STANFORD LINEAR ACCELERATOR CENTERSpring 1999, Vol. 29, No. 1

A PERIODICAL OF PARTICLE PHYSICS

SPRING 1999 VOL. 29, NUMBER 1

EditorsRENE DONALDSON, BILL KIRK

Contributing EditorsMICHAEL RIORDAN, GORDON FRASER

JUDY JACKSON, AKIHIRO MAKI,PEDRO WALOSCHEK

Editorial Advisory BoardGEORGE BROWN, PATRICIA BURCHAT,

LANCE DIXON, JOEL PRIMACK,ROBERT SIEMANN, GEORGE TRILLING

KARL VAN BIBBER

IllustrationsTERRY ANDERSON

DistributionCRYSTAL TILGHMAN

The Beam Line is published quarterly by theStanford Linear Accelerator Center,Box 4349, Stanford, CA 94309.Telephone: (650) 926-2585INTERNET: beamline@slac.stanford.eduFAX: (650) 926-4500Issues of the Beam Line are accessible electronically onthe World Wide Web athttp://www.slac.stanford.edu/ pubs/beamline.SLAC is operated by Stanford University under contractwith the U.S. Department of Energy.The opinions of the authors do not necessarily reflectthe policy of the Stanford Linear Accelerator Center.

Cover: A montage of microwave-guiding componentsused in SLAC’s Next Linear Collider Test Accelerator for transmitting high power microwaves from the klystronsto the accelerator.

Printed on recycled paper

FEATURES

2 PHYSICS AT LEP 2Particles colliding at CERN are changing the

character of electron-positron physics. Here’s

what it means for new particle searches and

precision measurements.

Michael Schmitt

8 LOOKING FOR COSMIC ANTIMATTERPhysicists have long puzzled about the

apparent absence of antimatter in a Universe

born in a Big Bang that presumably produced

equal amounts of matter and antimatter. A

new experiment seeks signs of antimatter in

space.

Maurice Bourquin and Gordon Fraser

13 TOWARD A TeV LINEAR COLLIDERTwo large R&D facilities at KEK and SLAC are

testing major subsystems for a next-

generation electron-positron linear collider.

Michael Riordanpage 14

BEAM LINE

14 THE JLC ACCELERATOR TEST FACILITYSeigi Iwata

17 THE NEXT LINEAR COLLIDER TEST ACCELERATORTheodore Lavine

DEPARTMENTS

21 THE UNIVERSE AT LARGECan’t You Keep Einstein’s Equations Out of MyObservatory? — Part II

Virginia Trimble

26 CONTRIBUTORS

DATES TO REMEMBER

CONTENTS

page 21

page 26

page 8

page 2

page 17

Particles colliding at

CERN with a center-

of-mass energy close

to 200 GeV are chang-

ing the character of

electron-positron

physics. Here's what it

means for new particle

searches and precision

measurements.

2 SPRING 1999

NEW ERA IN ELECTRON-POSITRONcollisions began four years ago at CERN. Thecenter-of-mass energy of the Large Electron-Positron Collider (LEP) increased by half to135 GeV—well above the peak of the Z bosonresonance around 91 GeV, where LEP and its

American cousin, the Stanford Linear Collider, have beentaking data for years. Since then the energy has risengradually to 189 GeV, making this collider, now called LEP 2,a unique high energy physics machine (see the article byDaniel Treille in the Fall 1992 issue of the Beam Line,Vol. 22, No. 3).

The first era of LEP physics began with the detailed studyof the Z boson. When early hopes of discovering new phe-nomena were not realized, the four LEP collaborations—ALEPH, DELPHI, L3, and OPAL—concentrated on precisionmeasurements and rare particle decays, leading to importantresults far beyond initial expectations. Perhaps the best ex-amples are the measurement of the Z boson mass—now oneof the best known quantities in all of particle physics—andthe isolation of a small sample of B0 → J/Ψ Ks

0 decays withwhich to examine CP-violation. The main areas of study in-cluded electroweak processes, tau physics, the physics of“beauty” mesons and baryons, and quantum chromodynam-ics. Searches for new particles such as the Higgs boson or su-persymmetric particles found nothing new within the limitsimposed by kinematics, namely, that the sum of the massesof the particles produced is less than the total beam energy.

PHYSICS AT LEP 2

Aby MICHAEL SCHMITT

BEAM LINE 3

γ/Ze–

e+

W–

W+

e–

e+

W–

W+

Feynman diagrams of processes whichdominated at LEP 1 (top), and those thatare relatively more important at LEP 2.

At LEP 2 the Z boson resonance is gone, and the previous-ly huge production of two-fermion final states (pairs of lep-tons or quarks) is now accompanied by the appearance offour-fermion final states made large by the production ofpairs of W bosons (see diagrams at right). The careful mea-surement of two-fermion cross sections and asymmetriescontinues, as tests of Standard Model predictions could wellturn up deviations suggesting new physics. In addition, LEP 2

presents a wonderful opportunity to study the W boson,which previously had been produced in large numbers onlyat Fermilab. Most important, however, is the direct searchfor new particles—for should any such thing be discovered,particle physics would enter a new age.

SEARCHES FOR NEW PARTICLES

Theorists have proposed many new particles that might bediscovered at LEP 2. Paramount among them is the Higgsboson, which in the Standard Model and its variants is theprincipal agent responsible for the masses of the knownparticles. This unique particle is not democratic with re-gard to the three generations of quarks and leptons: it cou-ples more strongly to the heavy than to the light. Conse-quently, it prefers to decay into a pair of heavy particlesthat together are lighter than it. Unless the Higgs boson is

e–γ/Z

e+

f

f–

W

4 SPRING 1999

only indirect supporting evidence.Its basic premise can be stated easi-ly: for every quark or lepton there aretwo new bosons, or scalar particles,and for every Standard Model bosonthere is a new fermion—including apartner for the as-yet unobservedHiggs boson! All these new particlesmust be very heavy, otherwise weshould have discovered one by now.

The supersymmetric partner ofthe W boson, a charged particle calledthe “chargino,” will be produced co-piously if it is light enough. In thesimplest scenario, the chargino de-cays the same way as a W, so one islooking for a second W-like particlethat decays into pairs of quarks orleptons. In other scenarios charginodecays into leptons may be en-hanced, but generally that poses noparticular problem for experimenters.Charginos with masses less than94 GeV have essentially been ex-cluded by now.

Neutral sisters of the charginos,called “neutralinos,” could enhancethe signal for supersymmetry; ifcharginos are produced, one mightexpect also to observe neutralinos.The lightest neutralino plays a spe-cial role as the lightest of all super-symmetric particles. If it is stable, asusually assumed, then neutralinosleft over from the Big Bang proba-bly comprise a large fraction of themysterious cold dark matter of theUniverse (see article by MichaelTurner in the Fall 1997 issue of theBeam Line, Vo. 27, No. 3) which isthought to clump together with thevisible galaxies. Indirect limits onsuch a particle require its mass to belarger than about 28 GeV.

Other supersymmetric particlesmight be produced at LEP 2, includ-

particularly heavy itself, this meansmostly a pair of b quarks, with cquarks and tau leptons showing up10 times less frequently.

A Higgs boson is thus expected tomaterialize most often as a pair ofhigh energy “b jets”—a bundle of or-dinary hadrons originating from a bquark. It would be produced when anelectron and positron annihilate tocreate a supermassive Z, whichwould immediately “decay” into anordinary Z and a Higgs boson. Al-though this would be a very rareprocess, it has advantageous prop-erties: the Z decays to a pair ofcharged leptons, quarks, or neutri-nos, all of which help physicists dis-tinguish Higgs events from standardprocesses. The two b-quark jetsemerging from the Higgs boson de-cay can be used to measure its mass;if Higgs bosons are produced at LEP2, a peak should appear in plots of thetwo-jet mass.

Standard searches for Higgsbosons have been developed and per-fected by all four collaborations, witheach group competing for even mod-est improvements in their analyses.Unfortunately, no hint of any telltaleexcess of b-quark jets has appearedyet, and the researchers have had tobe content with excluding ranges ofpossible Higgs boson mass. The com-bined data of all four experimentscurrently indicates that if the Higgsboson exists at all, its mass must begreater than 94 GeV (see figure on theleft). With an ultimate LEP 2 collisionenergy of 200 GeV, these experimentscan search for Higgs bosons up to amass of about 109 GeV.

Supersymmetry (SUSY) is a col-lection of theories with many un-determined parameters and so far

20

15

Rat

e L

imit

an

d E

xpec

ted

Rat

e

10

5

090 92 94

mH (GeV/c2)

96

Expected Signal Rate

Upper Limit (95% CL)

94.0 GeV

The expected rate of Higgs events as afunction of the Higgs mass, compared tothe upper limit derived from the directsearch. These preliminary data from theOPAL Collaboration exclude a Higgsboson lighter than 94 GeV.

BEAM LINE 5

ing scalar leptons and quarks, whichwould show up as ordinary leptonsand jets of hadrons plus missing en-ergy. Ironically, the scalar top quarkis the most promising of the scalarquarks; due to possible mixing of thetwo SUSY partners of the top quark,one light and one heavy mass eigen-state could result.

At the present time there is nowhiff of supersymmetry in the data.Does this mean that this theory willsoon be discarded? Probably not, asit does not specify precisely the mass-es of all the new particles—whichmight all be too heavy for LEP to pro-duce. Their masses should come inlargely below 1000 GeV, however, andthe Large Hadron Collider will havea mass reach nearly that high.Supersymmetry, however, does placeone important restriction on themass of the lightest Higgs boson: itmust weigh in at less than about135 GeV, which is not far above thereach of LEP 2. And, judging from theindications gleaned from all the pre-cision electroweak measurements,the Higgs boson may well fall with-in its reach.

PRECISION MEASUREMENTS

The triumph of the precision mea-surements at LEP 1 is the mass of theZ boson, known now to one part in30,000: MZ = 91.188±0.003 GeV. As themediator of the neutral weak force,the Z is produced naturally in elec-tron-positron collisions. The media-tor of the charged weak current is theW boson, which in electron-positroncollisions is usually produced inpairs. Since a higher energy—morethan 160 GeV—is required to producea pair of W’s than a single Z, the study

of the W has only become possible inthe LEP 2 era.

The extraction of the Z mass wasa question of measuring the eventrate as the center-of-mass energyswept across the Z resonance, but W’sare another story. Since each eventcontains a pair of W’s, the resonanceshape appears directly in the recon-struction of their masses. For exam-ple, if two W’s each decay to two jets,then in principle the masses of thosetwo W’s can be reconstructed directlyfrom the jet momenta. In realitythere is a problem with jet confusion(How do you know which jets comefrom which W boson?) and the re-construction of the jets themselves,so events that have a single chargedlepton and two jets are easier to an-alyze. From fits to the resonancepeaks measured so far (see graph be-low), the best W mass value is80.37±0.09 GeV. When the data takenin 1998 have been fully analyzed, the

60

40

20

050 60 70

Nu

mbe

r of

Eve

nts

/1.5

GeV

Minv (GeV)80 90 40 50 60 70

Minv (GeV)80 90

80

60

40

20

0

Reconstructed W mass peaks. On the left, one W decays to a charged lepton and aneutrino while the other decays into a pair of quarks. On the right, both W’s decayinto quark pairs. These data come from the L3 Collaboration; the solid lines are fitsto the data and the black areas represent background events.

6 SPRING 1999

than for the top quark, because theHiggs boson has a weaker impact onthings like the mass of the Z and themass of the W. Consequently, im-proving the accuracy of the W massmeasurement is vitally important.The bounds placed on the Higgs massby the current measurements of thetop quark and W masses (see graphon the left) are beginning to be sig-nificant. If the total error on the theW mass shrinks to 0.04 GeV, as an-ticipated, then the constraints on theHiggs boson mass will become muchmore stringent. If, for example, wefind the mass of the W equals80.48±0.04 GeV while the mass of thetop equals 174±5 GeV—and still noHiggs boson is found with a mass lessthan 109 GeV— then the StandardModel will be in jeopardy.

The calculation of these indirect“virtual” effects relies on field theorymethods that predict that the strengthof an interaction depends on its en-ergy. In the parlance of particlephysics, the “coupling constants run.”In fact they run at different rates: thecoupling constant for electromagnet-ism increases gradually with energy,the coupling for the weak force hard-ly changes at all, and the coupling con-stant for the strong force, αs, actual-ly decreases. One would like to knowwhether, at some high energy scale,all three have the same value. If theydo, then they may be viewed as threeaspects of a single, universal force, thatis, they will be “unified.” In the un-altered Standard Model, we alreadyknow that they do not unify, but inSUSY models, it looks like they do. Tomake a more stringent test, we needmore precise measurements of thecoupling constants, which in the caseof αS is experimentally challenging.

error will shrink substantially, sothat the mass of the W will be knownto one part in a thousand. By the endof LEP 2 running, this will be im-proved by another factor of two.

Of what use are very accurate val-ues for the W and Z masses? Isn’t itenough to know that they exist? Notin particle physics. We do not yethave a theory of everything; we havethe Standard Model—which is notfully validated until we find theHiggs boson—and speculative ex-tensions of it. In order to pick out themore worthy speculations, we needto “peer” from the energy scale of theexperimental phenomena we observe(roughly 100 GeV) to much higherscales (such as 1016 GeV), where newphenomena would dominate. Thisprocedure works because the newparticles that are active at those highenergy scales have indirect effects atthese low-energy scales; we can seetheir impact in very subtle shifts ofthe interactions and masses of theknown particles like the W and Z.

This procedure may seem ratherspeculative, but we know that itworks. The top quark was found atthe Tevatron in 1995 (see the articleby Bill Carrithers and Paul Grannis inthe Fall 1995 issue of the Beam Line,Vol. 25, No. 3), but before that no oneknew for sure what its mass was. Wecould make a serious estimate, how-ever, because it impacts the mea-surement of the mass of the Z bosonand its decays. As it turns out, the val-ue obtained indirectly agrees with theactual Tevatron measurement.

Now that we know the top quarkmass, we can use the precision mea-surements to try to deduce an indi-rect value for the Higgs boson mass.This turns out to be much harder

80.6

80.5

80.4

80.3

MH = 90

300

LEP1, SLD, νN DataLEP2, pp– Data

1000

80.2130 150 170

mt (GeV)

mw

(G

eV)

190 210

Bounds on the Higgs mass from preci-sion electroweak measurements fromcombined data presented at the 1998Vancouver conference. The two roundcurves indicate the (68% confidence)constraints from indirect measurementsof MW and Mt (solid curve) and directmeasurements (dashed curve). Thewhite band shows the theoretical calcu-lation. The data seem to favor low Higgsmasses, which are, however, graduallybeing excluded by direct searches atLEP.

BEAM LINE 7

Fortunately, physicists have iden-tified many indirect ways to measureαS. For example, the emission of glu-ons, which shows up as “extra”hadronic jets, depends directly on αS.A related feature is the overall“shape” of these hadronic events.They may be long and thin, indicat-ing only two energetic quarks in thefinal state; broad and flat, indicatingan additional gluon; or spherical, in-dicating two or more energetic glu-ons. It is impossible to make an ab-solute prediction for the shape ofevents or the number of jets, due toexperimental and theoretical ambi-guities. Fortunately, how these quan-tities change as a function of energyis well defined and easily measured.The doubling of LEP energy allowsexceptionally clean observations ofthe running of αS—better than 3 per-cent (see graph on the right).

There is a middle ground betweendirect searches for new particles andultra-precise measurements of mass-es and couplings, and that is the mea-surement of cross sections and an-gular distributions for which theStandard Model makes clear and def-inite predictions. For example, wecan easily calculate and cleanly mea-sure the total cross section forhadronic events with center-of-massenergy well above the mass of the Z.If the measured value comes in larg-er than predicted, it might be dueto the production of new particlessomehow missed in the directsearches, or perhaps a deviation ofthe coupling constants that wouldpoint to new virtual effects. Of par-ticular interest in this regard is thenumber of b-quark pairs produced,since this is the heaviest fermion

produced at LEP. Thus far no devi-ation has been spotted, although themeasurements have turned out to bemore challenging than anticipated.

A more exotic corner of cross-section measurements actually teststhe interactions among W’s, Z’s, andphotons, rather than just their cou-plings to quarks and leptons. Forexample, a W boson can turn the in-coming electron and positron intoa pair of neutrinos—which escapedetection—at the same time emit-ting a photon that generates a largesignal in the electromagnetic cal-orimeter. A contribution due to thisW-W-γvertex can be isolated on a sta-tistical basis, affording a direct testof the Standard Model in this im-portant aspect. Other kinds of eventstest the W-W-Z vertex that con-tributes to W-pair production, and theZ-Z-γ vertex that should vanish tolowest order. These vertices lie at thevery heart of electroweak symmetry.

The LEP collider will run throughthe year 2000, when its energy willreach 200 GeV. By the end of the pro-gram each experiment should haverecorded more than enough data tocomplete searches for Higgs bosonsand supersymmetric particles, andmeasure very precisely the proper-ties of the W boson. Perhaps a gen-uine discovery will be made, or a newvirtual effect uncovered. In eithercase the elucidation of any new phe-nomena would be carried out at fu-ture programs, such as future runs atthe Tevatron, or at the LHC, or per-haps best of all, at the next genera-tion of electron-positron collidersnow in the design stages.

0.14

0.12

0.10

0.0880 120

Center-of-Mass Energy (GeV)

α S(E

cm)

160 200

Variation of αs with collision energy. Thisplot shows preliminary results from theALEPH Collaboration.

8 SPRING 1999

HEN IT BLASTED OFF from NASA’sKennedy Space Center on June 2, 1998, theSpace Shuttle Discovery carried the three-tonAlpha Magnetic Spectrometer (AMS), the firstmajor particle physics experiment ever to go

into orbit around Earth.Despite its excitement and glamor, the 10-day flight of

the Alpha Magnetic Spectrometer aboard the Space Shuttlewas only a taste of bigger things to come. Although all AMSdetector systems were up and working, the mission was atrial run to provide operational experience before deployingAMS on the International Space Station in the first years ofthe new millennium. This milestone mission could revealthe first evidence for nuclear cosmic antimatter, a majorstep towards resolving a long-standing puzzle about the ap-parent absence of antimatter in a Universe created in a BigBang which supposedly produced matter and antimatter inequal amounts.

Stars and other powerful cosmic engines continuouslyblast out streams of high energy particles. These particlescrash into nuclei in the upper atmosphere, producing show-ers of secondary debris which rain down from the sky ascosmic rays. To see the primary cosmic particles, messen-gers from distant parts of the Universe, detectors have to beflown high up into the atmosphere in balloons, or above theatmosphere in satellites. The largest pieces of antimatterseen in cosmic rays so far are antiprotons. Cosmic raysappear to contain no antinuclei, suggesting that theirsources contain no nuclear antimatter.

LOOKING FOR COSMIC ANby MAURICE BOURQUIN and GORDON FRASER

W

BEAM LINE 9

This apparent absence of cos-mic antimatter has been under-lined in a careful appraisal of theimplications of a balancedmatter-antimatter Universe byAndy Cohen of Boston Univer-sity, Alvaro de Rújula of CERN,and Sheldon Glashow of Harvard,published in 1998 in the Astro-physical Journal. They look at the consequences of matterand antimatter confined in separate and distinct domains.

Such a balanced Universe should have produced matter-antimatter encounters wherever and whenever the bound-aries of the matter and antimatter domains touched. In suchencounters, the separate pieces of matter and antimattermutually annihilate to form bursts of energy in the same(but time-reversed) way that energy can create equalamounts of matter and antimatter. In the Universe, this mat-ter-antimatter annihilation would provide a source of highenergy cosmic radiation—gamma rays.

As the Universe expands, these annihilation radiationrelics cool, giving a diffuse cosmic gamma ray background.But detailed calculations by Cohen, de Rújula and Glashowshow that any such effect would be larger than currentlyobserved gamma ray signals, such as those from the EGRETtelescope aboard NASA’s Compton Gamma Ray Observatory.Today’s very low gamma background reveals no evidence forsuch annihilation processes ever having taken place on alarge scale.

TIMATTER

Space Shuttle astronauts at CERN. Leftto right are Mission Pilot CommanderDominic Gorie, Mission SpecialistFranklin Chang-Diaz, CommanderWendy Lawrence, Mission SpecialistJanet Kavandi, co-author MauriceBourquin of the University of Geneva,and Mission Commander ColonelCharles Precourt.

10 SPRING 1999

Detecting cosmic antimatterwould be a new Copernican revolu-tion, calling for a reappraisal of ourpicture of the Universe. However be-cause of the limited sensitivity of theexperiments, the existence of anti-matter somewhere in the Universecannot be completely ruled out. Inthe absence of any sighting, improv-ing the limits on how much anti-matter could exist and where it couldbe helps determine the fundamen-tal parameters of particle theory andits cosmological implications.

AN ANTIMATTER EXPERIMENTIN SPACE

On its own, antimatter should behavelike ordinary matter, with antiprotonsand antineutrons forming antinuclei,and then attracting orbital positronsto form anti-atoms. Paul Dirac, thespiritual father of antimatter, point-ed out that the spectra from atoms ofantimatter should be no differentfrom those of ordinary atoms, and an-timatter stars would shine in thesame way as ordinary ones. How thencan antimatter be detected?

The only direct way is to look forstray particles of antimatter, just asCarl Anderson did in 1932 when hesaw cosmic ray tracks bending the“wrong” way in a magnetic field andso discovered the anti-electron, bet-ter known as the positron. Howeverany primordial cosmic antinucleiwould be quickly mopped up by theearth’s atmosphere. To see themmeans sending a magnetic detectorinto space.

In 1994 Samuel Ting of Massa-chusetts Institute of Technologypresented an imaginative proposal toNASA. His Alpha Magnetic Spec-

trometer would be a space-borneequivalent of Anderson’s historic ex-periment. Instead of using a cloudchamber to track cosmic particles, itwould use sophisticated semicon-ductor technology.

Ting built up a diverse, skilledteam of scientists from the UnitedStates, China, Russia, Taiwan, Ger-many, Italy, Switzerland, and otherEuropean countries. The experimentbrings a novel symbiosis of space-borne and particle physics research.

Particle physicists are skilled atdesigning and building detectors torecord particle reactions under con-trolled laboratory conditions. ForAMS, the interactions would insteadbe supplied by Nature. However forAMS, the conditions are very differ-ent, calling for new solutions. Aswell as the size, weight, and elec-tric power restrictions of a space-borne experiment, the detector hasto respect stringent crew safety re-quirements and be compatible withdelicate space shuttle systems (evenin an airplane the use of electronicequipment by passengers is restricted!).AMS instrumentation has to with-stand the huge forces when the spaceshuttle blasts off and lands, whereaccelerations reach 15G and noise vi-bration levels attain 150 decibels. Inflight, the detector has to withstandlarge temperature swings, high ra-diation levels and the intensevacuum of outer space. This was newterritory for particle physicists usedto the relative calm of their terres-trial laboratories. Instead of being onhand in a nearby control room, theywould have to monitor their detec-tor from the remote ground station,with an astronaut mission specialistas their space-borne representative.

The AMS module installed in the SpaceShuttle’s payload bay. Above is theSpacehab module with supplies and lo-gistics for the Russian Mir space station.This Discovery mission was the last ofnine such dockings with Mir.

BEAM LINE 11

A VERY SPECIAL DETECTOR

The AMS detector contains the usualcomponents of a particle physicsexperiment—a central spectrometerwith a magnet to bend the particle tra-jectories and tracking to record thepaths of particles, particle identifica-tion capability, data acquisition sys-tems to filter, compress, and recordthe information, and monitoring andcontrol systems, as well as commu-nications with the mother craft. Al-most all these functions required spe-cial attention for a space environment.

To provide these capabilities, AMSuses a cylindrical permanent mag-net, a set of six silicon trackingplanes with double-sided readout, atime-of-flight measurement systemwith two pairs of scintillator arraysand an aerogel Cerenkov counter. Ananticoincidence counter systemaround the tracker helps distinguishparticles passing inside the detec-tor from those interacting in the sur-rounding material.

These components are assembledon an aluminum barrel structure,1.14 m in diameter and 0.80 m inheight, supporting the permanentmagnet. The outside of the barrel car-ries electronic crates for the powersupplies, trigger systems, data ac-quisition systems, monitoring andorbiter communication interfaces.

A charged antiparticle passingthrough the detector bends the“wrong” way in the magnetic field.However full identification comesfrom measuring the particle’s mo-mentum (from the exact curvatureof its trajectory), its velocity (mea-sured by the time-of-flight system)and its energy losses by ionization inthe tracker and scintillators.

The magnet uses a neodymium-iron-boron alloy to optimize its field-to-weight ratio. The ferromagneticmaterial is shaped into 6000 smallblocks (about 1 kg each) glued to-gether into prismatic bars with suit-ably oriented magnetization. In themagnet aperture, the highly uniformmagnetic field is of the order of 0.1tesla. With such a magnetic field, theflux leakage has to be very small tosafeguard the overall operation of thespacecraft. The magnet was built andspace qualified using carefully cho-sen components and with stringentacceleration and vibration tests inChina.

The core of the detector, the par-ticle tracker, is based on the SiliconMicrovertex Detector of the L3 ex-periment at CERN’s LEP electron-positron collider, but most modules

Honeycomb

Silicon Wafers

Honeycomb

Computers

PhotomultipliersAerogel

Magnet

Electronics

Foam

Scintillators Low Energy Particle Shield

Vet

o C

oun

ters

Cutaway view of the AMS detector asflown on the Space Shuttle in June 1998.To track cosmic particles, the interior ofthe AMS detector contains semiconduc-tor technology developed and perfectedfor the L3 experiment at CERN. As wellas pinpointing cosmic tracks with micronprecision, this instrumentation has to en-dure the extreme vibration and noiselevels during the launch and landing ofthe Space Shuttle. Only half the siliconsensors were installed for this flight, thusproviding valuable experience beforedeploying AMS on the InternationalSpace Station in the first years of thenew millennium.

12 SPRING 1999

are much longer. These detector el-ements using arrays of high puritysilicon wafers were developed to pin-point particle tracks and so detectthe decay products of very short-livedcharged particles, which even whenmoving almost at the speed of lighttravel only a fraction of a millimeterbefore decaying.

The AMS tracker is made of 41×72mm double-sided silicon sensors, 300microns thick. The arrangementgives measurements in three di-mensions. Charged particles can bepinpointed down to 10 microns. Themodules are mounted on disk-shapedhoneycomb supports. The front-endreadout electronics uses hybrid cir-cuits mounted perpendicular to themodule planes. Cooling bars conductthe heat produced to the magnet. Flatribbons of coaxial cables take the sig-nals to the analog to digital convert-ers and other data processing circuitsin the outside crates.

The time-of-flight system uses fourplanes of scintillators, two above and

two below the magnet. It has threetasks: to trigger the detector byselecting single particles traversingthe spectrometer; to measure theirvelocity and distinguish between up-ward and downward particles; and toperform four independent ionizationmeasurements to separate particlescarrying different electric charges.When the two independent mea-surements provided by the four planesare combined, the time-of-flight mea-surement is about 100 ps.

The threshold Cerenkov counterbelow the spectrometer uses a radi-ating medium made of 8 cm thickaerogel blocks optically connected tophotomultipliers. The blocks arearranged in two layers of 8×10 and8×11 matrices. As low energy protonsand antiprotons do not produceCerenkov light, they can easily bedistinguished from positrons andelectrons.

The anticoincidence counter sys-tem rejects sprays of neutrons andprotons coming from comic ray in-teractions in the magnet body or inthe detector material. This rejectionconsiderably reduces background in

the other systems and al-lows much more sensi-tive measurements. Thesystem consists of a cy-lindrical wall of 16 plas-tic scintillators betweenthe tracker and the in-ternal face of the magnet.

For the space stationmission, the silicontracker has to be aug-mented to reach a totalarea of about 6 squaremeters, and additionaldetectors have to bebuilt.

Space qualification tests on thestate-of-the-art detector were carriedout in specially-equipped space lab-oratories. After final assembly at theSwiss Federal Technical Polytechnic(ETH) in Zurich, the initial versionof the AMS detector was shipped tothe US for final integration aboardthe space shuttle.

Discovery’s flight crew brought thespace shuttle into land on scheduleJune 12, 1998. Although the orbiter’shigh speed data transmission link toearth failed during the flight, this didnot affect actual AMS data taking: alldata were safely recorded on board.Resourceful NASA communicationsspecialists and the astronauts werealso able to patch through some datavia a link normally reserved for videopictures, and this 10 percent sampleshowed that the detector performedflawlessly. The valuable 100 million-event trawl of physics from outerspace is being carefully analyzed.

Before publishing their final re-sults, AMS scientists have to calibrateall their detectors with benchmarkparticle beams, including helium andcarbon ions. This is being done at par-ticle accelerators at GSI, Germanyand CERN.

Is cosmic antimatter out there? Isit further away than we can currentlysee? As the curtain goes up on 21stcentury research, answers could soonbe within reach. As Cohen, de Rújulaand Glashow conclude in their mile-stone paper, “The detection of anti-nuclei among cosmic rays wouldshatter our current understanding ofcosmology.”

Front Side

l

A candidate cosmic antiproton recordedby the AMS tracker.

BEAM LINE 13

OST HIGH ENERGY PHYSICISTS agree that the next

major project after completing the Large Hadron

Collider at CERN is to build an electron-positron linear collider

operating at the trillion-volt (TeV) energy scale. Innovative designs

for such a machine, which will stretch tens of kilometers and cost

billions of dollars, have been evolving for over a decade. These

design efforts have converged on a few favored approaches (see the

article by Gregory Loew and Michael Riordan in the Winter 1997

issue of the Beam Line, Vol. 27, No. 4). An international collabora-

tion headquartered at DESY has pursued one avenue that uses

superconducting microwave cavities to accelerate electrons and

positrons (see the article by Reinhard Brinkmann in the Fall/

Winter 1998 issue of the Beam Line, Vol. 28, No. 3). Another

promising approach, which employs copper cavities operating at

close to ambient temperature, has been pioneered by the Stanford

Linear Accelerator Center and Japan’s High Energy Accelerator

Research Organization (KEK, formerly the National Laboratory for

High Energy Physics).

For more than a year, SLAC and KEK have been working closely

together toward achieving a single design for such a next-

generation linear collider. This joint R&D project occurs under an

inter-laboratory memorandum of understanding signed in

February 1998 by SLAC Director Burton Richter and KEK Director

Hirotaka Sugawara. Well before this agreement took effect, how-

ever, researchers from both laboratories had built extensive R&D

facilities to test some of the major subsystems required in such a

TeV-scale collider. In the following articles, Seigi Iwata of KEK and

Theodore Lavine of SLAC describe these facilities and the encour-

aging progress made with them to date.

—Michael Riordan

Toward a TeV Linear Collider. . .

The JLC Accelerator Test FacilitySEIGI IWATA

The NLC Test AcceleratorTHEODORE LAVINE

MTwo large R&D facilities

at KEK and SLAC are

testing major subsystems

for a next-generation

electron-positron

linear collider.

14 SPRING 1999

highest-priority project for the fu-ture. Research and development to-ward the design of such a machinehas occurred at KEK for more than adecade. Two major areas of currentaccelerator R&D at KEK include thetechnologies needed to generatehigh-quality beams in an injectorcomplex and high-power microwavetechnology required for the mainlinear accelerators.

Because its particle bunches en-counter another bunch only once, alinear collider must achieve very nar-row beams (several nanometers thickfor the JLC) at the interaction pointin order to provide sufficiently highluminosities required for the in-tended physics research. ThereforeKEK physicists made key contribu-tions to the Final Focus Test Beamproject at SLAC, which succeededin squeezing a 50 GeV electron beamdown to a thickness of only 60 nano-meters. Its advanced magnet systemperformed as designed by KatsunobuOide, while Tsumoru Shintake pio-neered a new technique to measuresuch narrow beams using laser in-terference fringes.

One major improvement remain-ing to be demonstrated before a TeVlinear collider can be built is thequality of the beams entering the

NE OF THE MOST important experimentalissues confronting physics today is the search forand study of the Higgs boson and other very heavyparticles thought to be responsible for imbuing

quarks, leptons, and gauge bosons with their various masses.This research can be done most effectively and efficiently ata high energy electron-positron collider. Thus the Japanese highenergy physics community chose the construction of a largelinear collider (called the Japan Linear Collider, or JLC) as its

The JLC Accelerator Test Facilityby SEIGI IWATA

O

S-Band Linac

Electron Source Damped Cavities

Modulators

Damping Ring

Magnet Power Supply

Beam Diagnostic System

Klystrons

RF Gun

Wiggler Magnets

RF Source

Positron Source

Layout of the Accelerator Test Facilitybuilt at KEK as a prototype injector sys-tem for the JLC. Accelerator physics re-search is being conducted on this facili-ty by an international collaboration withthe goal of developing the technology ofultralow-emittance beams.

BEAM LINE 15

final focus system. They must be suf-ficiently narrow and have smallenough angular divergence so thatthe final focus magnets can compressthe beams down to nanometer thick-nesses. The injector systems mustgenerate such high-quality beamsand the main linacs must acceler-ate them to their final energies whilemaintaining the beam quality.

Clearly the injector system willbe a key part of the JLC, determiningthe ultimate performance of its col-liding beams. The KEK AcceleratorTest Facility (ATF) was constructedin a large hall about the size of a foot-ball field; its purpose is to pioneer thestate-of-the-art techniques needed togenerate multibunch beam with un-precedentedly low emittance. (Thisis the conventional measure of beamquality, representing a one-standard-deviation divergence from the for-ward direction in the velocity vec-tors of individual particles in eachbunch.) The principal components ofthe ATF are an electron source, a1.54 GeV injector linac (operating ata microwave frequency of 2.9 giga-hertz), an injection beam transportline, a 1.54 GeV damping ring, and anextraction line. In addition, variousdiagnostic instruments are includedto measure beam performance. Muchof the work on the ATF—from its de-sign to current operations and re-search—has been done as an inter-national collaboration with SLAC,PAL (Korea), IHEP (China), DESY (Ger-many), CERN (Europe) and BINP (Rus-sia). In addition, university teamsfrom Tohoku, Tohoku-Gakuin,Tokyo-Metropolitan, Tokyo-Science,Yokohama-National, Nagoya andKyoto have been playing an increas-ingly important role in the project.

Since its speedy commissioningin autumn 1995, the injector linac hasserved as a facility for studying high-power microwave technology as wellas production, acceleration, handling,and monitoring of various kinds ofelectron beams. It now routinely op-erates at an accelerating gradient of30 MeV per meter—almost twicethat of the Stanford Linear Colliderand TESLA Test Facility. As an in-jector to the damping ring, this linacmust generate a beam that is stablein energy, intensity, trajectory andsize. Most of the early R&D workconcerned improvements on theseaspects. A drift in beam energy wassuppressed by introducing an ener-gy-feedback system based on infor-mation about the beam position atthe transport line, and by stabilizingthe temperature of the cooling waterused in some of the klystrons. In ad-dition, physicists and engineers havecompleted a systematic investigationof the stability of individual accel-erator elements.

Another important goal was toshow that one can stably compensatefor the energy spread caused by theeffects of multibunch beam loading.When a sequence of closely spacedbunches traverses an acceleratingstructure, each bunch carries away asmall amount of electromagnetic en-ergy; latecomers therefore suffer fromsuccessively larger losses in acceler-ation. In order to compensate forthese deficits, two short acceleratingstructures with slightly offset reso-nant frequencies were included inthe linac. The bunches passingthrough them are accelerated on theslope of the traveling electromagneticwave—not its crest—in such a waythat later bunches are accelerated

more strongly. Tests performed withtrains of 20 bunches spaced 2.8 nano-seconds apart successfully reducedthe bunch-to-bunch energy spread toonly 0.3 percent, thus verifying thatthis frequency-shift method workswell in practice. Another way tocompensate for beam loading is tofeed amplitude-modulated micro-wave power into ordinary acceler-ating structures; preliminary testshave shown that this principle workswell, too.

The goal of the damping ring is togenerate a beam with ultralow emit-tance within a storage time briefenough to handle the successivebeam trains coming from the injec-tor linac. Circulating electrons (andpositrons) exhibit oscillatory trans-verse motions known as betatron os-cillations. Upon deflection by bend-ing magnets, these particles emitsynchrotron-radiation photons, thuslosing a bit of their longitudinal andtransverse momentum. But in pass-ing through accelerating cavitiesevery orbit, they regain the longitu-dinal component, thus narrowing thebeam ever so slightly. After many or-bits, the subtle imbalance betweenthese losses and gains lowers theemittance exponentially to an equi-librium value independent of initialconditions.

The ATF damping ring was de-signed to reach a very low equilib-rium emittance—about a hundredththat of conventional storage rings anda tenth that of advanced synchrotron-light sources. Eight multipole wig-gler magnets were included in thering to boost radiation damping byforcing the beam to oscillate in shortsteps. All round the ring’s 140 me-ter circumference there are many

16 SPRING 1999

small magnets in addition to the wig-glers and microwave cavities. Its vac-uum chambers have inner diametersas small as 24 millimeters in the arcsections and only 12 millimeters highat the wigglers—considerably smallerthan the dimensions of convention-al storage rings. But they are stilllarge compared to the dynamic beamaperture arising due to nonlineareffects in an ultralow-emittance ring.Highly sophisticated beam control isnaturally called for in such a situa-tion, and the present ring is equippedwith almost a hundred button-electrode systems to measure thebunch positions at every turn.

A team of physicists led by JunjiUrakawa commissioned the ATFdamping ring (see photograph above)in January 1997. After dealing withinitial hardware problems, they es-tablished a sequence of successful op-erations from injection of a beam intothe ring, its storage with the mi-crowave cavities on, and extractionfrom the ring. So far, they have at-tempted only single-bunch operationat about 1 Hz repetition rate. Seriousaccelerator-physics research begansix months after commissioning end-ed. To model the ring precisely,physicists conducted a systematicstudy in which they gave the beamsmall electromagnetic kicks and

measured corresponding changes inthe downstream orbit; the latest mea-surements agree well with calcula-tions. These data are then used to ad-just the field intensity of individualmagnets. An automatic optimizationprocedure, using a similar beam-based alignment appoach, is about tobecome effective. Emittance damp-ing times of 19 milliseconds (hori-zontal component) and 30 msec (ver-tical) have been observed with thewigglers off, in good agreement withdesign values. The damping time isexpected to drop to 10 msec or lesswith the wigglers turned on.

A conventional synchrotron-lightmonitor was sufficient for measur-ing the beam profile in early stagesof the research. But as machinetuning improved, this approach hadto be abandoned due to the limitedcapability of such a monitor. Asthe beam gets very small, on the oth-er hand, its synchrotron light beginsto reveal a spatial coherence corre-sponding to the decreasing sourcesize. Thus Toshiyuki Mitsuhashi de-veloped and introduced a synchrotron-light interferometer that records theinterference fringes formed behind adouble slit. The size of the sourceis deduced from how the fringe con-trast varies with the slit separation.So far the beam size has been

measured (at a specific point in thering) to be 39 µm wide and 15 µmhigh, compared with 40 µm and 6 µmexpected from beam optics and thedesign emittance. Therefore the ringhas essentially reached its design goalof 1 nm-rad—at least as far as the hor-izontal emittance is concerned. Thisconclusion was confirmed by mea-surement of an extracted beam, aftercorrecting for beam jitter and spu-rious dispersion. The vertical emit-tance should be substantially small-er, as it is primarily determined bybetatron coupling associated withmagnet misalignments. Current esti-mates of this emittance, based on thebeam-size measurement mentionedabove, are about four times largerthan the design value, although witha fairly large uncertainty.

After the first, hectic year of re-search at ATF, physicists have comeclose to achieving the ultralow-emittance beam needed for the JLC.At the same time, it has become in-creasingly clear that, even with awell-designed accelerator, stabilityand resolution are key issues. Afterscheduled improvements in these as-pects, single-bunch operations willcontinue for some time before wemove on to multibunch operation.Our immediate objectives are toachieve reliable, high-precision, one-shot and turn-by-turn beam mea-surements and to reach ultralow ver-tical emittance.

A portion of the Accelerator Test FacilityDamping Ring. High precision alignmentof individual components is important,although an automatic beam-basedalignment system will ultimately be used.

BEAM LINE 17

THE NEXT LINEAR COLLIDER now under

design consists of two independent linear accelera-tors aimed head to head, one for electrons and the

other for positrons. Each accelerates its beam to hundreds ofGeV, providing sufficient collision energy to create exoticnew states of matter. The goal is to collide 250 GeV beams inthe first few years of operation and to be able to supporteventual upgrades to 500 GeV or more. To attain these beamenergies with linear accelerators of reasonable length, wemust achieve very high accelerating gradients. We need tolearn how to generate microwaves of sufficient peak powerto achieve the gradients, while keeping average power andoperating cost down. And we must preserve the high qualityand narrow energy spread of the beams during the accelera-tion process.

A linear-collider R&D program underway at SLAC and KEK

has developed a new generation of microwave power sourcesand high-gradient accelerators equal to these tasks. As partof this program, the Next Linear Collider Test Accelerator(NLCTA) has been operating at SLAC since 1996 as a full-system test bed for these technologies. NLCTA is a high-gradient linear accelerator (linac) with its own dedicatedelectron injector. The accelerator structures and the micro-wave power systems that energize them are engineering prototypes for the linacs of a full-scale collider.

The forefather of the NLCTA linac is the three-kilometer-long SLAC linac, built in the early 1960s utilizing the 10.5 cmwavelength (S-band) klystron amplifiers and accelerator

The Next Linear Collider Test Acceleratorby THEODORE LAVINE

18 SPRING 1999

same average electric power con-sumption (about 10 kW per meter ofaccelerator) at 120 pulses per second.

But many of the technical chal-lenges for building the new acceler-ator and its power sources grow withthe gradient because higher peakpower is needed. The SLAC linac re-quires peak power of about 12 MWper meter of structure, while theNLCTA requires 50 MW per meter toachieve a gradient of 50 MV/m—or100 MW/m to achieve 70 MV/m.

MAKING THE GRADIENT

The X-band klystrons for the NLCTAare the result of a decade of R&D onhigh-power klystron technology atSLAC. Each klystron generates 50 MWpulses of microwave radiation. Aspresently configured, the NLCTA op-erates with three 50 MW klystrons,each of which energizes a pair of ac-celerator structures. The energy ineach klystron pulse is compressed toproduce the full 200 MW required toachieve the 50 MV/m gradient in thepair.

The primary technical challengeof pulse compression is storing themicrowave energy with low loss forthe duration of the klystron pulse.The solution developed in the 1970sto boost peak power in the SLAClinac was to store the energy in over-sized, cylindrical copper cavities. Amajor difference in the NLCTA is thatthe shorter microwave pulse lengthnow makes it possible to use ex-tended microwave transmission linesfor low-loss energy storage. After theklystrons shut off, each storage linecontinues to discharge its pulse un-til the last part of the wave has tra-versed the entire line.

structures developed at Stanford dur-ing the 1940s and 1950s. Improve-ments in the klystrons powering thelinac led to a continuous series of up-grades from the original, 24 megawatt(MW) tubes to 67 MW tubes developedin the 1980s—boosting the SLACbeam energy from 16 GeV in 1966 to50 GeV today.

For the Next Linear Collider(NLC), accelerator designers haveelected to use an X-band wavelengthof 2.6 cm. The accelerator structuresin the NLCTA are prototypes devel-oped specifically for this shorterwavelength, which boosts theachievable gradient and reduces thecross-sectional area of the accelera-tor structure. The shorter wave-length also lowers the microwavefilling time of the structure from 1microsecond to 0.1 microsecond,reducing the needed microwavepulse length. While the SLAC linacachieves a gradient of 20 millionvolts per meter (20 MV/m), the NLClinacs will reach 50 MV/m for the

The NLCTA linac currently has four180 cm long accelerator structures installed between focusing magnets.

BEAM LINE 19

With higher peak power comesthe challenge of handling strongerelectromagnetic fields in the wave-guides and other components thatsupply power to the acceleratorstructure. The NLCTA componentsare able to handle the peak power asa result of years of careful microwaveengineering design and testing. Ini-tial prototypes that suffered fromelectrical breakdown were redesignedto reduce the field strengths.

The accelerator structures weredeveloped jointly by SLAC and KEK.Considerable attention went intocopper-processing and machiningtechniques needed to achieve clean,smooth surfaces capable of sustain-ing high field gradients without emit-ting stray electron currents that candisturb the primary beam. Even afterfabrication and installation, the in-ternal surfaces of the structures mustbe conditioned by an aggressive reg-imen of high-power microwave pro-cessing to reduce surface emission.

In 1997 we achieved the primarygoal for the NLCTA power system:the NLCTA linac operated stably atthe design gradient of 50 MV/m withtolerable electron emission. Themaximum beam energy at this gra-dient (with four 180 cm structuresand two 90 cm structures installed)is 450 MeV. The achievable energy atthe design current is only 350 MeVbecause of beam loading.

Our next goals are to generate andtest stable accelerating gradients upto 70 MV/m, which requires twice asmuch peak power. The first stepsof this program are complete. A sin-gle 50 MW klystron (and pulse com-pressor) has been used to push theaccelerating gradient in a singlestructure (not a pair) to 70 MV/m, and

one of the klystrons has been oper-ated at 75 MW (by increasing its highvoltage). The next steps will be touse two klystrons to generate the70 MV/m gradient simultaneously ina pair of structures and to test thestability of acceleration in that con-figuration.

BUNCH TRAINS

In order to accelerate enough currenton each machine pulse to create theevent rates required for high-energyphysics experiments, the NLC willaccelerate a long train of 100bunches, rather than a single bunch,on each microwave pulse. A chal-lenge that designers face is the trans-verse instability that arises becausea bunch can be deflected by the elec-tromagnetic fields (wake fields) cre-ated by slight but inevitable offsetsof preceding bunches from the cen-tral axis of the accelerator. The smallsize of the accelerator apertures forthe 2.6-cm wavelength exacerbatesthis problem.

All the NLCTA structures aredesigned to suppress these wakefields by varying, in a precise pattern,the internal dimensions of the 200cells that comprise a 180 cm struc-ture; this spoils the coherence of theset of microwave modes that wouldotherwise contribute to the wake-field. The diameters of the irises insuch a “detuned” structure vary from8 to 11 mm. Two of the structures fur-ther suppress beam deflection bydamping the undesirable modes bychanneling them through slots thatlead to microwave absorbers. Jointwork at SLAC and KEK has developedthe design techniques and manufac-turing methods necessary to achieve

Cutaway view of part of a damped anddetuned accelerator structure (top). Thestructure is fabricated from a stack ofcells similar to the one shown above.

20 SPRING 1999

the close tolerances needed for thesedamped and detuned structures.

The NLCTA injector makes trainsof 1400 electron bunches spaced2.6 cm apart. The strategies of de-tuning and damping have worked, forwithout them the accelerator struc-tures could not transmit these trains.The future NLC injectors will prob-ably distribute the same total chargeinto one-eighth as many bunches,spaced eight times further apart. Nev-ertheless, based on the stability ofbunch trains in the NLCTA, we canpredict stability in the NLC becausethe deflecting forces are proportionalto the ratio of bunch charge to spac-ing, which will be the same.

Uniform acceleration of all thebunches in each train is required be-cause only electrons and positronswithin a narrow energy range (tenthsof a percent wide) can be focused atthe collision point. One of the mostsignificant tests completed on theNLCTA has been to show that thebunch-to-bunch variation of energyalong the train can be kept this small.This is a significant issue since, un-der the wrong conditions, the lead-ing bunches in a train can extract toomuch microwave energy from the ac-celerator structure, and the trail-ing bunches will come up short. Onestrategy for achieving uniform ac-celeration is to fill the structure withmicrowave energy in a profile thatmatches what would occur behindan infinitely long train, so that all thebunches that follow get the same ac-celeration. The desired profile hasbeen obtained by modulating themicrowave pulses before the kly-strons amplify them. With thisapproach, the energies of the elec-trons along the entire bunch train

come out the same within a fewtenths of a percent, as desired.

There are other potential appli-cations for the NLCTA. A group atthe Stanford Synchrotron RadiationLaboratory and SLAC has consideredmodifying the NLCTA to drive an X-ray free-electron laser into self-amplified spontaneous emission.The NLCTA can also be used to gen-erate 2.6 cm microwaves or higherharmonics by decelerating the beamin resonant cavities or structures in-serted in the beam line. Physicistsfrom Harvard and SLAC are using thebeam to excite 3.3 mm waves (theeighth harmonic of 2.6 cm) in a cav-ity with an aperture 1 mm in diam-eter. In future experiments, they planto use a structure 25 mm long toexcite 3.3 mm waves at multi-megawatt peak power levels and usethem to generate accelerating gradi-ents perhaps greater than 100 MV/m.Such high gradients are possible atthese short wavelengths, but theyraise the challenges of power lev-els, field strengths, and instabilitiesto new heights. These experimentswith short, 3.3 mm wave structureswill test advanced concepts for ac-celerators in the era beyond the NLC.

The experience gained by operat-ing the NLCTA has been critical forunderstanding the performance andreliability of the complete systemsof power sources, microwave com-ponents, and structures to be used inthe NLC linacs. Future modificationwill continue to test new prototypecomponents. Other applications asan experimental tool for studying ac-celerator and beam physics are onlybeginning to be exploited.

Electron Energy

Position Along Bunch Train

Digitized images of NLCTA beam spotsshowing the energy variations along thetrain of electron bunches. The imagesshow the correlation of electron energywith position along the 36 meter longtrain. When the microwave pulses arenot modulated (upper image), the elec-tron energy drops off along the train byabout 15 percent, approaching an equi-librium value that corresponds tosteady-state beam loading. But whenthe microwave pulses are modulated tocompensate for the transient beam load-ing (lower image), the energies of theelectrons along the entire train are uni-form to withn a few tenths of a percent.

BEAM LINE 21

THE UNIVERSE AT LARGE

Part IICan’t You Keep Einstein’s EquationsOut of My Observatory?by VIRGINIA TRIMBLE

SPECTROSCOPY AND THE RISE OF ASTROPHYSICS

Our ability to recognize the chemical elements from theirpatterns of emission or absorption lines dates from 1859,when chemist Robert W. Bunsen (who had a burner) andphysicist Gustav R. Kirchhoff (who had an assortment oflaws) joined forces in Heidelberg to show that sodium in thelaboratory mimicked a pair of yellow lines in the spectrumof the sun (called “D” by Fraunhofer and also by modernastronomers).

The first people to aim their spectroscopes at the sun,stars, and nebulae were richly rewarded when(as described by Sir William Huggins) “nearlyevery new observation revealed a new fact,and almost every night’s work was red-lettered by some discovery.” (Huggins’ owndiscoveries included the gaseous nature ofmany nebulae that had formerly been regard-ed as dense crowds of stars.) This was very dif-ferent from the astrometry, celestial mechan-ics, and practical astronomy that generations

of classical astronomers had labored over their equations andtransit circles to accomplish. And it was correspondingly un-welcome among much of the existing community. Two reac-tions, one from each side of the Atlantic:

The Victorian astronomer royal, Sir George Biddle, declared thatwhat astronomy is expected to accomplish is evidently atall times the same . . . rules by which the movements of thecelestial bodies, as they appear to us upon the earth, can becomputed. . . . All else which we may learn respecting thesebodies . . . possesses no proper astronomical interest.

Gustav Kirchhoff, above, andSir William Huggins, right.(Courtesy Yerkes Observatoryand Lick Observatory, respectively)

Part I of Virginia Trimble’s two-part article was published in theSpring 1998 issue of Beam Line,Vol. 28, No. 2. It can be accessedfrom our Web site at http://www.slac.stanford.edu/pubs/beamline.

22 SPRING 1999

Seth Chandler, a member of the National Academy ofSciences and calculator of comet orbits and Earth’s po-lar motion, opined that the work of astrophysicists “willdisappear like smoke in the air” and its “authors will liein forgotten graves.” Both were pontificating in the 1890s,and the bitter feelings between practitioners of tradi-tional astronomy and of the new astrophysics nearly frac-tured the community and darkened efforts to form a sin-gle professional society (eventually the AmericanAstronomical Society, but only after several iterationson names).

It was a rather motley crew of people with back-grounds in medicine, brewing, chemistry, physics, thesilk trade, and occasionally even traditional astrono-my who came rather quickly to constitute the com-munity of solar physicists (by about 1870) and astro-physicists (by about 1890, with the Astrophysical Journalfounded in 1899). Among the items they offered backwere the discovery of the first element in a new columnof the periodic table (helium; Jules Janssen at the eclipseof 1868), the demonstration of the existence of metastable

atomic levels with exceedingly long radiative lifetimes(“nebulium” explained by Ira Bowen in 1927), and, downto the present time, accurate wavelengths and energy

*Personal taste still inclines very strongly to “heroine,” for I canremember when the girl cast as Louise in Carousel accidentallysaid to the barker, “I want to be an actor,” and it got a giantlaugh. But modern usage seems to have abandoned the feminineforms for authors, actors, poets, and so forth—so, I suppose, alsofor heroes.

Left, Jules Jansson (1824–1907) depended on temporary sci-entific jobs and stipends for his living until 1865, when his con-tributions to spectroscopy were recognized with an appoint-ment to the chair of physics at the Ecole Specialed’Architecture in Paris. He reached the eclipse of 1870 afterescaping by balloon from beseiged Paris, only to be cloudedout in Oran, North Africa. (Courtesy Yerkes Observatory)Right, Ira S. Bowen about 1968. For many years he was direc-tor of the Hale (Palomar and Mt. Wilson) observatories.(Courtesy Hale Observatories)

Left, Cecilia Payne Gaposchkin. The second edition of CeciliaPayne Gaposchkin (ed. K. Haramundanis, Cambridge Univer-sity Press, 1996) contains roughly equal numbers of her ownwords and those of friends and family. Right, Henry NorrisRussell, who was, of course, the R of the HR diagram as wellas the originator of the (very transient) giant and dwarf theoryof stellar evolution. Judging from a 1939 conference photo, thetwo were exactly the same height. (Courtesy Yerkes Observatory)

levels for molecules difficult or impossible to study inthe laboratory. HCO+ (initially X-ogen) and HC9N areamong those first seen in interstellar gas.

Quantitative spectroscopy was built upon two equa-tions, named for Ludwig Boltzmann and M. N. Saha.These describe the fraction of atoms that ought to be invarious states of excitation and ionization as a functionof kinetic temperature (yes, I also offer egg-suckinglessons for grandmothers). The astronomical hero* hereis Cecilia Payne (later Payne Gaposchkin), who, in her1925 Harvard PhD dissertation applied the equationsto the line spectra of stars of various colors (tempera-tures) and concluded, first, that nearly all stars have

BEAM LINE 23

much the same chemical composition, and, second, thatthis is heavily dominated by hydrogen and helium, atleast in the surface layers. So improbable did this dom-inance seem (remember Eddington) that the sun was al-lowed to have as much as 7 percent hydrogen only in1929 when Henry Norris Russell applied the same equa-tions and 75 percent (by mass) only in the late 1940s.Until then, the official excuse for the strong hydrogenlines was “anomalous excitation conditions,” that is,a flat refusal to believe Boltzmann and Saha.

RELATIVELY SPEAKING

Each year, the annual meeting of the American Asso-ciation for the Advancement of Sciences brings forth acoven of audience members who do not believe in spe-cial relativity. None, as far as I know, is currently em-ployed as an astronomer, though one or two were in the

past. Nor do I knowexactly what theymean, because eachyear one of themstarts the questionperiod by saying thatthere are alternativeexplanations of theMichelson-Morleyexperiment, and I, orwhoever is at bat atthe time, start by an-swering that our con-fidence in special rel-ativity does not todayrest primarily on ex-periments from thenineteenth century.They then say thatthere are other ex-planations for every-thing else as well.And the chairmanthen says that furtherdiscussion will haveto be deferred untilafter the session is

over. Anyhow, all recent considerations of accelera-tion of particles to high energy, whether in the lab orin cosmic sources, of radio emission from jets movingat close to the speed of light, and all the rest have spe-cial relativity built in. It is even done correctly much ofthe time.

General relativity has a more checkered history. Therewas initial enthusiasm from at least parts of the astro-nomical community. Karl Schwarzschild devised the so-lution of the Einstein equations that still bears his nameto describe the space-time around a spherical or pointmass. Eddington undertook to make sure there were as-tronomers at the right places in 1919 to look for thedeflection of light during a solar eclipse. There were andthey did. The precise quality of the data has been de-bated on and off ever since; but it doesn’t matter. Theobservation has been repeated and improved many times,especially at radio wavelengths, where you don’t haveto wait for an eclipse.

Thus, at the founding of the International Astro-nomical Union in 1919, one of the Commissions was de-voted to General Relativity. Its founding president wasLevi-Civita (who had a tensor). But, a couple of Gener-al Assemblies later, the Commission was disbanded forlack of need and interest, and GR did not come backto the IAU until 1970, with the establishment of Com-missions on Cosmology and High Energy Astrophysics(meaning quasars, pulsars, and such).

Between 1916 and 1929, solutions of the Einstein equa-tions to describe the Universe as a whole came from sev-eral people. Willem de Sitter (whose universe was ex-panding but empty) had previously worried about howto extract numbers for the mass of Mercury, the dipolemoment of the earth, and similar Newtonian quantitiesfrom astronomical objects. Alexander Friedman(n)(whose universes contained ordinary matter and still ex-panded) was a man of many parts, mostly meteorolog-ical, and all sadly short-lived. Georges Lemaitre inde-pendently found one of the expanding models and alsowrote down (as part of his PhD dissertation) what wenow call the Tolman-Oppenheimer-Volkoff relativis-tic equation of state, useful for neutron star models.

After the 1929 announcement of the redshift-distancerelation (Hubble’s law or “the expansion of the Uni-verse”) other physicists and mathematicians, including

Karl, the elder Schwarzschild, pub-lished fundamental work on theanalysis of stellar atmospheres, stel-lar kinematics, comet tails, and manyother subjects as well as deriving thesolution to the Einstein equations thatbear his name. He also inspiredH. Rosenberg (1910 Astron. Nach.186, 71) to draw the very first exampleof what we now call a Hertzsprung-Russell diagram. (Courtesy YerkesObservatory)

24 SPRING 1999

R. C. Tolman and H. P. Robertson, took up examinationof the cosmological solutions and their implications.But their work was not somehow perceived as being part

of mainstream astron-omy. Even some of thepeople whom youmight have expected todisplay great enthusi-asm were doubtful.Hubble himself nevertook a strong stand infavor of cosmic expan-sion over tired light(proposed by FritzZwicky in 1929). HeberDoust Curtis, defenderof the existence of ex-ternal galaxies in theCurtis-Shapley debateof 1920, “never hadmuch use for that fel-low Einstein,” accord-ing to Ralph Baldwin,one of his later stu-dents. And not all theearly conversions werepermanent. Sir William

H. McCrea, who was writing about relativistic cos-mology and Newtonian analogs as early as 1931, hasrecently expressed doubts about the correctness of thewhole picture of a relativistically expanding universe.

But an astronomer cannot evade GR forever. The firstrevival came with the discovery of quasars, leading tothe invention of a subdiscipline called “gravitationalcollapse and other topics in relativistic astrophysics.”Then came pulsars and X-ray emitting neutron stars.These forced us to think about how matter should be-have in deep gravitational potential wells and how theradiation would come out. Today there are binary pul-sars whose orbit evolution is precisely described by Ein-steinian relativity and by no other combination ofphysics. And, as our telescopes have seen to larger andlarger redshifts (5.64 is the record this afternoon), con-verting the fluxes and colors you see to energies and timescales is so model dependent that you have to assume

some particular version of an expanding universe tomake any sense at all of your data.

At the moment, the largest number of people earn-ing their precarious livings by checking each others’ rel-ativistic calculations are probably the students and post-docs attempting to predict the gravitational radiationsignal that should be seen by LIGO and its Europeancousins (a) when they are built and (b) when neutronstars collide. Most of these people do not think of them-selves as astronomers, or even astrophysicists. Rather,they are part of the subset of members of the Ameri-can Physical Society who recently formed a TopicalGroup on Gravitation, apparently because the largerDivision of Astrophysics did not feel like home.

POSTMODERN PHYSICS

Like most things, interaction between astronomy andparticle physics started just a little earlier than mostof us noticed.* The year 1965 saw not only the discov-ery of the cosmic microwave background (by radio en-gineers!) but also the first calculation of a limit on neu-trino rest masses from cosmological considerations andan explication of the conditions needed if we are to havemore baryons than anti-baryons in the Universe, withcredit to Zeldovich and Sakharov respectively.

An obituary of David Schramm mentioned that therehad been a time when he was just about the only as-tronomer in the world interested in neutral currents andthe correct form of the weak interaction (because of theirrole in driving supernova explosions). Still earlier in the1970s, however, came the realization that model starswould evolve to match a particular observed class ofhydrogen-poor red giants only if we included what wasthen called “the universal Fermi interaction.” The ideacame from Bohdan Paczynski, then in Warsaw, but Ico-authored one of the relevant 1973 papers, so the “we”for once does not mean Queen Victoria.

More recently, non-baryonic dark matter, GUTS, ax-ions, WIMPs, inflation, supersymmetry, and so forth haveglutted the literature to the point where your desire to

Heber D. Curtis at the Crossley tele-scope, some time before 1921, whenhe left Lick Observatory to take upthe directorship at Allegheny. Timehas shown that Shapley and Curtiswere right about roughly equalnumbers of issues in their 1920debate. Doust rhymes with“soused,” according to Ralph Baldwin. (Courtesy Lick Observatory)

*Would you believe that the first telecast from the Metro-politan Opera was March 10, 1940?! No, I wasn’t there; butI did see the first, 1951, televised Amahl.

BEAM LINE 25

read about it all again is surely even less than my desireto write about it (but see the Spring 1997 issue of BeamLine, Vol. 27, No. 1 if appetite should revive). And, nat-urally, mainstream astronomy has welcomed the col-laboration with the same enthusiasm it extended tospectroscopy, relativity, and all the rest. For instance,a generally outstanding 1991 encyclopedia of astronomymentions both dark matter and Io. But Io gets four pages,and dark matter only one.

The following paragraphs could have been written bymany of the people who are primarily interested in Io,stars, and other traditional subjects (but I think you willbe surprised at who actually wrote them).

We understand the concern of cosmologists that unbri-dled speculation should not take over the field, that itis better to persist with the standard model, warts andall, than for opinions to become splintered, with thedecline of professional standards which would then al-most inevitably ensue.

Our response to this point of view, with which wehave some sympathy, is that undesirable fragmenta-tion has been permitted already, through the invasionof cosmology by Particle physicists. If the invasion hadthe precision and the certainty of earlier invasions ofastrophysics by atomic theory and nuclear physics, theconsequences would obviously be positive. However,one can have reservations about the advantages of be-coming caught up in speculations from a differentfield, especially when those speculations are an-nounced with an air of authority that will probablyturn out to have been taken too seriously.

Notice that the earlier inputs, rejected by our astro-nomical ancestors, have been accepted. Only the mostrecent is being resisted. But the real startler is that theselines come from the pens or word processors of Sir FredHoyle, Geoffrey R. Burbidge, and Jayant Narlikar, who

have championed other ideas from outside the main-stream, particularly steady state cosmology and non-cosmological redshifts.

L’ENVOI

What should one make of these curious histories? Per-haps we have merely uncovered another of those “ir-regular verbs,” of the form, “I evaluate new ideas care-fully. You are a bit of a stick-in-the-mud. He is slightlyto the right of Genghis Khan.” Or perhaps the last wordbelongs to Darius Milhaud, who is supposed to have said(concerning music, of course) that the advance guardof today is the rear guard of tomorrow.

You might think that Io hasvery little to do with our topic.It is, however, the only bodyknown to be more volcanicallyactive than Earth and so mustbe useful at the interface be-tween astronomy and geo-physics, not otherwise men-tioned here.

(Courtesy NASA)

THE FACTALS in the preceding pages and Part I camefrom many sources, most lost in the mists of time, but thevictims of the most extensive plagiarism are the following:Henry Norris Russell, Raymond Smith Dugan, and

John Quincy Stewart, Astronomy (in two volumes)1926, Ginn. & Co. Boston. This was the standard as-tronomical textbook for about twenty years.

George Ogden Abell, Astronomy (4th edition) 1982, Saunders College Publishing, a second-generationtextbook.

Edward Harrison, Darkness at Night, 1987, Harvard University Press. Deals mostly with Olbers’ Paradox.

Stephen P. Maran (Ed.) The Astronomy and Astrophysics Encyclopedia, 1991, Van Nostrand.

John Lankford, American Astronomy, 1997, Harvard University Press. Mostly about people.

Karl Hufbauer, Exploring the Sun, 1991, Johns HopkinsPress. Recounts the development of solar physicsfrom Galileo to the present time.

More on the difficulties astronomers and astrophysicistshad in getting together in the 1890s to found a society willappear in the centenary volume of the American Astro-nomical Society edited by David DeVorkin and scheduledfor 1999 publication.

The quote from Hoyle, Burbidge, and Narlikar appearsin Monthly Notices of the Royal Astronomical Society 286,173 (1997).

SOURCES & SINKS

26 SPRING 1999

GORDON FRASER has beenEditor of the CERN Courier for a longtime. While a research student atLondon’s Imperial College in themid-1960s, he wrote short-story fic-tion as a respite from theoretical cal-culations and became side-trackedinto journalism. He returned tophysics as a science writer, eventu-ally transferring to CERN. He is co-author, with Egil Lillestøl and IngeSellevåg, of The Search for Infinity(New York, Facts on File, 1995) whichhas been translated into ten otherlanguages; author of The QuarkMachines (Bristol, Institute ofPhysics Publishing, 1997); and Editorof Particle Century (Bristol, Instituteof Physics Publishing, 1998). He iscurrently writing a new book—Antimatter—and is a visiting lec-turer in science communication atseveral UK universities.

MAURICE BOURQUIN isDirector of the University of Genevaparticle physics department. Afterinitial research at CERN for his PhDat Geneva, he became Research As-sociate at Columbia University anda member of Leon Lederman’s groupworking at Fermilab. From 1974–1983he returned to the University ofGeneva, where he worked on hyper-on experiments at CERN. Since be-coming a full Professor at Geneva, hehas worked on the MARKJ experi-ment at DESY, the L3 experiment atLEP, and most recently the AMS ex-periment for the Space Shuttle andthe International Space Station. Heis a member of the Swiss NationalScience Foundation’s National Re-search Council and a Swiss delegateto CERN Council.

CONTRIBUTORS

MICHAEL SCHMITT receivedhis PhD from Harvard Universityin 1991. He spent six years as a mem-ber of the ALEPH Collaboration atCERN, working first with Sau LanWu (Wisconsin) and later as a staffmember. His physics interests rangedfrom the properties of tau leptons andB hadrons to searches for supersym-metric particles. In 1998 he joined thefaculty at Harvard University and hasbecome an active member of the CDFCollaboration at Fermi National Ac-celerator Laboratory, where he pur-sues Higgs searches while prepar-ing a large section of the muonsystem for the next Tevatron run.

BEAM LINE 27

SEIGI IWATA is Deputy Direc-tor of the Institute of Particle andNuclear Studies and Head of the JLCPromotion Office at the High Ener-gy Accelerator Research Organiza-tion (KEK). He received his under-graduate degree from TohokuUniversity and his PhD from theUniversity of Tokyo. His involve-ment in collider physics began atCERN on the Intersecting StorageRings. Then he served as the leaderof the TPC subsystem and co-spokesperson for the TOPAZ exper-iment at the TRISTAN collider. Since1988 he has been Director of thePhysics Department at KEK.

THEODORE LAVINE hasworked on developing high-powermicrowave pulse compression andother approaches for generating highpeak power for linear colliders sincehe joined SLAC’s Technical Divisionin 1987. He was responsible for thedesign and construction of the mi-crowave energy compression systemfor the NLC Test Accelerator and hasbeen responsible for operations andsafety since the beginning of the pro-ject in 1993. He currently leads theProject Planning and CoordinatingGroup in the NLC design team.

He received his PhD from theUniversity of Wisconsin in experi-mental particle physics working atSLAC on a PEP experiment.

VIRGINIA TRIMBLE of theUniversity of California, Irvine andUniversity of Maryland has now beenwriting regularly for the Beam Linefor about 10 percent of her life. Thisspot has previously held picturestaken by her father, the late chemistLyne Starling Trimble. This one iscourtesy of her husband, physicistJoseph Weber. He is, as you can see,a “leg man.” You may also be able todeduce that she is fond of books andturtles and tends to hang on tothings, including father’s ROTC hat,grandmother’s papier maché ele-phant, and Mardi Gras beads fromthe New Orleans meeting of theSigma Xi (from whose board ofdirectors she has just retired).

DATES TO REMEMBER

Jun 7–18 Workshop on Physics at TeV Colliders, Les Houches, France (aurenche@lapp.in2p3.fr orbelanger@lapp.in2p3.fr)

Jun 7–Jul 9 ICTP Summer School in Particle Physics, Trieste Italy (ICTP, Box 586, Strada Costiera 11, I-34100, Trieste, Italy or smr1141@ictp.trieste.it)

Jun 14–19 7th International Conference on Supersymmetries in Physics (SUSY 99), Batavia, Illinois(Cynthia Sazama, MS 122, Fermilab, Box 500, Batavia, IL 60510 or sazama@fnal.gov)

Jun 14–25 US Particle Accelerator School at Argonne National Laboratory, Argonne, Illinois (USPAC atFermilab, MS 125, Box 500, Batavia, IL 60510 or uspas@fnal.gov)

Jun 27–30 12th IEEE International Pulsed Power Conference, Monterey, California (Teresa Monteroppc99@llnl.gov)

Jul 7–16 27th SLAC Summer Institute on Particle Physics: CP Violation in and Beyond the StandardModel, Stanford, California (Lilian DePorcel, SLAC, Box 4349, Stanford, CA 94309 orssi@slac.stanford.edu)

Aug 9–14 19th International Symposium on Lepton and Photon Interactions at High Energies, Stanford,California (by invitation only, Maura Chatwell, SLAC, MS 96, Box 4349, Stanford, CA 94309-4349or lp99@slac.stanford.edu)

Aug 30–Sep 10 CERN Accelerator School Course on Accelerator Physics, Benodet, France (CERN AcceleratorSchool, AC Division, 1211 Geneva 23, Switzerland, or suzanne.von.wartburg@cern.ch)

Sep 6–10 11th General Conference of the European Physical Society: Trends in Physics, London, England(Institute of Physics, Meetings and Conferences Dept, 76 Portland Pl, London W1N 4AAEngland or physics@iop.org)

Sep 12–25 CERN School of Computing, Stare Jabloniki, Poland (Jacqueline Turner, CERN School of Com-puting, CERN, 1211 Geneva 23, Switzerland, or computing.school@cern.ch)

Sep 21–24 International Workshop on Performance in Improvement of Electron-Positron Collider ParticleFactories, Tsukuba, Japan (Yoko Hayashi, Secretary, KEK, 1-1 Oho, Tsukuba-shi, Ibaraki-ken,305-0801 Japan or hayashiy@mail.kek.jp)

Oct 13–15 11th US National Synchrotron Radiation Instrumentation Conference, Stanford, California(Suzanne Barrett, Conference Administrator, SSRL, MS 99, Box 4349, Stanford, CA 94309-0210, or barrett@ssrl.slac.stanford.edu)