S T A N F O R D L I N E A R A C C E L E R A T O R C E N T E RSpring 1997, Vol. 27, No.1

FEATURES

4 THE DISCOVERY OF THE ELECTRONAbraham Pais

17 WHAT IS AN ELEMENTARY PARTICLE?Steven Weinberg

22 ELEMENTARY PARTICLES: YESTERDAY,TODAY, AND TOMORROWChris Quigg

30 THE INDUSTRIAL STRENGTH PARTICLEMichael Riordan

36 THE EVOLUTION OF PARTICLEACCELERATORS & COLLIDERSWolfgang K. H. Panofsky

A PERIODICAL OF PARTICLE PHYSICS

SPRING 1997 VOL. 27, NUMBER 1

Guest EditorMICHAEL RIORDAN

EditorsRENE DONALDSON, BILL KIRK

Editorial Advisory BoardJAMES BJORKEN, GEORGE BROWN

ROBERT N. CAHN, DAVID HITLIN

JOEL PRIMACK, NATALIE ROE

ROBERT SIEMANN

IllustrationsTERRY ANDERSON

DistributionCRYSTAL TILGHMAN

The Beam Line is published by the Stanford Linear Accelerator Center, PO Box 4349, Stanford, CA 94309. Telephone: (415) 926-2585 INTERNET: beamline@slac.stanford.eduFAX: (415) 926-4500 Issues of the Beam Line are accessible electronically onthe World Wide Web at http://www.slac.stanford.edu/pubs/beamlineSLAC is operated by Stanford University under contractwith the U.S. Department of Energy. The opinions of theauthors do not necessarily reflect the policy of theStanford Linear Accelerator Center.

Printed on recycled paper

Cover: A montage of images created by Terry Andersonfor this special Beam Line issue on the centennial of theelectron’s discovery. (Photograph of Electronics magazinecourtesy AT&T Archives.)

Pais Weinberg

45 THE UNIVERSE AT LARGE:THE ASTRO-PARTICLE-COSMO-CONNECTIONVirginia Trimble

DEPARTMENTS

2 FROM THE EDITOR’S DESK

52 CONTRIBUTORS

CONTENTS

Quigg Riordan Panofsky Trimble

HE ELECTRON—or at least our recognition ofits existence as an elementary particle—passesthe century mark this spring. On April 30, 1897,Joseph John Thomson reported the results of hisrecent experiments on cathode rays to a Fridayevening meeting of the Royal Institution, sug-gesting these rays were composed of negativelycharged pieces of atoms that he dubbed “corpuscles.” Six months later he published anextensive account of these experiments in thePhilosophical Magazine. One of the classicpapers of modern physics, it opened the doors ofhuman consciousness to a radically new andoften baffling world within atoms, one that hasprovided fertile ground for much of twentieth-century physics.

Together with the discovery of X rays andradioactivity during the preceding two years, andthe introduction of the quantum three yearslater, this breakthrough led to a revolutionaryconception of matter that has since had majorimpacts on other sciences, on modern tech-nology and art, and even on the way we talk andthink. The smooth, continuous, comfortableworld of late nineteenth-century Europe wasshattered into myriad bewildering fragments—some of which interact via forces that nobodyhad ever before encountered. Whether atomsthemselves existed or not was in hot dispute atthe time; among those who believed they didwere prominent physicists who regarded them asvortex rings in the luminiferous aether. Acentury later, despite many superb advances, weare still struggling to achieve a grand synthesis ofall the disparate shards encountered since.

To commemorate this pivotal breakthrough—and, in a more catholic sense, the discovery of

2 SPRING 1997

FROM THE EDITOR’S DESK

Introductory paragraphs from Thomson’s paper, as publishedin the May 21, 1897 issue of The Electrician.

BEAM LINE 3

subatomic particles—the Beam Line EditorialBoard organized this special anniversary issueand asked me to serve as its guest editor. It hasbeen a truly stimulating and rewarding experi-ence. I am privileged to have worked with someof the nation’s most literate physicists, who havecontributed perceptive essays in honor ofThomson’s fabulous discovery.

Three theorists open this issue by offering ustheir perspectives on the discovery, the meaningand the evolution of elementary particles. WhileAbraham Pais relates how the concept of theelectron emerged from nineteenth-centuryresearch on electrochemistry and vacuum-tubedischarges, Steven Weinberg and Chris Quigg

take more modern and personal viewpoints.They examine what it means to call a particle“elementary” and try to assess where our disci-pline is headed as its second century begins.

The final three articles concern “applications”of our knowledge of subatomic particles—in elec-tronics technology, in pushing back the frontiersof high-energy research itself, and in understand-ing the origin and evolution of the Universe. Myarticle indicates how our knowledge of the elec-tron as a particle has proved crucial to the surg-ing growth of what is now the world’s biggest in-dustry. Taking a retrospective look at particleaccelerators and colliders, Wolfgang Panofskyevaluates various avenues being considered forthe future of this technology. And Virginia Trimble closes this anniversary issue by survey-ing how the tiniest things in existence are close-ly linked to the structure and behavior of thelargest.

What will historians think, a hundred yearshence, when they gaze back upon our own time?What conceptions that we hold dear today willbe regarded then as we now regard the aether of1897? What will be the “elementary particles” ofthe late twenty-first century? We can only guess.Whatever the answers, however, there can belittle doubt that the hundred years that beganwith Thomson’s discovery will be viewed as aremarkable period of scientific, technological andcultural achievement.

Illustration from Thomson’s article showing luminous paths of cathode rays (lower trace) bending in a magnetic field. The upper trace is due to ionized atoms in the gas.

4 SPRING 1997

BEAM LINE 5

N THE EARLY YEARS

following the first ob-servation of the elec-

tron, a toast used to be of-fered at the CavendishLaboratory annual dinner:“The electron: may itnever be of use to any-body.”1 That wish has not

been fulfilled. The discovery

of the electron, the first par-

ticle in the modern sense of

the word, has brought about

profound changes in the world

at large. This essay is devoted

to the more provincial but not

less interesting question of

how this discovery came

about.

That event, occurring to-

ward the end of the nine-

teenth century, marks the end

of 2500 years of speculation

about the structure of matter

and the beginning of its cur-

rent understanding. In order

to lend perspective to this mo-

mentous advance, it will help

to begin with a look back to

earlier days—first, briefly to

the times of pure speculation,

then, in more detail, to earlier

nineteenth-century develop-

ments, and finally to the

decade of transition, the years

from 1895 to 1905.

I

J. J

. Tho

mso

n in

his

lab

orat

ory

at C

amb

ridg

e U

nive

rsity

. (C

ourt

esy

Sci

ence

Mus

eum

/Sci

ence

& S

ocie

ty P

ictu

re L

ibra

ry, L

ond

on)

6 SPRING 1997

THE ANCIENTS

THE NINETEENTH CENTURYREGARDING THE UNDERSTANDING of the basic structure of mat-ter, very little had changed between the days of speculation by theancient Greek philosophers and the beginning of the nineteenth cen-tury, when, in 1808, the British chemist and physicist John Dalton(1766–1844) commenced publication of his New System of Chemical

THE TERM atom, derived from the Greek α, a privative, and τεµειυ ,to cut, appears first, I am told, in the writings of Greek philoso-phers of the fifth century BC. Democritus (late fifth century BC) taughtthat atoms are the smallest parts of matter, though in his view theywere not necessarily minute. Empedocles (490–430 BC), physicist,physician, and statesman, held that there are four indestructible andunchangeable elements—fire, air, water and earth—eternally broughtinto union and eternally parted from each other by two divine forces,love and discord. Nothing new comes or can come into being. Theonly changes that can occur are those in the juxtaposition of elementwith element. Epicurus’ (341–270 BC) opinion that atoms cannot bedivided into smaller parts by physical means, yet that they have struc-ture, was shared by prominent scientists well into the nineteenthcentury AD. The Roman poet Lucretius (98–55 BC) was an eloquentexponent of the theory that atoms, infinite in number but limitedin their varieties, are, along with empty space, the only eternaland immutable entities of which our physical world is made. To-day’s scientist will not fail to note that in each of these specula-tive thinkers’ considerations one finds elements that sound curi-ously modern.

The opposite position, that matter is infinitely divisible and con-tinuous, likewise had its early distinguished proponents, notablyAnaxagoras (c 500–428 BC) and Aristotle (384–322 BC). The latter’sprestige eclipsed the atomists’ view until the seventeenth century.Even that late, Rene Descartes (1596–1650) pronounced that there can-not exist any atoms or parts of matter that are of their own natureindivisible; for though God had rendered a particle so small that itwas not in the power of any creature to divide it, He could not,however, deprive Himself of the ability to do so.2

BEAM LINE 7

Philosophy. He had of course illustrious precursors, no-tably Antoine-Laurent Lavoisier (1743–1794). Yet hisquantitative theory suddenly could explain or predictsuch a wealth of facts that he may properly be regard-ed as the founder of modern chemistry. In a sequel vol-ume Dalton expressed the fundamental principle of theyoungest of the sciences in these words:

I should apprehend there are a considerable number ofwhat may be properly called elementary principles,which can never be metamorphosed, one into another,by any power we can control. We ought, however, toavail ourselves of every means to reduce the numberof bodies or principles of this appearance as much aspossible; and after all we may not know what ele-ments are absolutely indecomposable, and what are re-fractory, because we do not know the proper means oftheir reduction. All atoms of the same kind, whethersimple or compound, must necessarily be conceived tobe alike in shape, weight, and every other particular.

These superb lines ushered in the intense nineteenth century dis-cussions on the nature of atoms and molecules. Perhaps the most re-markable fact about these debates is the great extent to whichchemists and physicists spoke at cross purposes when they did notactually ignore each other. This is not to say that there existed onecommon view among chemists, another among physicists. Rather,in either camp there were many and often strongly diverging opin-ions. The principal point of debate among chemists was whetheratoms were real objects or only mnemonic devices for coding chem-ical regularities and laws. The main issues for the physicists cen-tered around the kinetic theory of gases, in particular around themeaning of the second law of thermodynamics.

An early illustration of the dichotomies between chemists andphysicists is provided by the fact that Dalton did not accept thehypothesis put forward in 1811 by Amadeo Avogadro (1776–1856)that, for fixed temperature and pressure, equal volumes of gases con-tain equal numbers of molecules. Nor was Dalton’s position heldonly by a single person for a brief time. The tardiness with whichAvogadro’s law came to be accepted clearly indicates the widespreadresistance to the idea of molecular reality. As but one furtherillustration of this attitude I mention some revealing remarks by

John Dalton,whose NewSystem ofChemicalPhilosophyresurrected theatomic theory ofmatter. (CourtesyA. L. Smyth, JohnDalton: 1766–1844,a Bibliography ofWorks By and AboutHim and AIP EmilioSegrè VisualArchives)

8 SPRING 1997

Alexander Williamson (1824–1904), himself a convinced atomist. Inhis presidential address of 1869 to the London Chemical Society,he said:

It sometimes happens that chemists of high authority refer pub-licly to the atomic theory as something they would be glad to dis-pense with, and which they are ashamed of using. They seem tolook upon it as something distinct from the general facts of chem-istry, and something which the science would gain by throwing offentirely. . . . On the one hand, all chemists use the atomic theory,and . . . on the other hand, a considerable number view it withmistrust, some with positive dislike.3

On the whole, molecular reality met with less early resistance inphysics than it did in chemistry. That is not surprising. Physicistscould already do things with molecules and atoms at a time whenchemists could, for most purposes, take them to be real or leave themas coding devices.

The insight that gases are composed of discrete particles datesback at least to the eighteenth century. Daniel Bernoulli (1700–1782)may have been the first to state that gas pressure is caused by the col-lisions of particles with the walls that contain them. The nineteenth-century masters of kinetic theory were atomists—by definition, onemight say. In Rudolf Clausius’ (1822–1888) paper of 1857, entitled “Onthe kind of motion we call heat,” the distinction between solids, liq-uids, and gases is related to different types of molecular motion. Lud-wig Boltzmann (1844–1906) was less emphatic, but he could hardlyhave developed his theory of the second law of thermodynamics hadhe not believed in the particulate structure of matter.

Long before these learned fin-de-siecle discourses took place, infact long before the laws of thermodynamics were formulated, the-oretical attempts had begun to estimate the dimensions of molecules.As early as 1816, Thomas Young (1773–1829) noted that “the diame-ter or distance of the particles of water is between the two thou-sand and the ten thousand millionth of an inch.”4 In 1873 James ClerkMaxwell stated that the diameter of a hydrogen molecule is about6×10−8 cm.5 That same year Johannes Diderik van der Waals(1837–1923) reported similar results in his doctoral thesis.6 By 1890

the spread in these values, and those obtained by others, had narrowedconsiderably. A review of the results up to the late 1880s placed theradii of hydrogen and air molecules between 1 and 2×10−8 cm,7 aremarkably sensible range.

BEAM LINE 9

Until the very last years of the nineteenth century, most if notall scientists who believed in the reality of atoms shared the viewthat these particles cannot be decomposed further, as was eloquentlyexpressed by Maxwell in 1873:

Though in the course of ages catastrophes have occurred and may yetoccur in the heavens, though ancient systems may be dissolved andnew systems evolved out of their ruins, the molecules [i.e., atoms!]out of which these systems [the Earth and the whole solar system]are built—the foundation stones of the material universe—remainunbroken and unworn. They continue this day as they were creat-ed—perfect in number and measure and weight.8

ELECTROMAGNETISM BECAME A PART of science in the eighteenthcentury, largely due to rapid progress in the invention of new in-struments: the first condenser (the Leiden jar), the lightning rod, thefirst battery (the Voltaic pile), the first solenoid. These advances led,in turn, to the formulation of phenomenological laws based on newexperiments. Of interest here is the law of electrolysis, formulatedin the 1830s by Michael Faraday (1791–1867), one of the great ex-perimentalists of all time, who coined terms of lasting use: electrode,anode, cathode, electrolysis, ion, anion, cation. In modern language,his law can be stated like this:

The amount of electricity deposited at the anode bya gram mole of monovalent ions is a universal constant,the farad (F), given by F = Ne, where N, Avogadro’s num-ber, is the number of molecules per mole, and e is a uni-versal unit of charge.

What does this e signify? In 1881 Herman vonHelmholtz (1821–1894) put it like this in his Faradaylecture: “The most startling result of Faraday’s law isperhaps this. If we accept the hypothesis that the ele-mentary substances are composed of atoms, we cannotavoid concluding that electricity also, positive as wellas negative, is divided into definite elementary portions,which behave like atoms of electricity.”9 This state-ment explains why in subsequent years the quantitye was occasionally referred to in the German literatureas “das Helmholtzsche Elementarquantum.”

THE ATOMICITY OF CHARGE

Hermann vonHelmholtz, who in1881 speculatedon the atomicity ofcharge. (CourtesyAIP Emilio SegrèVisual Archives)

10 SPRING 1997

Even before Helmholtz’s memorable address, the Irish physi-cist George Johnstone Stoney (1826–1911) had reported to the 1874

meeting of the British Association for the Advancement of Sciencean estimate of e, the first of its kind, based on F = Ne. Values for F andN were reasonably well known by then. Stoney obtained e ~3×10−11 esu,too small by a factor of about 20, yet not all that bad for a first andvery early try.10 In 1891 he baptized the fundamental unit of charge,giving it the name “electron.”11 Thus the term was coined prior tothe discovery of the quantum of electricity and matter that now goesby this name.

DECADE OF TRANSITIONIN MARCH 1905 Ernest Rutherford (1871–1937) delivered the Silli-man lectures at Yale. He began the first of his talks as follows:

The last decade has been a very fruitful period in physical science, anddiscoveries of the most striking interest and importance have followedone another in rapid succession. . . . The march of discovery has beenso rapid that it has been difficult even for those directly engaged in theinvestigations to grasp at once the full significance of the facts thathave been brought to light. . . . The rapidity of this advance has sel-dom, if ever, been equaled in the history of science.12

The speed with which one important discovery followed another (seebox at left) was indeed breathtaking. It is natural to ask but noteasy to answer why so much novelty should be discovered in so shorta time span. It is clear, however, that a culmination of advances ininstrumentation was crucial. They include:

• Higher voltages. Higher voltages were the result of HeinrichRuhmkoff’s (1803–1874) work, beginning in the 1850s, on an improvedversion of the induction coil. These were the coils that in 1860 servedGustav Kirchhoff (1824–1887) and Robert Bunsen (1811–1899) in theiranalysis of spark spectra; Heinrich Hertz (1857–1894) in 1886–1888

in his demonstration of electromagnetic waves and his discoveryof the photoelectric effect; Wilhelm Roentgen in his discovery ofX rays; Guglielmo Marconi (1874–1937) in his transmission of tele-graph signals without wires; Pieter Zeeman in his discovery of theZeeman effect; and Thomson in his determination of e/m for elec-trons. By the turn of the century, voltages of the order of 100,000 voltscould be generated by these coils.

1895• Wilhelm Roentgen

(1845–1923) discovers X rays, for which he wouldreceive the first Nobel Prizein physics, in 1901.

1896• Antoine Henri Becquerel

(1852–1908) observes whathe called “uranic rays,” thefirst phenomenon thatopens a new field latercalled radioactivity.

• Wilhelm Wien (1864–1928)publishes his exponentiallaw for black-body radiation,the first quantum law everwritten down.

• Pieter Zeeman’s (1865–1934) first paper appears onthe influence of magneticfields on spectral lines.

1897• Determination of e/m for

cathode rays by J. J.Thomson and others.

• First mention of a particlelighter than hydrogen.

1898• Ernest Rutherford discov-

ers there are two speciesof radioactive radiations:α-rays and β-rays.

1899• Thomson measures the

electric charge of free elec-trons and realizes thatatoms are split in ionizationprocesses.

1900• Paul Villard (1860–1934)

discovers γ-rays.• First determination of a

half-life for radioactivedecay.

• Max Planck discoversthe quantum theory.

1905• Albert Einstein postulates

the light quantum (March). • Einstein’s first paper on

special relativity is pub-lished (June).

March of Discovery

BEAM LINE 11

• Improved vacua. Improved vacuawere achieved in the 1850s, when JohannGeissler (1815–1879) began developingthe tubes now named after him. Soonhe was able to reach and maintain pres-sures of 0.1 mm of mercury. Refined ver-sions of this tube were crucial to the dis-coveries of Roentgen and Thomson.

• Ionization chambers. Early ver-sions of the parallel-plate ionizationchamber were developed in Cambridgeduring the 1890s. They were used byRutherford and the Curies in the earli-est quantitative measurements of ra-dioactivity.

• Concave spectral gratings. Concave spectral gratings weredeveloped starting in the 1880s by Henry Rowland (1848–1901) atthe Johns Hopkins University. Their resolving power made Zeeman’sdiscovery possible.

• Cloud chambers. Work on the development of a cloud cham-ber was begun in Cambridge in 1895 by Charles T. R. Wilson(1869–1959). This instrument enabled Thomson to measure the elec-tron’s charge.

THE DISCOVERYALL RESEARCH THAT LED to the discovery of the electron deals withstudies of cathode rays, a subject that had already engaged Faraday,who in 1838 made this prophetic remark on its future: “The resultsconnected with the different conditions of positive and negative dis-charge will have a far greater influence on the philosophy of elec-trical science than we at present imagine.”13

J. J. Thomson discovered the electron. Numerous are the booksand articles in which one finds it said that he did so in 1897. I can-not quite agree. It is true that in that year Thomson made a gooddetermination of e/m for cathode rays, an indispensable step to-ward the identification of the electron, but he was not the only oneto do so. It is also true that in 1897 Thomson correctly conjecturedthat the large value for e/m he had measured indicated the existence

J. J. Thomson andErnest Rutherford(right) at theCavendish Lab in1934. (Courtesy AIPEmilio Segrè VisualArchives BainbridgeCollection)

12 SPRING 1997

WHEN J. J. THOMSON began his research on thecathode rays during the 1890s, there was great confu-sion about their exact nature. As he noted in the intro-duction to his paper, “On Cathode Rays,” [Phil. Mag.,Ser. 5, Vol. 44, No. 269 (1897), p. 293]:

The most diverse opinions are held as to these rays;according to the almost unanimous opinion of Germanphysicists they are due to some process in the ætherto which . . . no phenomenon hitherto observed isanalogous; another view of these rays is that, so farfrom being wholly ætherial, they are in fact whollymaterial, and that they mark the paths of particles ofmatter charged with negative electricity.

Following the lead of French physicist Jean Perrin,Thomson first satisfied himself that the rays were nega-tively charged, then addressed a quandary that hadbeen puzzling scientists on both sides of the Channel foryears. Although the rays were easily deflected by a mag-netic field, they were apparently not deflected by anelectric field between two plates. The absence of this deflection, he showed, was due to the ionization of thegas remaining in a cathode-ray tube, which permitted acurrent to flow between the plates and drastically reduced the field. This did not occur at high vacuum,however, and the rays were indeed deflected as expected for negatively charged particles. Thus he noted:

I can see no escape from the conclusion that they arecharges of negative electricity carried by particles ofmatter. The question next arises, What are these par-ticles? [A]re they atoms, or molecules, or matter ina still finer state of subdivision?

By simultaneously deflecting the rays in both electricand magnetic fields, Thomson was able to determinetheir velocity and the ratio m/e of the mass m to theelectric charge e carried by these (then) hypotheticalparticles. His result was startling:

From these determinations we see that the value of m/eis independent of the nature of the gas, and that its val-ue 10–7 [gram per emu] is very small compared withthe value 10–4, which is the smallest value of thisquantity previously known, and which is the value forthe hydrogen ion in electrolysis.

But he could not conclude from these data that m itselftherefore had to be very small. “The smallness of m/emay be due to the smallness of m or the largeness of e,”Thomson wrote. Because the values of m/e were inde-pendent of the nature and pressure of the gas, he began

Thomson’s Two Experimental Papers

BEAM LINE 13

to envision atoms as made of “primordial atoms, whichwe shall for brevity call corpuscles.” He went on:

The smallness of the value of m/e is, I think, due tothe largeness of e as well as the smallness of m. Thereseems to me to be some evidence that the charges car-ried by the corpuscles in the atom are large comparedwith those carried by the ions of an electrolyte.

Over the next two years, Thomson determined the massand charge of his corpuscles, but it took additional ex-periments culminating in a second paper, “On the Mass-es of the Ions in Gases at Low Pressures,” [Phil. Mag.,Ser. 5, Vol. 48, No. 295 (1899), p. 547]. Using a noveltechnique developed by his student C. T. R. Wilson, hemeasured both m/e and e for the negatively chargedparticles created by dissociation of atoms in ultravioletlight. He found m/e to be the same as for cathode raysand e to have the same absolute value as the hydrogenion in electrolysis. Thus he concluded:

The experiments just described, taken in conjunctionwith the previous ones on the value of m/e for the cath-ode rays . . . show that in gases at low pressures nega-tive electrification, though it may be produced by verydifferent means, is made up of units each having a chargeof electricity of a definite size; the magnitude of this neg-ative charge is about 6 x 10-10 electrostatic units, and isequal to the positive charge carried by the hydrogenatom in the electrolysis of solutions.

In gases at low pressures these units of negative elec-tric charge are always associated with carriers of a def-inite mass. This mass is exceedingly small, being onlyabout 1.4 x 10-3 of that of the hydrogen ion, the small-est mass hitherto recognized as capable of a separate

existence. The production of negative electrification thusinvolves the splitting up of an atom, as from a collectionof atoms something is detached whose mass is less thanthat of a single atom.

Thus was the first elementary particle finally discoveredand the field of particle physics born. Educated atCambridge as a mathematical physicist, Thomsonseems to have grasped the importance of his break-through almost immediately. For he ended his secondpaper with some bold speculations about its ultimatesignificance:

From what we have seen, this negative ion must be aquantity of fundamental importance in any theory ofelectrical action; indeed, it seems not improbable thatit is the fundamental quantity in terms of which allelectrical processes can be expressed. For, as we haveseen, its mass and its charge are invariable, indepen-dent both of the processes by which the electrificationis produced and of the gas from which the ions are setfree. It thus possesses the characteristics of being a fun-damental conception in electricity; and it seems desir-able to adopt some view of electrical action which bringsthis conception into prominence.

Within a few years most physicists recognized Thom-son’s new particle by the name “electron,” the termGeorge Stoney had coined for the fundamental unit ofcharge (see main text). But Thomson stuck resolutely byhis beloved “corpuscle” and still refused to call it any-thing else upon receiving the 1906 Nobel Prize inPhysics “in recognition of the great merits of his theoreti-cal and experimental investigations on the conduction ofelectricity by gases.” —M.R.

Figure from Thomson’s firstpaper (together with explana-tory text) illustrating the appa-ratus he used to measure e/m.

14 SPRING 1997

of a new particle with a very small mass onthe atomic scale. However, he was not thefirst to make that guess. In order to explain,I need to introduce two other players in thefield.

The first is Emil Wiechert (1861–1928), thena Privatdozent at the University of Konigs-berg. In the course of a lecture before Konigs-berg’s Physical Economical Society, onJanuary 7, 1897, he stated his conclusionabout cathode rays14 to which his experi-ments had led him: “It showed that we are

not dealing with the atoms known from chemistry, because the massof the moving particles turned out to be 2000–4000 times smaller thanthe one of hydrogen atoms, the lightest of the known chemicalatoms.” It was the first time ever that a subatomic particle is men-tioned in print and sensible bounds for its mass are given. Howev-er, these conclusions depended crucially on his assumption aboutthe charge. “Als Ladung ist 1 Elektron angenommen” (the chargeis assumed to be one electron) he stated, using Stoney’s terminology.

The second person is Walter Kaufmann (1871–1947), then Assis-tent at the University of Berlin, whose cathode-ray experimentshad taught him two crucial points.15 First, e/m for his rays was a con-stant, the same for whatever residual gas remained in his Geisslertube. That greatly puzzled him: “This assumption [of constant e/m]is physically hard to interpret; for if one makes the most plausibleassumption that the moving particles are ions [in the electrolyticsense] then e/m should have a different value for each gas.” Fur-thermore there was, as he perceived it, a second difficulty. Assum-ing e/m to be a constant, his measurements gave him about 107 emu/gfor the value of e/m, “while for a hydrogen ion [e/m] equals only 104.”Thus, he stated, “I believe to be justified in concluding that the hy-pothesis of cathode rays as emitted particles is by itself inadequatefor a satisfactory explanation of the regularities I have observed.”

Clearly Kaufmann was a fine experimentalist who, however,lacked the chutzpah of Thomson, who on August 7, 1897, submit-ted his memoir on cathode rays.16 His first determination of e/myielded a value 770 times that of hydrogen. He observed (see box

The vacuum tubeused by Thomsonin his discovery ofthe electron. (Courtesy ScienceMuseum/Science &Society Picture Library, London)

BEAM LINE 15

on pages 12 and 13) that, “The smallness of m/e may be due to thesmallness of m or the largeness of e, or to a combination of thesetwo.” He went on to argue in favor of the smallness of m, “Thuson this view we have in the cathode rays matter in a new state, a statein which the subdivision of matter is carried very much further thanin the ordinary gaseous state: a state in which all matter . . . is of oneand the same kind; this matter being the substance from which allthe chemical elements are built up.”

As I see it, Thomson’s finest hour as an experimentalist camein 1899 when he applied the methods just described to photoelec-trically produced particles and concluded—he was the first to do so!—that these particles were electrons: “The value of m/e in the caseof ultraviolet light . . . is the same as for cathode rays.”17 In the samepaper he announced his experimental results for the value of e, ob-tained by a method recently discovered by his student C. T. R. Wil-son, who had found that charged particles can form nuclei aroundwhich supersaturated water vapor condenses. Thomson’s measure-ment of e is one of the earliest applications of this cloud-chambertechnique. He determined the number of charged particles by dropletcounting, and their overall charge by electrometric methods, arriv-ing at e ~ 6.8×10-10 esu, a very respectable result in view of the nov-elty of the method. And that is why Thomson is the discoverer of theelectron.

When Thomson addressed a joint meeting of British and Frenchscientists in Dover in 1899, most doubts had been resolved. He quot-ed a mass of 3×10-26 g for the electron, the right order of magni-tude. The atom had been split. “Electrification essentially involvesthe splitting up of the atom, a part of the mass of the atom gettingfree and becoming detached from the original atom.”18

ENVOITO DEFINE the “birth of an era” is perhaps best left for parlor games.Let me write of the birth of particle physics nevertheless, define itto take place in 1897, and appoint Wiechert, Kaufmann and Thom-son as keepers at the gate. Their respective experimental arrange-ments are of comparable quality, their experimental results equal-ly good. Kaufmann’s observation that certain properties of cathode

16 SPRING 1997

rays are independent of the nature of the gas they traverse is, wewould say, a clear indication the universality of the constitution ofthese rays. The value for e/m he obtained is a good one. Had he addedone conjectural line to his paper, something like, “If we assume eto be the fundamental unit of charge identified in electrolysis, thencathode rays must be considered to be a new form of matter,” hewould have shared equal honors with Thomson for advances madein 1897. Perhaps the thought never struck him, perhaps it did but wasrejected as too wild. Perhaps also the Berlin environment was notconducive to uttering speculations of this kind, as is evidenced bya recollection about the year 1897: “I heard John Zeleny say that hewas in Berlin at that time, working in the laboratory of Warburg.When the discovery of the electron was announced, nobody in Berlinwould believe in it.”19 It may not have been known at that time whatwent through Kaufmann’s mind; it certainly is not known now.

It is fitting to conclude with a line from one of my favorite es-says: “On History,” by Thomas Carlyle20: “No hammer in theHorologe of Time peals through the universe when there is a changefrom Era to Era. Men understand not what is among their hands.”

1E. N. da C. Andrade, Rutherford and the Nature of the Atom (New York: Doubleday,1964), p. 48.

2R. Descartes, Principles of Philosophy, Part 2, Principle 20. Translated by E. Haldane andG. Ross (New York: Dover, 1955).

3A. Williamson, J. Chem. Soc. 22 (1869), 328.4T. Young, Miscellaneous Works, Vol. 9, (New York: Johnson Reprint, 1972), p. 461.5J. C. Maxwell, Collected Works, Vol. 2, (New York: Dover, 1952), p. 361.6J. D. van der Waals, (Ph.D. diss., Sythoff, Leiden 1873).7A. W. Rucker, J. Chem. Soc. 53 (1888), 222.8Maxwell, Collected Works, pp. 376–77.9H. von Helmholtz, in Selected Writings by Hermann von Helmholtz, ed. R.Kahl (Wesleyan

Univ. Press, 1971), p. 409.10G. J. Stoney, Phil. Mag. 11 (1881), 381.11———, Trans. Roy. Dublin Soc. 4 (1888–92), 563.12E. Rutherford, Radioactive Transformations (London: Constable, 1906), pp. 1, 16.13M. Faraday, Philos. Trans. Roy. Soc. 128 (1838), 125.14E. Wiechert, Schriften der Phys.-Okon. Ges. zu Konigsberg 38 (1897), 3.15W. Kaufmann, Ann. der Phys. und Chem. 61 (1897), 544.16J. J. Thomson, Phil. Mag. 44 (1897), 310–12.17———, Phil. Mag. 48 (1899), 547.18Ibid., p. 565.19G. Jaffe, J. Chem. Educ. 29 (1952), 230.20T. Carlyle, “On History,” in The Varieties of History, ed. F. Stern (New York: Vintage, 1973).

NOTES

BEAM LINE 17

HEN A STRANGER, hearing that I am a physi-

cist, asks me in what area of physics I work, I

generally reply that I work on the theory of elementary

particles. Giving this answer always makes me nervous.

Suppose that the stranger should ask, “What is an elemen-

tary particle?” I would have to admit that no one really

knows. Let me declare first of all that there is no difficulty in saying

what is meant by a particle. A particle is simply a physical system

that has no continuous degrees of freedom except for its total mo-

mentum. For instance, we can give a complete description of an

electron by specifying its momentum, as well as its spin around any

given axis, a quantity that in quantum mechanics is discrete rather

than continuous. On the other hand, a system consisting of a free

electron and a free proton is not a particle, because to describe it one

has to specify the momenta of both the electron and the proton—

not just their sum. But a bound state of an electron and a proton,

such as a hydrogen atom in its state of lowest energy, is a particle.

Everyone would agree that a hydrogen atom is not an elementary

particle, but it is not always so easy to make this distinction, or

even to say what it means.

What Is An Elementary Particle?by STEVEN WEINBERG

Copyright © 1996 by Steven Weinberg. Research supported in part by the Robert A. Welch Foundationand NSF Grant PHY 9511632.

W

18 SPRING 1997

FOR THE FIRST FEW decadesof this century there did notseem to be any trouble in say-

ing what is meant by an elementaryparticle. J. J. Thomson could use theelectric field in a cathode-ray tube topull electrons out of atoms, so atomswere not elementary. Nothing couldbe pulled or knocked out of electrons,so it seemed that electrons were el-ementary. When atomic nuclei werediscovered in Ernest Rutherford’s lab-oratory in 1911, it was assumed thatthey were not elementary, partly be-cause it was known that some ra-dioactive nuclei emit electrons andother particles, and also because nu-clear charges and masses could be ex-plained by assuming that nuclei arecomposed of two types of elementaryparticles: light, negatively chargedelectrons and heavy, positivelycharged protons.

Even without a definite idea ofwhat is meant by an elementary par-ticle, the idea that all matter consistsof just two types of elementary par-ticle was pervasive and resilient in away that is difficult to understand to-day. For instance, when neutronswere discovered by James Chadwickin 1932, it was generally assumedthat they were bound states of pro-tons and electrons. In his paper an-nouncing the discovery, Chadwickoffered the opinion: “It is, of course,possible to suppose that the neutronis an elementary particle. This viewhas little to recommend it at present,except the possibility of explainingthe statistics of such nuclei as N14.”(One might have thought this wasa pretty good reason: molecular spec-tra had revealed that the N14 nucle-us is a boson, which is not possibleif it is a bound state of protons and

electrons.) It was the 1936 discoveryof the charge independence of nuclearforces by Merle Tuve et al. thatshowed clearly that neutrons andprotons have to be treated in thesame way; if protons are elementary,then neutrons must be elementarytoo. Today, in speaking of protonsand neutrons, we often lump themtogether as nucleons.

This was just the beginning of agreat increase in the roster of so-called elementary particles. Muonswere added to the list in 1937 (thoughtheir nature was not understood un-til later), and pions and strange par-ticles in the 1940s. Neutrinos hadbeen proposed by Wolfgang Pauli in1930, and made part of beta-decaytheory by Enrico Fermi in 1933, butwere not detected until the Reines-Cowan experiment of 1955. Then inthe late 1950s the use of particle ac-celerators and bubble chambers re-vealed a great number of new parti-cles, including mesons of spin higherthan 0 and baryons of spin higherthan 1/2, with various values forcharge and strangeness.

On the principle that—even ifthere are more than two types of el-ementary particles—there reallyshould not be a great number oftypes, theorists speculated that mostof these particles are composites ofa few types of elementary particles. But such bound states would have tobe bound very deeply, quite unlikeatoms or atomic nuclei. For instance,pions are much lighter than nucle-ons and antinucleons, so if the pionwere a bound state of a nucleon andan antinucleon, as proposed by Fer-mi and Chen-Ning Yang, then itsbinding energy would have to belarge enough to cancel almost all of

James Chadwick who discovered theneutron in 1932. (Courtesy AIP MeggersGallery of Nobel Laureates)

BEAM LINE 19

the mass of its constituents. Thecomposite nature of such a particlewould be far from obvious.

How could one tell which of theseparticles is elementary and whichcomposite? As soon as this questionwas asked, it was clear that the oldanswer—that particles are elemen-tary if you can’t knock anything outof them—was inadequate. Mesonscome out when protons collide witheach other, and protons and antipro-tons come out when mesons collidewith each other, so which is a com-posite of which? Geoffrey Chew andothers in the 1950s turned thisdilemma into a point of principle,known as “nuclear democracy,”which held that every particle maybe considered to be a bound state ofany other particles that have the ap-propriate quantum numbers. Thisview was reflected decades later in a1975 talk to the German Physical So-ciety by Werner Heisenberg, whoreminisced that:

In the experiments of the fiftiesand sixties . . . many new particleswere discovered with long andshort lives, and no unambiguousanswer could be given any longerto the question about what theseparticles consisted of, since thisquestion no longer has a rationalmeaning. A proton, for example,could be made up of neutron andpion, or Lambda-hyperon and kaon,or out of two nucleons and an anti-nucleon; it would be simplest of allto say that a proton just consists ofcontinuous matter, and all thesestatements are equally correct orequally false. The difference be-tween elementary and compositeparticles has thus basically disap-peared. And that is no doubt themost important experimental dis-covery of the last fifty years.

LONG BEFORE Heisenbergreached this rather exaggerat-ed conclusion, a different sort

of definition of elementary particlehad become widespread. From theperspective of quantum field theory,as developed by Heisenberg, Pauli,and others in the period 1926–34, thebasic ingredients of Nature are notparticles but fields; particles such asthe electron and photon are bundlesof energy of the electron and the elec-tromagnetic fields. It is natural to de-fine an elementary particle as onewhose field appears in the funda-mental field equations—or, as the-orists usually formulate these the-ories, in the Lagrangian of the theory.It doesn’t matter if the particle isheavy or light, stable or unstable—ifits field appears in the Lagrangian, itis elementary; if not, not.

This is a fine definition if oneknows the field equations or the La-grangian, but for a long while physi-cists didn’t. A fair amount of theo-retical work in the 1950s and 1960swent into trying to find some objec-tive way of telling whether a givenparticle type is elementary or com-posite when the underlying theory is

Werner Heisenberg, left, talking withNeils Bohr at the CopenhagenConference, Bohr Institute, 1934.(Courtesy AIP Emilio Segrè Visual Archives)

P. E

hren

fest

, Jr.

20 SPRING 1997

becomes incorrect, and instead weget a formula for the scattering lengthin terms of the nucleon mass, thedeuteron binding energy, and thefraction of the time that the deuteronspends as an elementary particle (thatis, the absolute value squared of thematrix element between the physi-cal deuteron state and the elemen-tary free-deuteron state). Comparingthis formula with experimentshowed that the deuteron spendsmost of its time as a composite par-ticle. Unfortunately, arguments ofthis sort cannot be extended todeeply bound states, such as thoseencountered in elementary particlephysics.

The lack of any purely empiricalway of distinguishing composite andelementary particles does not meanthat this distinction is not useful. Inthe 1970s the distinction between el-ementary and composite particlesseemed to become much clearer,with the general acceptance of aquantum field theory of elementaryparticles known as the StandardModel. It describes quark, lepton, andgauge fields, so these are the ele-mentary particles: six varieties or“flavors” of quarks, each coming inthree colors; six flavors of leptons,including the electron; and twelvegauge bosons, including the photon,eight gluons, and the W+, W–, and Z0

particles. The proton and neutronand all of the hundreds of mesons andbaryons discovered after World War IIare not elementary after all; they are

not known. This turned out to bepossible in certain circumstancesin nonrelativistic quantum me-chanics, where an elementary par-ticle might be defined as one whosecoordinates appear in the Hamilton-ian of the system. For instance, atheorem due to the mathematicianNorman Levinson shows how tocount the numbers of stable non-elementary particles minus the num-ber of unstable elementary particlesin terms of changes in phase shiftsas the kinetic energy rises from zeroto infinity. The trouble with usingthis theorem is that it involves thephase shifts at infinite energy, wherethe approximation of nonrelativisticpotential scattering clearly breaksdown.

I worried about this a good deal inthe 1960s, but all I could come upwith was a demonstration that thedeuteron is a bound state of a protonand neutron. This was not exactlya thrilling achievement—everyonehad always assumed that thedeuteron is a bound state—but thedemonstration had the virtue of re-lying only on nonrelativistic quan-tum mechanics and low-energy neu-tron-proton scattering data, withoutany specific assumptions about theHamiltonian or about what happensat high energy. There is a classic for-mula that gives the spin triplet s-wave neutron-proton scatteringlength in terms of the nucleon massand the deuteron binding energy, butthe derivation of this formula actu-ally relies on the assumption that thedeuteron is a bound state. If we as-sume instead that the free-particlepart of the Hamiltonian contains anelementary deuteron state, then thisformula for the scattering length

composites of quarks and gluons, notbecause we can knock quarks andgluons out of them, which is believedto be impossible, but because that isthe way they appear in the theory.

The one uncertain aspect of theStandard Model is the mechanismthat breaks the electroweak gaugesymmetry and gives the W and Z par-ticles their masses, thereby addingan extra helicity state to what wouldhave been the two helicities of amassless W or Z particle of spin 1.Theories of electroweak symmetrybreakdown fall into two categories,according to whether these extra he-licity states are elementary, as in theoriginal form of the Standard Model,or composite, as in so-called tech-nicolor theories. In a sense, the primetask driving the design of both theLarge Hadron Collider and the ill-fated SSC was to settle the questionof whether the extra helicity statesof the W and Z particles are ele-mentary or composite particles.

THIS MIGHT have been theend of the story, but since thelate 1970s our understanding

of quantum field theory has takenanother turn. We have come to un-derstand that particles may be de-scribed at sufficiently low energiesby fields appearing in so-called ef-fective quantum field theories,whether or not these particles are tru-ly elementary. For instance, eventhough nucleon and pion fields donot appear in the Standard Model, wecan calculate the rates for processesinvolving low-energy pions and nu-cleons by using an effective quantumfield theory that involves pion andnucleon fields rather than quark and

We will not be

able to say

which particles

are elementary

until we have

a final theory

of force and matter.

BEAM LINE 21

gluon fields. In this field theory pi-ons and nucleons are elementary,though nuclei are not. When we usea field theory in this way, we are sim-ply invoking the general principlesof relativistic quantum theories, to-gether with any relevant symmetries;we are not really making any as-sumption about the fundamentalstructures of physics.

From this point of view, we are en-titled only to say that the quarks andgluons are more elementary than nu-cleons and pions, because their fieldsappear in a theory, the StandardModel, that applies over a muchwider range of energies than the ef-fective field theory that describes nu-cleons and pions at low energy. Wecannot reach any final conclusionabout the elementarity of the quarksand gluons themselves. The StandardModel itself is probably only aneffective quantum field theory,which serves as an approximation tosome more fundamental theorywhose details would be revealed atenergies much higher than thoseavailable in modern accelerators, andwhich may not involve quark, lep-ton, or gauge fields at all.

One possibility is that the quarksand leptons and other particles of theStandard Model are themselves com-posites of more elementary particles.The fact that we see no structure inthe quarks and leptons only tells usthat the energies involved in theirbinding must be quite large—largerthan several trillion electron volts.But so far no one has worked out aconvincing theory of this sort.

We will not be able to give a fi-nal answer to the question of whichparticles are elementary until wehave a final theory of force and

matter. When we have such a theo-ry, we may find that the elementarystructures of physics are not parti-cles at all. Many theorists think thatthe fundamental theory is somethinglike a superstring theory, in whichquarks, leptons, etc. are just differ-ent modes of vibration of the strings.It seems impossible in principle toidentify one set of strings as truly el-ementary, because, as recently real-ized, different string theories withdifferent types of strings are oftenequivalent.

There is a lesson in all this. Thetask of physics is not to answer a setof fixed questions about Nature, suchas deciding which particles are ele-mentary. We do not know in advancewhat are the right questions to ask,and we often do not find out until weare close to an answer.

Elementary particles today. There arethree known families of quarks andleptons in the Standard Model.

22 SPRING 1997

of my grandparents, there lived distin-guished scientists who did not believe in atoms. Within thelifetime of my children, there lived distinguished scientists whodid not believe in quarks. Although we can trace the notionof fundamental constituents of matter—minimal parts—to theancients, the experimental reality of the atom is a profoundlymodern achievement. The experimental reality of the quark ismore modern still.

Through the end of the nineteenth century, controversyseethed over whether atoms were real material bodies or merelyconvenient computational fictions. The law of multiple pro-portions, the indivisibility of the elements, and the kinetictheory of gases supported the notion of real atoms, but it waspossible to resist because no one had ever seen an atom. Oneof the founders of physical chemistry, Wilhelm Ostwald, wroteinfluential chemistry textbooks that made no use of atoms. Thephysicist, philosopher, and psychologist Ernst Mach likened“artificial and hypothetical atoms and molecules” to algebraicsymbols, tokens devoid of physical reality that could bemanipulated to answer questions about nature.

Atoms became irresistibly real when they began to comeapart, with the discovery of the electron that we celebrate in

BEAM LINE 23

this special anniversary issue. In the end the atomists won notbecause they could see atoms—atoms are far too small to see—but because they learned to determine the size and weight ofa single atom. In 1908 Jean-Baptiste Perrin established that theerratic “Brownian” movement of microscopic particles sus-pended in liquid was caused by collisions with molecules of thesurrounding medium. This demonstration of the mechanicaleffects of tiny atoms and molecules effectively ended skepti-cism about their physical reality. Ostwald announced his con-version in 1909, the year he won the Nobel Prize. Mach wentto his grave in 1916, still fighting a futile rear-guard action.

It is tempting to date the vanishing of resistance to the quarkmodel to the discovery of the J/ψ particle in November 1974,but a look at the theoretical papers in the famous January 6,1975, issue of Physical Review Letters will remind us thatthe epiphany wasn’t quite universal. The observation of the ψ',

24 SPRING 1997

what the particles were like. So thereI was, tabula rasa among the experts.

I could understand a little of theopening address by Murray Gell-Mann and a talk on symmetries byRichard Dalitz of Oxford. Both ofthem talked—rather cautiously, itseemed—about hypothetical objectscalled quarks as fundamental con-stituents ofthe protonand neu-tron and allthe others t r o n g l yinteractingparticles.Althoughthe ideathat threeq u a r k smade up aproton while a quark and antiquarkmade up a meson brought order toa lot of information, it was clear thatnobody had any idea how this couldhappen and whether there could bea self-consistent theory. And besides,no one had seen a quark.

Just as the Greek atomists hadtheir opponent in Anaxagoras, whoadvocated an infinite progression of

a second new particle that was ob-viously related to the J/ψ, made thenotion of quarks as mechanical ob-jects irresistible to all but an obdu-rate few. The holdouts were eitherconverted or consigned to a just ir-relevance by the discovery of charmeighteen months later.

MEETING THE QUARK

My first contact with quarks cameduring the summer of 1966, as I wasabout to begin graduate school inBerkeley. Before I had set foot in aclassroom, the Thirteenth Interna-tional Conference on High EnergyPhysics took place on campus, agathering of about four hundred sci-entists from around the world.Though attendance was by invita-tion, with strict national quotas, Icould present myself at the front doorof Wheeler Auditorium in the morn-ing and obtain a day pass that al-lowed me to sit inconspicuously inthe back of the room and watch theproceedings. Except for what I hadlearned that summer workingthrough two little books by RichardFeynman, I knew nothing of the in-teractions between particles, or even

Proceedings of the XIIIth InternationalConference on High-Energy Physics,

Berkeley, CaliforniaAugust 31–September 7, 1966

ContentsForeword

Preface

Conference Organizing Comitttee

Introductory SessionCURRENT TOPICS IN PARTICLE PHYSICS

M. Gell-Mann, Speaker

Session 1FUNDAMENTAL THEORETICAL QUESTIONS

M. Froissart, Rapporteur

Session 2WEAK INTERACTIONS

N. Cabibbo, Rapporteur

Session 3CURRENT COMMUTATORS

R. F. Dashen, Rapporteur

Session 4EXPERIMENTS IN T, C VIOLATION

V. L. Fitch, Rapporteur

Session 5OTHER ASPECTS OF WEAK INTERACTIONS: HIGH ENERGY NEUTRINO PHYSICS AND

QUESTIONS OF C, P, T NONINVARIANCE

T. D. Lee, Rapporteur

Session 6ELECTRODYNAMIC INTERACTIONS

S. Drell, Rapporteur

Session 7BOSON RESONANCES

G. Goldhaber, Rapporteur

Session 8PERIPHERAL COLLISIONS AT INTERMEDIATE

ENERGIES

J. D. Jackson, Rapporteur

Session 9aS = 0 BARYON RESONANCES

P. G. Murphy, Rapporteur

Session 9bBARYON RESONANCES WITH S ≠ 0

M. Ferro-Luzzi, Rapporteur

Session 10SYMMETRIES AND THE STRONG INTERACTIONS

R. H. Dalitz, Rapporteur

Session 11DYNAMICS OF STRONG INTERACTIONS

F. Low, Rapporteur

Session 12HADRON COLLISIONS AT VERY HIGH ENERGIES

L. Van Hove, Rapporteur

Richard Dalitz in 1961.

Cou

rtes

y G

.-C

. Wic

k an

d A

IP E

mili

o S

egrè

Arc

hive

s

BEAM LINE 25

seeds within seeds—and no minimalparts—the quark advocates had toface the challenge of “nuclear democ-racy.” Berkeley, it turned out, wasthe hotbed of an opposing point ofview: that there were no fundamen-tal constituents, that all the com-

posite “ele-mentary”p a r t i c l e swere some-how madeout of eachother in ani n t r i c a t einterplaycalled thebootstrap.Gell-Manndeflectedthis chal-lenge by re-p e a t e d l ys t ress ingthat quarksdidn’t have

to be real to be useful and that if themesons and baryons were made up of“mathematical quarks,” then thequark model might perfectly well becompatible with the bootstraphypothesis.

There was also the question ofhow to deal with interactions, withtheorists divided into sects promot-ing “S-matrix theory,” or “field the-ory,” or “Lagrangian field theory,” or“abstract field theory.” Gell-Mannurged the partisans to stop wastingtheir breath on sectarian quarrels andto pool their energies to campaign fora higher-energy accelerator thatwould enable us to really learn moreabout the basic structure of matter.That accelerator sweeps across theprairie outside my office window.

QUARKS IN BERKELEY?

Berkeley was indeed the MotherChurch of the S-matrix bootstrapdenomination. I don’t think quarkswere ever mentioned in GeoffChew’s course on the dynamics ofstrong interactions. Even in DaveJackson’s more catholic version ofthe course, quarks appeared onlyonce, on a list of term-paper topicsat the end of the year. But that wasonly part of the story. Learning aboutother approaches to particles and in-teractions was not only encouraged,it was obligatory. Berkeley graduatestudents were expected to follow twoyear-long courses in field theory. TheRad Lab was a center of hadron spec-troscopy where the quark model wasdiscussed as a classification tool. Inthe spring of 1968, George Zweig flew

Opposite: Physicists attending the 1966International Conference on HighEnergy Physics at Berkeley, California.(Courtesy Lawrence Berkeley NationalLaboratory)

Right: Geoffrey Chew in the 1960s.(Courtesy Lawrence Berkeley NationalLaboratory and AIP Emilio Segrè VisualArchives)

up from Caltech every Friday to teacha graduate seminar on the quark mod-el. Georgewas one ofthe inven-tors ofq u a r k s .He alsoknew ev-erythinga b o u tresonances p e c t r aand decays,and he gleefully showed us howmuch a simple quark model couldexplain.

What the quark model couldn’t ex-plain was itself: “How could this betrue?” was the question everyone hadto ask. Until the interactions ofquarks could be understood, the rules

The author in 1970, as afresh Ph.D. and researchassociate in the Institutefor Theoretical Physicsat the State University

of New York, Stony Brook.

George Zweig in 1965.

Cou

rtes

y G

. Zw

eig

Cou

rtes

y C

. Qui

gg

26 SPRING 1997

measure the number of colors of eachquark species, it really was three.And color would turn out to be thekey to explaining how the quarkmodel could be true.

The other evidence that drew at-tention to quarks arose from the MIT-SLAC experiments in which JerryFriedman, Henry Kendall, Dick Tay-lor, and their colleagues studied thestructure of the proton. To the pre-pared mind, the high rate of inelas-tic collisions they observed showedthat there were within the protontiny charged bodies. No mind wasmore prepared to take the leap thanFeynman’s. Feynman presented hisinterpretation at a SLAC colloquiumthat occasioned my first pilgrimageacross the Bay. The colloquium wasthen held in the evening after whathas been described to me as a vint-ner’s dinner. Whatever the reason, Iremember both speaker and audienceas extremely exuberant. If an elec-tron scattered from one of the hy-pothetical tiny charged bodies, notthe whole proton, it was easy to un-derstand why the inelastic cross sec-tion was so large. Instead of mea-suring the delicacy of the proton, theMIT and SLAC experimenters weremeasuring the hardness of the littlebits. Feynman wasn’t prepared to saywhat the tiny charged parts of theproton were, so he called them “par-tons.” Everyone in the room musthave thought, “Quarks?”

Before long, Bj Bjorken and Man-ny Paschos had worked out the con-sequences of the quark-parton mod-el for electron scattering and neutrinoscattering. The success of their pre-dictions added to a gathering crisis.If the quark-partons acted as if they

for combining quarks seemed arbi-trary, even baseless. Then there wasthe problem of the baryons, mostacute for the Ω −, a spin-3–2 particlemade of three strange quarks. Ac-cording to the quark model, the wavefunction of the Ω − was symmetric,whereas the Pauli exclusion princi-ple of quantum mechanics—the ba-sis for all of atomic spectroscopy—demanded that it be antisymmetric.Either there was something diceyabout the quark model, or there wasmore to quarks than met the eye.Wally Greenberg’s proposal that eachquark flavor (up, down, and strange)came in three distinguishable “col-ors,” and that antisymmetry in colorbrought the quark model into con-formance with the exclusion prin-ciple, seemed to many like invokingthe tooth fairy. But in one of thosedelicious ironies that make researchso interesting, when we learned to

Richard Feynman lecturing on his partonmodel at SLAC in October 1968.

Hen

ry W

. Ken

dal

l

BEAM LINE 27

were free, independent objects whenexamined by energetic electrons,why didn’t the quarks come out andshow themselves? Gell-Mann derid-ed Feynman’s picture as the “put-on”model. Many theorists of my gen-eration found great sport in showingthat Bjorken’s scaling law, which wasimplied by the parton model, wasn’tpossible in this or that interactingfield theory. Like the quark model ofthe hadron resonances, the partonmodel could explain many things,but it couldn’t explain itself.

DYNAMICS, DYNAMICS,

DYNAMICS!

Some of the reasons why it took solong for the idea of quarks to be ac-cepted have to do with the humanfrailties of obtuseness, or obstinacy,or preoccupation with other matters.But others, the reasons of real im-portance, reflect the standards ofscientific evidence. The repeated fail-ure to find any free quarks sustainedthe idea that quarks were computa-tional fictions. The main sticking-point was the absence of any under-standing of how quarks behave asfree and independent objects in hardcollisions, and yet form compositesin which they are permanently con-fined. Without an understanding ofdynamics, quarks were a story, not atheory.

The great illumination came in1973, when David Gross and FrankWilczek in Princeton and DavidPolitzer at Harvard found that, aloneamong field theories, non-Abeliangauge theories could reconcile thepermanent confinement of quarks

with the relative independencethe parton model presumes. Inthese theories the interactionbetween two quarks dimin-ishes when they are close to-gether, but becomes anineluctable pull when thequarks move apart. This “as-ymptotic freedom” of thestrong interaction is just whatwas needed to understand theMIT-SLAC results—not just ina useful cartoon, but in a realtheory.

In what seemed like theblink of an eye, a new theory ofthe strong interactions was codified.Gell-Mann named it quantum chro-modynamics (QCD) to celebrate thecentral role of color as the strong-in-teraction charge and perhaps to ex-press the hope that it would becomeas fertile and robust as quantum elec-trodynamics, the phenomenally suc-cessful theory of electrons and pho-tons. Soon precise predictions emergedfor the subtle deviations from Bjorkenscaling that QCD predicted.

Even before the scaling violationsimplied by QCD were establishedthrough painstaking experimental ef-fort, asymptotically free gauge theo-ries gave us license to take the quarkmodel and the parton picture seri-ously. All at once, what we had gin-gerly called “as-if” models took onnew meaning. Now, the J/ψ was sucha thunderbolt that it needed no the-oretical stage-dressing to help it setthe community of particle physicistson its ear. Yet it was the insight ofasymptotic freedom that prepared usto read the clues charmonium of-fered, and change forever the way wethink about the structure of matter.

Murray Gell-Mann in 1972. (Courtesy CERN)

0

–1

–2

–1 0

Electric Charge

Stra

nge

nes

s

1

n

Σ°Σ+Σ–

Ξ–

Λ°

Ξ°

p

28 SPRING 1997

QUARKS, LEPTONS,GAUGE FIELDS

Today’s elementary particles, the lep-tons (νe, e), (νµ ,µ ), (ντ ,τ ), and thequarks (u,d), (c,s), (t,b), form one ofthe pillars of our understanding ofmatter and energy. To the limits ofour resolution, they are all spin-1_

2

particles with no internal structure.The quarks are color triplets that ex-perience the strong interactions. Theleptons, which have no color charge,do not.

The top quark has so far been seenin such small numbers that wehaven’t yet examined it as closely asthe others. If top is as ephemeral aswe think, with a lifetime less thana trillionth of a trillionth of a second,it is the purest quark—the only onethat does not form mesons orbaryons. We know a great deal aboutthe tau neutrino from the study ofτ and Z decays, but it would still besatisfying to execute a “three-neutrino experiment,” in which abeam of tau neutrinos interacts witha target to produce tau leptons thatlive for a millimeter or two beforethey decay. The DONUT (Direct Ob-servation of NU-Tau) experiment be-ing commissioned at Fermilabshould observe about 150 examplesof the ντ → τ transition.

The other essential foundation forour current understanding is the no-tion that symmetries—gauge sym-metries—determine the character ofthe fundamental interactions. LikeQCD, the electroweak theory fash-ioned by Sheldon Glashow, StevenWeinberg, and Abdus Salam is a non-Abelian gauge theory. The elec-troweak theory got its own boost inthe summer of 1973 when André

Lagarrigue and his colleagues in theGargamelle bubble-chamber exper-iment at CERN announced the firstobservation of weak neutral-currentinteractions. Although it would takethe discovery of the weak-force par-ticles W and Z and many years ofstudy, culminating in the contribu-tions of the Z factories at CERN andSLAC, to show how successful a cre-ation the electroweak theory is, itwas clear very soon that the gauge-field-theory approach to the interac-tions of quarks and leptons was theright path.

The electroweak theory suppliesa clue of profound significance: ourworld must have both quarks andleptons. Unless each pair of leptons(like the electron and its neutrino) isaccompanied by a pair of quarks (likeup and down), quantum correctionswill clash with the symmetries fromwhich the electroweak theory is de-rived, leaving it inconsistent. I takethis constraint as powerful encour-agement for a family relationshipjoining quarks and leptons, and for aunified theory of the strong, weak,and electromagnetic interactions.

Have we found all the quarks andleptons? We do not really know. Pre-cision measurements of the width ofthe Z resonance assure us that thereare no more normal generations withvery light neutrinos. But there couldwell be new families of quarks andleptons in which all the members aretoo massive to be produced in Z de-cays. We don’t know yet whether theneutrinos have any mass. If they do,we need to learn whether each neu-trino is its own antiparticle.

Even if we have already met all thequarks and leptons, we have goodreason to be open to the possibility

The first single-electron event fromGargamelle. The electron’s trajectorygoes from left to right, beginning at thearrow’s tip. The haloed black circles arelights to illuminate the bubble-chamberliquid. (Courtesy CERN)

BEAM LINE 29

of new kinds of matter. The astro-physical case for dark matter in thegalaxy is persuasive, and the evidencethat most of the matter in the Uni-verse is both nonluminous andunlike the stuff we know is highlysuggestive. Supersymmetry, whichis for the moment the most popu-lar candidate to extend the elec-troweak theory, implies a greatly ex-panded list of elementary particles,including spin-zero partners of thequarks and leptons.

If we take as our goal not merelydescribing the world as we find it, butunderstanding why the Universe isthe way it is, the array of elementaryparticles presents us with many chal-lenges. What makes a top quark a topquark, or an electron an electron?Can we calculate the masses of thequarks and leptons and the relativestrengths of the weak transitions be-tween quark flavors? Why are therethree generations?

ANOTHER LAYER OF STRUCTURE?

No experimental evidence except thehistory of molecules and atoms andprotons suggests that quarks and lep-tons are composite. However, thereis an undeniable aesthetic allure tothe notion that a complex world mayarise out of the combinatoria of a fewsimple parts. If today’s elementaryparticles are composite, we might beable to compute their masses, un-derstand the trebling of generations,and decipher the relationship ofquarks to leptons.

Some specific currents in theo-retical research also lead towardcomposite quarks and leptons. In dy-namical theories of electroweak sym-metry breaking such as technicolor,

the longitudinal components of theweak gauge bosons are composite.Why not the quarks and leptons, too?And a new approach to supersym-metric model-building, in whichstrong gauge interactions break thesupersymmetry, suggests that someof the quarks may be composite.

Composite models of quarks andleptons must differ in a crucial wayfrom familiar dynamical pictures. InQCD the pions are the lightest—near-ly massless—particles, while the pro-ton mass is set by the scale of bind-ing energy. A theory of quark andlepton compositeness must deliverfermions much lighter than the (sev-eral TeV, at least) binding energy ofthe constituents. Without a specif-ic composite model, we have no the-oretical clue for the scale on whichwe might resolve structure in our el-ementary particles. Nevertheless, wecan characterize the experimentalsignatures that composite quarks andleptons would leave.

At energies approaching such acompositeness scale, quarks and lep-tons that have size will interact atsuch short distances that they inter-penetrate and rearrange, or even ex-change, their constituents. In quark-quark scattering, the conventional

gluon exchange of QCD would besupplemented by a contact interac-tion whose strength is determinedby the size of the quarks. In p–p col-lisions, this new contribution wouldlead to an excess of hadron jets atlarge values of the transverse energy,where quark-antiquark scattering isthe dominant reaction. Typically, theangular distribution of the jets willdiffer from the shape QCD predicts.If quarks and leptons have commonconstituents, a similar excess will beseen in dilepton production from theelementary process q–q → l+l−. At stillhigher energies, we would expect tosee the effects of excited quarks andleptons. Finally, at energies wellabove this compositeness scale,quarks and leptons would begin tomanifest form factors characteris-tic of their size.

Since I first met the quark, charm,beauty, top, and the rest have becomemy friends and teachers—in fact,they have taken over my life. Theidea that elementary constituents ofmatter interact according to the dic-tates of gauge symmetries has be-come the organizing principle of par-ticle physics, as important to ourfield as evolution is to biology. I don’tknow how far the revolution ofquarks and leptons and gauge bosonswill carry us, but I have a wish forthe decade ahead: that we will learnthe true nature of electroweak sym-metry breaking and begin to under-stand the individuality of the quarksand leptons. And I hope we will lookback with pleasure and satisfactionat how passionate and optimistic—and naïve—we were in 1997.

The idea that elementary

constituents of matter

interact according

to the dictates of gauge

symmetries has become

the organizing principle

of particle physics, as

important to our field as

evolution is to biology.

30 SPRING 1997

MORE THAN A DECADE before J. J. Thomson discovered the elec-tron, Thomas Edison stumbled across a curious effect, patentedit, and quickly forgot about it. Testing various carbon filaments

for electric light bulbs in 1883, he noticed a tiny current trickling in a single di-rection across a partially evacuated tube into which he had inserted a metalplate. Two decades later, British entrepreneur John Ambrose Fleming appliedthis effect to invent the “oscillation valve,” or vacuum diode—a two-termi-nal device that converts alternating current into direct. In the early 1900ssuch rectifiers served as critical elements in radio receivers, converting radiowaves into the direct current signals needed to drive earphones.

In 1906 the American inventor Lee de Forest happened to insert another elec-trode into one of these valves. To his delight, he discovered he could influ-ence the current flowing through this contraption by changing the voltage onthis third electrode. The first vacuum-tube amplifier, it served initially as animproved rectifier. De Forest promptly dubbed his triode the audion and ap-plied for a patent. Much of the rest of his life would be spent in forming a se-ries of shaky companies to exploit this invention—and in an endless series oflegal disputes over the rights to its use.

These pioneers of electronics understood only vaguely—if at all—thatindividual subatomic particles were streaming through their devices. For them,electricity was still the fluid (or fluids) that the classical electrodynamicistsof the nineteenth century thought to be related to stresses and disturbancesin the luminiferous æther. Edison, Fleming and de Forest might have been dim-ly aware of Thomson’s discovery, especially after he won the 1906 Nobel

INDUSTRIAL STRENGTHby MICHAEL RIORDAN

TH

E

Copyright © 1996 by Michael Riordan. Adapted in part from Crystal Fire: The Birth of the Information Age, byMichel Riordan and Lillian Hoddeson, to be published in 1997 by W. W. Norton & Co.

BEAM LINE 31

Prize in physics. But this knowledgehad yet to percolate out of academ-ic research labs such as the Caven-dish and into industrial workshops.Although he had earned a Ph.D. inphysics from Yale, in his daily prac-tice de Forest remained pretty mucha systematic tinkerer in the Edison-ian vein, trying endless variations onhis gadgets in his halting attempts toimprove their performance.

VOLUMES COULD be writ-ten about the practical appli-cations that owe their exis-

tence to the understanding ofelectricity as a stream of subatomicparticles rather than a continuous flu-id. While the telephone clearly an-tedated the discovery of the electron,for example, its modern manifesta-tions—cellular and touchtone phones,telefax machines, satellite communi-cations—would be utterly impossiblewithout such knowledge. And theubiquitous television set is of coursejust a highly refined version of thecathode-ray tube that Thomson usedto determine the charge-to-mass ra-tio of his beloved corpuscle. The fieldof electronics, a major subfield ofelectrical engineering today, grew upin the twentieth century around thisnew conception of electricity, even-tually taking its name in the 1920sfrom the particle at its core. (We areperhaps fortunate that Thomson didnot prevail in his choice of nomen-clature!)

In parallel with the upsurge ofelectronics, and in some part due toit, came a sweeping transformationof industrial research in America.Once the main province of highlyindividualistic inventors search-ing for a fruitful breakthrough,

PARTICLE

technology development slowly be-came an organized practice per-formed by multidisciplinary teamsof salaried scientists and engineersworking in well-equipped industriallabs. As the century waxed and quan-tum mechanics emerged to explainthe mysterious behavior of electrons,atoms and molecules, these re-searchers increasingly sported ad-vanced degrees in physics or chem-istry. A deeper understanding of thescientific principles governing thebehavior of matter gradually becameindispensable to the practice of in-dustrial research. As the noted his-torian of technology Thomas Hugh-es put it, “Independent inventors hadmanipulated machines and dynamos;industrial scientists would manip-ulate electrons and molecules.”

Few examples illustrate this evo-lutionary transformation better thanthe case of the vacuum-tube ampli-fier. For almost a decade after de For-est invented it, his audion found lit-tle use beyond low-voltage appli-cations in wireless receivers—as adetector of weak radio signals. Hesimply did not understand that thegas remaining in his tube was im-peding the flow of electrons from fil-ament to plate. At the higher volt-ages required for serious amplifica-tion, say in telephone communica-tions, the device began, as one ob-server noted, “to fill with blue haze,seem to choke, and then transmit nofurther speech until the incomingcurrent had been greatly reduced.”

One corporation extremely inter-ested in amplifying telephone signalswas the American Telephone andTelegraph Company, then seeking todevelop a suitable “repeater” fortranscontinental phone service.

32 SPRING 1997

Among its leading scien-tists was Frank Jewett,then working in the engi-neering department of itsWestern Electric Division.In 1902 he had earned aPh.D. in physics from theUniversity of Chicago, do-ing his research under Al-bert Michelson and be-friending Robert Millikan.Harboring a hunch that theelectrical discharges inevacuated tubes mightserve as the basis for a suit-able repeater, Jewett ap-proached his old chum,who in 1911 sent one of hisbrightest graduate stu-

dents, Harold Arnold, to WesternElectric. Here was a young mansteeped in the new thinking, who hadjust spent several years measuringthe charges of individual electronson oil droplets.

When de Forest demonstrated hisaudion to Western Electric scientistsand engineers in October 1912,Arnold was present. He diagnosedthe blue haze as due to the recom-bination of gas molecules that hadbeen ionized by energetic electrons.Then he solved its problems by useof high vacuum, an oxide-coated fil-ament, and other modifications dic-tated by a superior understandingof the electronic discharge. (A similardevelopment occurred simultaneu-osly at General Electric, but it lostthe ensuing patent fight to AT&T,which had wisely purchased theappropriate rights to de Forest’spatents.)

Within a year Western Electricwas making “high-vacuum thermi-onic tubes” that served as active

elements in excellent telephone re-peaters. At the grand opening of thePanama-Pacific Expositon held inSan Francisco on January 15, 1915,Alexander Graham Bell inauguratedthe nation’s first coast-to-coasttelephone service, talking to his for-mer assistant Thomas Watson inNew York. Recalling this event in hisautobiography, Millikan observedthat “the electron—up to that timelargely the plaything of the scien-tist—had clearly entered the fieldas a patent agent in the supplyingof man’s commercial and industrialneeds.”

Thus convinced of the value of sci-entific research in an industrial set-ting, Western Electric incorporatedits engineering department as a sep-arate entity—the Bell Telephone Lab-oratories—in 1925, naming Jewett itsfirst president. The very next year,as an outgrowth of their research onthe performance of vacuum tubes(also called electron tubes), ClintonDavisson and Lester Germer estab-lished the wave nature of electrons,which had been predicted a few yearsearlier by Louis de Broglie. For hispivotal work on electron diffraction,Davisson was to share the 1937Nobel Prize in physics with theBritish scientist George Thomson,son of J. J.

Quantum mechanics soon ex-plained the behavior not only of elec-trons in atoms but of the largeensembles of them that swarm aboutfreely within metals. Based on thetheoretical work of Enrico Fermi andPaul Dirac, Bell Labs physicists even-tually figured out why an oxide-coating worked so well on tungstenfilaments of vacuum tubes. It helpedto lower the work function of the

J. J. Thomson inspecting electron tubesin 1923 with Frank Jewett, the firstpresident of Bell Labs. (Courtesy AT&TArchives and AIP Niels Bohr Library)

BEAM LINE 33

metal, thereby making it easier forelectrons to escape from the sur-face—and substantially reducing theamount of power needed to heat a fil-ament. Such a fundamental under-standing of the physics of electronsproved crucial to further engineeringadvances in vacuum tubes that savedAT&T millions of dollars annually.

IN THE LATE 1920S and early1930s, Felix Bloch, RudolphPeierls, Alan Wilson and other

European physicists laid the foun-dations of modern solid-state physicsin their theoretical studies of howwaves of electrons slosh about with-in the periodic potentials encoun-tered inside crystalline materials.Their work resulted in a theory ofsolids in which there are specific al-lowed (or forbidden) energy levels—called “bands”—that electrons can(or cannot) occupy, analogous to theBohr orbitals of early quantum the-ory. Combined with practical meth-ods of calculating these band struc-tures in actual substances, pioneeredby Eugene Wigner, band theory fos-tered a better understanding of whycertain materials act as electricalconductors and others as insulators.And, in a decade when electrontubes reigned supreme as the activecomponents of electronic circuits,band theory began to elucidate theproperties of intermediate materialscalled semiconductors, whose myr-iad spawn would eventually sup-plant these tubes throughout elec-tronics.

World War II spurred tremendouspractical advances in the technolo-gy of semiconductors, largely due tothe fact that microwave receiversneeded rectifiers able to operate

above a few hundred megahertz,where electron tubes had proved use-less. Crystal rectifiers, with a deli-cate metal point pressed into a ger-manium or silicon surface, filled thegap nicely. By the end of the War,methods of purifying and dopingthese substances to make easily con-trolled, well-understood semicon-ductors had been perfected by sci-entists at such secret enclaves as theRad Lab at MIT and Britain’s Tele-communications Research Estab-lishment at Great Malvern.

No laggard itself in these pursuits,Bell Labs led the way during the post-war years in applying wartimeinsights and technologies to the cre-ation of practical new semiconduc-tor components. “The quantumphysics approach to structure of mat-ter has brought about greatly in-creased understanding of solid-statephenomena,” wrote its vice presidentMervin Kelly—another of Millikan’sgrad students—in 1945, authorizingformation of a solid-state physicsgroup. “The modern conception ofthe constitution of solids that has re-sulted indicates that there are greatpossibilities of producing new and

Right: Clinton Davisson and LesterGermer with the apparatus they used to

establish the wave nature of electrons.(Courtesy AT&T Archives)

Bottom: Graph from their 1927 Naturearticle showing diffraction peaks

observed in electron scattering from anickel crystal.

useful properties by finding physicaland chemical methods of controllingthe arrangement of the atoms andelectrons which compose solids.”

The most important postwarbreakthrough to occur at Bell Labswas the invention of the transistorin late 1947 and early 1948 by JohnBardeen, Walter Brattain, and Wil-liam Shockley. And a key to their in-terpretation of transistor action wasa new physical phenomenon Shock-ley dubbed “minority carrier injec-tion”—in which electrons and pos-itively charged quantum-mechanicalentities called “holes” can flow bydiffusion in the presence of one an-other. Once again, a detailed scien-tific understanding of how individ-ual subatomic particles (and what,in certain respects, act like theirantiparticles) behave proved crucialto a pivotal advance in electronics.

The transistor happened along ata critical juncture in technologicalhistory. For the electronic digitalcomputers that also emerged fromwartime research could not haveevolved much further without it.The thousands of bulky, fragileelectron tubes used in such early,

34 SPRING 1997

laboratory, Fermi replied, “Theyreally are very fine gadgets, and Ihope very much that they might beuseful in our work.”

THOMSON’S DISCOVERYtriggered a spectacular cen-tury of innovation in both

science and technology. Paced by in-creasingly detailed knowledge of theelectron’s properties and behavior,scientists and engineers developedmany other advanced devices—lasers, light-emitting diodes, mi-crowave tubes, solar cells and high-speed microprocessors, to nameseveral—that are essential to mod-ern computing and global commu-nications. Today we know the massand charge of the electron to betterthan seven significant figures. Aidedby quantum mechanics, we can ac-curately calculate its energy levelsin all kinds of atoms, molecules andsolid-state substances. Largely tak-en for granted, such information iscrucial for the precision control ofelectrons at the submicron scalesthat characterize many leading-edgetechnologies.

Of critical importance in attain-ing this deep understanding was theease with which electrons can bedetached from other forms of matterand manipulated using electromag-netic fields. Such features were read-ily apparent in Thomson’s landmarkexperiments, for example, andMillikan exploited them in his re-search. In certain key instances theenergy required corresponds to thatof photons in visible light. Thisunique partnership between the elec-tron and photon (whose centennialwe will also celebrate in the not-too-distant future) is central to much of

room-filling com-puters as the ENI-AC and UNIVACburned out withall-too-frustratingfrequency. Onlylarge corporations,the armed servicesand government

agencies could afford these massive,power-hungry monstrosities and thevigilant staff to keep them operating.“It seems to me,” Shockley con-jectured in December 1949, “thatin these robot brains the transistor isthe ideal nerve cell.”

But the transistor has proved to bemuch more than merely a replace-ment for electron tubes and electro-mechanical switches. Shrunk to lessthan a ten-thousandth of its originalsize and swarming by the millionsacross the surfaces of microchips,it has opened up entirely unexpect-ed realms of electronic possibility,which even the most farsighted couldnot have anticipated during thosebooming postwar years. The transis-tor was, as historians Ernest Braunand Stuart MacDonald observed, “theharbinger of an entirely new sort ofelectronics with the capacity not justto influence an industry or a scientificdiscipline, but to change a culture.”

Characteristically, particle physi-cists were among the first to glimpsethe potential ramifications of thisrevolutionary new solid-state am-plifier. “I would be very anxious todo some experimenting to learnabout the techniques of your newGermanium triods,” wrote Fermi toShockley in early January 1949(misspelling the final word). After re-ceiving a few samples and testingthem at his University of Chicago

Lee de Forest, inventor of the vacuum-tube amplifier, and Bell Labs PresidentMervin Kelly. (Courtesy AT&T Archives)

BEAM LINE 35

advanced technology. Perhaps thiscomplementary relationship is onereason why, during the past decadeor so, we have witnessed the emer-gence of a “photonics” industry thatthreatens to supplant electronics insome commercial sectors.

No comparable partnership exists,for example, between quarks and glu-ons. Quarks are not detachable, atleast not yet, and gluon fields do notextend to infinity like the electro-magnetic. It makes all the differencein the world. We still have only roughvalues for the quark masses, espe-cially the up and down quarks thatmake up the bulk of ordinary mat-ter. Only a few wild-eyed speculatorsdare to mention the possibility of“quarkonics” or “gluonics” indus-tries. Maybe one day soon we willmanipulate quarks for practical ben-efit, say by using high-energy elec-tron beams. Perhaps, but I seriouslydoubt it will ever amount to muchof an industry. For now, quarks andgluons remain the playthings of purephysics.

Just a hundred years after its dis-covery, the electron sits at the coreof modern life. Where previous gen-erations of writers, for example, usedchemical inks and mechanicaldevices to cast their ideas onto paper(and deliver them to readers), I have

composed this arti-cle while staring at aluminous screen—behind which a blaz-ing stream of elec-trons traces out myfumbling thoughtsas my fingertips tapkeys that activatetransistor-laden mi-croprocessors deepwithin my comput-er. And some of younow read my wordson luminous screensof your own, con-veyed to your desksby surging rivers of electrons andphotons pulsating as ones and zeroesthrough an intricate network thatstretches into almost every corner ofthe globe. The only paper involvedin the exchange is the pages markedwith red ink that now lie crumpledin my wastebasket.

We all owe a debt to J. J.Thomson—and the scientists andengineers who followed the path thathe pioneered—for taming the firstsubatomic particle and adapting itsunique properties for the practicalapplications that are relentlessly re-defining what it means to be human.

Left: John Bardeen, William Shockley, and WalterBrattain, who shared the 1956 Nobel Prize in physicsfor the invention of the transistor. (Courtesy AT&TArchives)

Bottom: Page from Brattain’s lab notebook thatrecords the 23 December 1947 demonstration of thefirst transistor. (Courtesy AT&T Archives)

36 SPRING 1997

WHEN J. J. THOMSON discovered the electron, he did not call theinstrument he was using an accelerator, but an accelerator itcertainly was. He accelerated particles between two electrodes

to which he had applied a difference in electric potential. He manipulated theresulting beam with electric and magnetic fields to determine the charge-to-mass ratio of cathode rays. Thomson achieved his discovery by studying theproperties of the beam itself—not its impact on a target or another beam, as wedo today. Accelerators have since become indispensable in the quest tounderstand Nature at smaller and smaller scales. And although they are muchbigger and far more complex, they still operate on much the same physical prin-ciples as Thomson’s device.

It took another half century, however, before accelerators became entrenchedas the key tools in the search for subatomic particles. Before that, experi-ments were largely based on natural radioactive sources and cosmic rays. ErnestRutherford and his colleagues established the existence of the atomic nucleus—as well as of protons and neutrons—using radioactive sources. The positron,muon, charged pions and kaons were discovered in cosmic rays.

One might argue that the second subatomic particle discovered at an accel-erator was the neutral pion, but even here the story is more complex. That itexisted had already been surmised from the existence of charged pions, and theoccurrence of gamma rays in cosmic rays gave preliminary evidence for sucha particle. But it was an accelerator-based experiment that truly nailed downthe existence of this elusive object.

Evolutionof Particle Accelerators& Collidersby WOLFGANG K. H. PANOFSKY

Th

e

BEAM LINE 37

There followed almost two decades of accelerator-based discoveries of othersubatomic particles originally thought to be elementary, notably theantiproton and the vector mesons. Most of these particles have since turnedout to be composites of quarks. After 1970 colliders—machines using twoaccelerator beams in collision—entered the picture. Since then most, butcertainly not all, new revelations in particle physics have come from thesecolliders.

IN CONSIDERING the evolution of accelerator and collider technology, weusually think first of the available energy such tools provide. Fundamen-tally, this is the way it should be. When the study of the atomic nucleus

stood at the forefront of “particle physics” research, sufficient energy was neededto allow two nuclei—which are positively charged and therefore repel oneanother—to be brought close enough to interact. Today, when the componentsof these nuclei are the main objects of study, the reasons for high energy aremore subtle. Under the laws of quantum mechanics, particles can be describedboth by their physical trajectory as well as through an associated wave whosebehavior gives the probability that a particle can be localized at a given pointin space and time. If the wavelength of a probing particle is short, matter canbe examined at extremely small distances; if long, then the scale of things thatcan be investigated will be coarser. Quantum mechanics relates this wavelengthto the energy (or, more precisely, the momentum) of the colliding particles: thegreater the energy, the shorter the wavelength.

38 SPRING 1997

This relationship can be expressedquantitatively. To examine matter atthe scale of an atom (about 10−8 cen-timeter), the energies required are inthe range of a thousand electronvolts. (An electron volt is the energyunit customarily used by particlephysicists; it is the energy a parti-cle acquires when it is accelerated

across a potential difference of onevolt.) At the scale of the nucleus, en-ergies in the million electron volt—or MeV—range are needed. To ex-amine the fine structure of the basicconstituents of matter requires en-ergies generally exceeding a billionelectron volts, or 1 GeV.

But there is another reason for us-ing high energy. Most of the objectsof interest to the elementary parti-cle physicist today do not exist as freeparticles in Nature; they have to becreated artificially in the laboratory.The famous E = mc2 relationship gov-erns the collision energy E requiredto produce a particle of mass m.Many of the most interesting parti-cles are so heavy that collisionenergies of many GeV are needed tocreate them. In fact, the key to under-standing the origins of many para-meters, including the masses of theknown particles, required to maketoday’s theories consistent is believedto reside in the attainment of colli-sion energies in the trillion electronvolt, or TeV, range.

Our progress in attaining everhigher collision energy has indeedbeen impressive. The graph on theleft, originally produced by M. Stan-ley Livingston in 1954, shows howthe laboratory energy of the parti-cle beams produced by acceleratorshas increased. This plot has been up-dated by adding modern develop-ments. One of the first things to no-tice is that the energy of man-madeaccelerators has been growing ex-ponentially in time. Starting fromthe 1930s, the energy has increased—roughly speaking—by about a fac-tor of 10 every six to eight years. Asecond conclusion is that this spec-tacular achievement has resulted

1000 TeV

100 TeV

10 TeV

1 TeV

100 GeV

10 GeV

1 GeV

100 MeV

10 MeV

1 MeV

1930 1950 1970

Year of Commissioning

1990

Par

ticl

e E

ner

gy

Proton Storage Rings (equivalent energy)

Proton Synchrotrons

Electron Linacs

Synchrocyclotrons

Proton Linacs

Cyclotrons

Electron Synchrotrons

Sector-Focused Cyclotrons

Electrostatic Generators

Rectifier Generators

Betatrons

A “Livingston plot” showing acceleratorenergy versus time, updated to includemachines that came on line during the1980s. The filled circles indicate new orupgraded accelerators of each type.

BEAM LINE 39

from a succession of technologiesrather than from construction of big-ger and better machines of a giventype. When any one technology ranout of steam, a successor technologyusually took over.

In another respect, however, theLivingston plot is misleading. Itsuggests that energy is the primary,if not the only, parameter that definesthe discovery potential of an accel-erator or collider. Energy is indeedrequired if physicists wish to cross anew threshold of discovery, provid-ed that this threshold is defined bythe energy needed to induce a newphenomenon. But there are severalother parameters that are importantfor an accelerator to achieve—for ex-ample, the intensity of the beam, orthe number of particles acceleratedper second.

When the beam strikes a target,its particles collide with those in thetarget. The likelihood of producing areaction is described by a numbercalled the cross section, which is theeffective area a target particle pre-sents to an incident particle for thatreaction to occur. The overall inter-action rate is then the product of thebeam intensity, the density of targetparticles, the cross section of the re-action under investigation, and thelength of target material the incidentparticle penetrates. This rate, andtherefore the beam intensity, is ex-tremely important if physicists areto collect data that have sufficientstatistical accuracy to draw mean-ingful conclusions.

Another important parameter iswhat we call the duty cycle—the per-centage of time the beam is actual-ly on. Unlike Thomson’s device,most modern accelerators do not

provide a steady flow of particles,generally because that would requiretoo much electric power; instead, thebeam is pulsed on and off. Whenphysicists try to identify what reac-tion has taken place, one piece of ev-idence is whether the different par-ticles emerge from a collision at thesame time. Thus electronic circuitsregister the instant when a particletraverses a detector. But if the ac-celerator’s duty cycle is small, thenall the particles will burst forth dur-ing a short time interval. Thereforea relatively large number of acci-dental coincidences in time will oc-cur, caused by particles emergingfrom different individual reactions,instead of from real coincidences dueto particles emerging from a singleevent. If time coincidence is an im-portant signature, a short duty cycleis a disadvantage.

Then there is the problem of back-grounds. In addition to the reactionunder study, detectors will registertwo kinds of undesirable events.Some backgrounds arise from parti-cles generated by processes otherthan the beam’s interaction with thetarget or another beam—such as withresidual gas, from “halo” particlestraveling along the main beam, oreven from cosmic rays. Other back-grounds stem from reactions that are

M. Stanley Livingston and Ernest O.Lawrence, with their 27-inch cyclotron atBerkeley Radiation Laboratory. (CourtesyLawrence Berkeley National Laboratory)

already well understood and containno new information. Acceleratorsdiffer in terms of the presence or ab-sence of both kinds of backgrounds;their discovery potential differs ac-cordingly. The amount and kinds ofbackground are directly related to theease of data analysis, the type ofdetector to be built, or whether thedesired results can be extracted at all.

In general, as the energy increases,the number of possible reactions alsoincreases. So does the burden on thediscriminating power of detectorsand on the data-analysis potential ofcomputers that can isolate the“wheat” from the “chaff.” With thegrowth in energy indicated by theLivingston plot, there had to be a par-allel growth in the analyzing poten-tial of the equipment required toidentify events of interest—as wellas a growth in the number of peo-ple involved in its construction andoperation.

And finally there is the matterof economy. Even if a planned ac-celerator is technically capable ofproviding the needed energy,intensity, duty cycle, and low back-ground, it still must be affordableand operable. The resources re-quired—money, land, electric pow-er—must be sufficiently moderatethat the expected results will have

40 SPRING 1997

the creation of new particles. Thiscollision energy is less than the lab-oratory energy of the particles in abeam if that beam strikes a station-ary target. When one particle hits an-other at rest, part of the available en-ergy must go toward the kineticenergy of the system remaining afterthe collision. If a proton of low en-ergy E strikes another proton at rest,for example, the collision energy isE/2 and the remaining E/2 is the ki-netic energy with which the protonsmove ahead. At very high energiesthe situation is complicated by rel-ativity. If a particle of total energyE hits another particle of mass M,then the collision energy is givenby Ecoll ~ (2Mc2E)11/2, which is muchless than E/2 for E much larger thanMc2.

If two particles of equal mass trav-eling in opposite directions collidehead on, however, the total kineticenergy of the combined system aftercollision is zero, and therefore theentire energy of the two particles be-comes available as collision energy.This is the basic energy advantage of-fered by colliding-beam machines, orcolliders.

The idea of colliding-beam ma-chines is very old. The earliest

reference to their possibility stemsfrom a Russian publication of the1920s; it would not be surprising ifthe same idea occurred indepen-dently to many people. The first col-lider actually used for particle-physics experiments, built atStanford in the late 1950s, producedelectron-electron collisions (see pho-tograph on the left). Other early ma-chines, generating electron-positroncollisions, were built in Italy, Siberiaand France. Since then there has beena plethora of electron-positron,proton-proton and proton-antiprotoncolliders.

There is another problem, how-ever. If the particles participatingin a collision are themselvescomposite—that is, composed ofconstituents—then the availableenergy must be shared among theseconstituents. The threshold for newphenomena is generally defined bythe collision energy in the con-stituent frame: the energy that be-comes available in the interactionbetween two individual con-stituents. Here there are major dif-ferences that depend on whether theaccelerated particles are protons,deuterons, electrons or somethingelse. Protons are composed of threequarks and surrounded by variousgluons. Electrons and muons, as wellas quarks and gluons, are consideredpointlike, at least down to distancesof 10−16 centimeter. Recognizingthese differences, we can translatethe Livingston plot into anotherchart (top right, next page) showingenergy in the constituent frame ver-sus year of operation for colliding-beam machines.

But the idea of generating highercollision energy via colliding beams

The first colliding-beam machine, adouble-ring electron-electron collider,built by a small group of Princeton andStanford physicists. (Courtesy StanfordUniversity)

commensurate value. Of course “val-ue” has to be broadly interpreted interms not only of foreseeable or con-jectured economic benefits but alsoof cultural values related to the in-crease in basic understanding. Inview of all these considerations, thechoice of the next logical step in ac-celerator construction is always acomplex and frequently a contro-versial issue. Energy is but one ofmany parameters to be considered,and the value of the project has to besufficiently great before a decision togo ahead can be acceptable to thecommunity at large.

All these comments may appearfairly obvious, but they are frequentlyforgotten. Inventions that advancejust one of the parameters—in par-ticular, energy—are often proposedsincerely. But unless the other pa-rameters can be improved at thesame time, to generate an overall ef-ficient complex, increasing theenergy alone usually cannot lead tofundamentally new insights.

THE ENERGY that reallymatters in doing elementaryparticle physics is the colli-

sion energy—that is, the energy avail-able to induce a reaction, including

BEAM LINE 41

is worthless unless (as discussedabove) higher interaction rates canbe generated, too. To succeed, thedensity of the two beams must behigh enough—approaching that ofatoms in ordinary matter—and theirinteraction cross sections must besufficient to generate an adequatedata rate. In colliding-beam machinesthe critical figure is the luminosityL, which is the interaction rate persecond per unit cross section. Thebottom graph on this page illustratesthe luminosity of some of these ma-chines. In contrast to the constituentcollision energy, which has contin-ued the tradition of exponentialgrowth begun in the Livingston plot,the luminosity has grown muchmore slowly. There are good reasonsfor this trend that I will discussshortly.

Naturally there are differencesthat must be evaluated when choos-ing which particles to use in accel-erators and colliders. In addition tothe energy advantage mentioned forelectrons, there are other factors. Asprotons experience the strong inter-action, their use is desirable, at leastin respect to hadron-hadron inter-actions. Moreover, the cross sectionsinvolved in hadron interactions aregenerally much larger than those en-countered in electron machines,which therefore require higher lu-minosity to be equally productive.

Proton accelerators are generallymuch more efficient than electronmachines when used to produce sec-ondary beams of neutrons, pions,kaons, muons, and neutrinos. Butelectrons produce secondary beamsthat are sharply concentrated in theforward direction, and these beamsare less contaminated by neutrons.

LHC (CERN)

TEVATRON (Fermilab)

SP–PS

(CERN)

LEP II

SLC (SLAC)

LEP (CERN)

TRISTAN (KEK)

PETRA (DESY)

PEP (SLAC)

CESR (Cornell)

VEPP IV (Novosibirsk)SPEAR II

SPEAR (SLAC)

DORIS (DESY)

VEPP III (Novosibirsk)

ISR (CERN)

ADONE (Italy)

PRIN-STAN (Stanford)

VEPP II (Novosibirsk)

ACO (France)

1 GeV

10 GeV

100 GeV

1 TeV

10 TeV

Hadron Colliders e+e– Colliders

1960 1970 1980 1990 2000 2010

Year of First Physics

Con

stit

uen

t C

ente

r-of

-Mas

s E

ner

gy NLC

Right: The energy in the constituentframe of electron-positron and hadroncolliders constructed (filled circles andsquares) or planned. The energy ofhadron colliders has here been deratedby factors of 6–10 in accordance withthe fact that the incident proton energyis shared among its quark and gluonconstituents.

e+e– Collider (operational)

e+e– Collider (planned)

1038

1036

1034

1032

1030

1028

1 10 100Center-of-Mass Energy (GeV)

1000 10000

Lu

min

osit

y

(cm

–2 s

–1)

AD

ON

E

SPE

AR

DO

RIS

PE

P-I

I/K

EK

B

BE

PC C

ESR

PE

TR

A

TR

IST

AN

ISR

LE

P

PE

P

SLC

NL

C

LE

P-I

I

HE

RA

TE

VA

TR

ON

LH

C

SP– P

S

Hadron Collider (operational)

Hadron Collider (planned)

e–p Collider

Peak luminosities achieved at existing colliders and valuesprojected for planned or upgraded machines. The dashed lineindicates luminosity increasing as the square of the center-of-mass energy Note that the rated machine energy has beenused in calculating the abscissa. (Data updated courtesyGreg Loew, SLAC)

42 SPRING 1997

Beyond this problem is the matterof interpretability. When we useelectrons to bombard hadron targets,be they stationary or contained in anopposing beam, we are exploring acomplex structure with an (as-yet)elementary object whose behavior iswell understood. Thus the informa-tion about the structure of the protonresulting from electron-proton colli-sions, for example, tends to be easi-er to interpret than the results fromproton-proton collisions. All theabove observations are generalities,of course, and there are numerousand important exceptions. For in-stance, if neutrinos or muons—copiously produced as secondarybeams from proton machines—areused to explore the structure ofhadrons, the results are comple-mentary to those produced by elec-tron beams.

Everything I have said about elec-trons is also true of muons. The useof muon beams offers significant ad-vantages and disadvantages rela-tive to electrons. The two lightestcharged leptons, the electron andmuon, experience essentially thesame interactions. But muons, beingheavier, radiate far less electromag-netic energy than do electrons ofequal energy; therefore backgroundsfrom radiative effects are muchlower. On the other hand, muonshave a short lifetime (about 2microseconds), whereas electrons arestable. Colliding-beam devices usingmuons must be designed to be

When discussing the relative meritsof electron and proton colliders, thebackground situation is complex be-cause the factors that cause them arequite different. When accelerated,and especially when their path isbent by magnets, electrons radiateX rays in the form of synchrotronradiation. Protons usually have moreserious interactions with residual gasatoms, and those that deviate fromthe nominal collider orbit are moreapt to produce unwanted backgroundsfrom such causes.

A much more difficult—and tosome extent controversial—subjectis the comparison of the complexi-ties of events initiated by electronswith those induced by hadrons ingeneral, and protons in particular.Today particle physicists are usually,but not always, interested in the re-sults of “hard” collisions betweenthe elementary constituents (bywhich I mean entities considered tobe pointlike at the smallest observ-able distances). Because protons arecomposite objects, a single hard col-lision between their basic con-stituents will be accompanied by amuch larger number of extraneous“soft” collisions than is the case forelectrons. Thus the fraction of in-teresting events produced in an elec-tron machine is generally much larg-er than it is for proton machines. Sothe analysis load in isolating the“needle” from the “haystack” tendsto be considerably more severe athadron machines.

compatible with this fact. In addi-tion, the remnants of the muons thatdecay during acceleration and stor-age constitute a severe background.Thus, while the idea of muon col-liders as tools for particle physics hasrecently looked promising, there isno example as yet of a successfulmuon collider.

BUT THERE is an overarchingissue of costs that dominatesthe answer to the question,

“How large can accelerators and col-liders become, and what energy canthey attain?” The relationship of sizeand cost to energy is determined bya set of relations known as scalinglaws. Accelerators and colliders canbe broadly classified into linear andcircular (or nearly circular) machines.With classical electrostatic acceler-ators and proton or electron radio-frequency linear accelerators, thescaling laws imply that the costs andother resources required should growabout linearly with energy. Althoughroughly true, linear scaling laws tendto become invalid as the machinesapproach various physical limits. Theold electrostatic machines becametoo difficult and expensive to con-struct when electrical breakdownmade it hard to devise acceleratingcolumns able to withstand the nec-essary high voltages. And radio-frequency linear accelerators indeedobey linear scaling laws as long asthere are no limits associated withtheir required luminosity.

The scaling laws for circular ma-chines are more complex. Ernest

Far left: William Hansen (right) andcolleagues with a section of his firstlinear electron accelerator, whichoperated at Stanford University in1947. Eventually 3.6 m long, it couldaccelerate electrons to an energy of6 MeV. (Courtesy Stanford University)

Left: Ernest Lawrence’s first successfulcyclotron, built in 1930. It was 13 cm indiameter and accelerated protons to80 keV. (Courtesy Lawrence BerkeleyNational Laboratory)

BEAM LINE 43

Lawrence’s cyclotrons obeyed anapproximately cubic relationship be-tween size or cost and the momen-tum of the accelerated particle. Themagnet’s radius grew linearly withthe momentum, and the magnet gapalso had to increase accordingly toprovide enough clearance for thehigher radio-frequency voltages re-quired to keep particles and the volt-age crest synchronized. All thischanged in 1945 with the inventionof phase stability by Edwin McMillanand Vladimir Vexler. Their indepen-dent work showed that only mod-erate radio-frequency voltages are re-quired in circular machines becauseall the particles can be “locked” insynchronism with the acceleratingfields.

Then came the 1952 invention ofstrong focusing, again independentlyby Nicholas Christophilos and byErnest Courant, Livingston, andHartland Snyder (see photograph onthe right). Conventional wisdom saysthat a magnetic lens to focus parti-cles both horizontally and verticallycannot be constructed— in contrastto optical lenses, which can. But theprinciple of strong focusing showedthat, while a magnetic lens indeedfocuses in one plane and defocusesin the orthogonal plane, if two suchlenses are separated along the beampath, then their net effect is to focusin both planes simultaneously. Thisbreakthrough made it possible tosqueeze beams in circular (and alsolinear!) accelerators to much tighterdimensions, thus reducing magnet-ic field volumes and therefore costs.

Because the basic linear scalinglaws apply to linear machines forboth electrons and protons, promi-nent physicists predicted that all

future acceleratorswould eventually belinear. But the questionremained, “Where isthe crossover in costsbetween circular andlinear machines?”New inventions, par-ticularly strong focus-ing, raised the predict-ed crossover to muchhigher energy. More-over, strong focusing also made thescaling law for high energy protonsynchrotrons almost linear. Thetransverse dimensions of the beamaperture do not need to grow verymuch with energy; thus the cost oflarge circular proton colliders growsroughly linearly with energy.

While the scaling laws for protonmachines are not affected signifi-cantly by radiation losses (althoughsuch losses are by no means negli-gible for the largest proton colliders),they become the dominant factor forcircular electron machines. Theradiation energy loss per turn of a cir-culating electron varies as the fourthpower of the energy divided by themachine radius. It is also inverselyproportional to the mass of the cir-culating particle, which tells youwhy electrons radiate much moreprofusely than protons. In an elec-tron storage ring, certain costs areroughly proportional to its radiuswhile others are proportional to theradiation loss, which must be com-pensated by building large and ex-pensive radio-frequency amplifiers.As the energy grows, it therefore be-comes necessary to increase theradius. The total cost of the radio-frequency systems and the ring itselfwill be roughly minimized if the

Ernest Courant, M. Stanley Livingston,and Hartland Snyder (left to right), who

conceived the idea of strong focussing.(Courtesy Brookhaven National Laboratory)

radius increases as the square of theenergy.

Such a consideration therefore in-dicates that linear electron machinesshould eventually become less ex-pensive than circular ones. But whatis meant by the word “eventually?”The answer depends on the details.As designers of circular electron ma-chines have been highly resource-ful in reducing the costs of compo-nents, the crossover energy betweencircular and linear colliders has beenincreasing with time. But it appearslikely that CERN’s Large Electron-Positron collider (LEP), with its 28kilometer circumference, will be thelargest circular electron-positron col-lider ever built.

The only reasonable alternative isto have two linear accelerators, onewith an electron beam and the oth-er with a positron beam, aimed at oneanother—thereby bringing thesebeams into collision. This is the es-sential principle of a linear collider;much research and development hasbeen dedicated to making such a ma-chine a reality. SLAC pioneered thistechnology by cheating somewhat onthe linear collider principle. Its lin-ear collider SLC accelerates both elec-tron and positron beams in the sametwo-mile accelerator; it brings these

44 SPRING 1997

one pinches the other, usually in-creasing its density; but if that pinch-ing action becomes too severe, thebeam blows up! In addition, theextremely high electric and magneticfields that arise in the process causethe particles to radiate; the energythereby lost diversifies the energy ofthe different particles in the bunch,which makes it less suitable forexperiments.

And there is an additional featurethat aggravates the problem. As theenergy of colliders increases, thecross sections of the interesting re-actions decrease as the square of theenergy. Therefore the luminosity—and therefore the density of the in-teracting bunches—must increasesharply with energy. Thus all theproblems cited above will becomeeven more severe.

As a result of all these factors, alinear collider is not really linear inall respects; in particular, the bright-ness of the beam must increase as ahigh power of its energy. This factis difficult to express as a simple cost-scaling law. It suffices to say thatall these effects eventually lead toa very serious limit on electron-positron linear colliders. Where thislimit actually lies remains in dispute.At this time an upper bound of sev-eral TeV per beam is a reasonableestimate. We can hope that humaningenuity will come to the rescueagain—as it has many times beforewhen older technologies appeared toapproach their limits.

THIS DISCUSSION of linearelectron-positron colliders isa part of a larger question:

“How big can accelerators and col-liders, be they for electrons and

beams into collision by swingingthem through two arcs of magnetsand then using other magnets tofocus the beams just before collision.In the SLC (and any future linear col-lider), there is a continuing struggleto attain sufficient luminosity. Thisproblem is more severe for a linearcollider than a circular storage ring,in which a single bunch of particlesis reused over and over again thou-sands of times per second. In a linearcollider the particles are thrownaway in a suitable beam dump aftereach encounter. Thus it is necessaryto generate and focus bunches of ex-ceedingly high density.

An extremely tight focus of thebeam is required at the point of col-lision. There are two fundamentallimits to the feasible tightness. Thefirst has to do with the brightnessof the sources that generate electronsand positrons, and the second isrelated to the disruption caused byone bunch on the other as they passthrough each other. According to afundamental physics principle thatis of great importance for the designof optical systems, the brightness (bywhich I mean the intensity that il-luminates a given area and is prop-agated into a given angular aperture)cannot be increased whatever you dowith a light beam—or, for that mat-ter, a particle beam. Thus even thefanciest optical or magnetic systemcannot concentrate the final beamspot beyond certain fundamentallimits set by the brightness of theoriginal source and the ability of theaccelerating system to maintain it.

The second limit is more complex.The interaction between one beamand another produces several effects.If beams of opposite charge collide,

positrons or for protons, become?”As indicated, the costs of electron-positron linear colliders may be lin-ear for awhile, but then costs increasemore sharply because of new physi-cal phenomena. The situation is sim-ilar for proton colliders. The costestimates for the largest proton col-lider now under construction—CERN’s Large Hadron Collider—andfor the late lamented SSC are rough-ly proportional to energy. But thiswill not remain so if one tries to buildmachines much larger than the SSC,such as the speculative Eloisatron,which has been discussed by certainEuropean visionaries. At the energyunder consideration there, 100 TeVper beam, synchrotron radiation be-comes important even for protonsand looms as an important cost com-ponent. Indeed, physical limits willcause the costs eventually to risemore steeply with energy than lin-early for all kinds of machines nowunder study.

But before that happens the ques-tion arises: “To what extent is soci-ety willing to support tools for par-ticle physics even if the growth ofcosts with energy is ‘only’ linear?”The demise of the SSC has not beena good omen in this regard. Hopefullywe can do better in the future.

BEAM LINE 45

The Astro-Particle-Cosmo-Connectionby VIRGINIA TRIMBLE

THE UNIVERSE AT LARGE

Observational astronomers and theoretical physicists

have been getting in each other’s hair since the time of

Newton and show no signs of letting up.

FOR ISAAC NEWTON (1642–1727), though there werelaboratory data from the work of Galileo (1564–1642),the British Union of Growers of Poorly-Attached

Apples (BUGPAA), and probably others, the real test of universal gravitation was its application to the lunar andplanetary orbits that Johannes Kepler (1571–1630) had managed to extract from the observations of his mentor Tycho Brahe (1546–1601). Looking at the various dates, youmight reasonably suppose that the planetary orbits wouldhave been somewhat improved by the time Principiaapproached publication (1687), but as the names of otherseventeenth-century astronomers will not be on the exam,you are not required to read or remember any of them.

Entering the twentieth century, we find the equally well-known example of Einstein’s theory of general relativity fac-ing reality in the form of the advance of the perihelion of Mercury* and the gravitational deflection of light by thesun.** From that day (1919) to this, GR has passed every testastronomy can throw at it, especially the correct descriptionof the changing orbits of binary pulsars (meaning neutronstars in orbits with other neutron stars or massive white

*Meaning that Mercury’s elliptical orbit rotates once every 3 million years relative to the distant stars.

**Meaning that the apparent positions of stars, and radio sources, have been seen to be shifted on the sky when your line of sight passes close to the limb of the sun.

46 SPRING 1997

dwarfs). In the pulsar case, for which Joseph Taylor andRussell Hulse shared the 1993 Nobel Prize in physics, thephysical processes include gravitational radiation andother strong-field effects, for which general relativitymakes different predictions from those of other theo-ries that would also fit the solar system, weak-field data.

Those pulsar orbits would be getting larger or small-er if the coupling constant, G, were changing with time.Non-zero dG/dt would also affect the lifetimes of stars(whose rate of energy generation scales like G5), the rangeof masses possible for old white dwarfs (supported againstgravity by degenerate electron pressure) and neutron stars(supported by degenerate neutron pressure), the dynam-ical evolution of clusters of stars, and distances withinthe solar system. Curiously, astronomical observationslead to just about the same limits on dG/dt from all ofthese systems: not more than about 10 percent eitherway in the 10–20 Gyr age of the universe. Such obser-vations, as well as the Mercurian orbit advance, also tellus that the speed of gravitons is very close to the speedof photons in a vacuum. One always writes the equa-tions with c, but one means c(gravity), not c(light).

OTHER HISTORICAL EXAMPLES AND FALSE ALARMS

Particle physics can perhaps be said to have begun withthe discovery of entities beyond the n, p, and e foundin ordinary atoms. The first were the positron, the mu(“Who ordered that?”) meson, and the pi (Yukawa par-ticle) meson. All first appeared as upper-atmosphere sec-ondary cosmic rays (ones produced when primary cos-mic ray protons hit atmospheric molecules—very hard).A convenient date to remember is 1937, when a show-er of papers by people you have heard of in other con-texts (Heitler, Oppenheimer, Serber, Homi Bhabha) clar-ified that these were indeed secondary products but alsonew particles with well-defined properties.

Astronomical considerations have also made occa-sional contributions to nuclear physics, most famouslyin 1953, when Fred Hoyle realized that the carbon-12nucleus must have a particular excited state, or we would

all be made out of pure hydrogen and helium, no furtherfusion being possible in stars. More recently, the needfor lifetimes, energy levels, and cross sections of nuclideslikely to form in exploding stars, but most unlikely inthe lab, have driven both calculations and experiments.

From time to time, astronomers have concluded thatsome set of observations simply could not be explainedin terms of the (then) standard model of physics and haveattempted to invent what they thought was needed. Likemany other examples of hubris, this has typically beenpunished, exile from the community being the most fre-quent outcome. Some cases are relatively well known,like the Steady State universe, invented to allow starsand galaxies to be older than the apparent cosmic ex-pansion time scale, but requiring the addition of a cre-ation field to general relativity or other theories of grav-ity. The suggestion that atomic spectral lines can beredshifted by something that is not a Doppler effect, notthe expansion of the universe, and not a strong gravi-tational field, at least when those lines come fromquasars, is another well known example.

Less famous, perhaps, are James Jeans’ proposal thatspiral nebulae represent new matter pouring into the uni-verse, “white hole” explanations of quasars, and the pre-stellar matter of Viktor Ambartsumian, who believedthat clusters of new stars expand out of regions of very

Russell Hulse,co-discoverer ofthe binary pulsar1913 + 16, whose

behavior in thedecades since

has provided themost stringent

available tests ofgeneral relativity.(It passed; Hulse

won a NobelPrize.) (Courtesy

AIP MeggersGallery of Nobel

Laureates)

BEAM LINE 47

dense, prestellar stuff, perhapsa bit like Gamow’s Ylem, but notconfined to the early universe,and then in turn expel gaseousnebulae from their surfaces toproduce configurations like thestars and gas of Orion. (Conven-tional stellar evolution tracks doroughly the reverse, beginningwith gas and ending with verydense remnants.)

As time goes on, the variouspossible interactions between as-tronomy, cosmology, particlephysics, and so forth that are dis-cussed in the following sectionswill move to this one. I am notprepared to guess which will then be seen as “interest-ing historical examples” and which as “that was an as-tronomer who thought he was Feynman.”*

THINGS THAT DO NOT GO BUMP IN THE NIGHT

You could write a whole book about the exotic particles,properties, and processes that we know do not existbecause they would violate some set of astronomical ob-servations. In fact someone has (Georg Raffelt; see thelist of “more reading” on page 51). The general idea isthat stars must be allowed to form from the interstel-lar medium, do their nuclear thing for millions or bil-lions of years, and die as planetary nebulae + white dwarfs(from low mass stars) or as supernovae + neutron starsor black holes (from stars of more than about 8 solarmasses), at all times adhering to a set of nonlinear dif-ferential equations that describe conservation laws, ratesof energy generation and transport, and equilibriumbetween pressure and gravity. The detection of neutri-nos from SN 1987A with very much the temperature,time scale, and flux that had been expected from neu-tron star formation brought this whole field into con-

siderable prominence. The constraints are sometimesquite tight simply because, on the whole, stars man-age pretty well with just the standard-model physics thatwe are all so tired of.

In addition, any new entities you might want to pos-tulate must not be so numerous and massive as to makethe average density of the universe big enough to slowthe expansion measurably today (since we see no suchslowing). Nor are they (or you) allowed to spoil the setof nuclear reactions at high density and temperature thatproduce deuterium, helium (3 and 4), and a bit of lithium-7in the early universe (“big bang nucleosynthesis”). I men-tion here only a small, representative set of examplesand urge you to peruse Raffelt’s book for many more andfor the corroborative details.

1. There must not be too many magnetic monopolesfloating around, or they will short out the large-scalemagnetic fields of Jupiter, pulsars, and the interstellargas. This Parker (for Eugene of Chicago) limit means thatsuch monopoles must have rest masses of at least 1016

GeV if they are to be dynamically important in the uni-verse.

2. The magnetic dipole moment of the electron neu-trino cannot be more than about 3×10−12 of the Bohr mag-neton, or the neutrinos from SN 1987A would never have

*The original image here was a New Yorker cartoon bearing thecaption: “That’s God. He thinks he’s a doctor.”

Academician ViktorAmbartsumian, who diedlast year, was among the

first astronomers topropose a specificmechanism for the

formation of expandingclusters of massive,

young stars. He laterextended the idea

(expansion from somekind of very dense, pre-stellar material, different

from known interstellar orlaboratory gases) into apossible explanation for

quasars. (Courtesy AIPEmilio Segrè Visual

Archives) Leo

Gol

db

erg

48 SPRING 1997

made it out of the parent star and through space to us.This is probably rather smaller than the best laboratorylimit, and the 1987A data also set limits to neutrinomasses, coupling to right-handed and Majorana neutri-nos, and such that are comparable to or better than thelaboratory numbers.

3. Quite a few of the things you might think of do-ing with neutrinos would mess up the early universe,including, in particular, adding to the “known” threeflavors. A fourth or fifth neutrino family would speed upthe early expansion so much that too many neutronswould survive to form more helium than we see. Thereis, of course, also a limit of roughly three neutrino fla-vors from the laboratory width of Z0 decay, but, becausewe do not know lifetimes or masses a priori, the two con-siderations rule out somewhat different volumes of pa-rameter space.

4. Any new bosons or weakly interacting massive par-ticles you might want to dream up must not couple toordinary or degenerate matter tightly enough to trans-port much energy in either normal stars or white dwarfsand neutron stars. If they do, you will cool off your WDsand NSs too fast (so we wouldn’t see the ones we see) andchange the internal density and temperature distribu-tion of nuclear-burning stars away from the ones neededto reproduce known correlations of stellar masses, lu-minosities, radii, and evolutionary phase.

A cross section of 10−36cm2 at stellar temperaturesborders on being “too big” for a number of these con-texts. Another false alarm was the attempt to reduceneutrino emission from the sun by cooling its interior

with WIMPs whose cross sections fell in the borderlinerange. At least two problems resulted. The interior dis-tribution of density no longer matched the one derivedfrom analysis of solar pulsation frequencies, and laterstages of evolution, like the horizontal branch phase, be-came so short-lived that you couldn’t account for thelarge numbers of stars seen in them.

There are also a few cases where something new un-der the sun might still improve agreement between mod-els and observations. One of these is the possible pres-ence of pion condensate or strange quark matter in theinteriors of neutron stars (which we should then callpion stars, quark stars, or some such). Either one willhasten cooling after nuclear reactions stop. This couldbe useful if existing upper limits on thermal emissionfrom the surfaces of neutron stars should ever get pushedlower than the predictions from conventional coolingcurves. In addition, each permits a given mass to be some-what more compact without collapsing. Thus the starcan rotate a bit faster without slinging mud in theorists’faces. At the moment (2:37 p.m. Wednesday, Septem-ber 25, 1996) the two shortest periods of rotation mea-sured for neutron stars are both quite close to 1.55 msecand are comfortably accommodated by most ordinaryequations of state for nuclear matter. The false alarm ofa 0.5 msec pulsar reported at the site of SN 1987A several

0.28

0.26

0.24

0.22

0.20

10–10

3 x 10–10

7 x 10–10

Nν = 4

3

2

Baryon-to-Photon Ratio

Hel

ium

Abu

nda

nce

(%

)

Plot of some of the consequences of nucleosynthesis duringthe hot dense (big bang) phase. Observed abundances oflithium-7 and deuterium in gas and stars that have experi-enced very little nuclear processing require that the real uni-versal density of baryonic material (the baryon-to-photon ratio)fall somewhere in the white stripe—corresponding to a baryondensity less than 10 percent of the closure density. Then thefact that the abundance of helium is, at very most, a little morethan 24 percent says that there can be at most three neutrinoflavors in the early universe. (Courtesy C. Copi and D. Schramm,University of Chicago)

Artist’s conception of the Sudbury Neutrino Observatory (SNO)detector. When fullyoperational, it willdetect all three fla-vors of neutrinosand give some indi-cation of the direc-tion from which theycome. Althoughsensitive only to thevery highest energy(boron-8) solarneutrinos, it shouldbe able to decide ifsome of themissing electronneutrinos haverotated into mu- ortau-neutrinos.(Courtesy LawrenceBerkeley NationalLaboratory)

BEAM LINE 49

of the other devices provides any directional informa-tion). And the SAGE and GALLEX gallium detectors alsosee about half the expected flux, mostly in the form oflower energy neutrinos from the proton-proton reaction(p + p → d + e+ + νe).

Third, it is rather difficult to make this combina-tion come out from any fiddle you can think of, mostlybecause it is the middle energy range that seems to bemost deficient. New weak interaction physics, along thelines of neutrino oscillations catalyzed by the presenceof nuclei (MSW effect), seems to work better than non-standard models of the solar interior. Fourth, even MSW-type oscillations are squeezed into a very narrow cornerof the space of neutrino masses and coupling constantswhen you also insist on accounting for the anomalousratio of neutrino flavors among cosmic-ray secondariesmade in the atmosphere. Fifth, new detectors under con-struction or planned (SNO, SuperKamiokande, Borexino) could sort things out (but need not), and I suspect thatthe last word has not been said on this topic, not evenmy last word.

years ago triggered a considerable flurry of preprintson quark (etc.) stars, some of which made it into printbefore the report was retracted—and a few afterwards!

Neutron stars remain, of course, the most extremeenvironment under which we can test pictures of howsuperfluids and superconductors behave. They alsoremain awkwardly refractory to experiment.

THERE’S GOT TO BE A PONY IN THERE SOMEWHERE*

The two topics on which nearly everybody agrees thatastronomers and particle physicists must cooperate ifanswers are ever to be found are “the solar neutrino prob-lem” and the complex of questions concerning the ex-istence and nature of dark matter, the origin of large-scale structure in the universe (formation anddistribution of galaxies and clusters of galaxies), andwhatever happened before big bang nucleosynthesis, in-cluding inflation, baryogenesis, phase transitions, andmiracles. Neither is at all new to regular readers of thesepages.

John Bahcall summarized the solar neutrino situationhere (see the Fall/Winter 1994 Beam Line, Vol. 24, No.3, page 10). I will summarize still further. First, RaymondDavis Jr.’s chlorine-37 experiment has been seeing abit less than a third of the predicted flux of high ener-gy neutrinos since before 1970, and the first generationof possible excuses already included many of the as-tronomical and weak-interaction fiddles that are stillwith us (for examples see the Trimble and Reines reviewmentioned under “more reading”). Second, three addi-tional experiments have not clarified things as much asone might have hoped. At the very highest energies thatcome only from boron-8 decay, the Kamiokande elec-tron-scattering detector has reported about half the num-ber of expected events from the direction of the sun (none

*Readers who remember the joke of which this is the punch lineare invited to share it with those who don’t, preferably keeping inmind that roughly half the preprint pile comes from my side ofthe interdisciplinary fence and half from yours—unless we are onthe same side.

50 SPRING 1997

Finally, we come to the constellation of issues asso-ciated with dark matter and the very early universe. Theobservational situation is quickly summarized: 90 per-cent or more of the stuff in the universe that contributesto gravitational potentials does not emit (or absorb) itsfair share of electromagnetic radiation. Dark matter un-questionably exists and outweighs the luminous matterin stars, galaxies, and the gas between them. But wehaven’t a clue what it is.

Colleagues often object to this second statement. Whatthey mean, however, is not that we have any very defi-nite information about what the dark matter is, but onlythat we know quite a lot of things it is not. This is prog-ress only if the number of ideas generated by theoristsis finite (not by any means a safe bet). For starters, therequirement of not messing up big bang nucleosynthe-sis almost certainly means that the dark matter can-not all be ordinary stuff made of protons, neutrons, andelectrons. Thus we are forced to hypothesize other stuffthat is capable of, at most, gravitational and weak in-teractions, and not of electromagnetic or nuclear ones(again a few colleagues would disagree at some level).

Dark matter, structure formation, inflation, phasetransitions, etc. get mixed up together in several ways.First, most obviously, galaxies and clusters live in po-tential wells made mostly of dark matter, and the na-ture of the stuff is bound to make a big difference to howgalaxies form (and whether we can model them at allsuccessfully, to which the present answer is no, notentirely). Second, galaxy formation might be aided (orimpeded) by various topological singularities (cosmicstrings, textures, . . .) left from the phase transitionsassociated with the four forces gradually separating them-selves. The supersymmetry arguments that go with theforces having once been the same more or less auto-matically imply the existence of several kinds of non-baryonic particles associated with assorted unfamiliarbut conserved quantum numbers.

Third, the “inflaton field” responsible for early, ex-ponential expansion of the universe (inflation) could pos-sibly leave behind a small ghost of itself to act as a cos-mological constant (Einstein’s unloved Λ). Fourth,

inflation, at least some kinds, is supposed to leave be-hind both the exact critical density required to stop uni-versal expansion in infinite time and a spectrum ofperturbations of that density with a definite form, wellshaped to grow into galaxies and clusters. No obviousastronomical observation would seem capable of prov-ing that inflation happened, but one could imagine de-finitive dynamical evidence for a total density less thanthe critical one or for a spectrum of not-yet-evolveddensity perturbations different from the inflationary pre-diction. But there are already variants of inflation in theliterature that can live with one or both anomalies.

In some ways, this mess looks slightly simpler fromthe astronomical side. As far as we can tell, for the pur-poses of galaxy formation and creation of large-scalestructure, everything nonbaryonic can be divided amongfour categories, and it doesn’t much matter which ex-ample nature has chosen to favor. The four categoriesare non-zero cosmological constant, seeds (like the topo-logical singularities), hot dark matter (consisting of par-ticles light enough that they are relativistic at T ≈ 3000Kwhen baryonic matter and light stop talking to eachother; ordinary neutrinos of 5–25 eV are the most ob-vious candidate), and cold dark matter (consisting of par-ticles massive enough to be non-relativistic at the sametemperature, like the lowest-mass supersymmetric par-ticle and its cousins; or axions which are low mass butform at rest; and no, I don’t know why).

You can, if you wish, have two of these or even three.I am not aware of any scenarios that involve all foursimultaneously, but this may well come. The varietyis welcomed because no current simulation of galaxy(etc.) formation simultaneously does a very good job ofaccounting for structures on relatively small linear scales(a megaparsec or less, promoted by CDM), the largestscales (up to 100 Mpc, promoted by HDM), the largest de-viations from smooth cosmic expansion that we see, andthe observed sizes of those deviations (for example, thedispersion of pair-wise velocity differences between near-by galaxies) as a function of scale length. Choosing aspectrum of initial density fluctuations different fromthe standard inflationary one allows yet another degree

BEAM LINE 51

MORE READING

For the multitude of limits on particle properties that arise from considerations of stellar structure, see G. G. Raffelt,Stars as Laboratories for Fundamental Physics,1996, University of Chicago Press.

Strange Quark matter is discussed in G. Vassiliadis et al. (eds) Proc. Int. Symp. Strangeness and Quark Matter, World Scientific Press, Singapore and in Nuclear Physics B (Proc. Supplement) 24B on Strange Quark Matter in Physics and Astrophysics, 1992.

Atmospheric neutrinos are featured in T. K. Gaiser et al. (1995) Phys. Reports 258, 173 and in M. Fukugita and A. Suzuki (Eds.) 1994, Physics and Astrophysics of Neutrinos (Springer-Verlag).

Various snapshots of the solar neutrino problem appear in V. Trimble and F. Reines, 1973, Rev. Mod. Phys. 45, 1; J. N. Bahcall, Neutrino Astrophysics (1989), Cambridge University Press; and Y. Susuki and K. Nakamura (Eds.) 1993, Frontiers of Neutrino Astrophysics (Universal Academy Press, Tokyo).

For the various kinds of WIMPs, inos, and other dark matter candidates implied by supersymmetry, see G. Jungman, M. Kamionkowski, and K. Griest 1995,Phys. Reports.

And, finally, inflation and other highlights of the early uni-verse appear in

A. Linde 1991, Particle Physics and Inflationary Cosmology, Harvard Univ. Press, E. W. Kolb and M. S. Turner 1990, The Early Universe, Addison-Wesley, and G. Boerner, The Early Universe, Fact and Fiction, 2nd ed. 1992, Springer-Verlag.

of freedom. It is not, I think, clear whether what is need-ed is just further exploration within the territory de-scribed above or whether there may still be some im-portant piece of physics missing from the simulations.

There is, however, one thing you can be sure of. Iam not going to be the person to holler that the astro-nomical observations require new physics (or new im-perial clothes, or whatever) or to suggest the form thatphysics should take.

52 SPRING 1997

CONTRIBUTORS

J. S

elsi

ng

L. W

einb

erg

ABRAHAM (BRAM) PAIS is theDetlev W. Bronk Professor Emeritusof The Rockefeller University. Bornin Amsterdam in 1918, he earned hisPh.D. degree from the University ofUtrecht in 1941 and emigrated to theUnited States after World War II. Inthe 1950s he introduced the conceptof associated production, which gov-erns the behavior of strange particles,and developed the idea of particlemixing. He resides in New York Cityand Copenhagen, Denmark.

Among his many publications isSubtle is the Lord: The Science and Lifeof Albert Einstein, which in 1983 wonthe American Book Award for Sci-ence and the American Institute ofPhysics Science Writing Award. Hisother books include Inward Bound,Niels Bohr’s Times and Einstein LivedHere. For his extensive contributionsto the public understanding of sci-ence, he received the 1993 GemantAward of the American Institute ofPhysics and the 1995 Lewis ThomasPrize.

STEVEN WEINBERG is a memberof the Physics and AstronomyDepartments at the University ofTexas, and the founding director ofits Theory Group. His work inphysics has been honored with theNobel Prize, the National Medal ofScience, the Heinemann Prize inMathematical Physics, 13 honorarydoctoral degrees, and election to theNational Academy of Sciences, theRoyal Society, the American Philo-sophical Society, and the AmericanAcademy of Arts and Sciences.

He is the author of over 200 sci-entific articles, one of which is themost cited article of the past 50 yearsin particle physics, and seven booksof which the most recent is TheQuantum Theory of Fields.

Educated at Cornell, Copenhagen,and Princeton, he taught at Colum-bia, Berkeley, MIT, and Harvard be-fore coming to Texas in 1982.

CHRIS QUIGG, left, is a memberof the Theoretical Physics Depart-ment at Fermilab and Visiting Pro-fessor at Princeton University. HisPh.D. research at Berkeley was moredistant from quarks and gauge fieldsthan today’s students can possiblyimagine. It nevertheless began hislifelong engagement with experi-ment and his close association withJ. D. Jackson (right), who is teachinghim still.

A recurring theme in Quigg’s re-search is the problem of electroweaksymmetry breaking and the explo-ration of the 1 TeV scale. His currentinterests include mesons with beau-ty and charm, the top quark, andneutrino interactions at ultrahigh en-ergies. He is also at work on a secondedition of Gauge Theories of the Strong,Weak, and Electromagnetic Interactions.

L. Q

uig

g

BEAM LINE 53

MICHAEL RIORDAN has spentmuch of his last twenty years writ-ing and editing general books on sci-ence, technology, and their impacton our lives. They include The SolarHome Book (1977), The Day After Mid-night (1981), The Hunting of the Quark(Simon & Schuster, 1987), The Shad-ows of Creation (W. H. Freeman, 1991)and the forthcoming Crystal Fire: TheBirth of the Information Age (W. W.Norton, 1997)—from which his ar-ticle in this issue has been adapted.

He currently divides his workingtime between SLAC, where he servesas Assistant to the Director and Con-tributing Editor of the Beam Line, andthe Santa Cruz Institute for ParticlePhysics, where he is researching ascholarly history of the Supercon-ducting Super Collider. For recrea-tion he can often be found paddlinghis kayak on the waters of MontereyBay or hiking near his home in theSanta Cruz Mountains.

WOLFGANG (PIEF) PANOFSKYis Professor and Director Emeritus,Stanford Linear Accelerator Center.A member of the National Academyof Sciences, he has served as Chair-man of its Committee on Interna-tional Security and Arms Controland currently heads its Weapons Plu-tonium Management and Disposi-tion Study Committee. He served onthe President’s Science AdvisoryCommittee in the Eisenhower andKennedy Administrations.

Born in Berlin in 1919, Pief earnedhis Ph.D. in 1942 from Caltech. Afterthe war, he joined the University ofCalifornia at Berkeley, then came toStanford as Professor of Physics in1951. He served as Director of itsHigh Energy Physics Laboratory from1953 to 1961 and was the foundingDirector of SLAC, continuing in thatposition until his “retirement” in1984. Among his many awards arethe National Medal of Science andthe Fermi and Lawrence Awards ofthe Department of Energy.

“The Universe at Large” appears thisissue largely thanks to the author’shusband, Dr. Joseph Weber (on herleft), who took her out to dinnerevery evening for a week to free upsome time for writing it. They con-tinue to oscillate between the Uni-versity of Maryland and the Univer-sity of California at 31.7 nHz, partlybecause of the interestingly differentsets of restaurants on the two coasts.The chap in front is Yakov B.Zeldovich, with whom they had avery memorable dinner in Jena in1982. He was among the pioneers inthe USSR in looking at the interfacebetween particle physics and cos-mology.

E. H

eito

wit