
turn, the program in the computer applies<br />
more refined filtering conditions so that<br />
events are written on magnetic tape at a rate<br />
of a few hertz.<br />
Each condition used to narrow down the<br />
event sample to those that are real candidates<br />
provides a suppression factor. The combined<br />
suppression factors must permit the desired<br />
sensitivity. The design of the apparatus<br />
begins with the required suppressions and<br />
applies the necessary technology to achieve<br />
them.<br />
A muon that stops in the target and decays<br />
by one of the subject decay modes produces<br />
only electrons, positrons or photons. The<br />
charged particles (hereafter referred to as<br />
139
electrons) are detected by an 8-layer wire<br />
drift chamber (Fig. 5 (b)) immediately surrounding<br />
the target. The drift chamber<br />
provides track information, pointing back at<br />
the origin of the event in the target and<br />
forward to the scintillators and crystals to<br />
follow. Its resolution and ability to operate in<br />
the high flux of electrons from ordinary<br />
muon decays in the target have pushed the<br />
performance limits of drift chambers; the<br />
chamber wires were placed accurately<br />
enough to achieve a final resolution of about<br />
1 millimeter (mm) in measuring the position<br />
ofa muon decay in the target.<br />
Electrons are detected again in the next<br />
shell out from the target-a set of 36 plastic<br />
scintillation counters surrounding the drift<br />
chamber. These counters provide a measurement<br />
of the time of passage of the electrons<br />
with an accuracy of approximately 350<br />
picoseconds. This accuracy is extraordinary<br />
for counters of the dimensions required (70<br />
cm Icing by 6 cm wide by 1 cm thick) but is<br />
crucial to suppressing the random trigger<br />
background for the p+ - e+ e+ e- reaction.<br />
This performance is achieved by using two<br />
photomultiplier tubes, one at each end of the<br />
scintillator, and two special electronic timing<br />
circuits developed by the collaborators.<br />
The electrons and photons that pass<br />
through the plastic scintillators deposit their<br />
energy in the next and outermost layer of the<br />
detector, a 396-crystal array of thalliumdoped<br />
sodium iodide crystals. These crystals,<br />
acting as scintillators, provide fast precision<br />
measurement of both electron and photon<br />
energy (providing the energy and momenturn<br />
filtering described earlier) and localize<br />
the interaction point of the photons with a<br />
position resolution of about 2 cm. The use of<br />
such large, highly segmented arrays of inorganic<br />
scintillator crystals was pioneered in<br />
high-energy physics in the late 1970's by the<br />
Crystal Ball detector at the Stanford Linear<br />
Accelerator Center. This technology is now<br />
widespread in particle physics research, with<br />
detectors planned that involve as many as<br />
12,000 crystals.<br />
The sodium iodide array also provides<br />
accurate time measurements on the photons.<br />
1410<br />
A fast photomultiplier tube and electronics<br />
with special pulse shaping, amplification,<br />
and a custom-tailored, constant-fraction timing<br />
discriminator were melded into a system<br />
that gives subnanosecond accuracy.<br />
The major detector elements-the drift<br />
chamber, plastic scintillators and sodium<br />
iodide crystals-are used in logical combinations<br />
to select events that may be of interest.<br />
A p'- e' e+ e- event is selected when three<br />
or more non-adjacent plastic scintillators are<br />
triggered and energy deposit occurs in the<br />
sodium iodide rows behind them. The<br />
special circuits developed for the scintillators<br />
are used for this selection: one high-speed<br />
circuit insures that the three or more triggers<br />
are coincident within a very tight time interval<br />
(approximately 5 nanoseconds), the second<br />
circuit requires the three or more hits to<br />
be in non-adjacent counters. The last requirement<br />
suppresses events in which low<br />
momentum radiative daughters trigger adjacent<br />
counters or when an electron crosses the<br />
crack between two counters.<br />
An even more sophisticated trigger processor<br />
was constructed to insure that the three<br />
particles triggering the apparatus conform to<br />
a topology consistent with a three-body<br />
decay of a particle at rest. Thus, a pattern of<br />
tracks that, say, necessarily has net momentum<br />
in one direction (Fig. 6 (a)) is rejected,<br />
but a pattern with the requisite symmetry<br />
(Fig. 6 (b)) is accepted. This "geometry box"<br />
is an array of programmable read-only-memory<br />
circuits loaded with all legal hit patterns<br />
as determined by a Monte-Carlo simulation<br />
of the p+ - e+ e+ e- experiment.<br />
Finally, the total energy deposited in the<br />
sodium iodide must be, within the real-time<br />
energy resolution, consistent with the rest<br />
energy of a muon.<br />
The p+ - e+ and p'- e+ y y reactions<br />
are selected by combining an identified electron<br />
(a plastic scintillator counter triggered<br />
coincident with sodium iodide signals) and<br />
one or more photons (a sodium iodide signal<br />
triggered with no count in the plastic scintillator<br />
in front ofit). Also, these events must<br />
be in the appropriate geometric pattern (for<br />
example, directly opposite each other for p+<br />
Fig. 6. (a) Apattern of tracks with net<br />
momentum is not consistent with the<br />
neutrinoless decay of a muon at rest,<br />
and such an event will be rejected,<br />
whereas an event with apattern such as<br />
the one in (b) will be accepted.<br />
- e+ y) and have the correct energy balance.<br />
The Crystal Box should report limits in the<br />
IO-" range on the three reactions of interest<br />
this calendar year. It will also be used during<br />
the next year in a search for the KO - y y y<br />
decay, which violates charge conjugation invariance.<br />
A search for only the p+ - e+ e+ e-
Experiments To Test Unification Schemes<br />
process is being carried out at the Swiss<br />
Institute for Nuclear Research with an ultimate<br />
sensitivity of IO-'* available in the<br />
next year.<br />
A third LAMPF p'- e' y experiment is<br />
planned after the Crystal Box experiment.<br />
With present meson factory beams and foreseeable<br />
detector technology, this next generation<br />
experiment may well be the final round.<br />
Neutrino-Electron Scattering<br />
e-<br />
e-<br />
/<br />
'e<br />
e-<br />
Weak (Neutral Current)<br />
\<br />
Weak (Charged Current)<br />
ig. 7. Examples of the electromagnetic and weak interactions in quantum field<br />
ieory.<br />
-<br />
e<br />
'*<br />
The unification of the electromagnetic and<br />
weak interactions is a treatment of physical<br />
processes described by the exchange of three<br />
fundamental bosons. The exchange of a<br />
photon yields an electromagnetic current,<br />
and the W' and Zo bosons are exchanged in<br />
interactions classified as charged and neutral<br />
weak currents, respectively. Figure 7 illustrates<br />
how quantum field theory represents<br />
these processes.<br />
A traditional method of probing electroweak<br />
unification in the standard model<br />
has been to determine the precise onset of<br />
weak effects in an interaction that is otherwise<br />
electromagnetic. Especially important<br />
are experiments-with polarized electron<br />
scattering at fixed target accelerators and<br />
more recent studies at electron-positron colliders-that<br />
probe the interference between<br />
the amplitudes of the electromagnetic and<br />
neutral-current weak interactions. Interference<br />
effects may be easier to observe than<br />
direct measurement of the small amplitudes<br />
of the weak interaction.<br />
An Irvine-<strong>Los</strong> <strong>Alamos</strong>-Maryland team is<br />
conducting a unique and novel search for<br />
another interference. They have set out to<br />
probe the pure1.v weak interference between<br />
the amplitudes of the charged and neutral<br />
currents. In the same way that electron scattering<br />
experiments search for interference<br />
between photon and Zo boson interactions,<br />
the <strong>Los</strong> Alarnos based experiment is searching<br />
for the interference between charged-current<br />
W interactions and neutral-current Zo<br />
interactions.<br />
This experiment is attempting a unique<br />
141
Fig. 8. The interaction between an electron and its neutrino<br />
can take place via either the neutral current (with a Z ') or<br />
the charged current (with a W-), which results in an interference<br />
term (2ANcur,,,,ACbrgrd) in the expression for the<br />
square of the total amplitude A,, An experiment at<br />
LAMPF willprobe this purely weak interference by studying<br />
v,-electron scattering.<br />
measurement because <strong>Los</strong> AIamos is currently<br />
the only laboratory in the world with<br />
the requisite source of electron neutrinos.<br />
Moreover, the experiment gains importance<br />
from the fact that comparatively little is<br />
known about the physics ofthe Zo relative to<br />
that of the W.<br />
The measurement is a simple variation on<br />
the electron-electron scattering experiments.<br />
To substitute the W current for the electromagnetic<br />
current, the experimenters<br />
substitute the electron neutrino v, as the<br />
projectile and set out to measure the frequency<br />
of electron-neutrino elastic scattering<br />
from electrons. While this is conceptually<br />
simple, it is, in fact, technically quite difficult.<br />
The experiment must yield a sufficiently<br />
precise measure of the frequency of<br />
these scatters to separate out theoretical<br />
predictions made with different assumptions.<br />
To illustrate how the experiment tests<br />
the standard model, we must examine the<br />
nature of the model's predictions for vde<br />
scal tering.<br />
142<br />
Electroweak theory obeys the group structure<br />
SU(2) X U(1). The SU(2) group has<br />
three generators, W', W-, and W3, which<br />
are the charged arid neutral vector bosons<br />
identified with the gauge fields. The U(1)<br />
group has a single neutral boson generator B.<br />
The familiar phenomenological neutral<br />
photon field is constructed from the linear<br />
combination<br />
A<br />
= W3 sin Ow + B cos OW,<br />
(where 9~ is the Weinberg angle, a measure<br />
of the ratio of the contributions of the weak<br />
and the electromagnetic forces to the total<br />
interaction). The phenomenological neutral<br />
current carried by the Zo is similarily constructed<br />
from<br />
Zo = W3<br />
cos Ow - B sin Ow.<br />
In the standard model the process<br />
v,+e--v,+e-<br />
can take place by the exchange of either the<br />
neutral-current boson Zo or the chargedcurrent<br />
boson W- (Fig. 8), resulting in the<br />
usual interference term for the probablity of<br />
a process that can take place in either of two<br />
ways. The question then is what form will<br />
this interference take.<br />
All models of the weak interaction that are<br />
currently considered viable predict a<br />
negative, or destructive, interference term. A<br />
model that can produce constructive interference<br />
is one that includes additional neutral<br />
gauge bosons beyond the 2'. Thus, the<br />
observation of a v,-e scattering cross section<br />
consistent with constructive interference<br />
would indicate a phenomenal change in our<br />
picture of electroweak physics. Since the<br />
common Zo with about the predicted mass<br />
was directly observed only last year, and<br />
since higher mass regions will be accessible<br />
during this decade, such a result would set off<br />
a vigorous search by the particle physics<br />
community.<br />
How will the traditional low-energy theory
Experiments To Test Unification Schemes<br />
/‘<br />
/<br />
\<br />
\<br />
\<br />
\<br />
\<br />
\<br />
\<br />
\<br />
1<br />
result of a two-body decay, is monoenergetic<br />
with an energy at about 30 MeV. The i,,<br />
spectrum has a cutoff energy at about 53<br />
MeV, and the v, spectrum peaks around 35<br />
or 40 MeV then falls off, also at about 53<br />
MeV. These three particles are the source of<br />
many possible measurements.<br />
The primary goal is the study of the v,-e<br />
elastic scattering already discussed. The detector,<br />
which we shall describe in more detail<br />
shortly, must detect electrons characteristic<br />
of the elastic scattering, that is, they should<br />
have energies between 0 and 53 MeV and lie<br />
within about 15 degrees of the forward direction<br />
(the tracks must point back to the neutrino<br />
source).<br />
Also, by selecting events with electrons<br />
below 35 MeV, the group will search for the<br />
first observation of an exclusive neutrinoinduced<br />
nuclear transition. The process<br />
V, + ”C -t e- + ‘*N<br />
0 10 20 30 40 50<br />
Neutrino Energy (MeV)<br />
Fig. 9. The energy spectra for the three types of neutrinos that result from the decay<br />
ofapositivepion (n+ - p+ + vP p+ - e+ + V, + ;,).<br />
of weak interactions (apparently governed by<br />
V - A currents) mesh with future observations<br />
at higher energies? The standard model<br />
prediction, which contains negative interference,<br />
is that the cross section for v,-e<br />
elastic scattering should be about 60 per cent<br />
of the cross section in the traditional V - A<br />
theory. The LAMPF experiment must<br />
measure the cross section with an accuracy of<br />
about 15 per cent to be able to detect the<br />
lower rate that would occur in the presence of<br />
interference and thus be able to determine<br />
whether interference effects are present or<br />
not.<br />
In addition, the magnitude of the interference<br />
is a function of sin20w, and a precise<br />
measurement of the interference constitutes<br />
a measurement of this factor. In fact, it is<br />
60<br />
statistically more efficient to do this with a<br />
neutral current process because the charged<br />
current contains sin2ew (= 0.25) summed<br />
with unity, whereas for the neutral current<br />
the leading term is sin20w.<br />
The Experiments. The LAMPF proton<br />
beam ends in a thick beam stop where pions<br />
(n+) are produced These pions decay by the<br />
process<br />
n+ - p+ + vp<br />
Le+ + v,. + Gp ,<br />
yielding three types of neutrinos exiting the<br />
beam stop. The v, and ;, are each produced<br />
with a continuous spectrum (Fig. 9) typical<br />
of muon decay, whereas the vp spectrum, the<br />
would produce electrons with less than 35<br />
MeV energy that lie predominantly outside<br />
the angular region for the elastic scattering<br />
events.<br />
Another important physics goal, neutrino<br />
oscillations, can be addressed simultaneously.<br />
A process, called an “appearance,” in<br />
which the ip species disappears from the<br />
beam and i, appears, can be probed by<br />
searching for the presence of v, in the beam.<br />
This type of neutrino does not exist in the<br />
original neutrino source, so its presence<br />
downstream could be evidence for the<br />
i,, -i, oscillation. The experimental<br />
signature for such a process is the presence of<br />
isotropic single positrons produced by the<br />
reaction<br />
combined with a selection in energy of more<br />
than 35 MeV, which can be used to isolate<br />
these candidate events from the nuclear transition<br />
process discussed above.<br />
In all three of the processes studied, the<br />
technical problem to be solved is the separation<br />
of the desired events from competing<br />
143
ackground processes. The properties of the<br />
detector (Fig. IO) needed to do this include:<br />
1. Passive shielding-Lead, iron, and concrete<br />
are used to absorb charged and neutral<br />
cosmic ray particles entering the detector<br />
volume. However, the shield is not thick<br />
enough to insure that events seen in the inner<br />
detector come only from neutrinos entering<br />
the detector and not from residual cosmic<br />
ray backgrounds. The outer shield merely<br />
reduces the flux, consisting mainly of muons<br />
and hadrons from cosmic rays and of neutrons<br />
from the LAMPF beam stop.<br />
The LAMPF beam is on between 6 and 10<br />
per cent of each second so that the periods<br />
between pulses will provide an important<br />
normalizing measurement indicating how<br />
well the passive shielding works.<br />
2. Active anti-coincidence shield-This<br />
multilayer device is an active detector that<br />
surrounds the inner detector and serves<br />
many purposes. For example, muons from<br />
cosmic rays that penetrate the passive shield<br />
are detected here by being coincident in time<br />
with an inner detector trigger. This allows the<br />
rejection of these “prompt” muons, with less<br />
than lone muon in IO4 surviving the rejection.<br />
Data acquisition electronics that store<br />
the history of the anti-coincidence shield for<br />
32 microseconds prior to an inner detector<br />
trigger serve an even more complex purpose.<br />
This information is used to reject any inner<br />
detector electrons coming from a muon that<br />
stopped in the outer shield and that took up<br />
to 32 microseconds to decay. The mean<br />
muon lifetime is only 2.2 microseconds, so<br />
this is a very satisfactory way to reject such<br />
events.<br />
3. Inner converter-Photons penetrating<br />
the anti-coincidence layer, produced perhaps<br />
by cosmic rays or particles associated with<br />
the beam, strike an additional layer of steel<br />
and are either absorbed or converted into<br />
electronic showers that are seen as tracks<br />
connected to the edge of the inner detector.<br />
Such events are discarded in the data analysis.<br />
4. Inner detector-This module’s primary<br />
role is to measure the trajectory and energy<br />
deposition of electrons and other charged<br />
Fig. 10. The detector for the neutrino-electron scattering experiments. The outer<br />
layer of passive shielding (mainly steel) cuts down the flux of neutral solar<br />
particles.<br />
The anti-coincidence shield rejects muons from cosmic rays and electrons<br />
coming from the decay of muons stopped in the outer shield. It consists of four<br />
layers of drvt tubes, totaling 603 counters, each 6 meters long. A total of 4824<br />
wires provides LI fine-grained, highly effective screen, with an inefficiency (and<br />
therefore a suppression) of 2 X 1r5.<br />
Another steel layer, the inner converter, is used to reject photons from cosmic<br />
rays or otherparticles associated with the beam.<br />
The inner detector consists of 10 tons ofplastic scintillators interleaved with 4.5<br />
tons of tracking chambers. The plastic scintillators sample the electron energy<br />
every 10 layers of track chamber. There are 160 counters, each 75 cm by 300 cm by<br />
2.5 cm thick, and they measure the energy to about 10 per cent accuracy. The track<br />
chambers are a classic technology: they are flash chambers that behave like neon<br />
lights when struck by an ionizing particle, discharging in a luminous and climactic<br />
way. There are a phenomenal 208,000Jlash tubes in the detector, and they measure<br />
the electron tracks and sort them into angular bins about 7 degrees wide.
Experiments To Test Unification Schemes<br />
particles. Electron tracks are the signature of<br />
the desired neutrino reactions, but recoil<br />
protons generated by neutrons from the<br />
beam stop and from cosmic rays must also be<br />
detected and filtered out in the data analysis.<br />
The inner detector contains layers of plastic<br />
scintillators that sample the particle energy<br />
deposited along its path for particle identification<br />
and also provide a calorimetric measurement<br />
of the total energy. Trajectory measurement<br />
is provided by a compact system of<br />
flash chambers interleaved with the plastic<br />
scintillators.<br />
When this detector is turned on, it counts<br />
about 10’ raw events per day, mostly from<br />
cosmic rays. To illustrate the selectivity required<br />
of this experiment, a recent data run<br />
ofa few months is expected to produce somewhat<br />
less than 50 v,-e elastic scattering<br />
events.<br />
This highly segmented detector is<br />
necessarily extremely compact. The neutrino<br />
flux produced in the beam stop is emitted in<br />
all directions and therefore has an intensity<br />
that falls off inversely with the square of the<br />
distance. Thus there was a strong design<br />
premium for developing a compact, dense<br />
detector and placing it as close to the source<br />
as feasible.<br />
The detector is now running around the<br />
clock, even when the LAMPF beam is off<br />
(to pin down background processes). The<br />
data already taken include many v,-e events<br />
that are being reported, as are preliminary<br />
results on lepton number conservation and<br />
neutrino oscillations. Data taken with additional<br />
neutron shielding during the next year<br />
or two are expected to provide the precision<br />
test of the standard model that the experimenters<br />
seek.<br />
Beyond this effort, the beginnings of a<br />
much larger and ambitious neutrino program<br />
at <strong>Los</strong> <strong>Alamos</strong> are evident. A group<br />
(<strong>Los</strong> <strong>Alamos</strong>; University of New Mexico;<br />
Temple University; University of California,<br />
<strong>Los</strong> Angeles and Riverside; Valparaiso University;<br />
University of Texas) working in a<br />
new LAMPF beam line are mounting the<br />
prototype for a much larger fine-grained neutrino<br />
detector. Currently, a focused beam<br />
source of neutrinos is being developed that<br />
will eventually employ a rapidly pulsed<br />
“horn” to focus pions that decay to neutrinos.<br />
This development will be used to<br />
provide neutrinos for a major new detector.<br />
The group is not content to work merely on<br />
developing the facility but is using a preliminary<br />
detector to measure some key cross<br />
sections and set new limits on neutrino oscillations<br />
as well.<br />
Another group (Ohio State, Louisiana<br />
State, Argonne, California Institute of Technology,<br />
<strong>Los</strong> <strong>Alamos</strong>) is assembling the first<br />
components ofan aggressive effort to search<br />
for the i, appearance mode. Other physicists<br />
at the laboratory are preparing a solar neutrino<br />
initiative.<br />
The exciting field of neutrino research,<br />
begun by <strong>Los</strong> <strong>Alamos</strong> scientists, is clearly<br />
entering a golden period.<br />
Precision Studies of Normal<br />
Muon Decay<br />
The measurement of the electron energy<br />
spectrum and angular distribution from ordinary<br />
muon decay,<br />
p +<br />
e+ i,+ V,, ,<br />
is one of the most fundamental in particle<br />
physics in that it is the best way to determine<br />
the constants of the weak interaction. These<br />
studies have led to limits on the V - A<br />
character of the theory.<br />
The spectrum of ordinary muon decay<br />
may be precisely calculated from the standard<br />
model. Built into the minimal standard<br />
model-consistent with the idea that everything<br />
in the model must be required by<br />
measurements-are the assumptions that<br />
neutrinos are massless and the only interactions<br />
that enter are of vector and axial vector<br />
form (that is, V- A, or equal magnitude and<br />
opposite sign). Lepton flavor conservation is<br />
also taken to be exact.<br />
This V - A structure of the weak interaction<br />
can be tested by precise measurements<br />
of the electron spectrum from ordinary<br />
muon decay. The spectrum is characterized<br />
(to first order in m,/m,, and integrated over<br />
the electron polarization) by<br />
dN a (3-2x)<br />
2 dx d (cos 0)<br />
+ -p-1 (4x-3)<br />
(: )<br />
+ 1 2<br />
m<br />
‘<br />
(I-x)<br />
-<br />
m,, x2<br />
where me is the electron mass, 8 is the angle<br />
of emission of the electron with respect to the<br />
muon polarization vector P,, m,, is the muon<br />
mass, and x’ is the reduced electron energy (x<br />
= 2E/m,, where E is the electron energy). The<br />
Michel parameters p, q, e, 6 characterize the<br />
spectrum.<br />
The standard model predicts that<br />
p=6=3/4, 5=1, and q=O.<br />
One can also measure several parameters<br />
characterizing the longitudinal polarization<br />
of the electron and its two transverse components.<br />
Table 2 gives the current world average<br />
values for the Michel parameters. These<br />
data have been used to place limits on the<br />
weak interaction coupling constants, as<br />
shown in Table 3. As can be seen, the current<br />
limits allow up to a 30 per cent admixture of<br />
something other than a pure V- A structure.<br />
Other analyses, with other model-dependent<br />
assumptions, set the limit below IO per cent.<br />
One of the extensions of the minimal standard<br />
model is a theory with left-right symmetry.<br />
The gauge symmetry group that embodies<br />
the left-handed symmetry would be<br />
joined by one for right-handed symmetry,<br />
and the charged-current bosons W’ and W-<br />
would be expanded in terms of a symmetric<br />
combination of fields W, and W,. Such an<br />
145
extension is important from a theoretical<br />
standpoint for several reasons. First, it<br />
restores parity conservation as a high-energy<br />
symmetry of the weak interaction. The wellknown<br />
observation of parity violation in<br />
weak processes would then be relegated to<br />
the status of a low-energy phenomenon due<br />
to the fact that the mass of the right-handed<br />
W is much larger than that of the left-handed<br />
W. Each lepton generation would probably<br />
require two neutrinos, a light left-handed one<br />
and a very heavy right-handed member.<br />
The dominance of the left-handed charged<br />
current at presently accessible energies<br />
would be due to a very large mass for W,,<br />
but the W, - WR mass splitting would still<br />
be small on the scale of the grand unification<br />
mass Mx. Thus the precision study of a weak<br />
decay such as ordinary muon decay or<br />
nucleon beta decay can be used to set a limit<br />
on the left-right symmetry of the weak interaction.<br />
With such plums as the V- A nature of<br />
the weak interaction and the existence of<br />
right-handed W bosons accessible to such<br />
precision studies, it is not surprising that<br />
several experimental teams at meson factories<br />
are carrying out a variety of studies of<br />
ordinary muon decay. One team working at<br />
the Canadian facility TRIUMF has already<br />
collected data and set a lower limit of 380<br />
GeV on the mass of the right-handed W.<br />
This was done with a muon beam of only a<br />
few MeV!<br />
Table 2<br />
Theoretical and1 experimental values for the weak-interaction Michel<br />
parameters.<br />
Michel<br />
Parameter<br />
P<br />
11<br />
5<br />
6<br />
_. .<br />
Table 3<br />
V-A<br />
Prediction<br />
Current<br />
Value<br />
3/4 0.752 k 0.003<br />
0 -0.12 k 0.21<br />
1 0.972 k 0.014<br />
Y4 0.755 k 0.008<br />
Expected<br />
<strong>Los</strong> <strong>Alamos</strong> Accuracy<br />
k 0.00023<br />
rt 0.006 1<br />
k 0.001<br />
rt 0.00064<br />
. . . . . . . ...-,<br />
- .. . .-<br />
Experimental limits on the weak-interaction coupling constants, including<br />
the expected limit for the <strong>Los</strong> <strong>Alamos</strong> Experiment.<br />
Constant Present Limit Expected Limit<br />
Axial Vector<br />
0.76 < gA < 1.20 0.988 < gA < 1.052<br />
Tensor gT < 0.28 gT < 0.027<br />
Scalar gs < 0.33 gs < 0.048<br />
Pseudo Scalar gp < 0.33 g~. < 0.048<br />
Vecto-axial Vector Phase (PVA = 180" Ifr 15" (PVA= 18Vrt2.6"<br />
i<br />
i<br />
I<br />
- I<br />
,<br />
The Time Projection Chamber. A <strong>Los</strong><br />
<strong>Alamos</strong> - University of Chicago-NRC Canada<br />
collaboration is carrying out a<br />
particularly comprehensive and sensitive<br />
study of the muon decay spectrum using a<br />
novel and elaborate device known as a time<br />
projection chamber (TPC).<br />
The TPC (Fig. 11) is a very large volume<br />
drift chamber. In a conventional drift<br />
chamber, an array of wires at carefully determined<br />
potentials collects the ionization<br />
left in a gas by a passing charged particle. The<br />
time of arrival of the packet of ionization in<br />
the cell near each wire is used to calculate the<br />
path of the particle through the cell. The gas<br />
and the field in the cell are chosen so that the<br />
ionization drifts at a constant terminal velocity.<br />
Thus the calculation of the position from<br />
the drift time can be done accurately. Many<br />
drift chambers provide coordinate measurements<br />
accurate to less than 100 micrometers.<br />
On the other hand, a TPC uses the same<br />
drift velocity phenomenon but employs it in<br />
a large volume with no wires in the sensitive<br />
region. The path iofionization drifts en masse<br />
under the influence of an electric field along<br />
the axis of the chamber. The ionization is<br />
collected on a series of electrodes, called<br />
pads, on the chamber endcaps, providing<br />
precision measurement of trajectory charge<br />
and energy. The pad signal also gives a time<br />
measurement, relative to the event trigger,<br />
that can be used to reconstruct the spatial<br />
coordinate of each point on the trajectory.<br />
The TPC in the <strong>Los</strong> <strong>Alamos</strong> experiment is<br />
placed in a magnetic field sufficiently strong<br />
that the decay electrons, whose energies<br />
range up to about 53 MeV, follow helical<br />
paths. The magnetic field is accurate enough<br />
to make absolute momentum measurements<br />
of the decay electrons.
Experiments To Test Unification Schemes<br />
I '<br />
Scintillator<br />
/<br />
Deflector 1<br />
Iron Yoke '<br />
Magnet<br />
Coils<br />
Removable<br />
Pole<br />
High-Voltage Electrodes<br />
Readout Plane<br />
(Series of Pads)<br />
Fig. 11. The time projection chamber (TPC), a device to study the muon decay<br />
spectrum. A beam of muons from LAMPF enters the TPC via a 2-inch beam pipe<br />
that extends through the magnet pole parallel to the magnetic field direction.<br />
Before entering the chamber, the muons pass through a IO-mil thick scintillator<br />
that serves as a muon detector. The scintillator is viewed, via fiber optic light<br />
gzides, by two photomultiplier tubes located outside the magnet. The thresholds for<br />
the discriminators on these photomultiplier channels are adjusted to produce a<br />
coincidence for the more heavily ionizing muons while the minimum-ionizing<br />
beam electrons are ignored. A deflector located in the beam line 2 meters upstream<br />
of the magnet produces a region of crossed electric and magnetic fields through<br />
which the beam passes. This device acts first as a beam separator, purifying the<br />
muon flux-in particular, reducing the number of electrons in the beam from<br />
about 200 to about 1.5 for every muon. The device also acts as a deflector, keeping<br />
additional particles out of the chamber by switching off the electric field once a<br />
muon has been observed entering the detector. The magnetic field in this detector is<br />
provided by an iron-enclosed solenoid, with the maximum field in the current<br />
arrangement being 6.6 kilogauss. The field has been carefully measured and found<br />
to be uniform to better then 0.6per cent within the entire TPC-sensitive volume of<br />
52 cm in length by 122 cm in diameter. The TPC readout, on the chamber endcaps,<br />
consists of 21 identical modules, each of which has 15 sense wires and 255 pads<br />
arranged under the sense wires in rows of I7pads each. The sense wires provide the<br />
high field gradient necessary for gas amplification of the track ionization. The 21<br />
modules are arranged to cover most of the 122-centimeter diameter of the chamber.<br />
A beam of muons from LAMPF passes<br />
first through a device that acts as a beam<br />
separator, purifying the muon flux<br />
(especially of electrons, which are reduced by<br />
this device from an electron-to-muon ratio of<br />
200:l to about 3:2). The device also acts as a<br />
deflector, keeping additional particles from<br />
entering the chamber once a muon is inside.<br />
With a proper choice of beam intensity, only<br />
one muon is allowed in the TPC at a time.<br />
Next the beam passes through a IO-mil thick<br />
scintillator (serving both as a muon detector<br />
and a device used to reject events caused by<br />
the remaining beam electrons) and continues<br />
into the TPC along a line parallel to the<br />
magnetic field direction.<br />
The requirement for an event to be triggered<br />
is that one muon enters the TPC during<br />
the LAMPF beam pulse and stops in the<br />
central lOcm ofthe drift region. The entering<br />
muon is detected by a signal coincidence<br />
from photomultipliers attached to the IO-mil<br />
scintillator (this signal operates the deflector<br />
that keeps other muons out). The scintillator<br />
signal must also be coincident-including a<br />
delay that corresponds to the drift time from<br />
the central IO cm of the TPC-with a high<br />
level signal from any of the central wires of<br />
the TPC. If no delayed coincidence occurs,<br />
indicating that the muon did not penetrate<br />
far enough into the TPC, or a high level<br />
output is detected before the selected time<br />
window, indicating that the muon<br />
penetrated too far, the event is rejected and<br />
all electronics are reset. Then 250 microseconds<br />
later (to allow for complete clearing<br />
of all tracks in the TPC) the beam is allowed<br />
to re-enter for another attempt. The event is<br />
also rejected if a second muon enters the<br />
TPC during the 200-nanosecond period required<br />
to turn off the deflector electric field.<br />
If the event is accepted, the computer<br />
reads 20 microseconds of stored data. This<br />
corresponds to five muon decay lifetimes<br />
plus the 9 microseconds it takes for a track to<br />
drift the full length of the TPC.<br />
The experiment is expected to collect<br />
about IOs muon decay events, at a trigger<br />
rate of 120 events per second, during the next<br />
year. Preliminary data have already been<br />
147
taken, showing that the key resolution for<br />
electron momentum falls in the target range,<br />
namely 4p/p is 0.7 per cent averaged over the<br />
entire spectrum. Figure 12 shows one of the<br />
elegant helica: tracks obtained in these early<br />
runs.<br />
Ultimately, this experiment will be able to<br />
improve upon the four parameters shown in<br />
Table 2, although the initial emphasis will be<br />
on p. In the context of left-right symmetric<br />
models, an improved measurement of p will<br />
place a new limit on the allowed mixing<br />
angle between W, and W, that is almost<br />
independent of the mass of the W,.<br />
Summary<br />
The particle physics community is aggressively<br />
pursuing research that will lead to<br />
verification or elaboration of the minimal<br />
standard model. Most of the world-wide activity<br />
is centered at the high-energy colliding<br />
beam facilities, and the last few years have<br />
yielded a bountiful harvest of new results,<br />
including the direct observation of the W*<br />
and Zo bosons. Many of the key measure-<br />
Fig. 12. An example of the typical helical track observed for a muon-decay event in<br />
an early run with the TPC. (The detector here is shown on end compared to Fig.<br />
11.)<br />
ments of the 1980s are likely to be made at<br />
the medium-energy facilities, such as<br />
LAMPF, or in experiments far from accelerators,<br />
deep underground and at reactors,<br />
where studies of proton decay, solar neutrino<br />
physics, neutrino oscillations, tritium beta<br />
decay, and other bellwether research is being<br />
camed out.<br />
Gary H. Sanders learned his physics on the east coast, starting at<br />
Stuyvesant High School in New York City, then Columbia and an A.B. in<br />
physics in 1967, and finally a Ph.D. from the Massachusetts Institute of<br />
Technology in 1971. The work for his doctoral thesis, which dealt with the<br />
photoproduction of neutral rho mesons on complex nuclei, was<br />
performed at DESY's electron synchrotron in Hamburg, West Germany<br />
under the guidance of Sam Ting. After seven years at Princeton University,<br />
during which time he used the beam at Brookhaven National<br />
Laboratory and Fermi National Accelerator Laboratory, he came west to<br />
join the Laboratory's Medium Energy <strong>Physics</strong> Division and use the<br />
beams at LAMPF. A great deal of his research has dealt with the study of<br />
muons and with the design of the beams, detectors, and signal processing<br />
equipment needed for these experiments.<br />
I 1413
Addendum<br />
An Experimental<br />
Update<br />
Neutrino-Electron Scattering. This experiment<br />
has now collected 121 f 25 v,e- scattering<br />
events, of which 9,' 25 are identified<br />
as v,e- scatterings. The resulting cross section<br />
agrees with the standard electroweak<br />
theory, rules out constructive interference<br />
between weak chargedcurrent and neutralcurrent<br />
interactions, and favors the existence<br />
of destructive interference between these two<br />
interactions. Additional data are being collected,<br />
and this result will be made considerably<br />
more precise.<br />
I<br />
n the two years since "Experiments to<br />
Test Unification Schemes" was written,<br />
each of the four experiments described<br />
has completed a substantial program of<br />
measurements and published its first results.<br />
In one case, the entire program is complete<br />
with final results submitted for publication.<br />
So far, all results are fblly consistent with the<br />
minimal standard model. Opportunities for<br />
theories with new physics have been substantially<br />
constrained.<br />
Tritium Beta Decay. Although dissociation<br />
into atomic tritium has not yet been employed<br />
to make a physics measurement, the<br />
study of the tritium beta decay spectrum<br />
using molelcular tritium has been completed.<br />
An upper limit of 26.8 eV (95 per cent confidence<br />
level) has been placed on the mass of<br />
the electron antineutrino, with a best fit<br />
value ofzero mass. This result is inconsistent<br />
with the best fit value most recently reported<br />
by the Russians (30 & 2 eV) and excludes a<br />
large fraction of their latest mass range of 17<br />
to 40 eV. Several other experiments, including<br />
those done by teams from Lawrence Livermore<br />
National Laboratory, the Swiss Institute<br />
for Nuclear Research, and a Japanese<br />
group, have also begun to erode the Russian<br />
claim of a nonzero neutrino mass. Improvements<br />
in these limits, including the <strong>Los</strong> <strong>Alamos</strong><br />
atomic tritium measurement, are expected<br />
soon.<br />
Rare Decays of the Muon. The Crystal Box<br />
detector has completed its search for rare<br />
decays of the muon and, for the three<br />
processes sought, has published (at the 90 per<br />
cent confidence level) the following new<br />
limits:<br />
B(p+- e'yy) < 7.2 X lo-'' .<br />
An experiment at the Swiss Institute for Nuclear<br />
Research has also obtained a limit on<br />
the first process of about 2.4 X : O-'*. These<br />
four results place severe lower limits on the<br />
masses of new objects that could produce<br />
nonconservation of lepton number. For example,<br />
in one analysis of deviations from the<br />
standard model, the Crystal Box limit on<br />
p+ - e'y sets the scale for new interactions<br />
to lo4 TeV or higher.<br />
A new and far more ambitious search for<br />
p'- e'y has been undertaken by some<br />
members of the Crystal Box group together<br />
with a large group of new collaborators.<br />
Their design sensitivity is set at less than 1 X<br />
and their new detector is under construction.<br />
In addition, four other groups, using<br />
rare decays of the kaon, are searching for<br />
processes that violate lepton number conservation,<br />
such as KL - pe and KL - npe.<br />
Normal Muon Decay. After preliminary<br />
studies, the team using the time projection<br />
chamber to study normal muon decay decided<br />
to concentrate on the measurement of<br />
the p parameter (Table 2). Since this parameter<br />
is measured by averaging over the muon<br />
spin, it was not necessary to preserve the spin<br />
direction of the muon stopping in the<br />
chamber. The researchers used an improved<br />
entrance separator to rotate the spin perpendicular<br />
to the beam direction, and precession<br />
in the chamber magnetic field then averaged<br />
the polarization. They also took advantage of<br />
a small entrance scintillator to trigger the<br />
apparatus on muon stops. This technique<br />
purified the experimental sample but<br />
perturbed the muon spin, which, however, is<br />
acceptable for the measurement of p. A<br />
higher event rate was possible because the<br />
entrance scintillator signal eliminated the<br />
need for the beam to be pulsed. The scintillator,<br />
by itself, effectively ensured a single<br />
stopping muon. In this mode, the group collected<br />
5 X lo7 events, which are now being<br />
analyzed, and this sample is expected to<br />
sharpen the knowledge of p by a factor of 5.<br />
Limits on charged right-handed currents<br />
from a related measurement have now been<br />
reported by the TRIUMF group.<br />
The minimal standard model has, to date,<br />
survived these demanding tests. Where is the<br />
edge of its validity? We shall have to wait a<br />
little longer to find the answer. Experimentalists<br />
are already mounting the next round<br />
of detectors in this inquiry.<br />
149
I<br />
h<br />
6.2<br />
J<br />
q<br />
P 0<br />
n<br />
f<br />
D<br />
@<br />
8<br />
by S. Peteg Rosen
When these questions have been<br />
answered, we may expect the cycle to repeat<br />
itself until we run out of resources-or out of<br />
space. So far the field of particle physics has<br />
been fixtunate: every time it seems to have<br />
reached the end of the energy line, some new<br />
technical development has come along to<br />
extend it into new realms. Synchrotrons such<br />
as the Bevatron and the Cosmotron, its sister<br />
and rival at Brookhaven, both represented<br />
an order-of-magnitude improvement over<br />
synchrocyclotrons, which in their time overcame<br />
relativistic problems to extend the<br />
energy of cyclotrons from tens of MeV into<br />
the hundreds. What allowed these developments<br />
was the synchronous principle invented<br />
independently by E. McMillan at<br />
Berkeley and V. Veksler in the Soviet Union.<br />
In a cyclotron a proton travels in a circular<br />
orbit under the influence of a constant magnetic<br />
field. Every time it crosses a particular<br />
diameter, it receives an accelerating kick<br />
from an rf electric field oscillating at a constanl<br />
frequency equal to the orbital frequency<br />
of the proton at some (low) kinetic<br />
energy. Increasing the kinetic energy of the<br />
proton increases the radius of its orbit but<br />
does not change its orbital frequency until<br />
the effects of the relativistic mass increase<br />
become significant. For this reason a<br />
cyclotron cannot efficiently accelerate<br />
proions to energies above about 20 MeV.<br />
The solution introduced by McMillan and<br />
Veksler was to vary the frequency of the rf<br />
field so that the proton and the field remained<br />
in synchronization. With such<br />
synchrocyclotrons proton energies of hundreds<br />
of MeV became accessible.<br />
In a synchrotron the protons are confined<br />
to a narrow range of orbits during the entire<br />
acceleration cycle by varying also the magnetic<br />
field, and the magnetic field can then be<br />
supplied by a ring of magnets rather than by<br />
the solid circular magnet of a cyclotron.<br />
Nevertheless, the magnets in early synchrotrons<br />
were still very large, requiring 10,000<br />
tons of iron in the case of the Bevatron, and<br />
for all practical purposes the synchrotron<br />
appeared to have reached its economic limit<br />
with this 6-GeV machine. Just at the right<br />
152<br />
time a group of accelerator physicists at<br />
Brookhaven invented the principle of<br />
“strong focusing,” and Ernest Courant, in<br />
May 1953, looked forward to the day when<br />
protons could be accelerated to 100<br />
GeV-fifty times the energy available from<br />
the Cosmotron--with much smaller<br />
magnets! In the meantime Courant and his<br />
colleagues contented themselves with building<br />
a machine ten times more energetic,<br />
namely, the AGS (Alternating Gradient<br />
Synchrotron).<br />
Courant proved to be most farsighted, but<br />
even his optimistic goal was far surpassed in<br />
the twenty years following the invention of<br />
strong focusing. The accelerator at Fermilab<br />
(Fermi National Accelerator Laboratory)<br />
achieved proton energies of 400 GeV in<br />
1972, and at CERN (Organisation Europeene<br />
pour Recherche NuclCaire) the SPS<br />
(Super Proton Synchrotron) followed suit in<br />
1976. Size is the most striking feature of<br />
these machines. Whereas the Bevatron had a<br />
circumference of 0.1 kilometer and could<br />
easily fit into a single building, the CERN<br />
and Fermilab accelerators have circumferences<br />
between 6 and 7 kilometers and are<br />
themselves hosts YO large buildings.<br />
Both the Fermilab accelerator and the SPS<br />
are capable of accelerating protons to 500<br />
GeV, but prolonged operation at that energy<br />
is prohibited by excessive power costs. This<br />
economic hurdle has recently been overcome<br />
by the successful development of superconducting<br />
magnets. Fermilab has now installed<br />
a ring of superconducting magnets in the<br />
same tunnel that houses the original main<br />
ring and has achieved proton energies of 800<br />
GeV, or close to 1 TeV. The success of the<br />
Tevatron, as it is called, has convinced the<br />
high-energy physics community that a 20-<br />
TeV proton accelerator is now within our<br />
technological grasp, and studies are under<br />
way to develop a proposal for such an accelerator,<br />
which would be between 90 and<br />
160 kilometers in circumference. Whether<br />
this machine, known as the SSC (Superconducting<br />
Super Collider), will be the terminus<br />
of the energy line, only time will tell; but if<br />
the past is any guide, we can expect some-<br />
thing to turn up. (See “The SSC-An Engineering<br />
Challenge.”)<br />
Paralleling the higher and higher energy<br />
proton accelerators has been the development<br />
of electron accelerators. In the 1950s<br />
the emphasis was on linear accelerators, or<br />
linacs, in order to avoid the problem of<br />
energy loss by synchrotron radiation, which<br />
is much more serious for the electron than<br />
for the more massive proton. The development<br />
of linacs culminated in the two-milelong<br />
accelerator at SLAC (Stanford Linear<br />
Accelerator Center), which today accelerates<br />
electrons to 40 GeV. This machine has had<br />
an enormous impact upon particle physics,<br />
both direct and indirect.<br />
The direct impact includes the discovery<br />
of the “scaling” phenomenon in the late<br />
1960s and of parity-violating electromagnetic<br />
forces in the late 1970s. By the<br />
scaling phenomenon is meant the behavior<br />
of electrons scattered off nucleons through<br />
very large angles: they appear to have been<br />
deflected by very hard, pointlike objects inside<br />
the nucleons. In exactly the same way<br />
that the experiments of Rutherford revealed<br />
the existence of an almost pointlike nucleus<br />
inside the atom, so the scaling experiments<br />
provided a major new piece of evidence for<br />
the existence of quarks. This evidence was<br />
further explored and extended in the ’70s by<br />
neutrino experiments at Fermilab and<br />
CERN.<br />
Whereas the scaling phenomenon opened<br />
a new vista on the physics of nucleons, the<br />
1978 discovery of parity violation in the<br />
scattering of polarized electrons by deuterons<br />
and protons closed a chapter in the history ol<br />
weak interactions. In 1973 the phenomenon<br />
ofweak neutral currents had been discovered<br />
in neutrino reactions at the CERN PS<br />
(Proton Synchrotron), an accelerator very<br />
similar in energy to the AGS. This discovery<br />
constituted strong evidence in favor of the<br />
Glashow-Weinberg-Salam theory unifying<br />
electromagnetic and weak interactions. During<br />
the next five years more and more<br />
favorable evidence accumulated until only<br />
one vital piece was missing-the demonstration<br />
of parity violation in electron-nucleon
the march toward higher energies<br />
The “string and sealing wax” version of a cyclotron. With this 4-inch device E. 0.<br />
Lawrence and graduate student M. S. Livingston successfully demonstrated the<br />
Peasibility of the cyclotron principle on January 2, 1931. The device accelerated<br />
wotons to 80 keV. (Photo courtesy of Lawrence Berkeley Laboratory.)<br />
eactions at a very small, but precisely<br />
iredicted, level. In a brilliant experiment C.<br />
’rescott and R. Taylor and their colleagues<br />
ound the missing link and thereby set the<br />
ea1 on the unification of weak and electronagnetic<br />
interactions.<br />
A less direct but equally significant impact<br />
f the two-mile linac arose from the electronositron<br />
storage ring known as SPEAR<br />
(Stanford Positron Electron Accelerating<br />
Ring). Electrons and positrons from the linac<br />
are accumulated in two counterrotating<br />
beams in a circular ring of magnets and<br />
shielding, which, from the outside, looks like<br />
a reconstruction of Stonehenge. Inside,<br />
enough rf power is supplied to overcome<br />
synchrotron radiation losses and to allow<br />
some modest acceleration from about 1 to 4<br />
GeV per beam. In the fall of 1974, the w<br />
particle, which provided the first evidence<br />
for the fourth, or charmed, quark was found<br />
among the products of electron-positron collisions<br />
at SPEAR; at the same time the J<br />
particle, exactly the same object as y, was<br />
discovered in proton collisions at the AGS.<br />
With the advent of J/w, the point of view<br />
that all hadrons are made of quarks gained<br />
universal acceptance. (The up, down, and<br />
strange quarks had been “found” experimentally;<br />
the existence of the charmed quark had<br />
been postulated in 1964 by Glashow and J.<br />
Bjorken to equalize the number of quarks<br />
and leptons and again in 1970 by Glashow, J.<br />
Iliopoulos, and L. Maiani to explain the apparent<br />
nonoccurrence of strangeness-changing<br />
neutral currents.<br />
The discovery of J/w, together with the<br />
discovery of neutral currents the year before,<br />
revitalized the entire field of high-energy<br />
physics. In particular, it set the building of<br />
electron-positron storage rings going with a<br />
vengeance! Plans were immediately laid at<br />
SLAC for PEP (Positron Electron Project), a<br />
larger storage ring capable of producing 18-<br />
GeV beams of electrons and positrons, and<br />
in Hamburg, home of DORIS (Doppel-Ring-<br />
Speicher), the European counterpart of<br />
SPEAR, a 19-GeV storage ring named<br />
PETRA (Positron Electron Tandem Ring<br />
Accelerator) was designed. Subsequently a<br />
third storage ring producing 8-GeV beams of<br />
positrons and electrons was built at Cornell;<br />
it goes by the name of CESR (Cornell Electron<br />
Storage Ring).<br />
Although the gluon, the gauge boson of<br />
quantum chromodynamics, was discovered<br />
at PETRA, and the surprisingly long lifetime<br />
of the b quark was established at PEP, the<br />
most interesting energy range turned out to<br />
be occupied by CESR. Very shortly before<br />
this machine became operative, L. Lederman<br />
and his coworkers, in an experiment at<br />
Fermilab similar to the J experiment at<br />
Brookhaven, discovered the T particle at 9.4<br />
GeV; it is the bquark analogue of J/w at 3.1<br />
GeV. By good fortune CESR is in just the<br />
right energy range to explore the properties of<br />
the T system, just as SPEAR was able to<br />
153
elucidate the I+I system. Many interesting results<br />
about T, its excited states, and mesons<br />
containing the b quark are emerging from<br />
this unique facility at Cornell.<br />
The next round for positrons and electrons<br />
includes two new machines, one a CERN<br />
storage ring called LEP (Large Electron-<br />
Positron) and the other a novel facility at<br />
SLAC called SLC (Stanford Linear Collider).<br />
LEP will be located about 800 meters under<br />
the Jura Mountains and will have a circumference<br />
of 30 kilometers. Providing 86-GeV<br />
electron and positron beams initially and<br />
later 130-GeV beams, this machine will be an<br />
excellent tool for exploring the properties of<br />
the W’ bosons. SLC is an attempt to overcome<br />
the problem of synchrotron radiation<br />
losses by causing two linear beams to collide<br />
head on. If successful, this scheme could well<br />
establish the basic design for future machines<br />
of extremely high energy. At present SLC is<br />
expected to operate at 50 GeV per beam, an<br />
ideal energy with which to study the Zo<br />
boson.<br />
High energy is not the only frontier against<br />
which accelerators are pushing. Here at <strong>Los</strong><br />
<strong>Alamos</strong> LAMPF (<strong>Los</strong> <strong>Alamos</strong> Meson <strong>Physics</strong><br />
Facility) has been the scene of pioneering<br />
work on the frontier of high intensity for<br />
more than ten years. At present this 800-<br />
MeV proton linac carries an average current<br />
of I milliampere. To emphasize just how<br />
great an intensity that is, we note that most of<br />
the accelerators mentioned above hardly<br />
ever attain an average current of 10 microamperes.<br />
LAMPF is one of three so-called<br />
meson factories in the world; the other two<br />
are highly advanced synchrocyclotrons at<br />
TRIUMF (Tri-University Meson Facility) in<br />
Vancouver, Canada, and at SIN (Schweizerisches<br />
Institut fur Nuklearforschung) near<br />
Zurich, Switzerland.<br />
The high intensity available at LAMPF<br />
has given rise to fundamental contributions<br />
in nuclear physics, including confirmation of<br />
the recently developed Dirac formulation of<br />
nucleon-nucleus interactions and discovery<br />
of giant collective excitations in nuclei. In<br />
addition, its copious muon and neutrino<br />
1 54<br />
A state-of-the art version of aproton synchrotron. Here at Fermilabprotons will be<br />
accelerated to an energy close to 1 TeV in a 6562-foot-diameter ring of superconducting<br />
magnets. Wilson Hall, headquarters of the laboratory and a fitting<br />
monument to a master accelerator builder, appears at the lower left. (Photo<br />
courtesy of Fermi National Accelerator Laboratory.)<br />
beams have been applied to advantage in<br />
particle physics, especially in the areas ofrare<br />
modes of particle decay and neutrino physics.<br />
The search for rare decay modes (such as<br />
p’ - e+ + y) remains high on the agenda of<br />
particle physics because our present failure<br />
to see them indicates that certain conservation<br />
laws seem to be valid. Grand unified<br />
theories of strong and electroweak interactions<br />
tell us that, apart from energy and<br />
momentum, the only strictly conserved<br />
quantity is electric charge. According to these<br />
theories, the conservation of all other quantities,<br />
including lepton number and baryon<br />
number, is only approximate, and violations<br />
of these conservation laws must occur, although<br />
perhaps at levels the minutest of the<br />
minute.<br />
Meson factories are ideally suited to the<br />
search for rare processes, and here at <strong>Los</strong><br />
<strong>Alamos</strong>, at TRIUMF, and at SIN plans are<br />
being drawn up to extend the range of present<br />
machines from pions to kaons. (See<br />
“LAMPF I1 and the High-Intensity Frontier.”)<br />
Several rare decays of kaons can<br />
provide important insights into grand uni-<br />
fied theories, as well as into theories that<br />
address the question of W’ and Zo masses,<br />
and so the search for them can be expected to<br />
warm up in the next few years.<br />
Another reason for studying kaon decays<br />
is CP violation, a phenomenon discovered<br />
twenty years ago at the AGS and still today<br />
not well understood. Because the effects of<br />
CP violation have been detected only in<br />
kaon decays and nowhere else, extremely<br />
precise measurements of the relevant<br />
parameters are needed to help determine the<br />
underlying cause. In this case too, kaon factories<br />
are very well suited to attack a fundamental<br />
problem of particle physics.<br />
In the area of neutrino physics, LAMPF<br />
has made important studies ofthe identity of<br />
neutrinos emitted in muon decay and is now<br />
engaged in a pioneering study of neutrinoelectron<br />
scattering. High-precision measurements<br />
of the cross section are needed as a test<br />
of the Glashow-Weinbcrg-Salam theory and<br />
are likely to be a major part of the experimental<br />
program at kaon factories.<br />
While the main thrust of particle physics<br />
has always been carried by accelerator-based
the march toward higher energies<br />
experiments, there are, and there have<br />
always been, important experiments performed<br />
without accelerators. The first<br />
evidence for strange particles was found in<br />
the late 1940s in photographic emulsions<br />
exposed to cosmic rays, and in 1956 the<br />
neutrino was first detected in an experiment<br />
at a nuclear reactor. In both cases accelerators<br />
took up these discoveries to explore<br />
and extend them as far as possible.<br />
Another example is the discovery of parity<br />
nonconservation in late 1956. The original<br />
impetus came from the famous 7-0 puzzle<br />
concerning the decay of K mesons into two<br />
and three pions, and it had its origins in<br />
accelerator-based experiments. But the definitive’<br />
experiment that demonstrated the<br />
nonconservation of parity involved the beta<br />
decay ofcobalt-60. Further studies of nuclear<br />
beta decay led to a beautiful clarification of<br />
the Fermi theory of weak interactions and<br />
laid the foundations for modern gauge theories.<br />
The history of this era reveals a remarkable<br />
interplay between accelerator and<br />
non-accelerator experiments.<br />
In more recent times the solar neutrino<br />
experiment carried out by R. Davis and his<br />
colleagues deep in a gold mine provided the<br />
original motivation for the idea of neutrino<br />
oscillations. Other experiments deep underground<br />
have set lower limits of order lo3’<br />
years on the lifetime of the proton and may<br />
yet reveal that “diamonds are not forever.”<br />
And the limits set at reactors on the electric<br />
dipole moment of the neutron have proved<br />
to be a most rigorous test for the many<br />
models of CP violation that have been<br />
proposed.<br />
In 1958, a time of much expansion and<br />
optimism for the future, Robert R. Wilson,<br />
the master accelerator builder, compared the<br />
building of particle accelerators in this century<br />
with the building of great cathedrals in<br />
12th and 13th century France. And just as<br />
the cathedral builders thrust upward toward<br />
Heaven with all the technical prowess at<br />
their command, so the accelerator builders<br />
strive to extract ever more energy from their<br />
mighty machines. Just as the cathedral<br />
builders sought to be among the Heavenly<br />
Hosts, bathed in the radiance of Eternal<br />
Light, so the accelerator builders seek to<br />
unlock the deepest secrets of Nature and live<br />
in a state of Perpetual Enlightenment:<br />
Ah, but a man’s reach should exceed<br />
his grasp,<br />
Or what s a heaven for?<br />
Robert Browning<br />
Wilson went on to build his great accelerator,<br />
and his cathedral too, at Fermilab<br />
near Batavia, Illinois. In its time, the early to<br />
mid 1970s, the main ring at Fermilab was the<br />
most powerful accelerator in the world, and<br />
it will soon regain that honor as the Tevatron<br />
begins to operate. The central laboratory<br />
building, Wilson Hall, rises up to sixteen<br />
stories like a pair of hands joined in prayer,<br />
and it stands upon the plain of northcentral<br />
Illinois much as York Minster stands upon<br />
the plain of York in England, visible for<br />
miles around. Some wag once dubbed the<br />
laboratory building “Minster Wilson, or the<br />
Cathedral of St. Robert,” and he observed<br />
that the quadrupole logo of Fermilab should<br />
be called “the Cross of Batavia.” But Wilson<br />
Hall serves to remind the citizens of northern<br />
Illinois that science is ever present in their<br />
.lives, just as York Minster reassured the<br />
peasants of medieval Yorkshire that God<br />
was always nearby.<br />
The times we live in are much less optimistic<br />
than those when Wilson first made<br />
his comparison, and our resources are no<br />
longer as plentiful for our needs. But we may<br />
draw comfort from the search for a few nuggets<br />
of truth in an uncertain world.<br />
To gaze upfiom the ruins of the<br />
oppressive present towards thestars is<br />
to recognise the indestructible world of<br />
laws, tostrengthen faith in reason, to<br />
realise the “harmonia mundi” that<br />
transfuses all phenomena, and that<br />
never has been, nor will be, disturbed.<br />
Hermann Weyl, 1919<br />
S. Peter Rosen, a native of London, was educated at Merton College, Oxford, receiving from that<br />
institution a B.S. in mathematics in 1954 and both an M.A. and D.Phil. in theoretical physics in 1957.<br />
Peter first came to this country as a Research Associate at Washington University and then worked as<br />
Scientist for the Midwestern Universities Research Association at Madison, Wisconsin. A NATO<br />
Fellowship took him to the Clarendon Laboratory at Oxford in 1961. He then returned to the United<br />
States to Purdue University, where he retains a professorship in physics. Peter has served as Senior<br />
Theoretical Physicist for the U.S. Energy Research and Development Administration’s High Energy<br />
<strong>Physics</strong> Program, as Program Associate for Theoretical <strong>Physics</strong> with the National Science Foundation,<br />
and as Chairman of the U.S. Department of Energy’s Technical Assessment Panel for Proton<br />
Decay and on the Governing Board for the Lewes Center for <strong>Physics</strong> in Delaware. His association<br />
with the Laboratory extends back to 1977 when he came as Visiting Staff Member. He has served as<br />
Consultant with the Theoretical Division and as a member of the Program Advisory Committee and<br />
Chairman of the Neutrino Subcommittee of LAMPF. He is currently Associate Division Leader for<br />
Nuclear and <strong>Particle</strong> <strong>Physics</strong> of the Theoretical Division. Peter’s research specialties are symmetries<br />
of elementary particles and the theory of weak interactions.
Addendum<br />
The Next Step in Energy<br />
T<br />
wo years have passed since this article<br />
was written, and the high-energy<br />
physics commuinity is now poised to<br />
take the next major step forward in energy.<br />
Between now and 1990 a progression of new<br />
accelerators (see table) will raise the centerof-mass<br />
energy of proton-antiproton collisions<br />
to 2 TeV and that of electron-positron<br />
collisions to 100 GeV-more than enough to<br />
produce Wf and Zo bosons in large quantities.<br />
In addition, the HERA accelerator at<br />
DESY will enable us to collide an electron<br />
beam with a proton beam, producing over<br />
300 (3eV in the center of mass. Thus over the<br />
next five years we can look forward to a<br />
wealth of new data and much new physics.<br />
If the past is any guide, we can anticipate<br />
many surprises and discoveries of new<br />
phenomena as the energy of accelerators<br />
marches upward. But even if we are surprised<br />
by a lack of surprises, there is still<br />
much important physics to be explored in<br />
this new domain. As explained in the article<br />
by Raby, Slansky, and West, we have a "standard<br />
model" of particle physics that does a<br />
beautiful job of describing all known<br />
phenomena but has the unsatisfactory feature<br />
of requiring far too many arbitrary<br />
parameters to be put in by hand. It is therefore<br />
important to look for physics beyond the<br />
standard model, the discovery of which<br />
could lead us to a more general, more highly<br />
unified model with far fewer arbitrary<br />
parameters.<br />
One avenue for searching beyond the standard<br />
model is the precise measurement of<br />
properties of the particles it includes. For<br />
example, the model provides a well-defined<br />
relationship between the masses of the W'<br />
arid Zo bosons on the one hand and the weal<br />
neutral-current mixing angle 8 on the other.<br />
The theoretical corrections to this relationship<br />
due to virtual quantum mechanical<br />
processes (the so-called radiative corrections)<br />
can be reliably calculated in perturbation<br />
theory. To achieve the experimental<br />
precision of 1 percent required to test these<br />
calculations, we must observe many<br />
thousands of events.<br />
Because of their higher energies, both the<br />
Tevatron and SLC are expected to be much<br />
more copious sources of electroweak bosons<br />
than the Spps facility at which they were<br />
discovered. Whereas the SppS has produced<br />
approximately 200 Zo - e'e- and 2000 W'<br />
-t e*v events in a period of three to four<br />
years, the design luminosity of the Tevatron<br />
is such that it should yield 1500 Zo -+ p'pand<br />
15,000 Wf - P'vp in a good year. Even<br />
more impressive is SLC. which will produce<br />
3,000,000 Zo bosons per year at its design<br />
luminosity! So we look forward to precise<br />
determinations of the properties of the Wf<br />
bosons from the Tevatron and of the Zo<br />
boson from SLC.<br />
Another hndamental test for the standard<br />
model is the existence of a neutral scalar<br />
boson, a component of the Higgs boson<br />
multiplet responsible for generating the<br />
masses of the W* and Zo gauge bosons.<br />
While theory imposes a lower limit of a few<br />
GeV on the mass of this Higgs particle, it
~ shall<br />
Addendum<br />
gives no firm prediction for its magnitude,<br />
nor even an upper bound, and so we have to<br />
conduct a systematic search over a wide<br />
band of energies. Should the mass of the<br />
Higgs particle be less than that of the Zo,<br />
then we have a good chance of finding it at<br />
SLC through such processes as Zo - Hoy,<br />
Hoe+e-, and Hop+p-. If the Higgs particle is<br />
more massive than the Zo, then it may show<br />
up at the Tevatron through such decays as<br />
Ho - Zoy and Zop+p-. Should it prove to<br />
be beyond the range of the Tevatron, then we<br />
have to wait until the SSC comes on line<br />
in the mid 1990s.<br />
Another avenue for exploration beyond<br />
the standard model is the search for particles<br />
it does not include. For example, are there<br />
more than three families of fermions?<br />
Precisions studies of the width for Zo -<br />
XpVpvp will enable us to count the number of<br />
neutrino species, while apparently<br />
anomalous decays of the W+ will enable us<br />
to detect new charged leptons, provided of<br />
course that the lepton mass is less than that<br />
of the W+. In a similar way, decays of W*<br />
into jets of hadrons may reveal the existence<br />
of new heavy quarks, including the top quark<br />
required to fill out the existing three families<br />
of elementary fermions. A hint of the top<br />
quark was found by the UAl detector at the<br />
SppS, but there were too few events for it to<br />
be convincing. The much higher event rates<br />
of the new accelerators will be extremely<br />
useful in these searches.<br />
Besides the Higgs scalar boson and further<br />
replications of the known fermion families,<br />
there are hosts of new particles predicted by<br />
theories that unify the strong and electroweak<br />
interactions with one another and<br />
with gravity. Some of these ideas are discussed<br />
in “Toward a Unified Theory” and<br />
“Supersymmetry at 100 GeV.” Perhaps the<br />
most prevalent of such ideas is that of supersymmetry,<br />
which predicts that for every particle<br />
there exists a sypersymmetric partner,<br />
or sparticle, differing in spin by ‘12 and obeying<br />
opposite statistics. Thus for each fermion<br />
there exists a scalar boson, an s fermion, and<br />
for each gauge boson there exists a fermion<br />
called a gaugino. Some of these sparticles<br />
could be suficiently light, between a few<br />
GeV and a few tens of GeV, to be in the mass<br />
range that can be explored with the new<br />
accelerators.<br />
There may also be new interactions that<br />
appear only at the higher mass scales open to<br />
the new accelerators. Right-handed currents<br />
are absent from known low-energy weak interactions,<br />
possibly because the corresponding<br />
gauge bosons are much heavier than the<br />
W* and Zo of the standard model. Depending<br />
upon their masses, these gauge bosons<br />
could be produced at LEP I1 (an energy<br />
upgrade of LEP) or at the Tevatron; the tails<br />
of their propagators might even show up at<br />
SLC and at LEP itself.<br />
Another way of searching for new interactions<br />
is provided by the electron-proton collider<br />
HEM, located at DESY in Hamburg.<br />
There collision of 820-GeV protons with 30-<br />
GeV electrons provides 314 GeV in the center<br />
of mass and momentum transfers as large<br />
as los GeV’. Furthermore the electrons are<br />
naturally polarized perpendicular to the<br />
plane of the ring around which they rotate,<br />
and this polarization can easily be converted<br />
to a longitudinal direction, left-handed or<br />
right-handed. We know how the left-handed<br />
state will interact through the known W*<br />
and 2’; with the right-handed state we can<br />
see if a new type of weak interaction comes<br />
into play at these high energies.<br />
As the energy of accelerators increases, so<br />
the resulting collisions become less like interactions<br />
of the complicated hadronic structures<br />
that make up protons and antiprotons<br />
and more like collisions between the elementary<br />
constituents of these structures, namely<br />
quarks and gluons. In electron-positron collisions<br />
we begin with what we believe are two<br />
elementary and point-like objects. (Think of<br />
them as the ideal mass points of mechanics<br />
rather than the heavy balls found upon the<br />
billiard table.) The hadrons produced in lowenergy<br />
collisions tend to emerge over large<br />
angles, but, as the incident energy increases,<br />
so does the tendency for the hadrons to<br />
become highly correlated in a small number<br />
of directions. These correlated groupings of<br />
hadrons are called jets, and we believe that<br />
they signify, as closely as is physically<br />
possible, the production of quarks and<br />
gluons. These elementary objects cannot<br />
emerge from the collision regions as free<br />
particles because of the confinement<br />
properties and the color neutrality of the<br />
strong force of quantum chromodynamics.<br />
Instead they “hadronize” into jets of highly<br />
correlated groupings of color-neutral<br />
hadrons. In practice gluon jets, which were<br />
first discovered at the Doris accelerator in<br />
Hamburg, are slightly fatter than quark jets,<br />
which were originally found at PEP.<br />
From such a point of view, the Tevatron<br />
becomes a quarkquark, quark-gluon, and<br />
gluon-gluon collider, while HERA is an electron-quark<br />
collider. The reduction to<br />
elementary fermions and bosons enables us<br />
to interpret events much more simply than<br />
might otherwise be possible, but it does reduce<br />
the effective energy available for collisions.<br />
At the SppS, for example, gluons take<br />
up half of the energy of the proton, and so<br />
each quark has, on the average, one-sixth of<br />
the energy of the parent proton. At the SSC<br />
the fraction will be somewhat lower. We can<br />
therefore anticipate that in the decades to<br />
come there will be a strong impetus to push<br />
available energies well beyond that of the<br />
ssc.<br />
Theoretical motivation for the continued<br />
thrust toward higher energies may come<br />
from the notion of compositeness. Fifty<br />
years ago the electron and proton were<br />
thought to be elementary objects, but we<br />
know today that the proton is far from<br />
elementary. It is possible that in the coming<br />
period of experimentation we will discover<br />
that electrons also are not elementary, but<br />
are made up of other, more fundamental<br />
entities. Indeed there are theories in which<br />
leptons and quarks are all composite objects,<br />
made from things called rishons, or preons.<br />
Should this be the case, we will need energies<br />
much higher than that of the SSC to explore<br />
their properties. The lesson is very simple: at<br />
whatever energy scale we may be located,<br />
there is always much more to learn. Today’s<br />
elementary particles may be tomorrow’s<br />
atoms.<br />
157
LAMPF HI and the<br />
High-Intensity Frontier<br />
by Henry A. Thiessen<br />
os group has spent the past two years plan-<br />
Why Do We Need LAMPF<br />
igun: 1 shows a layout of the proposed facility.<br />
future, the possibility of family-changing interactions can<br />
800-MeV H- Injection Line<br />
45-GeV Main Ring<br />
1<br />
Fig. 1. LAMPF 11, the proposed addition to LAMPe, is<br />
designed toproduceprotons beams with a maximum energy<br />
of 45 GeV and a maximum current of 200 microamperes.<br />
These proton beams will provide intense beams of aidprotons,<br />
kaons, muons, and neutrinos for use in experiments<br />
important to both particle and nuclearphysics. The addition<br />
consists of two synchrotrons, both located 20 meters below<br />
the existing LAMPF linac. The booster is a 9-GeV, 60-<br />
hertz, 200-microampere machine fed by UMPF, and the<br />
main ring is a 45-GeV, 6-hertz, 40-microampere m<br />
Proton beams will be delivered to the main experime<br />
area of LAMPF (Area A) and to an area for experiments<br />
with neutrino beams and short, pulsed beams of other<br />
secondary particles (Area C). A new area for experiments<br />
with high-energy secondary beams (Area H) will be constructed<br />
to make full use of the 45-GeVproton beam.
the march toward higher energies<br />
will include the search for quark effects with the Drell-Yan process,<br />
the production of quark-gluon plasma by annihilation of antiprotons<br />
in nuclei, the extraction of nuclear properties from hypernuclei, and<br />
low-energy tests of quantum chromodynamics.<br />
1.2<br />
th<br />
0 0.2 0.4 0.6 0.8<br />
ig. 2. The “EMC effect” was first observed in data on the<br />
:uttering of muons from deuterium and iron nuclei at high<br />
tomentum transfer. The ratio of the two nucleon structure<br />
rnctions (FF(Fe) and FF(D)) deduced from these data by<br />
?garding a nucleus as simply a collection of nucleons is<br />
kown above as a function of x, aparameter representing the<br />
*action of the momentum carried by the nucleon struck in<br />
be collision. The observed variation of the ratio from unity<br />
I quite contrary to expectations; it can be interpreted as a<br />
tanifetation of the quark substructure of the nucleons<br />
)?thin a nucleus. (Adapted from J. J. Aubert et al. (The<br />
‘uropean Muon Collaboration), <strong>Physics</strong> Letters<br />
23B(1983):175.)<br />
estigated only with high-intensity beams of kaons and muons. And<br />
udies of neutrino masses and neutrino-electron scattering, which<br />
re among the most important tests of possible extensions of the<br />
andard model, demand high-intensity beams of neutrinos to commate<br />
for the notorious infrequency of their interactions.<br />
Here I take the opportunity to discuss some of the experiments in<br />
&ear physics that can be addressed at LAMPF 11. The examples<br />
X<br />
Quark Effects, A major problem facing today’s generation of nuclear<br />
physicists is to develop a model of the nucleus in terms of its<br />
fundamental constituents-quarks and gluons. In terms of nucleons<br />
the venerable nuclear shell model has been as successful at interpreting<br />
nuclear phenomena as its analogue, the atomic shell model, has<br />
been at interpreting Ithe structure and chemistry of atoms. But<br />
nucleons are known to be made of quarks and gluons and thus must<br />
possess some additional internal degrees of freedom. Can we see<br />
some of the effects of these additional degrees of freedom? And then<br />
can we use these observations to construct a theory of nuclei based on<br />
quarks and gluons?<br />
Defining an experiment to answer the first question is difficult for<br />
two reasons. First, we know from the success of the shell model that<br />
nucleons dominate the observable properties of nuclei, and when this<br />
model fails, the facts can still be explained in terms of the exchange of<br />
pions or other mesons between the nucleons. Second, the current<br />
theory of quarks and gluons (quantum chromodynamics, or QCD) is<br />
simple only in the limit of extremely high energy and extremely high<br />
momentum transfer, the domain of “asymptotic QCD.” But the<br />
world of nuclear physics is very far from that domain. Thus, theoretical<br />
guidance from the more complicated domain of low-energy QCD<br />
is sparse.<br />
To date no phenomenon has been observed that can be interpreted<br />
unambiguously as an effect of the quark-gluon substructure of<br />
nucleons. However, the results of an experiment at CERN by the<br />
“European Muon Collaboration”’ are a good candidate for a quark<br />
effect, although other explanations are possible. This group determined<br />
the nuclear structure functions for iron and deuterium from<br />
data on the inelastic scattering of muons at high momentum transfers.<br />
(A nuclear structure function is a multiplicative correction to the<br />
Mott cross section; it is indicative of the momentum distribution of<br />
the quarks within the nucleus.) From these structure functions they<br />
then inferred values for the nucleon structure function by assuming<br />
that the nucleus is simply a collection of nucleons. (If this assumption<br />
were true, the inferred nucleon structure function would not vary<br />
from nucleus to nucleus.) Their results (Fig. 2) imply that an iron<br />
nucleus contains more high-momentum quarks and fewer lowmomentum<br />
quarks than does deuterium. This was quite unexpected<br />
but was quickly corroborated by a re-analysis2 of some ten-year-old<br />
electron-scattering data from SLAC and has now been confirmed in<br />
great detail by several new experiment^.^.^ The facts are clear, but<br />
how are they to be interpreted?<br />
The larger number of low-momentum quarks in iron than in<br />
deuterium may mean that the quarks in iron are shanng their<br />
momenta, perhaps with other quarks through formation of, say, six-<br />
159
ocess anses<br />
r (Fig. 3). Since valence and sea<br />
es make different contributions<br />
ry beams of pions, kaons,<br />
the possible explanations<br />
help solve these problems by providing large numbers of events<br />
e valencequark composition of a neutron is udd, an<br />
fact, the A plays a role in studies of the nuclear environment
~<br />
i 8<br />
! 2 13-<br />
er,<br />
K<br />
.-<br />
t<br />
t<br />
"<br />
I<br />
I<br />
;<br />
1<br />
I<br />
i<br />
I<br />
i<br />
8-<br />
and minimum disruption to the ongoing experimental programs at<br />
LAMPF. The designs of both of the new synchrotons reflect these<br />
goals.<br />
The booster, or first stage, will be fed by the world's best H-<br />
injector, LAMPF. This booster will provide a 9-GeV, 200-microamperr:<br />
beam ofprotons at 60 hertz. The 200-microampere current is<br />
the maximum consistent with continued use of the 800-MeV<br />
LAMPF beam by the Weapons Neutron Research Facility and the<br />
Protort Storage Ring. The 9-GeV energy is ideal not only for i<br />
into the second stage but also for production of neutrinos to be used<br />
in scattering experiments (Fig. 5). Eighty percent of the booster<br />
I<br />
I<br />
30 40<br />
Proton Momentum (GeV/c)<br />
I<br />
I<br />
all six dimensions than the injection requirements of LAMPF IT,<br />
lossless injection at a correct phase space is straightforward.<br />
The 45-GeV main ring is shaped like a racetrack for two reasons: it<br />
fits nicely on the long, narrow mesa site and it provides the<br />
straight sections necessary for efficient slow extraction. The<br />
solid curve is S<br />
based on various<br />
rate. The scatterz<br />
compromise minimizes the initial cost yet preserves<br />
kaons and antiprotons with energies up to 25 GeV. Such high<br />
energies should prove especially useful for the experiments<br />
tioned above on the Drell-Yan process and exclusive hadron interactions.<br />
The booster has a second operating mode: 12 GeV at 30 he<br />
100 microamperes with a duty factor of 30 percent. This 12-GeV<br />
mode will be useful for producing<br />
ring is delayed for financial reasons.<br />
The most difficult technical problem posed by LAMPF II is the if<br />
system, which must provide up to 10 megavolts at a peak power of 10<br />
megawatts and be tunable from 50 to 60 megahertz. Furthermore,<br />
tunxng must be rapid; that is, the band<br />
be on the order of 30 kilohertz. The ferrite-tuned rf systems used ia<br />
the past are typically capable of providing only 5 to 10 kilovolts pier<br />
gap at up to 50 kilowatts and, in addition, are limited by power<br />
dissipation in the ferrite tuners and plagued by strong, uncontrollable<br />
nonlinear effects. We have chosen to concentrate the modest development<br />
funds available at present on the rf system. A teststand is<br />
being built, and various ferrites are being studied to gain a beti.er<br />
understanding of their behavior.<br />
Cavity Frequency (MHz)<br />
applied in a buncher cavity developed by the Laboratory's Ac-<br />
perpendicular bias is a reduction in the ferrite losses by as much as<br />
similar results.<br />
162
the march toward higher energies<br />
two orders of magnitude (Fig. 6). Since the loss in the ferrite is<br />
proportional to the square of the voltage on each gap, reducing these<br />
losses is essential to achieving the performance required of the<br />
LAMPF 11 system.<br />
A collaboration led by R. Carlini and including the Medium<br />
Energy and Accelerator Technology divisions and the University of<br />
Colorado has made a number of tests of the perpendicular bias idea.<br />
Their results indicate that in certain ferrites the low losses persist at<br />
power levels greater than that needed for the LAMPF I1 cavities. A<br />
full-scale cavity is now being constructed to demonstrate that 100<br />
kilovolts per gap at 300 kilowatts is possible. This prototype will also<br />
help us make a choice of femte based on both rf performance and<br />
cost of the bias system. A full-scale, full-power prototype of the rf<br />
system is less than a year away.<br />
Conclusion<br />
This presentation of interesting experiments that could be camed<br />
out at LAMPF I1 is of necessity incomplete. In fact,’ the range of<br />
possibilities offered by LAMPF I1 is greater than that offered by any<br />
other facility being considered by the nuclear science community. Its<br />
funding would yield an extraordinary return.<br />
References<br />
I. J. J. Aubertfit al. (The European Muon Collaboration). “The Ratio of the Nucleon Structure<br />
Functions F2 for Iron and Deuterium.” <strong>Physics</strong> Letters 123B( 1983):275.<br />
2. A. Bodek et al. “Electron Scattering from Nuclear Targets and Quark Distributions in Nuclei.”<br />
Physical Revrew Letters 50( 1983): I43 1.<br />
3. A. Bodek et al. “Comparison of the Deephelastic Structure Functions of Deuterium and<br />
Aluminum Nuclei.” Physicd Review Letters 51( 1983):534.<br />
4. R. G. Arnold et al. “Measurements of the A Dependence of Deephelastic Electron Scattering<br />
from Nuclei.” Physical Review Letters 52(1984):727.<br />
5. T. A. Carey, K. W. Jones, J. B. McClelland, J. M. Moss, L. B. Rees, N. Tanaka, and A. D. Bacher.<br />
“Inclusive Scattering of 500-MeV Protons and Pionic Enhancement of the Nuclear Sea-Quark<br />
Distribution.” Physical Revrew Letters 53( 1984):144.<br />
6. I. R. Kenyon. “The Drell-Yan Process.” Reports on Progress in <strong>Physics</strong> 45( 1982):126l.<br />
7. D. Strottman and W. R. Gibbs. “High Nuclear Temperatures by Antimatter-Matter Annihilation.”<br />
Accepted for publication in <strong>Physics</strong> Letters.<br />
8. W. Briickner et al. “Spin-Orbit Interaction of Lambda <strong>Particle</strong>s in Nuclei.” <strong>Physics</strong> Letters<br />
79B( 1978): 157.<br />
9. I$, BeFini et al. “A Full Set of Nuclear Shell Orbitals for the A <strong>Particle</strong> Observed in i2S and<br />
A Ca.<br />
IO. A. Bouyssy. “Strangeness Exchange Reactions and Hypernuclear Spectroscopy.” <strong>Physics</strong> htters<br />
84B( 1979):41.<br />
11. M. May et ai. “Observation of Levels in fC, ;*N, and ,fO Hypernuclei.” Physical Review<br />
Letters 47( 1981): 1 106.<br />
12. A review of these experiments was presented by G. Bunce at the Conference on the Intersections<br />
between <strong>Particle</strong> and Nuclear <strong>Physics</strong>, Steamboat Springs, Colorado, May 23-30, 1984.and will<br />
be published in the conference proceedings.<br />
163
he march toward higher energies<br />
Features of the three SSC designs considered in the Reference Designs<br />
Study. The 6.5-tesla design involves a conductor-dominated field with<br />
both beam tubes in a common cold-iron yoke that contributes slightly to<br />
shaping the field. In this design the dipole magnet, beam tubes, and<br />
yoke are supported within a single cryostat. The 5-tesla design involves<br />
a conductor-dominated dipole field with a heavy-walled iron cryostat to<br />
attenuate the fringe field. This single-bore design requires two separate<br />
rings of dipole magnets. The 3-tesla design is similar to the 6.5 tesla<br />
design except that the field is shaped predominantly by the cold-iron<br />
yoke rather than by the conductor.<br />
Dipole<br />
Total<br />
Dipole Magnet Accelerator Estimated<br />
Field Length Diameter cost<br />
(TI (ft) (mi) 6)<br />
6.5 51<br />
5 46<br />
3 460<br />
. thermal contraction of the components<br />
ithin the cryostats must be accommodated.<br />
eat leaks from power and instrumentation<br />
ads must be minimized, as must those from<br />
e magnet supports. (What is needed are<br />
ipports with the strength of an ox yoke but<br />
e substance of a spider web.) Alignment<br />
ill require some means for knowing the<br />
,act location of the magnets within their<br />
yostats. And ifa leak should develop in any<br />
. the piping within a magnet’s cryostat,<br />
ere needs to be a method for locating the<br />
ick” magnet and determing where within it<br />
e problem exists.<br />
Questions of safety, also, must be adessed.<br />
For example, the refrigerator locains<br />
every 2 to 5 miles around the ring are<br />
gical sites for personnel acccss, but is this<br />
ten enough? What happens if a helium line<br />
ould rupture? After all, a person can run<br />
ly a few feet breathing helium. Will it be<br />
.-<br />
18 2.72 billion<br />
23 3.05 billion<br />
33 2.70 billion<br />
necessary to exclude personnel from the tunnel<br />
when the system is cold, or can this<br />
problem be solved with, say, supplied-air<br />
suits or vehicles?<br />
Achieving head-on collisions of the beams<br />
presents further challenges. Each beam must<br />
be focused down to 10 microns and, more<br />
taxing, be positioned to within an accuracy of<br />
about 1 micron. It takes a reasonably good<br />
microscope even to see something that small!<br />
Will a truck rumbling by shake the beams out<br />
of a collision course? What will be the effect<br />
of earth tides or earthquakes? Does the<br />
ground heave due to annual changes in temperature<br />
or water-table level? How stable is<br />
the ground in the first place? That is, does<br />
part of the accelerator move relative to the<br />
remainder? Will it be desirable, or necessary,<br />
to have a robot system constantly moving<br />
around the ringrweaking the positions ofthe<br />
magnets? What would the robot, or any<br />
surveyor, use as a reference for alignment?<br />
These are but a few of the many issues that<br />
have been raised about construction and<br />
operation of the SSC. Resolving them will<br />
require considerable technology and ingenuity.<br />
In April of this year, the Department of<br />
Energy assigned authority over the SSC effort<br />
to Universities Research Association<br />
(URA), the consortium of fifty-four universities<br />
that runs Fermilab. URA, in turn, assigned<br />
management responsibilities to a<br />
separate board of overseers under Boyce<br />
McDaniel of Cornell University. This board<br />
selected Maury Tigner as director and<br />
Stanley Wojcicki of Stanford University as<br />
deputy director for SSC research and development.<br />
A headquarters is being established<br />
at Lawrence Berkeley National Laboratory,<br />
and a team will be drawn together to define<br />
what the SSC must do and how best that can<br />
be done. Secretary of Energy Donald Hodel<br />
has approved the release of funds to support<br />
the first year of research and development.<br />
Since the $20 million provided was about<br />
half the amount felt necessary for progress at<br />
the desired rate, shortcuts must be taken in<br />
reaching a decision on magnet type so that<br />
site selection can begin soon.<br />
<strong>Los</strong> Mamos has been involved in the efforts<br />
on the SSC since the beginning. We<br />
have participated in numerous workshops,<br />
collated siting information and published a<br />
Site Atlas, and contributed to the portions of<br />
the Reference Designs Study on beam<br />
dynamics and the injector. We may be called<br />
upon to provide the injector linac, kicker<br />
magnets, accelcrating cavities, and numerous<br />
other accelerator components. Our research<br />
on magnetic refrigeration has the potential<br />
of halving the operating cost of the<br />
cryogenic system for the SSC. Although the<br />
results of this research may be too late to be<br />
incorporated in the initial design, magnetic<br />
refrigerator replacements for conventional<br />
units would quickly repay the investment.<br />
165
UNDERGROUND<br />
R<br />
emarkable though it may seem, some of our most direct<br />
information about processes involving energies far<br />
beyond those. available at any conceivable particle accelerator<br />
and far beyond those ever observed in cosmic'<br />
rays may come from patiently watching a large quantity of water,<br />
located deep underground, for indications of improbable behavior of<br />
3<br />
its constituents. Equally remarkable, our most direct informatio<br />
about the energy-producing processes deep in the cores of stars come<br />
not from telescopes or satellites but from carefblly sifting a larg<br />
volume of cleaning fluid, again located deep underground, for indic<br />
tions of rare interactions with messengers from the sun. In wi<br />
follows we will explore some of the science behind these statemen<br />
learn a bit about how such experiments are Carried out, and vent1<br />
into what the future may hold.<br />
The experiments that we will discuss, which can be characteriz<br />
as searches for exceedingly rare processes, have two features<br />
common: they are carried out deep below the surface ofthu earth, a<br />
they involve a large mass of material capable of undergoing<br />
participating in the rare process in question. The latter feature an!<br />
from the desire to increase the probability of observing the proci<br />
within a reasonable length oftime. The underground sire is necess;<br />
to shield the exDeriment from secondary cosmic rays. These produ<br />
of the interactions of primary cosmic rays within our atmosphi<br />
would create an overwhelming background of confusing, mislead!<br />
"noise." Since about 75 percent of the secondary cosmic rays I<br />
extremely penetrating muons (resulting from the decays of pions a<br />
kaons), effective shielding requires overburdens on the order o<br />
kilometer or so ofsolid rock (Fig. I).<br />
What arc the goals ofthe experiments that make worthwhile thi<br />
journeys into the hazardous depths of mines and tunnels w<br />
complex, sensitive equipment? The largest and in many ways<br />
ost spectacular experiments-the searches for decay of protons
the search for rare<br />
neutrons-are aimed at understanding the basic interactions of<br />
nature. The oldest seeks to verify the postulated mechanism of stellar<br />
energy production by detecting solar neutrinos-the lone truthful<br />
witnesses to the nuclear reactions in our star’s core. Smaller experiments<br />
investigate double beta decay, the rarest process yet observed<br />
in nature, to elucidate properties of the neutrino. Muon “telescopes”<br />
will observe the ndmbers, energies, and directions of cosmic-ray<br />
muons to obtain information about the composition and energy<br />
spectra ofprimary cosmic rays. Large neutrino detectors will measure<br />
the upward and downward flux of neutrinos through the earth and<br />
hence search for neutrino oscillations with the diameter of the earth<br />
sa baseline. These detectors can also serve as monitors for signals of<br />
are galactic events, such as the intense burst of neutrinos that is<br />
xpected to accompany the gravitational collapse of a stellar core.<br />
A site that can accommodate the increasingly sophisticated techiology<br />
required will encourage the mounting of underground experinents<br />
to probe these and other processes in ever greater detail.<br />
The Search for Nucleon Instability<br />
The universe is thought to be about ten billion (IO’O) years old, and<br />
if this unimaginable span of time, the life of mankind has occupied<br />
iut a tiny fraction. The lifetime of the universe, while immense on<br />
he scale of the lifetime of the human species, which is itself huge on<br />
he scale of our own lives, is totally insignificant when compared to<br />
he time scale on which matter is known to be stable. It is now certain<br />
hat protons and (bound) neutrons have lifetimes on the order of IO3’<br />
ears or more. Thus for all practical purposes these particles are<br />
otally stable. Why examine the issue any further?<br />
The incentive is one of principle. The mass of a proton or neutron,<br />
bout 940 MeV/c2, is considerably greater than that of many other<br />
iarticles: the photon (zero mass), the neutrinos (very small, perhaps<br />
ero mass), the electron (0.511 MeV/c’), the muon (IO6 MeV/c2), and<br />
the charged and neutral pions (140 MeV/cL and 135 MeV/c2), to<br />
name only the most familiar. Therefore, energy conservation alone<br />
does not preclude the possibility of nucleon decay. Bearing in mind<br />
Murray Gell-Mann’s famous dictum that “Everything not compulsory<br />
is forbidden,” we are obligated to search for nucleon decay<br />
unless we know ofsomething that forbids it.<br />
Conservation laws forbidding nucleon decay had been independently<br />
postulated by Weyl in 1929, Stueckelberg in 1938, and<br />
Wiper in 1949 and 1952. But Lee and Yang argued in 1955 that such<br />
laws would imply the existence of a long-range force coupled to a<br />
conserved quantum number known as baryon number. (The baryon<br />
number of a particle is the sum of the baryon numbers of its quark<br />
constituents, +% for each quark and -% for each antiquark. The<br />
proton and the neutron thus have baryon numbers of +I.) Lee and<br />
Yang’s reasoning followed the lines that lead to the derivation of the<br />
Coulomb force from the law of conservation of electric charge.<br />
However, no such long-range force is observed, or, more accurately,<br />
the strength of such a force, if it exists, must be many orders of<br />
magnitude weaker than that ofthe weakest force known, the gravitational<br />
force. Thus, although no information was available as to just<br />
how unstable nucleons might be, no theoretical argument demanded<br />
exact conservation of baryon number.<br />
<strong>Los</strong> <strong>Alamos</strong> has the distinction of being the site of the first searches<br />
for evidence of nucleon decay. In 1954 F. Reines, C. Cowan, and M.<br />
Goldhaber placed a scintillation detector in an underground room at<br />
a depth ofabout 100 feet and set a lower limit on the nucleon lifetime<br />
of lo2’ years. In 1957 Reines, Cowan, and H. Kruse deduced a greater<br />
limit of 4 X years from an improved version of the experiment<br />
located at a depth of about 200 feet (in “the icehouse,” an area<br />
excavated in the north wall of <strong>Los</strong> <strong>Alamos</strong> Canyon). Since these early<br />
<strong>Los</strong> <strong>Alamos</strong> experiments, the limit on the lifetime of the proton has
Nonconservation of baryon number is<br />
also favored as an explanation for a difficulty<br />
with the big-bang theory of creation of the<br />
universe. The difficulty is that the big bang<br />
supposedly created baryons and antibaryons<br />
in equal numbers, whereas today we observe<br />
a dramatic excess of matter over antimatter<br />
(and an equally dramatic excess of photons<br />
over matter). In 1967 A. Sakharov pointed<br />
out that this asymmetry must be due to the<br />
occurrence of processes that do not conserve<br />
baryon number; his original argument has<br />
since been elaborated in terms of grand unified<br />
theories by several authors. The very<br />
existence of physicists engaged in searches<br />
for nucleon decay is mute testimony to the<br />
baryon asymmetry of the universe and, by<br />
inference, to the decay of nucleons at some<br />
level.<br />
The recent resurgence of interest in the<br />
stability of nucleons arises in part from the<br />
success of the unified theory of electromagnetic<br />
and weak interactions by Glashow,<br />
Salam, and Weinberg. This non-Abelian<br />
gauge theory, which is consistent with all<br />
available data and correctly predicts the existence<br />
and strength of the neutral-current<br />
weak interaction and the masses of the Zo<br />
and W’ gauge bosons, involves essentially<br />
only one parameter (apart from the masses of<br />
the elementary particles). The measured<br />
value of this parameter (the Weinberg angle)<br />
is given by sin2Bw = 0.22 k 0.01. The success<br />
of the electroweak model gave considerable<br />
legitimacy to the idea that gauge theories<br />
may be the key to unifying all the interactions<br />
of nature.<br />
The simplest gauge theory to be applied to<br />
unifying the electroweak and strong interactions<br />
(minimal SU(5)) gave rise to two exciting<br />
predictions. One, that sin2ew = 0.2 I 5,<br />
agreed dramatically with experiment, and<br />
the other, that the lifetime of the proton<br />
against decay into a positron and a neutral<br />
pion (the predicted dominant decay mode)<br />
lay between 1.6 X IO2* and 6.4 X IO3’ years,<br />
implied that experiments to detect nucleon<br />
decay were technically feasible.<br />
Experimentalists responded with a series<br />
of increasingly sensitive experiments to test<br />
168<br />
this prediction of grand unification. What<br />
approach is followed in these experiments?<br />
Out of the question is the direct production<br />
of the gauge bosons assumed to mediate the<br />
interactions that lead to nucleon decay. (This<br />
was the approach followed recently and successfully<br />
to test the electroweak theory.) The<br />
grand unified theory based on minimal<br />
SU(5) predicts that the masses of these bosons<br />
are on the order of IOl4 GeV/c2, in<br />
contrast to the approximately 102-GeV/c2<br />
masses of the electroweak bosons and many<br />
orders of magnitude greater than the masses<br />
of particles that can be produced by any<br />
existing or conceivable accelerator or by the<br />
highest energy cosmic ray. Thus, the only<br />
feasible approach is to observe a huge number<br />
of nucleons with the hope of catching a<br />
few of them in the quantum-mechanically<br />
possible but highly unlikely act of decay.<br />
The largest of these experiments (the IMB<br />
experiment) is that of a collaboration including<br />
the University of California, Irvine, the<br />
University of Michigan, and Brookhaven<br />
National Laboratory. In this experiment<br />
(Fig. 2) an array of 2048 photomultipliers<br />
views 7000 tons of water at a depth of 1570<br />
meters of water equivalent (mwe) in the<br />
Morton-Thiokol salt mine near Cleveland,<br />
Ohio. The water serves as both the source of<br />
(possibly) decadent nucleons and as the medium<br />
in which the signal of a decay is generated.<br />
The energy released by nucleon decay<br />
would produce a number ofcharged particles<br />
with so much energy that their speed in the<br />
water exceeds that of light in the water (about<br />
0.75c, where c is the speed of light in<br />
vacuum). These particles then emit cones of<br />
Cerenkov radiation at directions characteristic<br />
of their velocities. The photomultipliers<br />
arrayed on the periphery of the water detect<br />
this light as it nears the surfaces. From the<br />
arrival times of the light pulses and the patterns<br />
of their intersections with the planes of<br />
the photomultipliers, the directions of the<br />
parent charged particles can be inferred.<br />
Their energies can be estimated from the<br />
amount of light observed, in conjunction<br />
with calibration studies based on the vertical<br />
passage of muons through the detector. (The<br />
Fig. 1. For some experiments the only<br />
practical way to sufficiently reduce the<br />
background caused by cosmic-ray<br />
muons is to locate the experiments deep<br />
underground. Shown above is the number<br />
of cosmic-ray muons incident per<br />
year upon a cube 10 meters on an edg<br />
as a function of depth of burial. B<br />
convention depths of burial in rocks a<br />
various densities are normalized t<br />
meters of water equivalent (mwe). Th<br />
depths of some of the experiments dis<br />
cussed in the text are indicated.<br />
impressive sensitivity of such an experimer<br />
is well illustrated by the information that th<br />
light from a charged particle at a distance c<br />
10 meters in water is less than that on th<br />
earth from a photoflash on the moon.)<br />
This “water Cerenkov” detection schem<br />
was chosen in part for its simplicity, in pal<br />
for its relatively low cost, and in part for il
.---<br />
Science Underground<br />
\<br />
Fig. 2. Schematic view of the IMB nucleon-decay detector. A total of 2048 5-inch<br />
photomultipliers are arrayed about the periphery of 7000 tons of water contained<br />
within a plastic-lined excavation at a depth of 1570 mwe in a salt mine near<br />
Cleveland, Ohio. The photomultipliers monitor the water for pulses of Cerenkov<br />
radiation, some of which may signal the decay of aproton or a neutron. (From R.<br />
M. Bionta et al., “IMB Detector-The First 30 Days,” in Science Underground<br />
(<strong>Los</strong> <strong>Alamos</strong>, 1982) (American Institute of <strong>Physics</strong>, New York, 1982)).<br />
high efficiency at detecting the electrons that<br />
are the ultimate result of the p - e+ + no<br />
decay. (The neutral pion immediately decays<br />
to two photons, which produce showers of<br />
electrons in the water.) Note, however, that<br />
although this two-body decay is especially<br />
easy to detect because of the back-to-back<br />
orientation of the decay products, it must be<br />
distinguished, at the relatively shallow depth<br />
of the IMB experiment, among a background<br />
of about 2 X IOs muon-induced events per<br />
day. (The lower limit on the proton lifetime<br />
predicted by minimal SU(5) implies a maximum<br />
rate for p - e+no of several events per<br />
day.)<br />
Another experiment employing the water<br />
Cerenkov detection scheme is being carried<br />
out at a depth of 2700 mwe by a collaboration<br />
including the University ofTokyo, KEK<br />
(National Laboratory for High-Energy <strong>Physics</strong>),<br />
Niigata University, and the University<br />
of Tsukuba. The experiment is located under<br />
Mt. lkenayama in the deepest active mine in<br />
Japan, the Kamioka lead-zinc mine of the<br />
Mitsui Mining and Smelting Co. Although<br />
the mass of the water viewed in this experiment<br />
(3000 tons) is substantiallly less than<br />
that in the IMB experiment, its greater depth<br />
of burial results in lower background rates.<br />
More important, 1000 20-inch photomultipliers<br />
are deployed at Kamioka (Fig. 3), in<br />
contrast to the 2048 5-inch photomultipliers<br />
at IMB. As a result, a ten times greater fraction<br />
of the water surface at Kamioka is covered<br />
by photocathode material, and the lightcollection<br />
efficiency is greater by a factor of<br />
about 12. Thus the track detection and<br />
identification capabilities of the Kamioka<br />
experiment are considerably better.<br />
To date neither the IMB experiment nor<br />
the Kamioka experiment has seen any candidate<br />
for p - e+lro. These negative results<br />
yield a proton lifetime greater than 3 X lo3’<br />
years for this decay mode, well outside the<br />
range predicted by the grand unified theory<br />
based on minimal SU(5). Since this theory<br />
has a number of other deficiencies (it fails to<br />
predict the correct ratio for the masses of the<br />
light quarks and predicts a drastically incorrect<br />
ratio for the number of baryons and<br />
169
.<br />
-.<br />
photons produced by the big bang), it is<br />
therefore now thought to be the wrong unification<br />
model. Other models, at the current<br />
stage of their development, have too little<br />
predictive power to yield decay rates that can<br />
be uriambiguously confronted by experiment.<br />
The question of nucleon decay is now<br />
a purely experimental one, and theory awaits<br />
the guidance of present and future experiment!;.<br />
The cosmic rays that produce the interfering<br />
muons also produce copious quantities of<br />
neutrinos (from the decays of pions, kaons,<br />
and muons). No amount of rock can block<br />
these neutrinos, and some of them interact in<br />
the water, mimicking the effects of proton<br />
decay. Estimates of this background as a<br />
function of energy are based on calculations<br />
of the flux of cosmic-ray-induced neutrinos<br />
from the known flux of primary cosmic rays.<br />
Although these calculations enjoy reasonable<br />
confidence, no accurate experimental data<br />
are available as a check. Full analyses of the<br />
neutrino backgrounds in the proton-decay<br />
experiments will provide the first such verification.<br />
Whether new effects in neutrino astronomy<br />
will be discovered from the spectrum<br />
of neutrinos incident on the earth remains<br />
to be seen. Thus nucleon-decay experiments<br />
may open a new field, that of<br />
neutrino astronomy.<br />
The water Cerenkov experiments have detected<br />
several events that could possibly be<br />
interpreted as nucleon decays by modes<br />
other than e+no (Table I). It is also possible<br />
that these events are induced by neutrinos.<br />
Although their configurations are not easily<br />
explained on that basis, their total number is<br />
consistent with the rate expected from the<br />
calculated neutrino flux.<br />
A perusal of Table 1 shows that the IMB<br />
and Kamioka experiments yield different<br />
lifetime limits and do not see the same number<br />
of candidate events for the various decay<br />
modes. This is not surprising since the two<br />
also differ in aspects other than those already<br />
mentioned. The Kamioka experiment can<br />
more easily distinguish events with multiple<br />
tracks, such as p - p'q, which is immediately<br />
followed by decay of the q meson<br />
Fig. 3. Photograph of the Kamioka nucleon-decay detector under construction a? a<br />
depth of 2700 mwe in a lead-zinc mine about 300 kilometers west of Tokyo.<br />
Already instal/ed are the bottom layer of photomultipliers and two ranks of<br />
photomultipliers on the sides of the cylindrical volume. The wire guards around the<br />
photomultipliers protect the workers from occasional implosions. The upper ranks<br />
and top layer of photomultipliers were installed from rafts as the water level was<br />
increased. The detector contains a total of 1000 20-inch photomultipliers. (Photo<br />
courtesy of the Kamioka collaboration.)<br />
+<br />
p-+ VK<br />
p -+ vn+<br />
n - r+n-<br />
n-vK<br />
0<br />
& .I . Y.<br />
[Ol 8 .,' roi<br />
I<br />
170
Science Underground<br />
(99.75%) p + p ---+ d + e' + v, o - O.Q WV, 607 x la'/cm2 * s<br />
or<br />
(0.25%) p + p + e-- d + v, 1.44 MeV, 1.5 X la'/cd * s<br />
d+p+'He+y<br />
(86%) 3 ~ + e 3 ~ - e 2p + 4 ~ e<br />
(1 4%)<br />
or<br />
3He + 4He -+ 'Be<br />
+ y<br />
(99.89~0) 7Be + e- --* 7Li + v, 0.86 MeV, 43 X l$/cm2 * s<br />
7~i + p-+ z4tie + y<br />
(O.O.li%)<br />
or<br />
7Bi? + p -. 'B + y<br />
'6 -+ 'Be* + e' + v,<br />
'Be* - 24He<br />
0 - 14.0 MeV, 0.056 X 1a'/em2 ' S<br />
Fig. 4. The proton-proton chain postulated by the standard solar model as the<br />
principal mechanism of energyproduction in the sun. The net result of this series of<br />
nuclear reactions is the conversion of fourprotons into a helium-4 nucleus, and the<br />
energy released is carried off by photons, positrons, and neutrinos. Predicted<br />
branching ratios for competing reactions are fisted. Some of the reactions in this<br />
chain produce neutrinos; the energies of these particles and theirpredictedjluxes at<br />
the earth are listed at the right.<br />
by a number of modes. On the other hand,<br />
the IMB experiment has been in progress for<br />
a longer time and is thus more sensitive to<br />
decay modes with long lifetimes.<br />
The IMB collaboration has recently installed<br />
light-gathering devices around each<br />
photomultiplier and will soon double the<br />
number of tubes with the goal of increasing<br />
the lightcollection efficiency by a factor of<br />
about 6. At Kamioka accurate timing circuits<br />
are being installed on each photomultiplier<br />
to record the exact times of arrival of the<br />
light signals. As a result, more and better data<br />
can be expected from both experiments.<br />
What else does the future hold? The European<br />
F6jus collaboration (Aachen, Orsay,<br />
Palaiseau, Saclay, and Wuppertal) has completed<br />
construction of a 912-ton modular<br />
fine-grained tracking calorimeter. This detector<br />
is located at a depth of 4400 mwe in a<br />
3300-cubic-meter laboratory excavated near<br />
the middle of the Frkjus Tunnel connecting<br />
Modane, France and Bardonnecchia, Italy.<br />
Its 114 modules consist of 6-meter by 6-<br />
meter planes of Geiger and flash chambers<br />
interleaved with thin iron-plate absorbers.<br />
The detector can pinpoint particle tracks<br />
with a resolution on the order of 2 millimeters,<br />
a 250-fold greater resolution than<br />
that of the water Cerenkov detectors. Data<br />
about energy losses of the particles along<br />
their tracks distinguish electrons and muons.<br />
To date the Frbjus collaboration has observed<br />
no candidate proton-decay events.<br />
The Soudan I1 collaboration (Argonne National<br />
Laboratory, the University of Minnesota,<br />
Oxford University, Rutherford-Appleton<br />
Laboratory, and Tufts University) has<br />
excavated an 1 1,000cubic-meter laboratory<br />
at 2200 mwe in the Soudan iron mine in<br />
northern Minnesota and is now constructing<br />
an 1 100-ton dense fine-grained tracking calorimeter.<br />
The detector will contain 256 modules,<br />
each 1 meter by 1 meter by 2.5 meters,<br />
incorporating thin steel sheets and high-resolution<br />
drift tubes in hexagonal arrays. The<br />
spatial resolution of the detector will be<br />
about 3 millimeters. Information about the<br />
ionization deposited along the track lengths<br />
will provide excellent particle-identification<br />
capabilities. Completion of the detector is<br />
scheduled for 1988, but data collection will<br />
begin in 1987.<br />
Because the Frkjus and Soudan I1 detectors<br />
view relatively small numbers of<br />
nucleons (fewer than 6 X lo3'), they can<br />
record reasonable event rates only for those<br />
decay modes (if any) with lifetimes considerably<br />
less than 10'' years. On the other hand,<br />
they have good resolution for high-energy<br />
cosmic-ray muons, and this feature will be<br />
put to good use in experiments of astrophysical<br />
interest.<br />
Despite the hopes for these newer experiments,<br />
the IMB and Kamioka results to date<br />
imply that accurate investigation of most<br />
nucleon decay modes demands multikiloton<br />
detectors with very fine-grained resolution.<br />
These second-generation detectors will be<br />
multipurpose devices, sensitive to many<br />
other rare processes. Realistically, they can<br />
be operated to greatest advantage only in the<br />
environment of a dedicated facility capable<br />
of providing major technical support.<br />
The Solar Neutrino Mystery<br />
The light from the sun so dominates our<br />
existence that all human cultures have<br />
marveled at its life-giving powers and have<br />
concocted stories explaining its origins.<br />
Scientists are no different in this regard. How<br />
do we explain the almost certain fact that the<br />
sun has been radiating energy at essentially<br />
the present rate of about 4 X loz6 joules per<br />
second for some 4 to 5 billion years? Given a<br />
solar mass of 2 X lom kilograms, chemical<br />
means are wholly inadequate, by many orders<br />
of magnitude, to support this rate of<br />
energy production. And the gravitational<br />
171
energy released in contracting the sun to its<br />
present. radius of about 7 x 10’ kilometers<br />
could provide but a tiny fraction of the<br />
radiated energy. The only adequate source is<br />
the conversion of mass to energy by nuclear<br />
reactions.<br />
This answer has been known for a generation<br />
or two. Through the work of Hans Bethe<br />
and others in the 1930s and of many workers<br />
since, we have a satisfactory model for solar<br />
energy production based on the thermonuclear<br />
fusion of hydrogen, the most abundant<br />
element in the universe and in most stars.<br />
The product of this proton-proton chain<br />
(Fig. 4) is helium, but further nuclear reactions<br />
yield heavier and heavier elements.<br />
Detailed models of these processes are quite<br />
successful at explaining the observed abundances<br />
of the elements. Thus it is possible to<br />
say (with W. A. Fowler) that “you and your<br />
neighbor and I, each one of us and all of us,<br />
are truly and literally a little bit of stardust.”<br />
The successes of the standard solar model<br />
may:, however, give us misplaced confidence<br />
in its reality. It is all very well to study<br />
nuclear reactions and energy transport in the<br />
laboratory and to construct elaborate computational<br />
models that agree with what we<br />
observe of the exteriors of stars. But what is<br />
the direct evidence in support of our story of<br />
what goes on deep within the cores of stars?<br />
The difficulties presented by the demand<br />
for direct evidence are formidable, to say the<br />
least. Stars other than our sun are hopelessly<br />
distant, and even that star, although at least<br />
reasonably typical, cannot be said to lie conveniently<br />
at hand for the conduct of experiments.<br />
Moreover, the sun is optically so<br />
thick that photons require on the order of 10<br />
million years to struggle from the deep interior<br />
to the surface, and the innumerable<br />
interactions they undergo on the way erase<br />
any memory of conditions in the solar core.<br />
Thus, all conventional astronomical obseiivations<br />
of surface emissions provide no<br />
direct information about the stellar interior.<br />
The situation is not hopeless, however, for<br />
several of the nuclear reactions in the protonproton<br />
chain give rise to neutrinos. These<br />
pihcles interact so little with matter that<br />
Fig. 5. A view of the solar neutrino experiment located at a depth of 4850 feet in the<br />
Homestake gold mine. The steel tank contains 380,000 liters of perchloroethylene,<br />
which serves as a source of chlorine atoms that interact with neutrinos from the<br />
sun. Nearby is a small laboratory where the argon atoms produced are counted.<br />
(Photo courtesy of R. Davis and Brookhaven National Laboratory.)<br />
they provide true testimony to conditions in<br />
the solar core.<br />
The parameters incorporated in the standard<br />
solar model (such as nuclear cross sections,<br />
solar mass, radius, and luminosity,<br />
and elemental abundances, opacities (from<br />
the <strong>Los</strong> <strong>Alamos</strong> Astrophysical Opacity<br />
Library), and equations of state) are known<br />
with such confidence that a calculation of the<br />
solar neutrino spectrum is expected to be<br />
reasonably accurate. At the moment only<br />
one experiment in the world-that of Raymond<br />
Davis and his collaborators from<br />
Brookhaven National Laboratory-attempts<br />
to measure any portion of the solar neutrino<br />
flux for comparison with such a calculation.<br />
Located at a depth of 4400 mwe in the<br />
Homestake gold mine in Lead, South Dakota,<br />
this experiment (Fig. 5) detects solar<br />
neutrinos by counting the argon atoms from<br />
the reaction<br />
V,+~~C~+~’A~+B ,<br />
which is sensitive primarily to neutrinos<br />
from the beta decay of boron-8 (see Fig. 4).<br />
Since chlorine-37 occurs naturally at an<br />
abundance of about 25 percent, any compound<br />
containing a relatively large number<br />
of chlorine atoms per molecule and satisfying<br />
cost and safety criteria can serve as the<br />
target. The Davis experiment uses 380,000<br />
liters of perchloroethylene (CzCl4).<br />
You might well ask why this reaction occurs<br />
at a detectable rate. All the solar neutrinos<br />
incident on the tank of percholoroethylene<br />
have made the journey from the<br />
solar core to the earth and then through 4850<br />
feet of solid rock with essentially no interactions,<br />
and the neutrinos from the boron-8<br />
decay constitute but a small fraction of the<br />
total neutrino flux. What is the special feature<br />
that makes this experiment possible?<br />
Apart from the large number of target<br />
chlorine atoms, it is the existence of an excited<br />
state in argon-37 that leads to an excep<br />
tionally high cross section for capture by<br />
chlorine-37 of neutrinos with energies<br />
greater than about 6 MeV. Figure 4 shows<br />
that the only branch of the proton-proton<br />
chain producing neutrinos with such energies<br />
is the beta decay of boron-8. The standard<br />
solar model predicts a rate for the reaction<br />
of about 7 x per target atom per<br />
1’72
Science Underground<br />
Fig. 6. Yearly averages of the flux of boron-8 solar neutrinos,<br />
as measured by the Homestake experiment. The<br />
discrepancy between the experimental results and thepredictions<br />
of the standard solar model has not yet been explained.<br />
(From R. Davis, Jr., B. T. Cleveland, and J. K. Rowley,<br />
“Report on Solar Neutrino Experiments, ” in Intersections<br />
Between <strong>Particle</strong> and Nuclear <strong>Physics</strong> (Steamboat Springs,<br />
Colorado), New York American Institute of <strong>Physics</strong>, 1984.)<br />
second (7 solar neutrino units, or SNUs), nique has been verified by continual scrutiny<br />
which corresponds in the Davis experiment over more than meen years.<br />
to an expected argon-37 production rate of The Homestake experiment has provided<br />
about forty atoms per month.<br />
the scientific world with a long-standing<br />
It may seem utterly miraculous that such a mystery: its results are significantly and consmall<br />
number of argon-37 atoms can be de- sistently lower than the predictions of the<br />
tected in such a large volume of target mate- standard solar model (Fig. 6). So what’s<br />
rial, but the technique is simple. About every wrong?<br />
two months helium is bubbled through the The first possibility that immediately sugtank<br />
to sweep out any argon-37 that has been gests itself, that the Davis experiment con-<br />
Formed. The resulting sample is purified and tains some subtle mistake, cannot be<br />
:oncentrated by standard chemical tech- eliminated. But it must be dismissed as uniiques<br />
and is monitored for the 35-day decay likely because of the careful controls in-<br />
If argon-37 by electron capture. Great care is corporated in the experiment and because of<br />
&en to distinguish these events by pulse the years of independent scrutiny that the<br />
ieight, rise time, and half-life from various experiment has survived. The possibility<br />
)ackground-induced events. As part of the that the parameters employed in the calculatxovery<br />
technique argon-36 and -38 are in- tion might be in error has been repeatedly<br />
erted into the tank in gram quantities or less examined by careful investigators seeking to<br />
o monitor the recovery efficiency (about 95 explain the mystery (and thereby make reercent).<br />
An artifidly introduced sample of putations for themselves). However, no one<br />
00 argon-37 atoms has also been recovered has suggested corrections that are large<br />
uccessfully. Indeed, the validity of the tech- enough to explain the discrepancy.<br />
Another possibility is that the standard<br />
solar model is wrong. The reaction that gives<br />
rise to boron-8 is inhibited substantially by a<br />
Coulomb barrier and is thus extraordinarily<br />
sensitive to the calculated temperature at the<br />
center of the sun. A tiny change in this<br />
temperature or a small deviation from the<br />
standard-model value of the solarcore composition<br />
would be sufficient to change the<br />
rate of production of boron-8 and thus the<br />
neutrino flux to which the Davis experiment<br />
is primarily sensitive. Although many<br />
“nonstandard” solar models predict lower<br />
boron-8 neutrino fluxes, none of these are<br />
widely accepted. In general, the only experimentally<br />
testable distinction among the<br />
nonstandard models lies in their predictions<br />
of neutrino fluxes. A complete characterization<br />
of the solar neutrino spectrum is needed<br />
to provide quantitative constraints on the<br />
standard solar model of the future.<br />
The explanation of the solar neutrino<br />
puzzle quite possibly lies in the realm of<br />
173
particle physics rather than solar physics,<br />
nuclear physics, or chemistry. The results of<br />
the Hornestake experiment have generally<br />
been interpreted on the basis of conventional<br />
neutrino physics. It is, however, not known<br />
with certainty how many species of neutrinos<br />
exist, whether they are massless, or whether<br />
they are stable. New information about these<br />
issues could drastically influence the interpretation<br />
of solar neutrino experiments.<br />
For example, Bahcall and collaborators<br />
have pointed out that it is possible for a more<br />
massive neutrino species to decay into a less<br />
massive neutrino species and a scalar particle<br />
(such as a Goldstone boson arising from<br />
spontaneous breaking of the symmetry associated<br />
with lepton number conservation).<br />
If a neutrino species less massive than the<br />
electron neutrino exists and if the lifetime of<br />
the electron neutrino is such that those with<br />
an energy of 10 MeV have a mean life of 500<br />
seconds (the transit time to the earth), then<br />
lower-energy electron neutrinos would decay<br />
before reaching the earth. The resulting reduction<br />
in the solar neutrino flux could be<br />
sufficient to explain the Davis results. Note<br />
that this explanation for the solar neutrino<br />
puzzle, in direct contrast to explanations<br />
based on nonstandard solar models, involves<br />
a great reduction in the flux of essentially all<br />
but the boron-8 neutrinos.<br />
Several other explanations of the solar<br />
neutrino puzzle are also based on speculated<br />
features of neutrino physics. One of these,<br />
“oscillations” among the various neutrino<br />
species, is discussed in the next section.<br />
Future Solar Neutrino<br />
Experiments<br />
Among the nonstandard solar models alluded<br />
to above are some that allow long-term<br />
variations in the rate of energy production in<br />
the solar core. Such variations violate the<br />
constraint on steady-state solar models that<br />
hydrogen be burned in the core at a rate<br />
commensurate with the currently observed<br />
solar luminosity. To test the validity of these<br />
models, a <strong>Los</strong> <strong>Alamos</strong> group has devised an<br />
experiment for determining an average of the<br />
174<br />
solar neutrino flux over the past several<br />
million years.<br />
The experiment, likle Davis’s, is based on<br />
an inverse beta decay induced by boron-8<br />
solar neutrinos, namely,<br />
The molybdenum target atoms must be<br />
located at depths such that the cosmic-rayinduced<br />
background of technetium isotopes<br />
is low compared to the solar neutrino signal.<br />
This condidtion is satisfied by a molybdenite<br />
ore body 1 100 to 1500 meters below Red<br />
Mountain in Clear Creek County, Colorado.<br />
The ore is currently being mined by AMAX<br />
Inc. at a depth in excess of about 1150<br />
meters. The long half-lives of technetium-97<br />
and -98 (2.6 million and 4.2 million years,<br />
respectively) have permitted their accumulation<br />
to a level (calculated on the basis of the<br />
standard solar model) of about 10 million<br />
atoms each per 2000 metric tons of ore.<br />
Fortuitously, the initial large-scale concentration<br />
of the technetium (into a rheniumselenium-technetium<br />
sludge) occurs during<br />
operations involved in producing<br />
molybdenum trioxide from the raw ore. The<br />
<strong>Los</strong> <strong>Alamos</strong> group has developed chemical<br />
and mass-spectrographic techniques for<br />
isolating and counting the technetium atoms<br />
in the sludge. The first results from the experiment<br />
should be available in late 1987.<br />
Much more remains unknown about solar<br />
neutrinos. In particular, we completely lack<br />
information about the flux of neutrinos from<br />
other reactions in the proton-proton chain.<br />
According to the standard solar model, the<br />
preponderance of solar neutrinos arises from<br />
the first reaction in the chain, the thermonuclear<br />
fusion of two protons to form a deuteron.<br />
A thorough test of the solar model<br />
must include measurement of the neutrino<br />
flux from this reaction, the rate of which,<br />
although essentially independent of the details<br />
of the model (varying by at most a few<br />
percent), involves the basic assumption that<br />
hydrogen burning is the principal source of<br />
solar energy.<br />
The preferred reaction for investigating<br />
the initial fusion in the proton-proton chain<br />
is<br />
v, + 7’Ga --+ 7’Ge + e- ,<br />
which has a threshold of 233 keV, well below<br />
the maximum energy of the pp neutrinos.<br />
Calculations based on the standard solar<br />
model and the relevant nuclear cross sections<br />
predict a capture rate of about 110 SNU, of<br />
which about two-thirds is due to the pp reaction,<br />
about one-third to the electron-capture<br />
reaction of beryllium-7, and a very small<br />
fraction to the other neutrino-producing reactions.<br />
Several years ago members of the Homestake<br />
team, in collaboration with scientists<br />
from abroad, carried out a pilot experiment<br />
to assess a technique suggested for a solar<br />
neutrino experiment based on this reaction.<br />
Germanium-71 was introduced into a solution<br />
of over one ton of gallium (as GaC13) in<br />
hydrochloric acid. In such a solution<br />
germanium forms the volatile compound<br />
GeQ, which was swept from the tank with a<br />
gas purge. By fairly standard chemical techniques,<br />
a purified sample of GeH4 was<br />
prepared for monitoring the 1 l-day decay of<br />
germanium-7 1 by electron capture. The pilot<br />
experiment clearly demonstrated the feasibility<br />
of the technique.<br />
Why has the full-scale version of this important<br />
experiment not been done? The<br />
trouble, as usual, is money. The original<br />
estimates indicated that achieving an amp<br />
table accuracy in the measured neutrino flux<br />
would require about one neutrino capture<br />
per day, which corresponded to 45 tons of<br />
gallium as a target. Gallium is neither common<br />
nor easy to extract, and the cost ofb5<br />
tons was about $25,000,000, a sum that<br />
proved unavailable. Nor did the suggestion<br />
to “borrow” the required amount of gallium<br />
succeed (despite the fact that only one gallium<br />
atom per day was to be expended), and<br />
the collaboration disbanded.<br />
The chances of mounting a gallium experiment<br />
seem brighter today, however, since<br />
recent Monte Carlo simulations have shown<br />
that an accuracy of 10 percent in the
Science Underground<br />
measured neutrino flux is possible from a<br />
four-year experiment incorporating improved<br />
counting efficiencies and reduced<br />
background rates and involving only 30 tons<br />
of gallium.<br />
The European GALLEX collaboration<br />
(Heidelberg, Karlsruhe, Munich, Saclay,<br />
Paris, Nice, Milan, Rome, and Rehovot) has<br />
received approval to install a 3(rton gallium<br />
chloride experiment in the Gran Sasso Laboratory<br />
(this and other dedicated underground<br />
science facilities are described in the next<br />
section) and sufficient funding to acquire the<br />
gallium. The collaboration has achieved the<br />
low background levels required for monitoring<br />
the decay of germanium-71 and has the<br />
counting equipment in hand. Progress awaits<br />
acquisition of the gallium, which will take<br />
several years.<br />
The Institute for Nuclear Research of the<br />
Soviet Academy of Sciences has 60 tons of<br />
gallium available for an experiment, and a<br />
chamber has been prepared in the Baksan<br />
Laboratory. As planned, this experiment<br />
uses metallic gallium as the target rather than<br />
GaCl3. However, after a novel initial extraction<br />
of the germanium, the experiment is<br />
similar to the gallium chloride experiment.<br />
Pilot studies have demonstrated the chemical<br />
techniques necessary for separating the<br />
germanium from the gallium, and counters<br />
are being prepared. In November 1986 the<br />
Soviet group and scientists from <strong>Los</strong> <strong>Alamos</strong><br />
and the University of Pennsylvania agreed to<br />
collaborate on the experiment, which will<br />
begin in late 1987.<br />
The INR also plans to repeat the Davis<br />
experiment, increasing the target volume of<br />
perchloroethylene by a factor of 5. This will<br />
increase the signal proportionally.<br />
As mentioned above, a gallium experiment<br />
detects neutrinos from both proton<br />
fusion and beryllium-7 decay. To determine<br />
the individual rates of the two reactions requires<br />
a separate measurement of the neutrinos<br />
from the latter. A reaction that satisfies<br />
the criterion of being sensitive primarily<br />
to the beryllium-7 neutrinos is<br />
v, + 'lBr - "Kr + e-<br />
Results from this bromine experiment are<br />
important to an unambiguous test of the<br />
standard solar model.<br />
The chemical techniques needed for the<br />
bromine experiment are substantially identical<br />
to those employed in the chlorine-37<br />
experiment, and therefore the feasibility of<br />
this aspect of the experiment is assured.<br />
However, since krypton-81 has a half-life of<br />
200,000 years, counting a small number of<br />
atoms by radioactive-decay techniques is out<br />
of the question. Fortunately, another technique<br />
has recently been developed by G. S.<br />
Hurst and his colleagues at Oak Ridge National<br />
Laboratory. In barest outline the technique<br />
involves selective ionization of atoms<br />
of the desired element by laser pulses of the<br />
appropriate frequency. The ionized atoms<br />
can then readily be removed from the sample<br />
and directed into a mass spectrometer, where<br />
the desired isotope is counted. Repetitive<br />
application of the technique to increase the<br />
selection efficiency has been demonstrated.<br />
The standard solar model predicts that a<br />
few atoms of krypton-8 1 would be produced<br />
per day in a volume of bromine solution<br />
similar to that of the chlorine solution in the<br />
Davis experiment. This is a sufficient number<br />
for successful application of resonance<br />
ionization spectroscopy. However, two other<br />
problems must be addressed. Protons<br />
produced by muons, neutrons, and alpha<br />
particles may introduce a troublesome background<br />
via the "Br(p,n)"Kr reaction, and<br />
naturally occurring isotopes of krypton may<br />
leak into the tank of bromine solution and<br />
complicate the mass spectrometry. Davis,<br />
Hurst, and their collaborators have undertaken<br />
a complete assessment of the feasibility<br />
of the bromine-8 1 experiment.<br />
Other inverse beta decays have been suggested<br />
as bases for detecting solar neutrinos<br />
by radiochemical techniques. An experiment<br />
based on one such reaction,<br />
v,+~L~ -7Be+e- ,<br />
is being actively developed in the Soviet<br />
Union by the INR. According to the standard<br />
solar model, the observed rate of the<br />
reaction will be about 46 SNU.<br />
Particularly appealing is the inverse beta<br />
decay<br />
~,+~~~~n--.~~~~n'+e- ,<br />
which has an enormous predicted rate (700<br />
SNU according to the standard solar model)<br />
and is dominated by pp, ppe, and beryllium-7<br />
neutrinos. Moreover, the 3-microsecond<br />
half-life of the product, an excited state of<br />
tin-I 15, implies that the reaction could be<br />
the basis for real-time measurements of the<br />
solar neutrino flux. Unfortunately, indium-1<br />
15 is not completely stable, decaying<br />
by beta emission with a half-life of about 5 X<br />
lOI4 years. Electrons from the beta decay of<br />
indium-1 15 give rise to signals that can<br />
mimic the signature of its interaction with a<br />
solar neutrino (a prompt electron followed 3<br />
microseconds later by two coincident<br />
gamma rays). This background is difficult to<br />
overcome, and such an experiment has not<br />
yet been fully developed.<br />
As mentioned above, the source of the<br />
solar neutrino puzzle may lie not in imperfections<br />
of solar models but in our limited<br />
knowledge of neutrino physics. Neutrino oscillations,<br />
for example, could provide an explanation<br />
for the Davis results. This phenomenon<br />
is a predicted consequence of<br />
nonzero neutrino rest masses, and no theory<br />
compels an assignment of zero mass to these<br />
particles.<br />
If neutrinos are massive, the flavor<br />
eigenstates that participate in weak interactions<br />
need not be the same as the mass<br />
eigenstates that propagate in free space. The<br />
two types of eigenstates are related by a<br />
unitary matrix that mixes the various neutrino<br />
species. For the case of only two neutrino<br />
species, say electron and muon neutrinos,<br />
this relation is<br />
where v, and v,, and VI and v2 are flavor and<br />
mass eigenstates, respectively. According to<br />
175
the SchriMinger equation, the wave functions<br />
of vl and v2 acquire phase factors e8’1‘<br />
and e-iE2‘ as they propagate. Therefore a pure<br />
Ve state (created by, say, the beta decay of<br />
boron-8) evolves with time (“oscillates”)<br />
into a state with a nonzero vp component.<br />
The probability Pve that v, remains at time t<br />
is given by<br />
Pve= 1 - sin2 2q3 sin2[(E2 - El) 2/21 ,<br />
where El and E2 are the energies of v1 and v2.<br />
Thus P., differs from unity if and only if ml<br />
+ m2, since, in units such that the speed of<br />
light and Planck’s constant are unity, Ef=p2<br />
+ mf ForpB m2 > ml,<br />
rnt-rnt Am2 Am2<br />
E2-El<br />
--- x-<br />
2P 2p 2Ev ’<br />
and the characteristic oscillation length (the<br />
distance over which P., undergoes one cycle<br />
of its variation) is proportional to Ev/Am2.<br />
The failure of numerous experiments to<br />
detect neutrino oscillations in terrestrial neutrino<br />
sources places an upper limit on Am2 of<br />
about 0.02 (eV)2. (The precise limits are joint<br />
limits on Am2 and the mixing angle e.) However,<br />
if Am2 -c 0.02 (eV2 (as some theoretical<br />
considerations suggest), oscillations<br />
would be undetectable in most terrestrial<br />
experiments and would most profitably be<br />
sought in lowenergy neutrinos at large distances<br />
from the source (distances comparable<br />
to the oscillation length). Unlike the<br />
terrestrial oscillation experiments to date,<br />
experiments designed to characterize the<br />
solar neutrino spectrum could effectively<br />
search for oscillations in solar neutrinos and<br />
be capable of lowering the upper limit on<br />
Am2 to perhaps lo-’’ (eV)2.<br />
Vacuum oscillations consistent with the<br />
standard solar model and the Davis experiment<br />
would require a rather large value of<br />
the mixing angle 8. However, Wolfenstein,<br />
Mikhaev, and Smirnov have recently<br />
pointed out a feature of neutrino oscillations,<br />
namely, their amplification by matter, that<br />
could accommodate the Davis results even if<br />
e is small, since it would greatly increase the<br />
probability that an electron neutrino<br />
176<br />
produced in the highdensity core of the sun<br />
emerge as a muon neutrino. The amplification<br />
is due to scattering by electrons and is<br />
therefore dependent upon electron density.<br />
(Scattering changes the phase of the<br />
propagating neutrino; its effect can be viewed<br />
as a change in either the index of refraction of<br />
the matter for neutrinos or in the potential<br />
energy (that is, effective mass) of the neutrino.)<br />
Observation of matterenhanced oscillations<br />
should be possible for values of<br />
Am2 between lo4 and lo-* (eV)2, a range<br />
inaccessible to experiments on terrestrial<br />
neutrino sources.<br />
The importance of the solar neutrino<br />
puzzle and the exciting possibility that its<br />
solution may involve fundamental prop<br />
erties of neutrinos have led to a number of<br />
recent proposals for real-time flux measurements.<br />
The Japanese protondecay group,<br />
together with researchers From Caltech and<br />
the University of Pennsylvania, is improving<br />
the Kamioka detector to observe the most<br />
energetic of the boron-8 solar neutrinos. The<br />
signal detected will be the Cerenkov radiation<br />
emitted by electrons in the water that<br />
recoil from neutrino scattering, receiving on<br />
average about half the neutrino energy. If the<br />
goal of a 7-MeV threshold for the detector is<br />
achieved, about 1 ‘scattering event should be<br />
observed every 2 days (as predicted on the<br />
basis of the Davis flux measurements). The<br />
directionality of the signal relative to the sun<br />
wil help distinguish scattering events from<br />
the isotropic background. Similar real-time<br />
flux measurements will also be possible with<br />
several of the second-generation detectors<br />
being built or planned for the Gran Sasso<br />
Laboratory.<br />
The Sudbury Neutrino Observatory collaboration<br />
(Queen’s, Irvine, Oxford, NRCC,<br />
Chalk River, Guelph, Laurentian, Princeton,<br />
Carleton) has proposed installing a 1000-ton<br />
heavy-water Cerenkov detector in the Sudbury<br />
Facility fix real-time flux measurements<br />
of a different type. Here the source of<br />
the Cerenkov radiation will be the electrons<br />
produced in the inverse beta decay<br />
ve + d - p -I- p + e-. Since the energy<br />
imparted to the electron is Ev - 1.44 MeV<br />
and the hoped-for threshold of the detector is<br />
about 7 MeV, the experiment will provide<br />
data on the higher energy portion of the<br />
boron-8 spectrum. About 8 events per day<br />
are expected to be recorded. The detector will<br />
be sensitive also to proton decay and to<br />
events induced by neutrinos from astrophysical<br />
sources and by muon neutrinos.<br />
Dedicated Underground Science<br />
Facilities<br />
For at least two decades scientists with<br />
experiments demanding the enormous<br />
shielding from cosmic rays afforded by deep<br />
underground sites have been setting up their<br />
apparatus in working mines. We owe a great<br />
debt to the enlightened mine owners who<br />
have allowed this pursuit of knowledge to<br />
take place alongside their search for valuable<br />
minerals. However, as the experiments increase<br />
in complexity, the need for more sup<br />
portive, dedicated facilities becomes more<br />
obvious.<br />
One argument in favor of a dedicated facility<br />
is simple but compelling: the need to<br />
have access to the experimental area controlled<br />
not by the operation of a mine or a<br />
tunnel but by the schedules of the experiments<br />
themselves. Another is the need for<br />
technical support facilities adequate to experiments<br />
that will rival in complexity those<br />
mounted at major accelerators. And not to<br />
be ignored is the need for accommodations<br />
for the scientists and graduate students from<br />
many institutions who will participate in the<br />
experiments.<br />
What should such a facility be like? The<br />
entryway should be large, and the experimental<br />
area should include at least several<br />
rooms in which different experiments<br />
can be in progress simultaneously.<br />
Provisions for easy expansion, ideally not<br />
only at the principal depth but also at greater<br />
and lesser depths, should be available. Another<br />
aspect that must be carefully planned<br />
for is safety. The underground environment<br />
is intrinsically hostile, and in addition some<br />
experiments may, like the Homestake experiment,<br />
involve large quantities of
Science Underground<br />
4-<br />
To L'Aquila<br />
0 5<br />
Kilometers<br />
Fig. 7. The three large (-35,000-<br />
cubic-meter) experimental halls<br />
planned for the Gran Sass0 Laboratory<br />
are shown in afloorplan of the facility<br />
(top). Two of the three halls are now<br />
excavated. Also shown are the locations<br />
of the laboratory ofl the highway tunnel<br />
under the Gran Sass0 d'Italia and<br />
of the tunnel in central Italy.<br />
CORN0 GRAND€<br />
(2912 rnl-<br />
10<br />
materials that pose hazards in enclosed<br />
spaces. Materials being considered for the<br />
bromine experiment, for example, include<br />
dibromoethane, and other experiments being<br />
planned involve cryogenic materials under<br />
high pressure and toxic or inflammable<br />
materials. Excellent ventilation and gas-tight<br />
entries to some areas are obvious requirements.<br />
Such dreams of dedicated facilities for underground<br />
science are now being realized.<br />
Italy, for example, recognized the op<br />
portunity offered by the construction several<br />
years ago of a new highway tunnel in the<br />
Apennines and incorporated a major underground<br />
laboratory (Fig. 7) under the Gran<br />
Sasso d'Italia near L'Aquila, which is about<br />
80 kilometers east of Rome. This location<br />
offers an overburden of about 5000 mwe in<br />
rock of high strength and low background<br />
radioactivity. Two of three large rooms (each<br />
about 120 meters by 20 meters by 15 meters)<br />
have been completed. Support laboratories<br />
and offices are located above ground at the<br />
west end of the tunnel.<br />
Because of its size, depth, support facilities,<br />
and ready access by superhighway, the<br />
Gran Sasso Laboratory is unrivaled as a site<br />
for underground science. In the spring of<br />
1985, about a dozen new experiments were<br />
approved for installation. Among these are<br />
experiments on geophysics, gravity waves,<br />
and double beta decay; the GALLEX solar<br />
neutrino experiment; the large-area (1400-<br />
square-meter) MACRO detector, which can<br />
be used in studies of rare cosmic-ray<br />
phenomena, high-energy neutrino and<br />
gamma-ray astronomy, and searches for<br />
magnetic monopoles; and the 6500-ton<br />
liquid-argon ICARUS detector, which will<br />
have unprecedented sensitivity to neutrinos<br />
of solar and galactic origin, proton decay,<br />
high-energy muons, and many other rare<br />
phenomena. As an example of the<br />
capabilities of ICARUS, in one year of operation,<br />
it will detect, with an accuracy of 10<br />
percent, a flux of boron-8 neutrinos more<br />
than twenty times smaller than the Davis<br />
limit, far below that allowed by any nonstandard<br />
solar model.<br />
177
For more than ten years the Soviet Union<br />
has maintained an underground laboratory<br />
for cosmic-ray experiments in the Baksan<br />
River valley near Mt. Elbrus, the highest<br />
peak of the Caucasus Mountains. A 460-ton<br />
cosmic-ray telescope and a double beta decay<br />
experiment are in place at about 800 mwe.<br />
This laboratory is being greatly expanded<br />
(Fig. 8). The horizontal entry has been extended<br />
3.6 kilometers under Mt. Andyrchi.<br />
There a 60-meter by 5-meter laboratory has<br />
been constructed to accommodate the Soviet<br />
gallium solar neutrino experiment and other<br />
smaller experiments. Further excavations<br />
are in progress to extend the adit an additional<br />
700 meters and provide a large room<br />
for the 3000-ton chlorine solar neutrino experiment.<br />
On a more modest scale Canada has<br />
proposed creation of an underground laboratory<br />
within the extensive and very deep excavations<br />
of the INCO Creighton No. 9<br />
nickel mine near Sudbury, Ontario. The<br />
company has suggested available sites at<br />
about 2100 meters where rooms as large as<br />
20 meters in diameter can be constructed.<br />
Within the United States all underground<br />
experiments are in working or abandoned<br />
mines. None of these sites offers any prospect<br />
for expansion into a full-scale underground<br />
laboratory to rival Gran Sasso, Baksan,<br />
or even Sudbury. In 198 1 and 1982 <strong>Los</strong><br />
<strong>Alamos</strong> conducted a site survey and developed<br />
a detailed proposal to create a dedicated<br />
National Underground Science Facility<br />
at the Department of Energy’s Nevada<br />
Test Site. The proposal called for vertical<br />
entry by a l4foot shaft extending initially to<br />
3600 feet (approximately 2900 mwe) and<br />
optionally to 6000 feet, excavation of two<br />
large experimental chambers, and provision<br />
of surface laboratories and offices. The<br />
proposal was not funded, and there is no<br />
other plan to provide a dedicated site in the<br />
United States for the next generation of underground<br />
searches for rare events.<br />
Conclusion<br />
We have touched in detail upon only two<br />
of the fascinating experiments that drive<br />
Kilometers<br />
Fig. 8. The main experimental areas of the Baksan Laboratory are shown in a<br />
profile of Mt. .Andyrchi through the adit (top). Area A houses a large cosmic-ray<br />
telescope, area BI has been excavated for the gallium solar neutrino experiment,<br />
and area B2, when excavated, will house the 3000-ton chlorine solar neutrino<br />
experiment. Also shown is the location of the facility near Mt. Elbrus in the<br />
Kabardino- Balkarian Autonomous Soviet Socialist Republic.<br />
scientists deep underground. Such experiments<br />
are not new on the scene, but the large<br />
and sophisticated second-generation detectors<br />
being built open up a new era. These<br />
devices should not be regarded as apparatus<br />
for a single experiment but as facilities useful<br />
for a variety of observations. They may be<br />
able to monitor continuously the galaxy for<br />
rare neutrino-producing events or the sun for<br />
variations in neutrino flux and hence in<br />
energy production. The day may be approaching,<br />
as Alfred Mann is fond of saying,<br />
where we will be able, from underground<br />
laboratories, to take the sun’s temperature<br />
each morning to see how our nearest star is<br />
feeling.<br />
1178
Science Underground<br />
AUTHORS<br />
L. M. Simmons, Jr., is Associate Division Leader for Research in the Laboratory’s Theoretical<br />
Division and was until 1985 Program Manager for the proposed National Underground Science<br />
Facility. He received a B.A. in physics from Rice University in 1959, an M.S. from Louisiana State<br />
University in 1961, and, in 1965, a Ph.D. in theoretical physics from Cornel1 University, where he<br />
studied under Peter Carmthers. He did postdoctoral work in elemetary particle theory at the<br />
University of Minnesota and the University of Wisconsin before joining the University of Texas as<br />
Assistant Professor. In 1973 he left the University of New Hampshire, where he was Visiting<br />
Assistant Professor, to join the staff of the Laboratory’s Theoretical Division Office.There he worked<br />
closely with Carmthers, as Assistant and as Associate Division Leader, to develop the division as an<br />
outstanding basic research organization while continuing his own research in particle theory and the<br />
quantum theory of coherent states. He has been, since its inception, coeditor of the University of<br />
California’s “<strong>Los</strong> <strong>Alamos</strong> Series in Basic and Applied Sciences.” In 1979 he originated the idea for<br />
the Center for Nonlinear Studies and was instrumental in its establishment. In 1980 he took leave, as<br />
Visiting Professor of <strong>Physics</strong> at Washington University, to work on strongcoupling field theories<br />
and their large-order behavior, returning in 1981 as Deputy Associate Director for <strong>Physics</strong> and<br />
Mathematics. While in that position, he developed an interest in underground science and began<br />
work as leader of the NUSF project. He is President of the Aspen Center for <strong>Physics</strong> and has also<br />
served that organization as Trustee and Treasurer.<br />
Further Reading<br />
Michael Martin Nieto, W. C. Haxton, C. M. Hoffman, E. W. Kolb, V. D., Sandberg, and J. W. Toevs,<br />
editors. Science Underground (<strong>Los</strong> Alarnos, 1982). New York Amencan Institute of <strong>Physics</strong>, 1983.<br />
F. Reines. “Baryon Conservation: Early Interest to Current Concern.” In Proceedings of the 8th International<br />
Workshop on Weak Interactions and Neutrinos. A Morales, editor. Singapore: World Scientific, 1983.<br />
D. H. Perkins. “Proton Decay Experiments.” Annual Review of Nuclear and <strong>Particle</strong> Science 34( 1984): 1.<br />
R. Bionta et al. “The Search for Proton Decay.” In Intersections Between <strong>Particle</strong> and Nuclear <strong>Physics</strong><br />
(Steamboat Springs, 1984). New York American Institute of <strong>Physics</strong>, 1984.<br />
R. Davis, Jr., B. T. Cleveland, and J. K. Rowley, “Report on Solar Neutrino Experiments.” In Intersections<br />
Beiween <strong>Particle</strong> and Nuclear <strong>Physics</strong> (Steamboat Springs, 1984). New York American Institute of <strong>Physics</strong>,<br />
1984.<br />
J. N. Bahcall, W. F. Huebner, S. H. Lubow, P. D. Parker, and R. K. Ulrich. “Standard Solar Models and the<br />
Uncertainties in Predicted Capture Rates of Solar Neutrinos.” Reviews of Modern <strong>Physics</strong> 54( 1982):767.<br />
R. R. Shafp, Jr., R. G. Warren, P. L. Aamodt, and A. K. Mann. “Prelifninary Site Selection and Evaluation<br />
for a National Underground <strong>Physics</strong> Laboratory.” <strong>Los</strong> <strong>Alamos</strong> National Laboratory unclassified release<br />
LAUR-82-556.<br />
S. P. Rosen, L. M. Simmons, Jr., R. R. Sharp, Jr., and M. M. Nieto. “<strong>Los</strong> <strong>Alamos</strong> Proposal for a National<br />
Underground Science Facility. In ICOBAN 84: Proceedings of the International Conference on Baryon<br />
Nonconservation (Park City, January I984), D. Cline, editor. Madison, Wisconsin: University of Wisconsin,<br />
1984.<br />
R. E. Mischke, editor. Intersections Between <strong>Particle</strong> and Nuclear <strong>Physics</strong> (Steamboat Springs. 1984). New<br />
York American Institute of <strong>Physics</strong>, 1984.<br />
C. Castagnoli, editor. “First Symposium on Underground <strong>Physics</strong> (St.-Vincent, 1985):’ I1 Nuovo Cimento<br />
9C( 1986): 1 I 1-674.<br />
J. C. Vander Velde. “Experimental Status of Proton Decay.” In First Aspen Winter <strong>Physics</strong> Conference, M.<br />
Block, editor. Annals of the New York Academy of Sciences 461( 1986).<br />
M. L. Cherry, K. Lande, and W. A. Fowler, editors. Solar Neutrinos and Neutrino Astronomy (Homestake,<br />
1984) AIP Conference Proceedings No. 126. New York American Institute of <strong>Physics</strong>, 1984.<br />
G. A. Cowan and W. C. Haxton. “Solar Neutrino Production of Technetium-97 and Technetium-98.’’<br />
Science216(1982):SI.<br />
179
6<br />
hat could be worse<br />
than a bunch of<br />
physicists gathering<br />
in a corner at a<br />
cocktail party to discuss physics?’ asks Pete<br />
Carruthers. We at <strong>Los</strong> <strong>Alamos</strong> Science<br />
frankly didn’t know what could be<br />
worse. . .or better, for that matter. However<br />
we did find the idea of “a bunch of physicists<br />
gathering in a corner to discuss physics”<br />
quite intriguing. We felt we might gain some<br />
insight and, at the same time, provide them<br />
with an opportunity to say things that are<br />
never printed in technical journals. So we<br />
gathered together a small bunch of four, Pete<br />
Carruthers, Stuart Raby, Richard Slansky,<br />
and Geoffrey West, found them a corner in<br />
the home of physicist and neurobiologist<br />
George Zweig and turned them loose. We<br />
knew it would be informative; we didn’t<br />
know it would be this entertaining.<br />
WEST: Z have here a sort of “jiractalized”<br />
table of discussion, ihe,first topic being,<br />
“What isparticlephysr’cs, and what are its<br />
origins?” Perhaps the older gentlemen among<br />
us might want to answer that.<br />
CARRUTHERS: Everyone knows that older<br />
gentlemen don’t know what particle physics<br />
is.<br />
ZWEIG: <strong>Particle</strong> physics deals with the<br />
structure of matter. From the time people<br />
began wondering what everything was made<br />
of, whether it was particulate or continuous,<br />
from that time on we had particle physics.<br />
WEST: In that sense ofwondering about the<br />
nature of matter, particle physics started<br />
with the Greeks, if not observationally, at<br />
least philosophically.<br />
ZWEIG: I think one ofthe first experimental<br />
contributions to particle physics came<br />
around 1830 with Faraday’s electroplating<br />
experiments, where he showed that it would<br />
take certain quantities ofelectricity that were<br />
integral multiples ofeach other to plate a<br />
mole ofone element or another onto his<br />
electrodes.<br />
An even earlier contribution was Brown’s<br />
observation ofthe motion ofminute particles<br />
suspended in liquid. We now know the<br />
chaotic motion he observed was caused by<br />
the random collision ofthese particles with<br />
liquid molecules.<br />
RABY So Einstein’s study of Brownian motion<br />
is an instance of somebody doing particle<br />
physics?<br />
ZWEIG: Absolutely. There’s a remarkable<br />
description of Brown’s work by Darwin, who<br />
was a friend of his. It’s interesting that<br />
Darwin, incredible observer of nature<br />
though he was, didn’t recognize the chaotic<br />
nature of the movement under Brown’s<br />
microscope; instead, he assumed he was see-<br />
180
181
ing “the marvelous currents of protoplasm in<br />
some vegetable cell.” When he asked Brown<br />
what he was looking at, Brown said, “That is<br />
my little secret.”<br />
SLANSKY: Quite a bit before Brown, Newton<br />
explained the sharp shadows created by<br />
light as being due to its particulate nature.<br />
That’s really not the explanation from our<br />
present viewpoint, but it was based on what<br />
he saw.<br />
CARRUTHERS: Newton was only half<br />
wrong. Light, like everything else, does have<br />
its particulate aspect. Newton just didn’t<br />
have a way ofexplaining its wave-like<br />
behavior. That brings us to the critical concept<br />
offield, which Faraday put forward so<br />
clearly. You can speak ofparticulate structure,<br />
but when you bring in the field concept,<br />
you have a much richer, more subtle structure:<br />
fields are things that propagate like<br />
waves but materialize themselves in terms of<br />
quanta. And that is the current wisdom of<br />
what particle physics is, namely, quantized<br />
fields.<br />
Quantum field theory is the only conceptual<br />
framework that pieces together the concepts<br />
of special relativity and quantum theory,<br />
as well as the observed group structure<br />
of the elementary particle spectrum. All these<br />
things live in this framework, and there’s<br />
nothing to disprove its structure. Nature<br />
looks like a transformation process in the<br />
framework ofquantum field theory. Matter<br />
is not just pointy little particles; it involves<br />
the more ethereal substance that people<br />
sometimes call waves, which in this theory<br />
are subsumed into one unruly construct, the<br />
quantized field.<br />
ZWEIG: <strong>Particle</strong> physics wasn’t always<br />
quantized field theory. When I was a graduate<br />
student, a different philosophy governed:<br />
S-matrix theory and the bootstrap<br />
hypothesis.<br />
CARRUTHERS: That was a temporary<br />
aberration.<br />
ZWEIG: But a big aberration in our lives! S-<br />
matrix theory was not wrong, just largely<br />
irrelevant.<br />
RABY: If particle physics is the attempt to<br />
understand the basic building blocks of<br />
182<br />
nature, then it’s not a static thing. Atomic<br />
physics at one point was particle physics, but<br />
once you understood the atom, then you<br />
moved down a level to the nucleus, and so<br />
forth.<br />
WEST: Let’s bring it up to date, then. When<br />
would you say particle physics turned into<br />
high-energy physics?<br />
ZWEIG: With accelerators.<br />
SLANSKY: Well, it really began around<br />
19lOwith theuseofthecloudchambers to<br />
detect cosmic rays; that is how Anderson<br />
detected the positron in 1932. His discovery<br />
straightened out a basic concept in quantized<br />
field theory, namely, what the antiparticle is.<br />
CARRUTHERS: Yes, in 1926 Dirac had<br />
quantized the electromagnetic field and had<br />
given wave/particle duality a respectable<br />
mathematical framework. That framework<br />
predicted the positron because the electron<br />
had to have a positively charged partner.<br />
Actually, it was Oppenheimer who predicted<br />
the positron. Dirac wanted to interpret the<br />
positive solution of his equation as a proton,<br />
since there were spare protons sitting around<br />
in the world. To make this interpretation<br />
plausible, he had to invoke all that hankypanky<br />
about the negative energy sea being<br />
filled-you could imagine that something<br />
was screwy.<br />
SLANSKY: Say what you will, Dirac’s idea<br />
was a wonderful unification of all nature,<br />
much more wonderful than we can envisage<br />
today.<br />
ZWEIG: Ignorance is bliss.<br />
SLANSKY: There were two particles, the<br />
proton and the electron, and they were the<br />
basic structure ofall matter, and they were,<br />
in fact, manifestations ofthe same thing in<br />
field theory. We have nothing on the horizon<br />
that promises such a magnificent unification<br />
as that.<br />
RABY Weren’t the proton and the electron<br />
supposed to have: the same mass according to<br />
the equation?<br />
WEST: No, the negative energy sea was supposed<br />
to take care ofthat.<br />
CARRUTHERS: It was not unlike the present<br />
trick ofexplaining particle masses<br />
through spontaneous symmetry breaking.<br />
Dirac’s idea ofviewing the proton and the<br />
electron as two different charge states ofthe<br />
same object was a nice idea that satisfied all<br />
the desires for symmetries that lurk in the<br />
hearts oftheorists, but it was wrong. And the<br />
reason it was wrong, ofcourse, is that the<br />
proton is the wrong object to compare with<br />
the electron. It’s the quark and the electron<br />
that may turn out to be different states ofa<br />
single field, a hypothesis we call grand unification.<br />
WEST: Well, it is certainly true that highenergy<br />
particle physics now is cloaked in the<br />
language ofquantized field theory, so much<br />
so that we call these theories the standard<br />
model.<br />
CARRUTHERS: But I think we’re overlooking<br />
the critical role of Rutherford in inventing<br />
particle physics.<br />
WEST: The experiments ofalpha scattering<br />
on gold foils to discern the structure ofthe<br />
atom.<br />
ZWEIG: Rutherford established the<br />
paradigm we still use for probing the structure<br />
of matter: you just bounce one particle<br />
off another and see what happens.<br />
CARRUTHERS: In fact, particle physics is a<br />
continuing dialogue (not always friendly) between<br />
experimentalists and theorists. Sometimes<br />
theorists come up with something that<br />
is interesting but that experimentalists<br />
suspect is wrong, even though they will win a<br />
Nobel prize if they can find the thing. And<br />
what the experimentalists do discover is frequently<br />
rather different from what the<br />
theorists thought, which makes the theorists<br />
go back and work some more. This is the way<br />
the field grows. We make lots of mistakes,we<br />
build the wrong machines, committees decide<br />
to do the wrong experiments, and<br />
journals refuse to publish the right theories.<br />
The process only works because there are so<br />
many objective entrepreneurs in the world<br />
who are trying to find out how matter<br />
behaves under these rather extreme conditions.<br />
It is marvelous to have great synthetic<br />
minds like those of Newton and Oalileo, but<br />
they build not only on the work ofunnamed<br />
thousands of theorists but also on these<br />
countless experiments.
“To understand the universe that we feel and touch,<br />
even down to its minutiae, you don’t have to know a<br />
damn thing about quarks.”<br />
ROUND TABLE<br />
WEST: Perhaps we should tell how wepersonally<br />
got involved inphysics, what drives us,<br />
why we stay with it. Because it is an awfully<br />
difficult field and a very frustrating field.<br />
How do wefind the reality of it compared to<br />
our early romantic images? Let’s start with<br />
Pete, who’s been interviewed many times and<br />
should be in practice.<br />
CARRUTHERS: I was enormously interested<br />
in biology as a child, but I decided that<br />
it was too hard, too formless. So I thought I’d<br />
do something easy like physics. Our town<br />
library didn’t even have modern quantum<br />
mechanics books. But I read the old quantum<br />
mechanics, and I read Jeans and Eddington<br />
and other inspirational books filled<br />
with flowery prose. I was very excited about<br />
the mysteries of the atom. It was ten years<br />
before I realized that I had been tricked. I had<br />
imagined I would go out and learn about the<br />
absolute truth, but aftera little bit ofexperience<br />
I saw that the “absolute truth” of<br />
this year is replaced next year by something<br />
that may not even resemble it, leaving you<br />
with only some small residue of value.<br />
Eventually I came to feel that science, despite<br />
its experimental foundation and reference<br />
frame, shares much with other intellectual<br />
disciplines like music, art, and literature.<br />
WEST Dick, what about you?<br />
SLANSKY In college I listed myself as a<br />
physics major, but I gave my heart to<br />
philosophy and writing fiction. I had quite a<br />
hard time with them, too, but physics and<br />
mathematics remained easy. However, since<br />
I didn’t see physics as very deep, I decided<br />
after I graduated to look at other fields. I<br />
spent a year in the Harvard Divinity School,<br />
where I found myself inadvertently a<br />
spokesman for science. I took Ed Purcell’s<br />
quantum mechanics course in order to be<br />
able to answer people’s questions, and it was<br />
there that I found myself, for the first time,<br />
absolutely fascinated by physics.<br />
During that year I had been accepted at<br />
Berkeley as a graduate student in philosophy,<br />
but in May I asked them whether I could<br />
switch to physics. They wrote back saying it<br />
GEOFFREY B. WEST: “One of the great things that has happened in particle<br />
physics is that some of. . . the wonderful, deep questions. . . are being asked<br />
again . . . Somehow we have to understand why there is a weak scale, why there is<br />
an electromagnetic scale, why a strong scale, and ultimately why a grand scale.”<br />
would be fine. I don’t know that one can be<br />
such a dilettante these days.<br />
SCIENCE: Why were people in Divinity<br />
School asking about quantum mechanics?<br />
SLANSKY People hoped to gain some insight<br />
into the roles of theology and<br />
philosophy from the intellectual framework<br />
of science. In the past certain philosophical<br />
systems have been based on physical theories.<br />
People were wondering what had really<br />
happened with quantum mechanics, since no<br />
philosophical system had been built upon it.<br />
Efforts have been made, but none so successful<br />
as Kant’s with Newtonian physics, for<br />
example.<br />
CARRUTHERS: <strong>Particle</strong> physics doesn’t<br />
stand still for philosophy. The subject is such<br />
that as soon as you understand something,<br />
you move on. I think restlessness<br />
characterizes this particular branch of science,<br />
in fact.<br />
SLANSKY I never looked at science as<br />
something I wanted to learn that would be<br />
absolutely permanent for all the rest ofthe<br />
history of mankind. I simply enjoy the doing<br />
of the physics, and I enjoy cheering on other<br />
people who are doing it. It is the intellectual<br />
excitement of particle physics that draws me<br />
to it.<br />
ZWEIG: Dick, was there some connection,<br />
in your own mind, between religion and<br />
physics?<br />
183
“The realproblem was that you had a zoo ofparticles,<br />
with none seemingly more fundamental than any<br />
other.”<br />
SLANSKY: Some. One of the issues that<br />
concerned me was the referential<br />
mechanisms oftheological language. How<br />
we refer to things. In science we also have<br />
that concern, very much so.<br />
ZWEIC;: What do you mean by “How we<br />
refer to things”?<br />
SLANSKY When we use a word to refer to<br />
God or to refer to great generalizations in our<br />
experience, how does the word work to refer<br />
beyond the language? Language isjust a<br />
sound. How does the word refer beyond just<br />
the mere word to the total experience? I’ve<br />
never really solved that problem in my own<br />
mind.<br />
CARRUTHERS. When you mention the<br />
word God, isn’t there a pattern of signals in<br />
your mind that corresponds to the pattern of<br />
sound? Doesn’t God have a peculiar pattern?<br />
SLANSKY The referential mechanisms of<br />
theological language became a major concern<br />
around 1966, after I’d left Harvard Divinity.<br />
Before that the school was under the influerice<br />
of the two great theologians Paul<br />
Tillich and Reinhold Niebuhr. Their concern<br />
was with the eighteenth and nineteenth century<br />
efforts to put into some sort of theoretical<br />
or logical framework all of man and his<br />
nature. I found myselfswept up much more<br />
into theological and philosophical issues<br />
than into the study ofethics.<br />
ZWEIG: Do you think these issues lie in the<br />
domain of science now? Questions about<br />
what man is, what his role in nature is, and<br />
what nature itself is, are being framed and<br />
answered by biologists and physicists.<br />
SLANSKY: I don’t view what I am trying to<br />
do in particle physics as finding man’s place<br />
in nature. I think of it as a puzzle made ofa<br />
lot ofexperimental data, and we are trying to<br />
assemble the pieces.<br />
CARRUTHERS: But the attitudes are very<br />
theological, and often they tend to be<br />
dogmatic.<br />
SLANSKY: I would like to make a personal<br />
statement here. That is, when I go out for a<br />
walk in the mountains, enjoying the beauties<br />
of nature with a capital N, I don’t feel that<br />
that has any very direct relationship to formulating<br />
a theory of nature. While my per-<br />
184<br />
sonal experience may set my mind in motion,<br />
may provide some inspiration, I don’t<br />
feel that seeing the Truchas peaks or seeing<br />
wild flowers in the springtime is very closely<br />
related to my efforts tCJ build a theory.<br />
WEST: Along that line I have an apocryphal<br />
story about Hans and Rose Bethe. One summer’s<br />
evening when the stars were shining<br />
and the sky was spectacular, Rose was exclaiming<br />
over their beauty. Allegedly Hans<br />
replied, “Yes, but yoii know, I think I am the<br />
only man alive that knows why they shine.”<br />
There you have the difference between the<br />
romantic and the scientific views.<br />
RABY: <strong>Particle</strong> physics to me is a unique<br />
marriage ofphilosophy and reality. In high<br />
school I read the philosopher George<br />
Berkeley, who discusses space and time and<br />
tries to imagine what space would be like<br />
were there nothing in it. Could there be a<br />
force on a particle were there nothing else in<br />
space? Obviously a particle couldn’t move<br />
because it would have nothing to move with<br />
respect to. <strong>Particle</strong> physics has the beauty of<br />
philosophy constrained by the fact you are<br />
working with observable reality. For a science<br />
fair in high school I built a cloud<br />
chamber and tried to observe some alpha<br />
particles and beta particles. That’s the reality<br />
part: you can actually build an experiment<br />
and actually see some of these fundamental<br />
objects. And there are people who are<br />
brilliant enough, like Einstein, to relate ideas<br />
and thought to reality and then make predictions<br />
about how the world mu$t be. Special<br />
relativity and all the Gedanken experiments,<br />
which are basically philosophical, say how<br />
the world is. To me what particle physics<br />
means is that you can have an idea, based on<br />
some physical fact, that leads to some experimental<br />
prediction. That is beautiful, and<br />
I don’t know how you define beauty except<br />
to say that it’s in the eye ofthe beholder.<br />
ZWEIG How was science viewed in your<br />
family?<br />
RABY: No one understood science in my<br />
family.<br />
ZWEIG: Well, did they respect it even if they<br />
STUART A. RABY: “I think what particle physics means to me is this unique<br />
intermarriage of philosophy and reality. . . . <strong>Particle</strong> physics has the beauty of<br />
philosophy constrained by the fact that you are working with observable reality.<br />
. . . If you have a beautiful idea and it leads to a prediction that, in fact,<br />
comes true, that would be the most amazing thing. That you can understand<br />
something on such a fundamental level!”
“. b b b one thing that distinguishesphysics from<br />
philosophy is predictive powerb The quark model had a<br />
lot ofpredictive powerb”<br />
ROUND TABLE<br />
didn’t understand it?<br />
RABY: I guess they accepted the fact that I<br />
would pursue what interested me. I’m the<br />
first one in my family to finish college, and<br />
that in itself is something big to them. My<br />
grandfather, who does understand a little,<br />
has read about Einstein. My grandfather’s<br />
interest in science doesn’t come from any<br />
particular training, but from the fact that he<br />
is very inventive and intuitive and puts<br />
radios together and learns everything by<br />
himself.<br />
ZWEIG: Was he respected for it?<br />
RABY By whom? My grandfather owned a<br />
chicken market, so he did these things in his<br />
spare time.<br />
WEST: That’s interesting. I have to admit I<br />
am another person who got into physics in<br />
spite of himself. I was facile in mathematics<br />
but more keen on literature. I turned to<br />
natural sciences when I went to Cambridge<br />
only because I had begun reading Jeans and<br />
Eddington and all those early twentieth century<br />
visionaries. They were describing that<br />
wonderful time of the birth of quantum mechanics,<br />
the birth of relativity, the beginning<br />
of thinking about cosmology and the origin<br />
ofthe universe. Wonderful questions! Really<br />
important questions that dovetailed into the<br />
big questions raised by literature. What is it<br />
all really about, this mysterious universe?<br />
The other crucial reason that I went into<br />
science was that I could not stand the world<br />
ofbusiness, the world of the wheelerdealer,<br />
that whole materialistic world. Somehow I<br />
had an image of the scientist as removed<br />
from that, judged only by his work, his only<br />
criteria being proof, knowledge, and wisdom.<br />
I still hold that romantic image. And that has<br />
been my biggest disappointment, because, of<br />
course, science, like everything else that involves<br />
millions ofdollars, has its own<br />
wheeler-dealers and salesmen and all the rest<br />
of it.<br />
My undergraduate experience at Cambridge<br />
was something of a disaster in terms<br />
ofphysics education, and I was determined<br />
to leave the field. I had become very interested<br />
in West Coastjazzand managed to<br />
obtain a fellowship to Stanford where, for a<br />
- . *.<br />
RICHARD A. SLANSKY: “It is the intellectual excitement of particle physics<br />
that draws me to it, really . . . . Ifindparticlephysics an intriguing effort to try to<br />
explain and understand, in a very special way, what goes on in nature . . . . I enjoy<br />
the effort. . . . I enjoy cheering on otherpeople who are trying. . . . I think of it as<br />
a puzzle made of a lot of experimental data, and we are trying to assemble all the<br />
pieces. ”<br />
year, I could be near San Francisco, North<br />
Beach, and that whole scene. Although at<br />
first I hated Palo Alto, my physics courses<br />
were on so much more a professional level,<br />
so much more an exciting level, that my<br />
attitude eventually changed. Somehow the<br />
whole world opened up. But even in graduate<br />
school I would go back to reading Eddington,<br />
whether he were right or not, because his<br />
language and way ofthinking were inspirational,<br />
as ofcourse, were Einstein’s.<br />
CARRUTHERS: Do you think our visions<br />
have become muddied in these modern<br />
times?<br />
WEST: I don’t think so at all. One ofthe<br />
great things that has happened in particle<br />
physics is that some of the deep questions are<br />
being asked again. Not that I like the<br />
proposed answers, particularly, but the questions<br />
are being asked. George, what do you<br />
say to all this? You often have a different<br />
slant.<br />
ZWEIG: My parents came from eastern<br />
Europe-they fled just before the second<br />
World War. I was born in Moscow and came<br />
to this country when I was less than two years<br />
old. Most of my family perished in the war,<br />
probably in concentration camps. I learned<br />
185
“It is an old Jewish belief that ideas are what really<br />
matter. If you want to create things that will endure,<br />
you create them in the mind of man.”<br />
at a very early age from the example of my<br />
father, who was wise enough to see the situation<br />
in Germany for what it really was, that<br />
it is very important to understand teality.<br />
Reality is the bottom line. Science deals with<br />
reality, and psychology with our ability to<br />
accept it.<br />
I grew up in a rough, integrated<br />
neighborhood in Detroit. Much of it subsequently<br />
burned down in the 1967 riot. I<br />
hated school and at first did very poorly. I<br />
was placed in a “slower” non-college<br />
preparatory class and took a lot of shop<br />
courses. Although I did not like being viewed<br />
as a second class citizen, I thought that operating<br />
machines was a hell of a lot more<br />
interesting than discussing social relations<br />
with my classmates and teachers.<br />
Eventually I was able to do everything that<br />
was asked of me very quickly, but the teachers<br />
were not knowledgeable, and classes were<br />
boring. In order to get along I kept my mouth<br />
shut. Occasionally I acted as an expediter,<br />
asking questions to help my classmates.<br />
At that time science and magic were really<br />
one and the same in my mind, and what<br />
child isn’t fascinated by magic? At home I<br />
did all sorts of tinkering. I built rockets that<br />
flew and developed my own rocket fuels. The<br />
ultimate in magic was my tesla coil with a six<br />
foot corona emanating from a door knob.<br />
College was a revelation to me. I went to<br />
the University of Michigan and majored in<br />
mathematics. For the first time I met teachers<br />
who were smart. And then I went to<br />
Caltech, a place I had never even heard of six<br />
months before I arrived. At Caltech I was<br />
very fortunate to work with Alvin<br />
Tollestrup, an experimentalist who later designed<br />
the superconducting magnets that are<br />
used at Fermilab. And I was exposed to<br />
Feynman and Cell-Mann, who were unbelievable<br />
individuals in their own distinctive<br />
ways. That was an exciting time.<br />
Shelly Glashow was a postdoc. Ken Wilson,<br />
Hung Cheng, Roger Dashen, and Sidney<br />
Coleman were graduate students. Rudy<br />
Mossbauer was down the hall. He was still a<br />
research fellow one month before he got the<br />
Nobel prize. The board of trustees called a<br />
GEORGE ZWEIG: “I learned at a very early age. . .that reality is the bottom line.<br />
Science deals with reality, andpsychology with our ability to accept it.”<br />
crash meeting and promoted him to full<br />
professorjust before the announcement. I<br />
remember pleading with Dan Kevles in the<br />
history department to come over to the physics<br />
department and record the progress, because<br />
science history was in the making, but<br />
he wouldn’t budge. “You can never tell what<br />
is important until many years later,” he said.<br />
CARRUTHERS: I’ve forgotten whether I<br />
first met all ofyou at Cal Tech or at Aspen.<br />
ZWEIG: Wherever Pete met us, I know<br />
we’re all here because of him. He was always<br />
very gently asking me, “How about coming<br />
to <strong>Los</strong> <strong>Alamos</strong>?’ Eventually I took him up<br />
on his offer.<br />
WEST: Before we leave this more personal<br />
side of the interview, I want to ask a question<br />
or two about families. Is it true that<br />
physicists generally come from middle class<br />
and lower backgrounds? Dick, what about<br />
your family?<br />
SLANSKY My father came from a farming<br />
family. Since he weighed only ninety-seven<br />
pounds when he graduated from high school,<br />
farm work was a little heavy for him. He<br />
entered a local college and eventually earned<br />
a graduate degree from Berkeley as a physical<br />
chemist. My mother wanted to attend<br />
medical school and was admitted, but back<br />
in those days it was more important to have<br />
186
ROUND TABLE<br />
children. So I am the result rather than her<br />
becoming a doctor.<br />
CARRUTHERS: My father grew up on a<br />
farm in Indiana, was identified as a bright<br />
kid, and was sent off to Purdue, where he<br />
became an engineer. So I at least had somebody<br />
who believed in a technical world.<br />
However, when I finally became a professor<br />
at Cornell, my parents were a bit disappointed<br />
because in their experience only<br />
those who couldn’t make it in the business<br />
world became faculty members.<br />
WEST What about your parents, George?<br />
ZWEIG: Both my parents are intellectuals,<br />
people very much concerned with ideas. To<br />
me one ofthe virtues ofdoing science is that<br />
you contribute to the construction ofideas,<br />
which last in ways that material monuments<br />
don’t. It is an old Jewish belief that ideas are<br />
what really matter. If you want to create<br />
things that will endure, you create them in<br />
the mind of man.<br />
WEST What did your parents do?<br />
ZWEIG: My mother was a nursery school<br />
teacher. She studied in Vienna in the O OS, an<br />
exciting time. Montessori was there; Freud<br />
was there. My father was a structural engineer.<br />
He chose his profession for political<br />
reasons, because engineering was a useful<br />
thing to do.<br />
WEST Then all three ofyou have scientific<br />
or engineering backgrounds. My mother is a<br />
dressmaker, and my father was a professional<br />
gambler. But he was an intellectual<br />
in many ways, even though he left school at<br />
fifteen. He read profusely, knew everything<br />
superficially very well, and was brilliant in<br />
languages. He wasted his life gambling, but it<br />
was an interesting life. I think I became facile<br />
in mathematics at a young age just because<br />
he was so quick at working out odds, odds on<br />
dogs and horses, how to do triples and<br />
doubles, and so on.<br />
CARRUTHERS: Are we all firstborn sons? I<br />
think we are, and that’s an often quoted<br />
statistic about scientists.<br />
WEST Have we all retreated into science for<br />
solace?<br />
RABY: It’s more than that. At one time I felt<br />
divided between going into social work in<br />
PETER A. CARRUTHERS: “There’s no point in a full-blown essay on quantum<br />
field theory because it’s probably wrong anyway. That’s what fundamental science<br />
is all about-whateveryou’re doing is probably wrong. That’s how you know when<br />
you’re doing it. Once in a while you’re right, and then you’re a great man, or<br />
woman nowadays. I’ve tried to explain this before to people, but they’re very slow to<br />
understand. What you have to do is look back and find what has been filtered out as<br />
correct by experiments and a lot of subsequent restructuring. Right? But when<br />
you’re actually doing it, almost every time you’re wrong. Everybody thinks you sit<br />
on a mountaintop communing with Jung’s collective unconscious, right? Well you<br />
try, but the collective unconscious isn’t any smarter than you are.”<br />
187
“Why do the forces in nature have different<br />
strengths . That’s one of those wonderful deep questions<br />
that has come back to haunt us.”<br />
order to be involved with people or going<br />
into science and being involved with ideas. It<br />
was continually on my mind, and when I<br />
graduated from college, I took a year off to do<br />
social work. I worked in a youth house in the<br />
South Bronx as a counselor for kids between<br />
the ages of seven and seventeen. They were<br />
all there waiting to be sentenced, and they<br />
were very self-destructive kids. The best<br />
thing you could do was to show them that<br />
they should have goals and that they<br />
shouldn’t destroy themselves when the goals<br />
seemed out ofreach. For example, a typical<br />
goal was to get out of the place, and a typical<br />
reaction was to end up a suicide. I kept trying<br />
to tell these kids, “Do what you enjoy doing<br />
and set a goal for yourselfand try to fulfill<br />
that goal in positive ways.” In the end I was<br />
convinced by my own logic that I should<br />
return to physics.<br />
WEST Let’s discuss the way physics affects<br />
our personal lives now that we are grown<br />
men. Suppose you are at a cocktail party, and<br />
someone asks, “What do you do?’ “I am a<br />
physicist,” you say, “High-energy physics,”<br />
or “<strong>Particle</strong> physics.” Then there is a silence<br />
and it is very awkward. That is one response,<br />
and here is the other. “Oh, you do particle<br />
physics? My God, that’s exciting stuff! I read<br />
about quarks and couldn’t understand a<br />
word of it. But then I read this great book,<br />
The Tu0 oSphysics. Can you tell me what you<br />
do?’ I groan inwardly and sadly reflect on<br />
how great the communication gap is between<br />
scientists such as ourselves and the general<br />
public that supports us. We seem to have<br />
shirked our responsibility in communicating<br />
the fantastic ideas and concepts involved in<br />
our enterprise to the masses. It is a sobering<br />
thought that Capra’s book, which most of us<br />
don’t particularly like because it represents<br />
neither particle physics nor Zen accurately, is<br />
probably unique in turning on the layman to<br />
some aspects ofparticle physics. Whatever<br />
your views ofthat book may be, you’ve<br />
certainly got to appreciate what he’s done for<br />
the publicity ofthe field. As for me, I find it<br />
difficult to talk about this life that I love in<br />
two-line sentences.<br />
Now, the cocktail party is just a superficial<br />
188<br />
aspect of my social life, but the problem<br />
enters in a more crucial way in my relationship<br />
with my family, the people dear to<br />
me. Here is this work which I love, which I<br />
spend a majority of my time in doing, and<br />
from which a large number ofthe frustrations<br />
and disappointments and joys in my<br />
life come, and I cannot communicate it to<br />
my family except in an incredibly superficial<br />
way.<br />
SLANSKY: The cocktail party experiences<br />
that Geoffrey describes are absolutely<br />
perfect, and I know what he means about the<br />
family. Now that my children are older, they<br />
are into science, and sometimes they ask me<br />
questions at the dinner table. I try to give<br />
clear explanations, but I’m never sure I’ve<br />
succeeded even superficially. And my wife,<br />
who is very bright but has no science background,<br />
doesn’t hesitate to say that science in<br />
more than twenty-five words is boring.<br />
Sometimes, in fact, I feel that my doing<br />
physics is viewed by them as a hobby.<br />
CARRUTHERS: Socially, what could be<br />
worse than a bunch ofphysicists gathering in<br />
a corner at a cocktail party to discuss physics?<br />
RABY: I find there are two types ofpeople.<br />
There are people who ask you a question just<br />
to be polite and who don’t really want an<br />
answer. Those people you ignore. Then there<br />
are people who are genuinely interested, and<br />
you talk to them. Ifthey don’t understand<br />
what a quark is, you ask them if they understand<br />
what a proton or an electron is. If they<br />
don’t understand those, then you ask them if<br />
they know what an atom is. You describe an<br />
atom as electrons and a nucleus of protons<br />
and neutrons. You go down from there, and<br />
you eventually get to what you are studying-particle<br />
physics.<br />
WEST: Does particle physics affect your relationship<br />
with your wife?<br />
RABY: My wife is occasionally interested in<br />
all this. My son, however, is genuinely interested<br />
in all forms of physical phenomena and<br />
is constantly asking questions. He likes to<br />
hear about gravity, that the gravity that pulls<br />
objects to the earth also pulls the moon<br />
around the earth. I have to admit that I find<br />
his interest very rewarding.<br />
WEST: Maybe, since we’ve been given the<br />
opportunity today, we should start talking<br />
aboutphysics. <strong>Particle</strong>physics has gone<br />
through a minirevolution since the discovery<br />
of thepsi/Jparticle at SLACand at<br />
Brookhaven ten years ago. Although not important<br />
in itself; that discovery confirmed a<br />
whole way of thinking in terms of quarks,<br />
symmetry principles, gauge theories, and<br />
unification. It was a bolt out of the blue at a<br />
time when the direction ofparticlephysics was<br />
uncertain. From then on, it became clear that<br />
non-Abelian gauge theories and unification<br />
were going to form the fundamental<br />
principles for research. Sociologically, there<br />
developed a unanimity in thefield, a unanimity<br />
that has remained. This has led us to<br />
the standard model, which incorporates the<br />
strong, weak, and electromagnetic interactions.<br />
SLANSKY: Yes, the standard model is a<br />
marvelous synthesis of ideas that have been<br />
around for a long time. It derives all interactions<br />
from one elegant principle, the principle<br />
oflocal symmetry, which has its origin<br />
in the structure ofelectromagnetism. In the<br />
1950s Yang and Mills generalized this structure<br />
to the so-called non-Abelian gauge the-
ROUND TABLE<br />
ories and then through the '60s and '70s we<br />
learned enough about these field theories to<br />
feel confident describing all the forces of<br />
nature in terms ofthem.<br />
RABY: 1 think we feel confident with Yang-<br />
Mills theories because they are just a sophisticated<br />
version of our old concept of force.<br />
The idea is that all of matter is made up of<br />
quarks and leptons (electrons, muons, etc.)<br />
and that the forces or interactions between<br />
them arise from the exchange of special kinds<br />
ofparticles called gauge particles: the photon<br />
in electromagnetic interactions, the W' and<br />
Zo in the weak interactions responsible for<br />
radioactive decay, and the gluons in strong<br />
interactions that bind the nucleus. (It is believed<br />
that the graviton plays a similar role in<br />
gravity.) [The local gauge theory ofthe strong<br />
forces is called quantum chromodynamics<br />
(QCD). The local gauge theory that unifies<br />
the weak force and the electromagnetic force<br />
is the electroweak theory that predicted the<br />
existence ofthe W'and Zo.]<br />
ZWEIG: You are talking about a very limited<br />
aspect ofwhat high-energy physics has<br />
been. Our present understanding did not develop<br />
in an orderly manner. In fact, what<br />
took place in the early '60s was the first<br />
revolution we have had in physics since<br />
quantum mechanics. At that time, ifyou<br />
were at Berkeley studying physics, you<br />
studied S-matrix, not field theory, and when<br />
I went to Caltech, I was also taught that field<br />
theory was not important.<br />
SLANSKY Yes, the few people that were<br />
focusing on Yang-Mills theories in the '50s<br />
and early '60s were more or less ignored.<br />
Perhaps the most impressive ofthose early<br />
papers was one by Julian Schwinger in which<br />
he tried to use the isotopic spin group as a<br />
local symmetry group for the weak, not the<br />
strong, interactions. (Schwinger's approach<br />
turned out to be correct. The Nobel prizewinning<br />
SU(2) X U( 1) electroweak theory<br />
that predicted the W'and Zo vector mesons<br />
to mediate the weak interactions is an expanded<br />
version of Schwinger's SU(2)<br />
model.)<br />
WEST In retrospect Schwinger is a real hero<br />
in the sense that he kept the faith and made<br />
some remarkable discoveries in field theory<br />
during a period when everybody was<br />
basically giving him the finger. He was completely<br />
ignored and, in fact, felt left out ofthe<br />
field because no one would pay any attention.<br />
SCIENCE: Why was field theory dropped?<br />
CARRUTHERS: Now we reach a curious<br />
sociological phenomenon.<br />
RABY Sociological? I thought the theory<br />
wasjust too hard to understand. There were<br />
all those infinities that cropped up in the<br />
calculations and had to be renormalized<br />
away.<br />
CARRUTHERS I am afraid there is a phase<br />
transition that occurs in groups of people of<br />
whatever IQ who feverishly follow each new<br />
promising trend in science. They go to a<br />
conference, where a guru raises his hands up<br />
and waves his baton; everyone sits there,<br />
their heads going in unison, and the few<br />
heretics sitting out there are mostly intimidated<br />
into keeping their heresies to<br />
themselves. After a while some new religion<br />
comes along, and a new faith replaces the<br />
old. This is a curious thing, which you often<br />
see at football games and the like.<br />
WEST The myth perpetrated about field<br />
theory was, as Stuart said, that the problems<br />
were too hard. But if you look at Yang-Mills<br />
and Julian Schwinger's paper, for example,<br />
there was still serious work that could have<br />
been done. Instead, when I was at Stanford,<br />
Sidney Drell taught advanced quantum mechanics<br />
and gave a whole lecture on why you<br />
didn't need field theory. All you needed were<br />
Feynman graphs. That was the theory.<br />
RABY: The real problem was that you had a<br />
zoo of particles, with none seemingly more<br />
fundamental than any other. Before people<br />
knew about quarks, you didn't feel that you<br />
were writing down the fundamental fields.<br />
WEST In 1954 we had all the machinery<br />
necessary to write down the standard model.<br />
We had the renormalization group. We had<br />
local gauge theories.<br />
SLANSKY But nobody knew what to apply<br />
them to.<br />
SCIENCE: George, in 1963, when you came<br />
up with the idea that quarks were the constit-<br />
uents of the strongly interacting particles, did<br />
you think at all about field theory?<br />
ZWEIG: No. The history I remember is<br />
quite different. The physics community<br />
responded to this proliferation ofparticles by<br />
embracing the bootstrap hypothesis. No particle<br />
was viewed as fundamental; instead,<br />
there was a nuclear democracy in which all<br />
particles were made out ofone another. The<br />
idea had its origins in Heisenberg's S-matrix<br />
theory. Heisenberg published a paper in 1943<br />
reiterating the philosophy that underlies<br />
quantum mechanics, namely, that you<br />
should only deal with observables. In the<br />
case ofquantum mechanics, you deal with<br />
spectral lines, the frequencies of light emitted<br />
from atoms. In the case ofparticle physics,<br />
you go back to the ideas of Rutherford.<br />
Operationally, you study the structure of<br />
matter by scattering one particle off another<br />
and observing what happens. The experimental<br />
results can be organized in a kind<br />
of a matrix that gives the amplitudes for the<br />
incoming particles to scatter into the outgoing<br />
ones. Measuring the elements of this<br />
scattering, or S-matrix, was the goal ofexperimentalists.<br />
The work oftheorists was to<br />
write down relationships that these S-matrix<br />
elements had to obey. The idea that there<br />
was another hidden layer of reality, that there<br />
were objects inside protons and neutrons<br />
that hadn't been observed but were responsible<br />
for the properties ofthese particles, was<br />
an idea that wasjust totally foreign to the S-<br />
matrix philosophy; so the proposal that the<br />
hadrons were composed of more fundamental<br />
constituents was vigorously resisted. Not<br />
until ten years later, with the discovery ofthe<br />
psi/Jparticle, did the quark hypothesis become<br />
generally accepted. By then the<br />
evidence was so dramatic that you didn't<br />
have to be an expert to see the underlying<br />
structure.<br />
RABY: The philosophy ofthe bootstrap,<br />
from what I have read ofit, is a very beautiful<br />
philosophy. There is no fundamental particle,<br />
but there are fundamental rules ofhow<br />
particles interact to produce the whole spectrum.<br />
But one thing that distinguishes physics<br />
from philosophy is predictive power. The<br />
189
“It’s important to pick one fundamental question, push<br />
99<br />
on it, and get the right answer .<br />
quark model had a lot of predictive power. It<br />
predicted the whole spectrum of hadrons<br />
observed in high energy experiments. It is<br />
not because of sociology that the bootstrap<br />
went out; it was the experimental evidence of<br />
J/psi that made people believe there really<br />
are objects called quarks that are the building<br />
blocks ofall the hadrons that we see. It is this<br />
reality that turned people in the direction<br />
they follow today.<br />
CARRUTHERS: And because of the very<br />
intense proliferation ofunknowns, it is unlikely<br />
that the search for fundamental constituents<br />
will stop here. In the standard<br />
model you have dozens of parameters that<br />
are beyond any experimental reach.<br />
SCIENCE: But you have fewer coordinates<br />
now than you had originally, right?<br />
CARRUTHERS: If you are saying the<br />
coordinates have all been coordinated by<br />
group symmetry, then ofcourse there are<br />
many fewer.<br />
WEST: I think the deep inelastic scattering<br />
experiments at SLAC played an absolutely<br />
crucial role in convincing people that quarks<br />
are real. It was quite clear from the scaling<br />
behavior of the scattering amplitudes that<br />
you were doing a classic Rutherford type<br />
scattering experiment and that you were literally<br />
seeing the constituents ofthe nucleon.<br />
I think that was something that was extremely<br />
convincing. Not only was it<br />
qualitatively correct, but quantitatively<br />
numbers were coming out that could only<br />
come about if you believed the scattering was<br />
taking place from quarks, even though they<br />
weren’t actually being isolated. But let me<br />
say one other thing about the S-matrix approach.<br />
That approach is really quantum<br />
mechanics in action. Everything is connected<br />
with everything else by this principle of unitarity<br />
or conservation of probability. It is a<br />
very curious state ofaffairs that the quark<br />
model, which requires less quantum mechanics<br />
to predict, say, the spectrum ofparticles,<br />
has proven to be much more useful.<br />
SLANSKY Remember, though, there were<br />
some important things missing in the<br />
bootstrap approach. There was no natural<br />
way to incorporate the weak and the elec-<br />
190<br />
tromagnetic forces.<br />
WEST That picks up another important<br />
point; the S-matrix theory could not cope<br />
with the problem ofscale. And that brings us<br />
back to the standard model and then into<br />
grand unification. The deep inelastic scattering<br />
experiments focused attention on the<br />
idea that physical theories exhibit a scale<br />
invariance similar to ordinary dimensional<br />
analysis.<br />
One of the wonderful things that happened<br />
as a result was that all of us began to accept<br />
renormalization (the infinite rescaling of<br />
field theories to make the answers come out<br />
finite) as more than just hocus-pocus. Any<br />
graduate student first learning the renormalization<br />
procedure must have thought<br />
that a trick was being pulled and that the<br />
procedure for getting finite answers by subtracting<br />
one infinity from another really<br />
couldn’t be right. An element ofhocus-pocus<br />
may still remain, but the understanding that<br />
renormalization was just an exploitation of<br />
scale invariance in the very complicated context<br />
offield theory has raised the procedure<br />
to the level ofa principle.<br />
The focus on scale also led to the feeling<br />
that somehow we have to understand why<br />
the forces in nature have different strengths<br />
and become strong at different energies, why<br />
there are different energy scales for the weak,<br />
for the electromagnetic, and for the strong<br />
interactions, and ultimately whether there<br />
may be a grand scale, that is, an energy at<br />
which all the forces look alike. That’s one of<br />
those wonderful deep questions that has<br />
come back to haunt us.<br />
RABY: I guess we think ofquantum electrodynamics<br />
(QED) as being such a successful<br />
theory because calculations have been<br />
done to an incredibly high degree of accuracy.<br />
But it is hard to imagine that we will<br />
ever do that well for the quark interactions.<br />
The whole method of doing computations in<br />
QED is perturbative. You can treat the electromagnetic<br />
interaction as a small perturbation<br />
on the free theory. But, in order to<br />
understand what is going on in the strong<br />
interactions of quantum chromodynamics<br />
(QCD), you have to use nonperturbative<br />
methods, and then you get a whole new<br />
feeling about the content of field theory.<br />
Field theory is much richer than a<br />
perturbative analysis might lead one to believe.<br />
The study of scaling by M. Fisher, L.<br />
Kadanoff, and K. Wilson emphasized the
interrelation of statistical mechanics and<br />
field theory. For example, it is now understood<br />
that a given field theoretic model may,<br />
as in statistical mechanics systems, exist in<br />
several qualitatively different phases.<br />
Statistical mechanical methods have also<br />
been applied to field theoretic systems. For<br />
example, gauge theories are now being<br />
studied on discrete space-time lattices, using<br />
Monte Carlo computer simulations or analog<br />
high temperature expansions to investigate<br />
the complicated phase structure. There has<br />
now emerged a fruitful interdisciplinary<br />
focus on the non-linear dynamics inherent in<br />
the subjects offield theory, statistical mechanics,<br />
and classical turbulence.<br />
ZWEIG: Isn’t it true to say that the number<br />
ofthings you can actually compute with<br />
QCD is far less than you could compute with<br />
S-matrix theory many years ago?<br />
WEST: I wouldn’t say that.<br />
ZWEIG: What numbers can be experimentally<br />
measured that have been computed<br />
cleanly from QCD?<br />
RABY: What is yourdefinition ofclean?<br />
ZWEIG: A clean calculation is one whose<br />
assumptions are only those ofthe theory. Let<br />
me give you an example. I certainly will<br />
accept the numerical results obtained from<br />
lattice gauge calculations ofQCD as definitive<br />
ifyou can demonstrate that they follow<br />
directly from QCD. When you approximate<br />
space-timeasadiscreteset ofpoints lying in<br />
a box instead ofan infinite continuum, as<br />
you do in lattice calculations, you have to<br />
show that these approximations are legitimate.<br />
For example, you have to show that<br />
the effects of the finite lattice size have been<br />
properly taken into account.<br />
RABY: To return to the question, this is the<br />
first time you can imaginecalculating the<br />
spectrum ofstrongly interacting particles<br />
from first principles.<br />
ZWEIG: The spectrum of strongly interacting<br />
particles has not yet been calculated in<br />
QCD. In principle it should be possible, and<br />
much progress has been made, but operationally<br />
the situation is not much better than<br />
it was in theearly ’60s when the bootstrap<br />
was gospel.<br />
SLANSKY Yes, but that was a very dirty<br />
calculation. The agreement got worse as the<br />
calculations became more cleverly done.<br />
WEST The numbers from latticegauge theory<br />
calculations of QCD are not necessarily<br />
meaningful at present. There is a serious<br />
question whether the lattice gauge theory, as<br />
formulated, is a real theory. When you take<br />
the lattice spacing to zero and go to the<br />
continuum limit, does that give you the theory<br />
you thought you had?<br />
RABY: That’s the devil’sadvocate point of<br />
view, the view coming from the mathematical<br />
physicists. On the other hand, people<br />
have made approximations, and what you<br />
can say is that any approximation scheme<br />
that you use has given the same results. First,<br />
there are hadrons that are bound states of<br />
quarks, and these bound states have finite<br />
size. Second, there is no scale in the theory,<br />
but everything, all the masses, for example,<br />
can be defined in terms ofone fundamental<br />
scale. You can get rough estimates of the<br />
whole particle spectrum.<br />
WEST: You can predict that from the old<br />
quark model, without knowing anything<br />
about the local color symmetry and the eight<br />
colored gluons that are the gauge particles of<br />
the theory. There is only one clean calculation<br />
that can be done in QCD. That is the<br />
calculation of scattering amplitudes at very<br />
high energies. Renormalization group analysis<br />
tells us the theory is asymptotically free at<br />
high energies, that is, at very high energies<br />
quarks behave as free point-like particles so<br />
the scattering amplitudes should scale with<br />
energy. The calculations predict logarithmic<br />
corrections to perfect scaling. These have<br />
been observed and they seem to be unique to<br />
QCD. Another feature unique to quantum<br />
chromodynamics is the coupling ofthegluon<br />
to itselfwhich should predict the existence of<br />
glueballs. These exotic objects would provide<br />
another clean test of QCD.<br />
ZWEIG: I agree. The most dramatic and<br />
interesting tests ofquantum chromodynamics<br />
follow from those aspects of the<br />
theory that have nothing to do with quarks<br />
directly. The theory presumably does predict<br />
the existence ofbound states ofgluons, and<br />
ROUND TABLE<br />
furthermore, some of those bound states<br />
should have quantum numbers that are not<br />
the same as those ofparticles made out of<br />
quark-antiquark pairs. The bound states that<br />
I would like to see studied are these “oddballs,”<br />
particles that don’t appear in the<br />
simple quark model. The theory should<br />
predict quantum numbers and masses for<br />
these objects. These would be among the<br />
most exciting predictions of QCD.<br />
RABY: People who are calculating the<br />
hadronic spectrum are doing those sorts of<br />
calculations too.<br />
ZWEIG: It’s important to pick one fundamental<br />
question, push on it, and get the right<br />
answer. You may differ as to whether you<br />
want to use the existence ofoddballs as a<br />
crucial test or something else, but you should<br />
accept responsibility for performing calculations<br />
that are clean enough to provide meaningful<br />
comparison between theory and experiment.<br />
The spirit ofempiricism does not<br />
seem to be as prevalent now as it was when<br />
people were trying different approaches in<br />
particle physics, that is, S-matrix theory,<br />
field theory, and the quark model. The development<br />
of the field was much more Danvinian<br />
then. Peopleexplored many different<br />
ideas, and natural selection picked the winner.<br />
Now evolution has changed; it is<br />
Lamarckian. People think they know what<br />
the right answer is, and they focus and build<br />
on one another’s views. The value of actually<br />
testing what they believe has been substantially<br />
diminished.<br />
SLANSKY I don’t think that is true. The<br />
technical problems of solving QCD have<br />
proved to be harder than any other technical<br />
problems faced in physics before. People<br />
have had to back offand try to sharpen their<br />
technical tools. I think, in fact, that most do<br />
have open minds as to whether it is going to<br />
be right or wrong.<br />
WEST: What do you think about the rest of<br />
the standard model? Do we think the electroweak<br />
unification is a closed book,<br />
especially now that W’and Zo vector bosons<br />
have been discovered?<br />
SLANSKY: It is to a certain level ofaccuracy,<br />
but the theory itself is just a<br />
191
“It may be that all this matter is looped together in<br />
some complex topological web and that ifyou tear<br />
apart the Gordian knot with your sword of Damocles,<br />
something really strange will happen. ”<br />
phenome:nology with some twenty or so free<br />
parameters floating around. So it is clearly<br />
not the final answer.<br />
SCIENCE: What are these numbers?<br />
RABY: All the masses of the quarks and<br />
leptons are put into the theory by hand. Also,<br />
the mixing angle, the so-called Cabibbo<br />
angle, which describes how the charmed<br />
quark decays into a strange quark and a little<br />
bit of the down quark, is not understood at all.<br />
ZWEIC;: Operationally, the electroweak theory<br />
is solid. It predicted that the W’and Zo<br />
vector bosons would exist at certain masses,<br />
and they actually do exist at those masses.<br />
SLANSKY: The theory also predicted the<br />
coupling ofthe Zo to the weak neutral current.<br />
People didn’t want to have to live with<br />
neutral currents because, to a very high<br />
degree ofexperimental accuracy, there was<br />
no evidence for strangeness-changing weak<br />
neutral currents. The analysis through local<br />
symmetry seemed to force on you the existence<br />
ofweak neutral currents, and when<br />
they were observed in ’73 or whenever, it was<br />
a tremendous victory for the model. The<br />
electron has a weak neutral current, too, and<br />
this current has a very special form in the<br />
standard model. (It is an almost purely axial<br />
current.) This form ofthe current was established<br />
in polarized electron experiments at<br />
SLAC. Very shortly after those experimental<br />
results, Glashow, Weinberg, and Salam received<br />
the Nobel prize for their work on the<br />
standard model of electroweak interactions. I<br />
think that was the appropriate time to give<br />
the Nobel prize, although a lot of my colleagues<br />
felt it was a little bit premature.<br />
RABY: However, the Higgs boson required<br />
for the consistency of the theory hasn’t been<br />
seen yet.<br />
SLANSKY: A little over a year ago there<br />
were four particles that needed to be<br />
seen-now there is only one. The standard<br />
model theory has had some rather impressive<br />
successes.<br />
WEST Can we use this as a point of departure<br />
to talk about grand unification? Unification<br />
ofthe weak and electromagnetic<br />
interactions, which had appeared to be quite<br />
separate forces, has become the prototype for<br />
attempts to unify those two with the strong<br />
interactions.<br />
RABY In the standard model of the weak<br />
interactions, the quarks and the leptons are<br />
totally separate even though phenomenologically<br />
they seem to come in families. For<br />
example, the up and the down quarks seem<br />
to form a family with the electron and its<br />
neutrino. Grand unification is an attempt to<br />
unify quarks and leptons, that is, to describe<br />
them as different aspects of the same object.<br />
In other words, there is a large symmetry<br />
group within which quarks and leptons can<br />
transform into each other. The larger group<br />
includes the local symmetry groups of the<br />
strong and electroweak interaction and<br />
thereby unifies all the forces. These grand<br />
unified theories also predict new interactions<br />
that take quarks into leptons and vice versa.<br />
One prediction ofthese grand unified theories<br />
is proton decay.<br />
WEST: The two most crucialpredictions of<br />
grand unified theories are, first, rhat prorons<br />
are not perfectly stable and can decay and,<br />
second, that magnetic monopoles exist.<br />
Neither of these has been seen so far. Suppose<br />
rhey are never seen. Does fhaf mean the question<br />
of grand unification becomes merely<br />
philosophical? Also, how does that bear on the<br />
idea of building a very high-energy accelerator<br />
like fhe SSC (superconducting super<br />
collider) rhar will cost the taxpayer $3 billion?<br />
CARRUTHERS: Why should we build this<br />
giant accelerator? Because in our theoretical<br />
work we don’t have a secure world view; we<br />
need answers to many critical questions<br />
raised by the evidence from the lower<br />
energies. Even though I know that as soon as<br />
you do these new experiments, the number<br />
ofquestions is likely to multiply. This is part<br />
ofmy negative curvature view ofthe progress<br />
of science. But there are some rather<br />
primitive questions which can be answered<br />
and which don’t require any kind ofsophistication.<br />
For instance, are there any new particles<br />
ofwell-defined mass ofthe oldfashioned<br />
type or new particles with different<br />
properties, perhaps? Will we see the Higgs<br />
particle that people stick into theories just to<br />
make the clock work? If you talk to people<br />
who make models, they will give you a panorama<br />
of predictions, and those predictions<br />
will become quite vulnerable to proof if we<br />
increase the amount ofaccelerator energy by<br />
a factor of 10 to 20. Those people are either<br />
going to be right, or they’re going to have to<br />
retract their predictions and admit, “Gee, it<br />
didn’t work out, did it?”<br />
There is a second issue to be addressed,<br />
and that is the question ofwhat the fundamental<br />
constituents of matter are. We<br />
messed up thirty years ago when we thought<br />
protons and neutrons were fundamental. We<br />
know now that they’re structured objects,<br />
like atoms: they’re messy and squishy and all<br />
kinds of things are buzzing around inside.<br />
Then we discovered that there are quarks<br />
and that the quarks must be held together by<br />
glue. But some wise guy comes along and<br />
says, “How do you know those quarks and<br />
gluons and leptons are not just as messy as<br />
those old protons were’?’’ We need to test<br />
whether or not the quark itselfhas some<br />
composite structure by delivering to the<br />
quarks within the nucleons enough energy<br />
and momentum transfer. The accelerator<br />
acts like a microscope to resolve some fuzziness<br />
in the localization of that quark, and a<br />
whole new level ofsubstructure may be discovered.<br />
It may be that all this matter is<br />
looped together in some complex topological<br />
web and that ifyou tear apart the Gordian<br />
knot with your sword ofDamocles. something<br />
really strange will happen. A genie may<br />
pop out ofthe bottle and say, “Master, you<br />
have three wishes.”<br />
A third issue to explore at the SSC is the<br />
dynamics of how fundamental constituents<br />
interact with one another. This takes you<br />
into the much more technical area ofanalyzing<br />
numbers to learn whether the world view<br />
you’ve constructed from evidence and theory<br />
makes any sense. At the moment we have<br />
no idea why the masses of anything are what<br />
they are. You have a theory which is attractive,<br />
suggestive, and can explain many, many<br />
things. In the end, it has twenty or thirty
~<br />
“If we can get people to agree on why we should be<br />
doing high-energy physics, then 1 think we can solve the<br />
problem ofprice. ’’<br />
ROUND TABLE<br />
I<br />
parameters. You can’t be very content that<br />
you’ve understood the structure of matter.<br />
SLANSKY: To make any real progress both<br />
in unification of the known forces and in<br />
understanding anything about how to go<br />
beyond the interactions known today, a machine<br />
ofsomething like 20 to 40 TeV center<br />
of mass energy from proton-proton collisions<br />
absolutely must be built.<br />
SCIENCE: Will these new machines test<br />
QCD at the same time they test questions of<br />
unification?<br />
SLANSKY The pertinent energy scale in<br />
QCD is on the order of GeV, not TeV, so it is<br />
not clear exactly what you could test at very<br />
high energies in terms of the very nonlinear<br />
structure of QCD. Pete feels differently.<br />
CARRUTHERS: All I say is that you may be<br />
looking at things you don’t think you are<br />
looking at.<br />
WEST: Obviously all this is highly speculative.<br />
A question you are obligated to ask is at<br />
what stage do you stop the financing. I think<br />
we have to put the answer in terms ofa<br />
realistic scientific budget for the United<br />
States, or for the world for that matter.<br />
CARRUTHERS: Is there a good reason why<br />
the world can’t unify its efforts to go to higher<br />
energies?<br />
WEST: Countries are mostly at war with one<br />
another. They couldn’t stop to have the<br />
Olympic games together, so certainly not for<br />
a bloody machine.<br />
SLANSKY The Europeans themselves have<br />
gotten together in probably one ofthe most<br />
remarkable examples of international collaboration<br />
that has ever happened.<br />
WEST: Yes, I think the existence ofCERN is<br />
one ofthe greatest contributions ofparticle<br />
physics to the world.<br />
RABY: But the next step isgoing to have to<br />
be some collaborative effort of CERN, the<br />
U.S., and Japan.<br />
WEST: Our SSC is going to be the next step.<br />
But you are still not answering the question.<br />
Should we expect the government to support<br />
this sort of project at the $3 billion level?<br />
SLANSKY That’s $3 billion over ten years.<br />
RABY You can ask that same question of<br />
any fundamental research that has no direct<br />
application to technology or national security,<br />
and you will get two different answers.<br />
The “practical” person will say that you do<br />
only what you conceive to have some<br />
benefits five or ten years down the line,<br />
whereas the person who has learned from<br />
history will say that all fundamental research<br />
leads eventually either to new intellectual<br />
understanding or to new technology.<br />
Whether technology has always benefited<br />
mankind is debatable, but it has certainly<br />
revolutionized the way people live. I think<br />
we should be funded purely on those<br />
grounds.<br />
WEST Where do you stop? If you decide<br />
that $3 billion is okay or $ IO billion, then do<br />
you ask for $100 billion?<br />
ZWEIG: This is a difficult question, but if<br />
we can get people to agree on why we should<br />
be doing high-energy physics, then I think we<br />
can solve the problem of price. Although<br />
what we have been talking about may sound<br />
very obscure and possibly very ugly to an<br />
outside observer (quantum chromodynamics,<br />
grand unification, and twenty or<br />
thirty arbitrary parameters), the bottom line<br />
is that all ofthis really deals with a fundamental<br />
question, “What is everything made<br />
Of?”<br />
It has been our historical experience that<br />
answers to fundamental questions always<br />
lead to applications. But the time scale for<br />
those applications to come forward is very,<br />
very long. For example, we talked about<br />
Faraday’s experiments which pointed to the<br />
quantal nature of electricity in the early<br />
1800s; well, it was another half century<br />
before the quantum of electricity, the electron,<br />
was named and it was another ten years<br />
before electrons were observed directly as<br />
cathode rays; and another quarter century<br />
passed before the quantum of electric charge<br />
was accurately measured. Only recently has<br />
the quantum mechanics of the electron<br />
found application in transistors and other<br />
solid state devices.<br />
Fundamental laws have always had application,<br />
and there’s no reason to believe<br />
193
“I groan inwardly and sadly reflect on how great the<br />
communication gap is between scientists such as<br />
ourselves and the general public that supports us.”<br />
this will aot hold in the future. We need to<br />
insist that our field be supported on that<br />
basis. We need ongoing commitment to this<br />
potential for new technology, even though<br />
technology’s future returns to society are difficult<br />
to assess.<br />
CARRIJTHERS: Whenever support has to<br />
be ongoing, that’sjust when there seems to be<br />
a tendency to put it off.<br />
WEST: What’s another few years, right?<br />
Now I would like to play devil’s advocate.<br />
One of the unique things about being at <strong>Los</strong><br />
<strong>Alamos</strong> is that you are constantly being<br />
asked to justify yourself. In the past, science<br />
has dealt with macroscopic phenomena and<br />
natural phenomena. (I am a little bit on<br />
dangerous ground here.) Even when it dealt<br />
with the quantum effects, the effects were<br />
macroscopic: spectroscopic lines, for example,<br />
and the electroplating phenomena. The<br />
crucial difference in high-energy physics is<br />
that what we do is artificial. We create rare<br />
states of matter: they don’t exist except<br />
possibly in some rare cosmic event, and they<br />
have little impact on our lives. To understand<br />
the universe that we feel and touch,<br />
even down to its minutiae, you don’t have to<br />
know a damn thing about quarks.<br />
ZWEIG: Maybe our experience is limited.<br />
Let me give you an example. Suppose we had<br />
stable heavy negatively-charged leptons, that<br />
is, heavy electrons. Then this new form of<br />
matter would revolutionize our technology<br />
because it would provide a sure means of<br />
catalyzing fusion at room temperature. So it<br />
is not true that the consequences ofour work<br />
are necessarily abstract, beyond our experience,<br />
something we can’t touch.<br />
WEST: This discussion reminds me of<br />
something I believe Robert Wilson said during<br />
his first years as director of Fermilab. He<br />
was before a committee in Congress and was<br />
asked by some aggressive Congressman,<br />
“What good does the work do that goes on at<br />
your lab? What good is it for the military<br />
defense ofthis country?” Wilson replied<br />
something to the effect that he wasn’t sure it<br />
helped directly in the defense ofthe country,<br />
but it made the country worth defending.<br />
Certainly, finding applications isn’t<br />
1194<br />
predominantly what drives people in this<br />
field. People don’t sit there trying to do grand<br />
unification, saying to themselves that in a<br />
hundred years’ time there are going to be<br />
transmission lines of Higgs particles. When I<br />
was a kid, electricity was going to be so cheap<br />
it wouldn’t be metered. And that was the<br />
kind ofattitude the AEC took toward science.<br />
I, at least, can’t work that way.<br />
SCIENCE: George, do you work that way?<br />
ZWEIG: I was brought up, like Pete, at a<br />
time when the funding for high-energy physics<br />
was growing exponentially. Every few<br />
years the budget doubled. It was absolutely<br />
fabulous. As a graduate student Ijust<br />
watched this in amazement. Then I saw it<br />
turn off, overnight. In 1965, two years after I<br />
got my degree from Caltech, I was in Washington<br />
and met Peter Franken. Peter said,<br />
“It’s all over. High-energy physics is dead.” I<br />
looked at him like he was crazy. A year later I<br />
knew that, in a very real sense, he was<br />
absolutely right.<br />
It became apparent to me that if I were<br />
going to get support for the kind of research I<br />
was interested in doing, I would have to<br />
convince the people that would pay for it that<br />
it really was worthwhile. The only common<br />
ground we had was the conviction that basic<br />
research eventually will have profound applications.<br />
The same argument I make in high-energy<br />
physics, I also make in neurobiology. lfyou<br />
understand how people think, then you will<br />
be able to make machines that think. That, in<br />
turn, will transform society. It is very important<br />
to insist on funding basic research on<br />
this basis. It is an argument you can win.<br />
There are complications, as Pete says; if<br />
applications are fifty years off, why don’t we<br />
think about funding twenty-five years from<br />
now? In fact, that is what we havejust heard:<br />
they have told us that we can have another<br />
accelerator, maybe, but it is ten or fifteen<br />
years down the road.<br />
SLANSKY: We really can’t build the SSC<br />
any faster than rhat.<br />
ZWEIG: They could have built the machine<br />
at Brookhaven.<br />
WEST: Let’s talk about that. How can you<br />
explain why a community who agreed that<br />
building the Isabelle machine was such a .<br />
great and wonderful thing decided, five years<br />
later, that it was not worth doing.<br />
SLANSKY: It is easy to answer that in very<br />
few words. The Europeans scooped the US.<br />
when they got spectacular experimental data<br />
confirming the electroweak unification. That<br />
had been one of our main purposes for building<br />
Isabelle.<br />
CARRUTHERS: If you want to stay on the<br />
frontier, you have to go to the energies where<br />
the frontier is going to be.<br />
ZWEIG: Some interesting experiments were<br />
made at energies that were not quite what<br />
you would call frontier at the time. CP violation<br />
was discovered at an embarrassingly low<br />
energy.<br />
SLANSKY: The Europeans already have the<br />
possibility ofbuilding a hadron collider in a<br />
tunnel already being dug, the large electronpositron<br />
collider at CERN. It is clear that the<br />
U.S., to get back into the effort, has to make a<br />
bigjurnp. Last spring the High Energy <strong>Physics</strong><br />
Advisory Panel recommended cutting off<br />
Isabelle so the U.S. could go ahead in a<br />
timely fashion with the building of the SSC.<br />
WEST: If you were a bright young scientist,<br />
would you go into high-energy physics now?<br />
I think you could still say there is a glamour<br />
in doing theory and that great cosmic questions<br />
are being addressed. But what is the<br />
attraction for an experimentalist, whose<br />
talents are possibly more highly rewarded in<br />
Silicon Valley?<br />
RABY: It will become more and more difficult<br />
to get people to go into high-energy<br />
physics as the time scale for doing experiments<br />
grows an order of magnitude equal to<br />
a person’s lifetime.<br />
ZWEIG: Going to the moon was a successful<br />
enterprise even though it took a long time<br />
and required a different state of mind for the<br />
participating scientists.<br />
WEST Many of the great creative efforts of<br />
medieval life went into projects that lasted<br />
more than one generation. Building a great<br />
cathedral lasted a hundred, sometimes two<br />
hundred years. Some of the great craftsmen,<br />
the great architects, didn’t live to see their
“I consider doing physics something that causes me an<br />
enormous amount of emotional energy. Iget upset. I<br />
get depressed. I get joyful. ”<br />
ROUND TABLE<br />
work completed.<br />
As for going to high energies, I see us<br />
following Fermi’s fantasy: we will find the<br />
hydrogen atom of hadronic physics and<br />
things will become simpler. It is a sort of<br />
Neanderthal approach. You hit as hard as<br />
you can and hope that things break down<br />
into something incredibly small. Somewhere<br />
in those fragments will be the “hydrogen”<br />
atom. That’s the standard model. Some people<br />
may decide to back off from that<br />
paradigm. Lower energies are actually<br />
amenable.<br />
CARRUTHERS: I think that people have<br />
already backed OK Wasn’t Glashow going<br />
around the country saying we should do lowenergy<br />
experiments?<br />
WEST: Just to bring it home, the raison<br />
d’etre for LAMPF I1 is to have a low-energy,<br />
high-intensity machine to look for interesting<br />
phenomena. It is again this curious thing.<br />
We are looking at quantum effects by using a<br />
classical mode-hitting harder. The idea of<br />
high accuracy still uses quantum mechanics.<br />
I suppose it is conceivable that one would<br />
reorient the paradigm toward using the<br />
quantum mechanical nature of things to<br />
learn about the structure ofmatter.<br />
SLANSKY: Both directions are very important.<br />
SCIENCE: Is high-energy physics still attracting<br />
the brightest and the best?<br />
SLANSKY: Some of the young guys coming<br />
out are certainly smart.<br />
CARRUTHERS I think there is an increasing<br />
array of very exciting intellectual<br />
challenges and new scientific areas that can<br />
be equally interesting. Given a limited pool<br />
of intellectual talent, it is inevitable that<br />
many will be attracted to the newer disciplines<br />
as they emerge.<br />
ZWEIG: Computation, for example. Stephen<br />
Wolfram is a great example of someone<br />
who was trained in high-energy physics but<br />
then turned his interest elsewhere, and<br />
profitably so.<br />
CARRUTHERS: Everything to do with conceptualization-computers<br />
or theory of the<br />
mind, nonlinear dynamics advances. All of<br />
these things are defining new fields that are<br />
very exciting-and that may in turn help us<br />
solve some of the problems in particle physics.<br />
ZWEIG: That’s optimistic. What would<br />
physics have been like without your two or<br />
three favorite physicists? I think we would all<br />
agree that the field would have been much<br />
the poorer. The losses ofthe kind we are<br />
talking about can have a profound effect on a<br />
field. Theoretical physics isn’tjust the<br />
cumulative efforts ofmany trolls pushing<br />
blocks to build the pyramids.<br />
WEST: But my impression is that the work<br />
is much less individualized than it ever was.<br />
The fact that the electroweak unification was<br />
shared by three people, and there were others<br />
who could have been added to that list, is an<br />
indication. If you look at QCD and the standard<br />
model, it is impossible to write a name,<br />
and it is probably impossible to write ten<br />
names, without ignoring large numbers of<br />
people who have contributed. The grand unified<br />
theory, ifthere ever is one, will be more<br />
the result ofmany people interacting than of<br />
one Einstein, the traditional one brilliant<br />
man sitting in an armchair.<br />
SCIENCE: Was that idea ever really correct?<br />
WEST: It was correct for Einstein. It was<br />
correct for Dirac.<br />
SCIENCE: Was their thinking really a total<br />
departure?<br />
ZWEIG: The theory ofgeneral relativity is a<br />
great example, and almost a singular example,<br />
of someone developing a correct theoretical<br />
idea in the absence ofexperimental information,<br />
merely on the basis ofintuition. I<br />
think that is what people are trying to do<br />
now. This is very dangerous.<br />
RABY: Another point is that Einstein in his<br />
later years was trying to develop the grand<br />
unified theory ofall known interactions, and<br />
he was way off base. All the interactions<br />
weren’t even known then.<br />
WEST: Theorizing in the absence ofsupportive<br />
data is still dangerous.<br />
CARRUTHERS <strong>Particle</strong> physics, despite all<br />
ofits problems, remains one ofthe principal<br />
frontiers of modern science. As such it combines<br />
a ferment of ideas and speculative<br />
thoughts that constantly works to reassess<br />
the principles with which we try to understand<br />
some of the most basic problems in<br />
nature. If you take away this frothy area in<br />
which there’s an enormous interface between<br />
the academic community and all kinds of<br />
visitors interacting with the laboratory, giving<br />
lectures on what is the latest excitement<br />
in physics, then you won’t have much left in<br />
the way of an exciting place to work, and<br />
people here won’t be so good after awhile. W<br />
195
Index<br />
accelenitors 150-1 57<br />
AGS 52,150,152,153,158,188,194<br />
Bevatron 150,152<br />
CEIW 130,150,152,156,157,159,193,<br />
194<br />
CESR 153<br />
Cosmotron 152<br />
cyclotron 152, 153<br />
DORIS 153<br />
Fermilab 152,155<br />
HERA 156,157<br />
LAMPF 136, 137, 141-145, 147, 148, 154,<br />
158-163,195<br />
LEP 154,156,157<br />
linac 152, 154<br />
PEP 153<br />
PETRA 153<br />
SIN 137,141,154<br />
SLAC 18,140,152,159,188,190, 192<br />
SLC 154,156,157<br />
SPEAR 153<br />
SSC 152,157,164-165<br />
synchrocyclotron 152<br />
synchrotron 152<br />
Tevatron 152,155,156,157<br />
Tristan 156<br />
TRIUMF 137,146,154<br />
antiparticles 108, 131, 182<br />
antilepton 160<br />
antimuon 160<br />
antineutrino 100,112,132<br />
antiproton 28, 109,150, 158-160, 162<br />
antiquark 108,160,191<br />
antisquark 108,112,113<br />
asymmetry, matter-antimatter 168<br />
asymptotic freedom 16-18, 19,40,41<br />
axial vector 145.146<br />
bare parameters 16,17<br />
bare theory 19<br />
baryon 161<br />
Lambda 160,161<br />
number 37,167<br />
conservation of 34, 154, 167, 168<br />
Sigma 161<br />
beta decay 43,128,130,146,167<br />
intrinsic linewidth 132<br />
betaelectron spectrometer 133, 134<br />
toroidal 132<br />
resolution finctiori 132- 134<br />
beta-function 17<br />
big bang theory 168<br />
birds' eggs 5-7<br />
bone structure 4-5<br />
bosonicfield 102,103,105<br />
boson 25,28,40-43,69,76,79,98-112,13 1,<br />
141,142,146,153,158<br />
charged-current 141,142,145<br />
charged-vector 141,142<br />
heavy vector 130<br />
Goldstone 59,63, 121<br />
Higgs47,50,66, 101, 105-107, 131,<br />
156,157,192,194<br />
intermediate vector 136<br />
massless gauge 131<br />
neutralvector 131,141,142<br />
vector 25,40,43,69,76,130, 131, 145,<br />
158,191,192<br />
W*44,45,50,67,77-78, 101, 105-107,<br />
112, 130, 141, 142, 146, 148, 150,<br />
154,156,157,189,191,192<br />
2'46-48,50,67,77-78, 112, 130, 141,<br />
142, 148, 150, 154, 156, 157, 189,<br />
191,192<br />
brain size 10<br />
Cabibbo<br />
angle 51,71<br />
matrix 120<br />
chargedcurrents 130,131,141<br />
interference effects 130,141,142<br />
chirality 44<br />
color30,39,40, 131, 191<br />
charge 69,131<br />
force 131<br />
gluons 130<br />
conservation 34,100,136,137<br />
of energy loo., 132<br />
ofleptons 128,136,137,145<br />
continuity equation 57<br />
correlation fimctions 14, 18<br />
cosmic ray 144,145<br />
cosmic-ray astronomy 167<br />
cosmological constant problem 84,9 1<br />
CPviolation, 33, 121, 123, 131, 154-157, 194<br />
in B '-B system 123<br />
inKo-Rosystem 121, 156, 157<br />
relation to family symmetry breaking<br />
121,123<br />
Crystal Ball 140<br />
Crystal BOX 137-141<br />
deBroglie relation 27<br />
decuplet 37<br />
detector properties 138,139<br />
dimension 3<br />
anomalous 17<br />
dimensional analysis 4,6-14,17, 190<br />
dimensionless variables 7- 19<br />
doublets 131<br />
drag, V~SCOUS 7- 10<br />
Eightfold way 37,38<br />
electrodynamics 26<br />
electromagnetic<br />
coupling constant 13 1<br />
current 142<br />
field 14,34<br />
force 19,23,24, 135<br />
interaction 25, 128<br />
shower 139<br />
electromagnetism 23,28, 188<br />
electron 27,29, 100, 128, 131<br />
number 34,136<br />
scattering<br />
deep inelastic 17, 18,42, 190<br />
electronelectron I42<br />
electron-neutrino elastic 142<br />
polarized 14 1<br />
electron-positron colliders 141<br />
electron-positron pair I5<br />
electroweak theory 19,45-50,65-68,76-79,<br />
128, 130, 131, 141, 142, 168, 189, 192,<br />
194,195. See also forces and<br />
interactions, basic<br />
elementary particles, representations of<br />
in quantum chromodynamics 1 15<br />
in electroweak theory 1 15,116<br />
EMC effect 159,160<br />
end-point energy 132,133<br />
endocranial volume 10<br />
1196
families, quark-lepton 5 1,71, 115, 117, 136, gauge<br />
157,158<br />
fields 142<br />
familychanging interactions 116-122<br />
invariance 36<br />
family problem 31, 128, 130, 136<br />
prticles 189,191<br />
family-symmetry breaking 116- 122<br />
theory25,39,45,130,155,188,191<br />
Fermi 130<br />
non-Abelian (Yang-Mills) 40,45, 188,<br />
constant 43<br />
189<br />
theory 43<br />
gauginos 107<br />
fermionicfield 102, 103, 105<br />
geometry box 140<br />
fermion 28,98- 112, 131<br />
general relativity. See gravity, Einstein’s<br />
generations 136<br />
theory of<br />
Goldstone 101<br />
global invariance 34<br />
Feynman diagram 14- 16,29,189<br />
glueballs41,161,191<br />
fields in higher dimensions 86-87<br />
gluino 108, 109,112,131<br />
fine-structure constant 130<br />
gluons 19,25,40,42,79, 101, 108, 153,<br />
fixed points 18<br />
159-161, 189, 191, 192<br />
flavor 130<br />
goldstino 105, 107, 110<br />
symmetry 39<br />
Goldstone fermion 101<br />
forces and interactions<br />
grandunifiedtheory81-82,106,110,131,<br />
basic 24,74,78-79<br />
146, 154, 168- 170, 182, 190, 192- 195<br />
electromagnetic 19,23-26,28-30,31,36, graviton 80,82,84,93,189<br />
42,98, 110, 128, 130,135, 141,142, gravitino 105,107,110<br />
182, 188-190, 192<br />
massive 110<br />
electroweak 65-68,70-71, 128,130, 131,<br />
gravity<br />
189,192<br />
Einstein’s theory of 74,75,80-8 1,82,84<br />
gravitational 23,24,26,59,98, 110, 130, unification with other forces 82-95<br />
188,189<br />
group multiplets 31<br />
neutral current 46<br />
strengths of 24,28,30,40,43-45,66,70 hadron 36,80,91-92, 109, 144,160-162,<br />
strong 19,23-25,28,30,36,38-40,69-70, 189-191,195<br />
98, 110,128,130, 136, 188-192, 192 Heisenberg uncertainty principle 13, 14,29<br />
unification of23,25,28,30,44-46,53, helium-3 132<br />
72-95, 152, 153,168<br />
Hierarchy problem 106<br />
weak 19,23,24,28,42-45,98, 107,110, Higgs<br />
112, 128, 130, 136, 141-143, 145, bosons47,50,101,105-107,131,192,194<br />
146, 188-190, 192<br />
mechanism 131<br />
beta decay 42-43<br />
Higgsino 107,108<br />
Fermi theory of 43-44,76,77,79,155 hypercharge 37<br />
chargedcurrent 43,46-49,7 1<br />
hypernucleus 159- 161<br />
neutralcurrent 46-47,48,49,68,7 1,<br />
152,153<br />
interference effects<br />
right-handed 157<br />
between neutral and charged weak<br />
form invariance 11,14<br />
currents 130,141-143<br />
fundamental<br />
isotopic spin 30,37,189<br />
constants, 12,13<br />
scales 11<br />
J/Psi 52, 153, 154, 188-189<br />
symmetries 136<br />
jets, hadronic 109, 112, 157<br />
Kaluza-Klein theories 83,84,87<br />
kaons 158- 162<br />
Kobayashi-Maskawa matrix 5 1,123<br />
Lagrangian 14,17,34,100<br />
complex scalar field 55<br />
electroweak theory 65-68<br />
quantum chromodynamics 69-70<br />
quantum electrodynamics 63<br />
quantum field theory 25<br />
real vector fields 56-57<br />
standard model 71<br />
weak interactions of quarks 70-71<br />
Yang-Mills theories 64<br />
Lambda 160,161<br />
lepton 19,25,31,36,51, 107,131, 160, 189,<br />
192,194<br />
conservation 128,136,137,145,154<br />
families 128, 135, 136, 146<br />
flavor 133,136<br />
multiplets 136<br />
local gauge<br />
invariance 36<br />
theory 25,189<br />
transformations 39<br />
local symmetry 25,30,34,40<br />
magnetic<br />
monopoles 110, 192<br />
pinch 134<br />
Majorana Fields 34<br />
mass<br />
neutrino 128,130,133,135,146<br />
scale 13,15,16,18,101,106,107,131<br />
massive gravitino 110<br />
meikton 161<br />
meson 111, 135, 136,161,189<br />
metabolic rate 5-7<br />
Michelparameters 145, 146, 148<br />
modeling theory 9,18,19<br />
Monte Carlo<br />
calculation 162<br />
sampling 41<br />
simulation 137, 140,191<br />
modeling 134<br />
multiplets 136<br />
muon 111, 128,131, 135-141, 143-147,<br />
158-160,189<br />
branching ratio 132,136,137<br />
daughter 137<br />
decay 136, 138-141, 145-148, 155<br />
discovery 135<br />
lifetime 138,144<br />
Michelparameters 145,146,148<br />
number 34,136,137<br />
prompt 144<br />
range 136<br />
197
neutralweakcurrent46,48,130,131,141,<br />
X 92<br />
neutrino 100, 111, 112, 128, 130,136-146,<br />
155,158,159,162,192<br />
appearancemode 143,144<br />
astronomy 167, 170<br />
decay 173- 174<br />
electronneutrino 137,142,143,145<br />
flavors 130<br />
mass 122,128,130-133,135,159<br />
oscillations 122, 133, 143, 148, 155, 167,<br />
176<br />
physics 154<br />
right-handed 131<br />
solar 155,167<br />
flux measurements of 172- 177<br />
neutrinoelectron scattering 130,136,<br />
141-145, 159,162<br />
detector 144<br />
background 144<br />
neutron decay. See proton decay<br />
non-Abelian gauge theories 16,18,40,188<br />
nucleon decay 111,146<br />
octets 31<br />
!2-37,150<br />
Parity<br />
conservation 128,146<br />
violation 49,69,146<br />
Pauli exclusion principle 28, 104<br />
photinos 106,109,112<br />
photon 14, 16, 19,29,36,76-78, 100, 135<br />
pion 100,110, 136, 137, 143, 145,159,160,<br />
162<br />
decay 140,143,145<br />
pion dynamics 59, 152, 155<br />
Planck mass, length 80,82,83<br />
positrons 29,109- 111,138,139,143,182<br />
preons 53,157<br />
propagator 14- 19<br />
proton<br />
beam 137,158,160-162<br />
decay81-82, 110, 111, 148, 155,168, 192<br />
searches for 166-17 1<br />
pyrgons 84-87,95<br />
quantum chromodynamics 18,19,39-42,<br />
69-70,79-80, 130, 131, 159, 161, 190,<br />
191,193. See abo forces and<br />
interactions, basic<br />
quantum electrodynamics 14, 16-19,28-30,<br />
31,34,36,40,55,62-63,76,79, 130,<br />
189,190. See also forces and<br />
interactions, basic<br />
quantum field theory, 4,11,13,14,16,18,<br />
25,27,28,30,24,64, 141, 182, 187<br />
quark3, 17, 19,25,31,38,39,42,51, 100,<br />
107-110, 112, 131, 136, 150, 152-155,<br />
159-161, 180, 182-184, 188-192, 194<br />
confinement of 42<br />
families 5 1<br />
flavors 39<br />
masses of 69-71<br />
mixing of 5 1,71<br />
transitions between states 136<br />
rare decays 128<br />
limits 138<br />
ofthemuon 135,137<br />
Rayleigh-Ruabouchinsky Paradox 12- 13<br />
Regge<br />
recurrences 95<br />
trajectories 92<br />
renormalization'll, 14-18,28, 189-191<br />
group 11,16,18,19,189,191<br />
group equation 13,14,17<br />
rotation group 32<br />
rowing 9,lO<br />
scalar 146<br />
pseudo 146<br />
particles 100, 103, 105, 107, 108<br />
scale2-21,101, 107, 131,183,190,191<br />
energy 19<br />
invariance 11,13,190<br />
scaling 4-2 1,28,42,190<br />
classical 4<br />
curve 9<br />
scattering experiments 130,189,191<br />
inelastic 17,18,42,79-80, 190<br />
selectron 100, 109<br />
similitude $7<br />
singlets 131<br />
slepton 107<br />
!+matrix theory 189- 191<br />
sneutrino 100<br />
solar energy-production models 166, 167,<br />
171-172, 173, 174<br />
solar neutrino 145<br />
physics 148<br />
space-time manifold 82<br />
extension to higher dimensions, 76,83,<br />
84,86<br />
special relativity 27<br />
spin 30,87, 100, 131, 189<br />
spin-polarized hydrogen 133<br />
spin-statistics theorem 100<br />
quark 100,107-109,11,113<br />
standardmodel 18, 19,23,25,30,42,50-51,<br />
53,54,71,74,76-80, 100, 106,107,<br />
109, 110, 115, 130, 141, 142, 145, 159,<br />
182, 188- 190, 192<br />
minimal 130,131,136,145,148<br />
strangeness 30,37<br />
supergap 101<br />
supergravity 76,88-91,93,94, 101, 110<br />
superstring theories 76,82,91-95<br />
superspaces 93<br />
supersymmetry 74,76,88-90,98, 100-113,<br />
157<br />
in quantum mechanics 102- 105<br />
interaction 1 19<br />
spontaneous breaking 100,105,107,110<br />
supersymmetry rotation 106, 108<br />
symmetries; symmetry groups, multiplets,<br />
and operations 30-36,38-39,45-47,<br />
61,64,75,88,90,100,101,110,111,<br />
136,142,189<br />
Eightford Way 69,70<br />
of electroweak interactions 65,77<br />
ofhgrangians 56-57<br />
of quark-lepton interactions 8 1<br />
of strong interactions 69-70<br />
strong isospin 60-6 1,69<br />
weak isospin 6 1<br />
198
symmetry30,39,128,145,146,182,188-192<br />
boson-fermion 74<br />
broken 31,32,33,39,45,48, 100,<br />
105-107, 131, 182<br />
continuous 31,56<br />
CP33,131<br />
discrete 31<br />
exact 34 4<br />
external 100-102<br />
left-handed 145<br />
local 25,30,3436,39-40,46-47,54-56<br />
7475,188,192<br />
Lorentz in variance 56,59<br />
phase invariance 56,59,62-63,76<br />
Poincak 56,80<br />
right-handed 145<br />
spontaneous breaking of47-48,54,58-59<br />
62-63,66-68,78,81,83,88<br />
tau 30,128,136<br />
particle number 34,131<br />
time projection chamber (TPC) 146, 147<br />
tritium 128,132-135<br />
beta decay 128,130,148<br />
betadecay spectrometer 128, 132-135<br />
resolution function 132-134<br />
end-point energy 132, 133<br />
final state spectrum 132<br />
molecular 133-135<br />
recombination 133<br />
source 133-135<br />
W*. Seeboson.<br />
weak charge 131<br />
weak force 19,23-25,28<br />
currents49,131,141,192<br />
weakinteraction 42,107,110,112,128,136,<br />
141-143, 145, 146, 192<br />
constructive and destructive interference<br />
142,143<br />
coupling constant 131,145,146<br />
weak mixingangle 48,66,67,110,111,131,<br />
142,143<br />
weak scale 101<br />
Weinbergangle48,110,111,131,142,143<br />
Weinberg-Salam-Glashow model 130<br />
Yang-Mills theories 40,45, 188, 189<br />
Yukawa's theory 135<br />
Zo. Seeboson.<br />
zero modes 85,86,95<br />
ultraviolet laser technology 133<br />
uncertainty principle 13, 14,29<br />
underground science facilities 178-1 79<br />
unification26, 101, 141, 188, 190, 192-195<br />
T 53. 153-154<br />
vacuum state 58<br />
vector-axial 13 1, 145<br />
currents 143<br />
vector potential 28<br />
199