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 RFall 1995, Vol. 25, No. 3

A PERIODICAL OF PARTICLE PHYSICS

FALL 1995 VOL. 25, NUMBER 3

EditorsRENE DONALDSON, BILL KIRK

Editorial Advisory BoardJAMES BJORKEN, ROBERT N. CAHN,

DAVID HITLIN, JOEL PRIMACK, NATALIE ROE, ROBERT SIEMANN,

HERMAN WINICK

IllustrationsTERRY ANDERSON

DistributionCRYSTAL TILGHMAN

The Beam Line is published quarterly by the Stanford Linear Accelerator Center, PO Box 4349, Stanford, CA 94309. Telephone: (415) 926-2585 INTERNET: beamline@slac.stanford.eduBITNET: beamline@slacvm FAX: (415) 926-4500 Issues of the Beam Line are accessible electronically onthe World Wide Web from the SLAC home page athttp://www.slac.stanford.edu.SLAC 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: Display of a top quark candidate event with anelectron (the large magenta arrow), a muon (green tube),two jets (gray and pink arrows), and the missing trans-verse momentum carried by neutrinos (large light arrow).The arrow widths are proportional to the energy carried bythe object. The lattice work shows the cells of thecalorimeter showing significant energy deposit. Thespheres represent the amount of energy deposited in thecalorimeter cells (orange and white for electromagneticdeposits and blue for hadronic deposits). The faint shad-ings represent the edges of the sub-detectors. The samedilepton event is shown as a lego plot on page 12. Moresuch displays can be found on the World Wide Web athttp://d0sgi0.fnal.gov/art_souvenir/art_souvenir.html.

page 4

page 17

FEATURES

4 DISCOVERY OF THE TOP QUARKTwo leaders of the Fermilab experiments that isolated the top quark tell the adventureof its discovery.Bill Carithers & Paul Grannis

17 IN WATER OR ICE?Physicists are building four huge particle-detectorarrays to do high energy neutrino astronomy.John Learned & Michael Riordan

25 OSTEOPOROSIS—NEW INSIGHTSSynchrotron radiation is giving us new insights into causes and treatments for osteoporosis.John Kinney

31 LASER-ELECTRON INTERACTIONSHigh power lasers combined with SLAC’s uniquehigh energy electron beam are openinga new field of research.Adrian Melissinos

DEPARTMENTS

2 GUEST EDITORIALJames Bjorken

38 CONTRIBUTORS

40 DATES TO REMEMBER

CONTENTS

page 25

Polarized Gun

e–

Left or Right Circularly Polarized Light

Unpolarized Gun

Mirror Box

e–

Ti:Sapphire LasersBunch 1

Bunch 2Combiner

Circular Polarizer

Load Lock

Bunchers

Accelerator

Linearly Polarized Light

page 31

James Bjorken (Bj), who played a central role in thedevelopment of the quark-parton model at SLAC from1963–1979, spent ten years at Fermilab in 1979–1989,including a period as Associate Director of Physics.He is presently co-spokesperson of the Fermilabtest/experiment T864 in the Tevatron collider whichhe describes as being “closer to Roentgen in stylethan to CDF or to DØ.” Since returning to SLAC in1989, he has been active in theoretical and experi-mental activities.

2 FALL 1995

GUEST EDITORIAL

HE FEATURE ARTICLE in the summerissue of the Beam Line was theserendipitous discovery of X rays 100years ago by Wilhelm Roentgen. Upon

observing fluorescence occurring some distanceaway from a Crookes tube he was operating, heset aside his duties as university rector andteacher and retreated into his laboratory to workalone, sharing nothing with colleagues. Hereemerged six weeks later with his discoveriesready for publication and for the subsequentpress conferences.

In this issue is featured the particle-physicsdiscovery of the year, that of the top quark. It isat an opposite extreme, a programmed discoveryunder way for most of a decade at Fermilab, fea-turing two competing experimental groups, theCDF and DØ collaborations, each consisting of400 to 500 physicists, plus countless engineers,technicians, and support personnel, especiallythe many members of the Fermilab AcceleratorDivision who provide the essential ingredient ofthe colliding beams. This article is written by thespokespersons of the two competing collabora-tions and tells much of how scientific discoveryis done in such an extreme social environment.Each collaboration is a community in itself, withthree branches of governance, robust bureaucracyand organization charts, and a considerable levelof specialization.

It happens that over the last year or two Ihave spent a lot of time at Fermilab, much of itprecisely halfway between these behemoth de-tectors. So I watched with interest and bemuse-ment the excitement of the chase, along with thefrustrations, the impatience, the fatigue, the

T

BEAM LINE 3

anxieties, and some of the inevitable politics. Iwitnessed the awkward first stage of CDF evi-dence, when four short draft papers were by com-promise merged into a very long paper. I evenpromised CDF collaborators to read the thing anddid, carefully, beginning to end, and found it aclassic—a masterful exhibit of how science re-sults should (but seldom do) get reported.Throughout I maintained a somewhat dour pos-ture toward those CDF/DØ crowds, baiting themregarding their obsessive pursuit of top to the ap-parent neglect of all else (not quite true ofcourse), including that hidden piece of NewPhysics Beyond the Standard Model lurking un-detected in their data. But when the top quark re-sults eventually bubbled up to the surface, Icould not do anything but be bouyed by them aswell, joyful in the occasion and deeply respectfulof the magnitude of the accomplishment.

The history of physics is full of near-simultaneous discoveries by separate individualsor groups, and with that often has come acrimo-ny and controversy, from Newton and Leibnitz toRichter and Ting, and down to the present time.There has been competition between CDF andDØ as well. In fact, it was built in from the be-ginning by then-director Leon Lederman, whovisited CERN’s big collaborations, UA1 and UA2,while they were discovering intermediate bosonsW and Z and searching for the top quark. AtCERN, it was vital to have two collaborations aschecks and balances, and Lederman upon his re-turn strongly encouraged the creation of the pre-sent DØ collaboration, something which was notin the works prior to that. And the ensuingCDF/DØ competition has served for constructive

purposes; I have never seen this competitivenessto be corrosive. The evidence is in these pages forthe reader to see, in the very fact of co-authorship and in the nature of the interactionsbetween the collaborations as described in the ar-ticle. This piece of competition has been a classact.

Not only has this been true between the col-laborations, but it seems also to have been thecase within them. This is no mean feat, sinceharmony within a big group of strong individual-istic physicists of great talent and often evengreater ego is not easy to maintain. I can do nobetter than quote here what is found near the endof the article, and I do this without regrets forcreating some redundancy:

In the end, the chief necessity for the conver-gence on the top discovery was the willingnessof a collaboration to abide by a majority view.Securing this willingness requires extensive at-tention to the process—of being sure that allshades of opinion, reservations, and alternateviewpoints are fully heard and understood. It ismore important perhaps that each point of viewis carefully listened to than that it be heeded.A fine line in resolving these viewpoints mustbe drawn between autocracy and grass-rootsdemocracy. The process must have the confi-dence of the collaboration, or its generaleffectiveness can diminish rapidly.

These are not mere words, but an account ofsuccessful actions. In this increasingly fractiousworld of ours, it should be read and taken toheart by all those who despair of progress beingmade through reasoning and consensus.

James Bjorken

4 FALL 1995

TOPQUARK

by BILL CARITHERS& PAUL GRANNIS

Two leaders of

the Fermilab experiments

that isolated the top quark

tell the adventure of its discovery.

Ferm

ilab

Vis

ual M

edia

Ser

vice

s

of theDISCOVERY

BEAM LINE 5

MANKIND has sought the elementary building blocks ofmatter ever since the days of the Greek philosophers. Overtime, the quest has been successively refined from the

original notion of indivisible “atoms” as the fundamental elementsto the present idea that objects called quarks lie at the heart of allmatter. So the recent news from Fermilab that the sixth—andpossibly the last—of these quarks has finally been found may signalthe end of one of our longest searches.

Ferm

ilab

Vis

ual M

edia

Ser

vice

s

6 FALL 1995

But the properties of this funda-mental constituent of matter arebizarre and raise new questions. Inparticular, the mass of the top quarkis about forty times that of anyother—a fact which suggests that per-haps it plays a fundamental role inthe question of how the mass of anyobject arises.

IN 1964 Murray Gell-Mann andGeorge Zweig proposed the quarkhypothesis to account for the ex-

plosion of subatomic particles dis-covered in accelerator and cosmic-ray experiments during the 1950s andearly 1960s. Over a hundred new par-ticles, most of them strongly inter-acting and very short-lived, had beenobserved. These particles, calledhadrons, are not elementary; theypossess a definite size and internalstructure, and most can be trans-formed from one state into another.The quark hypothesis suggested thatdifferent combinations of threequarks—the up (u), down (d), andstrange (s) quarks—and their an-tiparticles could account for all of thehadrons then known. Each quark hasan intrinsic spin of 1/2 unit and ispresumed to be elementary, like theelectron. So far, quarks appear tohave no size or internal structure andthus represent the smallest knownconstituents of matter. To explainthe observed spectrum of hadrons,quarks had to have electric chargesthat are fractions of the electroncharge. The u quark has charge 2/3while the d and s quarks have charges−1/3 (in units where the electroncharge is −1).

The observed hadron spectrumagreed remarkably well with theexpected states formed from

combinations of three quarks or aquark-antiquark pair. Quarks alsoseemed to form a counterpart to theother class of elementary particles,the leptons, which then included theelectron (e) and muon (µ) (both withunit charge) and their companionchargeless neutrinos, νe and νµ. Theleptons do not feel the strong inter-action, but they do participate in theelectromagnetic interactions and theweak interaction responsible for ra-dioactive decays. They have thesame spin as the quarks and alsohave no discernible size or internalstructure.

But most physicists were initiallyreluctant to believe that quarks wereanything more than convenient ab-stractions aiding particle classifica-tion. The fractional electric chargesseemed bizarre, and experiments re-peatedly failed to turn up any indi-vidual free quarks. And—as becameapparent from studies of fundamen-tal theories of quarks and leptons—major conceptual problems arise ifthe numbers of quarks and leptonsare not the same.

Two major developments estab-lished the reality of quarks duringthe 1970s. Fixed-target experimentsdirecting high energy leptons at pro-tons and neutrons showed that thesehadrons contain point-like internalconstituents whose charges andspins are just what the quark mod-el had predicted. And in 1974 ex-periments at Brookhaven NationalLaboratory in New York and Stan-ford Linear Accelerator Center(SLAC) in California discovered astriking new hadron at the then verylarge mass of 3.1 GeV—over threetimes that of the proton. This hadron(called the J/ψ after its separate

CDF Collaboration

Argonne National LaboratoryIstituto Nazionale di Fisica Nucleare,

University of BolognaBrandeis University

University of California at Los AngelesUniversity of Chicago

Duke UniversityFermi National Accelerator LaboratoryLaboratori Nazionali di Frascati, Istituto

Nazionale di Fisica NucleareHarvard University

Hiroshima UniversityUniversity of Illinois

Institute of Particle Physics, McGillUniversity and University of Toronto

The Johns Hopkins UniversityNational Laboratory for High Energy

Physics (KEK)Lawrence Berkeley Laboratory

Massachusetts Institute of TechnologyUniversity of Michigan

Michigan State UniversityUniversity of New Mexico

Osaka City UniversityUniversità di Padova, Instituto Nazionale

di Fisica NucleareUniversity of Pennsylvania

University of PittsburghIstituto Nazionale di Fisica Nucleare,

University and Scuola NormaleSuperiore of PisaPurdue University

University of RochesterRockefeller University

Rutgers UniversityAcademia Sinica

Texas A&M UniversityTexas Tech UniversityUniversity of Tsukuba

Tufts UniversityUniversity of Wisconsin

Yale University

names in the two experiments) wasfound to be a bound state of a newkind of quark, called charm or c, withits antiquark. The c quark has amuch greater mass than the firstthree, and its charge is 2/3. With twoquarks of each possible charge, asymmetry could be established be-tween the quarks and the leptons.Two pairs of each were then known:(u,d) and (c,s) for quarks and (e,νe) and(µ, νµ) for leptons, satisfying theo-retical constraints.

But this symmetry was quicklybroken by unexpected discoveries. In1976 experiments at SLAC turned upa third charged lepton, the tau lep-ton or τ. A year later at FermiNational Accelerator Laboratory inIllinois a new hadron was discoveredcalled the upsilon or ϒ, at the hugemass of about 10 GeV; like the J/ψ, itwas soon found to be the bound stateof yet another new quark—the bot-tom or b quark—and its antiparticle.Experiments at DESY in Germanyand Cornell in New York showedthat the b quark has spin 1/2 and acharge of −1/3, just like the d and squarks.

With these discoveries, andthrough the development of the Stan-dard Model, physicists now under-stood that matter comes in twoparallel but distinct classes—quarksand leptons. They occur in “gener-ations” of two related pairs with dif-fering electric charge—(+2/3, −1/3)for quarks and (−1, 0) for leptons (see

BEAM LINE 7

+2/3 up (u) charm (c) top (t)-1/3 down (d) strange (s) bottom (b)

-1 electron (e) muon (µ) tau (τ)0 electron neutrino (νe) muon neutrino (νµ) tau neutrino (ντ)

Electric First Second ThirdCharge Family Family Family

QUARKS

LEPTONS

DØ Collaboration

Universidad de los AndesUniversity of Arizona

Brookhaven National LaboratoryBrown University

University of California, DavisUniversity of California, Irvine

University of California, RiversideLAFEX, Centro Brasileiro de Pesquisas

FísicasCentro de Investigacion y de

Estudios AvanzadosColumbia University

Delhi UniversityFermi National Accelerator Laboratory

Florida State UniversityUniversity of Hawaii

University of Illinois, ChicagoIndiana University

Iowa State UniversityKorea University

Kyungsung UniversityLawrence Berkeley Laboratory

University of MarylandUniversity of Michigan

Michigan State UniversityMoscow State UniversityUniversity of Nebraska

New York UniversityNortheastern University

Northern Illinois UniversityNorthwestern UniversityUniversity of Notre Dame

University of PanjabInstitute for High Energy Physics,

ProtvinoPurdue University

Rice UniversityUniversity of Rochester

Commissariat à l’Energie Atomique,Saclay

Seoul National UniversityState University of New York,

Stony BrookSuperconducting Super Collider

LaboratoryTata Institute of Fundamental Research

University of Texas, ArlingtonTexas A&M University

chart above). Ordinary matter is com-posed entirely of first-generation par-ticles, namely the u and d quarks,plus the electron and its neutrino.But the third-generation quark dou-blet seemed to be missing its charge+2/3 member, whose existence wasinferred from the existing pattern. Inadvance of its sighting, physicistsnamed it the top (t) quark. Thus be-gan a search that lasted almosttwenty years.

USING THE RATIOS of the ob-served quark masses, somephysicists naively suggested

that the t might be about three timesas heavy as the b, and thus expect-ed that the top would appear as aheavy new hadron containing a tt–

pair, at a mass around 30 GeV. Theelectron-positron colliders then un-der construction (PEP at SLAC andPETRA at DESY) raced to capture theprize, but they found no hint of thetop quark.

In the early 1980s a new class ofaccelerator came into operation atCERN in Switzerland, in whichcounter-rotating beams of protonsand antiprotons collided with an en-ergy of about 600 GeV. The protonsand antiprotons brought their con-stituent quarks and antiquarks intocollision with typical energies of 50to 100 GeV, so the top quark searchcould be extended considerably. Be-sides the important discovery of theW and Z bosons that act as carriers

8 FALL 1995

rely on the production of separate tand t̄ quarks from annihilation of in-coming quarks and antiquarks in theproton and antiproton, with subse-quent decays into observable parti-cles (see box on the right). The designof DØ stressed recognition of the tell-tale leptons and jets over as large asolid angle as possible. MeanwhileCDF had installed a new vertex de-tector of silicon microstrips near thebeams intended to detect short-livedparticles that survive long enough totravel a millimeter or so from the in-teraction point. This detector wasparticularly good at sensing the pres-ence of the b quarks characteristic oftop decay. Another method of taggingb quarks, by detecting their decaysinto energetic leptons, was used byboth experiments. Thus the two ex-periments, while searching for thesame basic decay sequence, hadrather complementary approaches.

With the data from this 1992–93run, progress accelerated. First, DØpublished a new lower limit of131 GeV on the possible top massfrom the absence of events with thecharacteristic dilepton or single lep-ton signatures. This paper turned outto be the end of the line in excludingtop masses. Up to then, with theabsence of an excess of candidateevents, the analyses had tended to re-strict the event sample, so as to setas high a limit on the mass as pos-sible. But by summer of 1993, CDFnoticed that the top mass limits werenot improving with additional data,due to a growing handful of eventsthat passed all preassigned criteria.Naive (in retrospect) estimates oftheir statistical significance triggeredintense activity to review the resultsand prepare a draft publication with

of the unified electroweak force, theCERN experiments demonstrated an-other aspect of quarks. Thoughquarks had continued to elude directdetection, they can be violently scat-tered in high energy collisions. Thehigh energy quarks emerging fromthe collision region are subject to thestrong interaction as they leave thescene of the collision, creating ad-ditional quark-antiquark pairs fromthe available collision energy (us-ing E = mc2). The quarks and anti-quarks so created combine intoordinary hadrons that the experimentcan detect. These hadrons tend tocluster along the direction of the orig-inal quark, and are thus recorded asa “jet” of rather collinear particles.Such quark jets, previously sensed atSLAC and DESY, were clearly ob-served at CERN and became a key in-gredient in the next round of topquark searches.

With the advent of the CERN col-lider, and in 1988 the more powerful1800 GeV collider at Fermilab, thesearch for the top quark turned tonew avenues. At the large massesnow accessible, the tt– bound statewas unlikely to form and isolated topquarks were expected. For masses be-low that of the W boson, W decayinto a t and b̄

–could predominate.

Some indication of this process wasreported in 1984 by the CERN UA1experiment, but it was later ruled outby the CERN UA2 and Fermilab CDFexperiments. By 1990 CDF had ex-tended the top mass limit to 91 GeV,thus eliminating the possibility forW decay to top.

IN 1992, the DØ detector joinedCDF as a long Tevatron run began.Further searches would have to

π+

π–

K+

K–

u

u

p

p–

u

u

d

s–

u–

u–u–

d

g

dd–

d–

d–

s

Jet

ACOLLIDING PROTON (p) andantiproton (p−) bring togetherquarks uud and uud. In this

example, a u and u− scatter byexchanging a gluon (g), the carrierof the strong nuclear force. Like allquarks, the scattered u quark pos-sesses the strong interaction“charge” called color. The strongattraction of the u quark color to theother quark color charges preventsit from escaping freely. Energy fromthe collision region is converted tomatter in the form of quark-antiquark pairs. The antiquark (with anti-color) joins with a quark to produce a colorless meson whichis free to escape.

In this example, the primaryscattered u quark is accompaniedby (dd

−), (uu

−), (ss

−), and (dd

−) crea-

ted pairs, leading to the formation ofπ+, π−, K +, and K − mesons, all trav-eling in similar directions and form-ing a “jet” of particles that may bedetected in the experiment. (The re-maining d quark will be joined toone of the other spectator quarks inthe collision.)

Jet Production

BEAM LINE 9

the goal of publishing the results bythe October 1993 pp− Workshop inJapan. As the analysis was too com-plex for a single journal letter, CDFplanned a series of four papers inPhysical Review Letters (PRL)—twodevoted to the counting experimentsin the two search topologies; one de-scribing the kinematics of the eventsand giving an estimate of the topmass; and a fourth to bring it all to-gether with the overall significance,conclusions, and top productioncross section. In a heated argumentat the October collaboration meet-ing, it became clear both that themass and kinematics sections need-ed more work and that the “fourPRL” format was not working well.CDF decided instead to prepare a sin-gle long Physical Review article, toconcentrate on the remaining holesin the analysis, and to forego any pub-lic discussion of new results.

By the next CDF collaborationmeeting in late January, the count-ing experiments were complete, themass and kinematics analyses hadmade real progress, and the singlelong draft was in reasonable shape.The big remaining issues were thewordings of the title and conclusionssection and the “between the lines”message that they sent to the com-munity. Traditionally, scientific pub-lications about new phenomena usecertain key phrases: “Observationof” indicates a discovery when thecase is unassailable; “Evidence for”means the authors believe they areseeing a signal but the case is notiron-clad; and “Search for” impliesthe evidence is weak or non-existent.Within every collaboration, there isa wide spectrum of comfort levels inclaiming a new result. On one end of

CDF, a few physicists felt that nopublication should occur until thecollaboration consensus was for “ob-servation.” The other end of the spec-trum felt that the data were alreadyin that category. Finally, CDF settledon a conservative interpretation andwas careful to disclose even those re-sults that contra-indicated top.

The pressure within CDF to pub-lish increased enormously as thewinter 1994 conferences approached.In addition, pirated copies of thePhysical Review D manuscript werecirculating widely and rumors (someastoundingly accurate and otherscomically off-base) were flying aboutthe globe. The rumor mill churnedonce again when CDF withdrew itstop talks at the LaThuile andMoriond Conferences. An April CDFCollaboration meeting was the wa-tershed event, as the group ham-mered out the final wordings. Aftera virtually unanimous agreement onthe final draft, the collaborationspokespersons finally notified Fer-milab Director John Peoples of theirintent to publish.

This CDF analysis found 12 can-didate events that could not be eas-ily explained on the basis of knownbackgrounds (which were predictedto be about six events). The odds forbackground fluctuations to yield theobserved events were low—about 1in 400—but not small enough to war-rant the conclusions that the top hadbeen found. Under the hypothesisthat the data did include some tt

−

events, the mass of the top was es-timated to be 175 GeV±20 GeV. Thecross section for tt

−production was

determined to be about 13.9 pico-barns, larger than the theoretical pre-diction of 3–7 picobarns.

Top Production and Decay

FOR TOP QUARK MASSESabove MW + Mb, top productionproceeds mainly through anni-

hilation of incoming quarks and anti-quarks from the beam particles.

The q and q− fuse briefly into a gluon,the carrier of the strong force, andthen rematerialize as t and t− quarkstraveling in roughly opposite direc-tions. The top quarks have a veryshort lifetime (about 10−24 seconds)and decay almost always into a Wboson and a b quark (t → W +b;t− → W −b

−). The b quarks evolve into

jets seen in the detectors. The W +

and W − have several decay possi-bilities:

W ± → e±ν (fraction = 1/9)→ µ±ν (fraction = 1/9)→ τ±ν (fraction = 1/9)→ qq− (fraction = 2/3).

The q and q− from W decays appearas jets. The final states containingτ’s are difficult to isolate and werenot sought in the experiments. Thefinal state with both W ’s decayinginto qq−, though relatively copious,are buried in huge strong interactionbackgrounds. The remaining de-cays, studied by both CDF and DØ,can be categorized by the numberleptons (l = e or µ) from the W de-cays:

Dilepton channel:tt−→W +bW −b

−→ l +l−ννbb

−

Single lepton channel:tt−→W +bW −b

−→ l+νqq

−bb

−

or l −ν−qq−bb

−

The dilepton channels are relativelyfree from background but compriseonly 5 percent of all tt

−decays. The

single lepton channels give largeryields (30 percent of all decays) buthave substantial backgrounds fromW + jets production processes whichmust be reduced.

q t

t–q–

gluon

The mass of the top quark can bereconstructed from the energies anddirections of its decay products asmeasured in the detectors using theconservation laws for energy andmomentum. Since the top quark hasa unique mass, the data (indicatedby the black histograms) shouldshow a “peak” in the reconstructeddistribution. The non-top back-ground (the red dashed curve forDØ and the red dotted curve forCDF) has very different shapes. Asimulation is required to provide thecorrespondence between the mea-sured jet energies and the parentquark momenta. The red dottedcurve for DØ shows the expectedcontribution from top for the best fitvalue of the top mass. The solid redcurve shows what a simulated topquark mass distribution would looklike when added to the background,and these curves should be com-pared to the actual data.

10 FALL 1995

DØ used the period while CDF waspreparing its publication to re-optimize its own earlier analysis,focusing on higher mass top. By April1994, DØ was also in a position toshow some of its new results, whichwere subsequently updated in thesummer Glasgow conference andpublished in Physical Review Let-ters. DØ also had a small excess oftop-like events but with smaller sta-tistical significance; its analysis wasbased upon the dilepton and single-lepton channels with a combinationof lepton tags and topological sup-pression techniques to reduce back-ground. Nine events were observed,compared with an expected back-ground of about four, giving odds fora background fluctuation of about1 in 40. The excess events corre-sponded to a cross section of 8.2±5.1picobarns. The expected yield of tt

−

events was virtually the same for DØand CDF. Taken together, these re-sults from CDF and DØ were not suf-ficient to establish conclusively theexistence of the top quark.

The final chapter in finding thetop quark began with the resumptionof the collider run in late summer1994. The performance of the Teva-tron was the key to the success. TheTevatron involves a collection of sev-en separate accelerators with a com-plex web of connecting beam lines.Many technical gymnastics are re-quired to accelerate protons, producesecondary beams of antiprotons froman external target, accumulate andstore the intense antiproton beams,and finally inject the counter-rotat-ing beams of protons and antiprotonsinto the Tevatron for acceleration to900 GeV. Enormous effort had beenpoured into understanding and

tuning each of the separate elementsof the process, but until summer1994 the intensity of the collider wasdisappointing. During a brief mid-summer break, however, one of theTevatron magnets was found to havebeen inadvertently rotated. With thisproblem fixed, beam intensities roseimmediately by a factor of 2. Withthe now good understanding of theaccelerator, a further doubling of theevent rate was accomplished byspring 1995. In a very real sense, thefinal success of CDF and DØ in dis-covering top rested upon the superbachievements of the Fermilab Ac-celerator Division. The improved op-erations meant that the data samplesaccumulated by early 1995 were ap-proximately three times larger thanthose used in the previous analyses,and both experiments were nowpoised to capitalize on the increase.

By December, both collaborationsrealized that the data now on tapeshould be enough for a discovery, ifthe earlier event excess had been ap-proximately correct. In fact, the ex-periments do not keep daily talliesof the number of events in their sam-ples. The physicists prefer to refinetheir analysis techniques and selec-tion parameters in order to optimizethe analysis on simulated events be-fore “peeking” at the data. This ret-icence to check too often on the realdata stems from the desire to avoidbiasing the analysis by the idiosyn-crasies of the few events actuallyfound. At the beginning of January,DØ showed a partial update in theAspen Conference using some newdata but retaining previous selectioncriteria; these results had increasedsignificance, with only a 1 in 150chance of background fluctuations.

2

4

6

0100 200

Eve

nts

/(20

GeV

/c2 )

Fitted Mass (GeV/c2)

DO

5

4

3

2

1

080 120 160 200 240 280

Eve

nts

/(10

GeV

/c2 )

Reconstructed Mass (GeV/c2)

CDF

Top Mass Distribution

World Wide Web Statistics

BOTH THE CDF and DØ WorldWide Web pages can beaccessed from the Fermilab

home page at http://www.fnal.gov/(click there for top quark discovery).Since the announcement of the dis-covery of the top quark last March,the CDF and DØ home pages havereceived over 100,000 hits each.There were 2279 hits on the CDF

page alone on the day after theannouncement. Six months after thediscovery, the Fermilab top quarkpage is receiving about 200 hits aday.

BEAM LINE 11

Its simulations showed that for thelarge-mass top now being sought,even better significance could be ob-tained. Recognizing that the statis-tics were nearly sufficient, the col-laboration began working in highgear to finalize the selection crite-ria and obtain the last slug of data be-fore a late-January Tevatron shut-down. In mid-January, CDF updatedall of its data; with all of the piecesassembled side by side, it becameclear that the collaboration had thenecessary significance in hand forclaiming a discovery. The CDFprocess was now considerablystreamlined by the experience of pro-ducing the earlier top quark papers.

By February, both CDF and DØ rec-ognized that convergence was im-minent. Both worked around theclock to complete all aspects of theanalyses. Even though each collab-oration knew that the other was clos-ing fast and in the process of writingits paper, there were still no formalexchanges of results, nor of intend-ed timing. Some copies did leakout—for example, drafts left on print-ers in institutions with physicists onboth collaborations and computerhacking from one collaboration toanother. The one avenue of cross-fertilization that seems not to haveoperated was through couples withone person in each group!

On February 24, CDF and DØ“Observation” papers were submit-ted to Physical Review Letters (PRL);public seminars were scheduled atFermilab for March 2. By agreement,the news of the discovery was not tobe made available to the physicscommunity or news media until theday of the seminars. In spite of all ef-forts, word did leak out a few days

before. (Los Angeles Times reporter,K. C. Cole, called a distinguished Fer-milab physicist to get an explanationof statistical evidence presented inthe O.J. Simpson trial. The physicistused, as illustration, “the recent sta-tistical evidence on the top quarkfrom CDF and DØ,” and Cole swift-ly picked up the chase.)

In its paper, CDF reported findingsix dilepton events plus 43 single-lepton events (see box on the nextpage for more details); it concludedthat the odds were only one in a mil-lion that background fluctuationscould account for these events. DØ,in its paper, observed three dileptonevents plus 14 single-lepton eventsand concluded that the odds weretwo in a million that these couldhave been caused by backgrounds.The top quark masses reported by thetwo experiments were 176±13 GeVfor CDF and 199±30 GeV for DØ. Andboth experiments gave consistent re-sults, although somewhat belowCDF’s earlier value for the tt

−pro-

duction cross section (see box on thenext page for more detailed infor-mation).

The top quark appears to be apoint-like particle; it has no internalstructure that we can discern. As oneof the six fundamental constituentsof matter, it has properties very sim-ilar to the up and charm quarks, withthe exception of its remarkable mas-siveness and its very short lifetime.The top quark is about 200 timesmore massive than the proton, aboutforty times that of the second heav-iest quark (the b), and roughly thesame as the entire mass of the goldnucleus! Surely this striking obesityholds an important clue about howmass originates.

by the need to identify the correct combination of jets withparent quarks in the decay and to accommodate the ten-dency of the strong interaction to generate additional jets.The two experiments obtained consistent results for thismass measurement: 176±13 GeV for CDF and 199±30 GeVfor DØ.

The rate for observing tt−

pairs is controlled by the strong-interaction production cross section. Each experiment evalu-ated this cross section at their mean mass value; CDFobtained 6.8+3.6

-2.4 picobarns, and DØ obtained 6.4±2.2 pico-barns. These values are consistent with the theoretical pre-diction within the combined experimental and theoreticalerror, although the tendency for the measurements to fall above theory is interesting and if it persists could indicate that new physics is at play.

12 FALL 1995

THE RESULTS FROM THE TWO COLLABORATIONSwere remarkably similar. CDF found 6 dilepton eventswith a background of 1.3; 21 single-lepton events in

which 27 cases of a b quark tag by the vertex detector (with6.7 background tags expected); and 22 single-leptonevents with 23 cases of a b tag through leptonic decay (with15.4 background tags expected). DØ found 3 dileptonevents (0.65 background events); 8 single-lepton eventswith topological tagging (1.9 background events); and 6 sin-gle-lepton events with b-to-lepton tags (1.2 backgroundevents). A particularly striking example of a dilepton eventwith very energetic electron, muon, and missing ET (due tothe neutrinos), plus two jets, is shown below from the DØdata. The plot shows the detector unfolded on to a plane,with the energy of the various objects indicated by theheight of the bars. This event has a very low probability tobe explained by any known background. The probabilitythat background fluctuations could explain the observedsignal was one-in-a-million for CDF and two-in-a-million forDØ—sufficiently solid that each experiment was able toclaim the observation of the top independently.

Additional studies helped to establish that the new signalwas indeed the top quark. Both experiments were able toshow that the candidate events were consistent with theexpected presence of b quarks. The single-lepton channelevents should have one W boson that decays to a qq− finalstate, and the data showed the expected enhancement inthe di-jet mass.

Finally, both experiments made a measurement of thetop quark mass, using the single-lepton events. In thisstudy, the missing neutrino from one W decay must beinferred and the mass of the parent top quark deduced froma constrained-fit procedure. This mass fitting is complicated

CDF AND DØ RESULTS

3 meters

Primary Vertex Secondary

Vertex

5 millimeters

Jet #2Jet #3

e+

Jet #1

Jet #4

CDF

6.4

3.2φ 0.1

3.7

–3.7

0η

100

10

1

ET

(G

eV)Missing ET

Electron Muon

JET

JET

DO

BEAM LINE 13

MODERN HIGH energy phys-ics experiments are uniqueinstitutions; though large

enough to require formal organiza-tions and governance procedures,they remain voluntary collaborationsof individual scientists who havebanded together to permit achieve-ments that a few physicists could notaccomplish by themselves. But lack-ing the usual structures of largeorganizations—the ability to hire andfire, or to rigorously assign jobs to in-dividual workers—they rely to a verylarge degree on the existence of ashared purpose and consensus fortheir constructive operation.

Although about 430 physicistsparticipate in each collaboration, thenumber active in the top studies ismuch less—roughly 50 directly en-gaged at a significant level. Thisresults from the broad scope of col-lider experiments; the range of tasksneeding attention is very large. Thephysics research being done by thecollaborations is widely dispersedover five major areas (top search,bottom-quark physics, strong-interaction studies, electroweak bo-son measurements, and searches fornew phenomena). Within thesephysics groups there are perhaps 50active analyses under study at anygiven time. Often these other analy-ses contributed synergistic tech-niques to aid the top search; for ex-ample, studies of jets by thestrong-interaction group entereddirectly into measurements of thetop mass.

A diversity of tasks occupy themembers of the collaborations. In ad-dition to running the experimentaround the clock and attending to itsmaintenance, constant attention

must be given to optimizing theevent selection process (only ten in-teractions out of every million canbe kept for analysis) and to improv-ing the software algorithms for se-lecting particles in the data. Workmust continue to improve the over-all computing and software environ-ment. New detector developmentis conducted with an eye to futureupgrades, and major design and build-ing projects are underway for futureruns. The experiments are so big thatreviews by the collaborations them-selves, the Laboratory, and externalcommittees continually occur.Finally, with collaborations of thesesizes, group psychologists, financialanalysts, and tea-leaf readers areneeded!

The techniques for the top quarksearch and measurement of its massdeveloped over several years. Formost important aspects (and someminor ones) several different ap-proaches were developed by rival sub-groups or individuals. This multi-plicity is valuable for cross-checkingresults and as a tool for natural se-lection of the strongest methods, butsuch competitiveness and rivalry canalso be debilitating. In some cases,the winner in these rivalries can be

Ferm

ilab

Vis

ual M

edia

Ser

vice

s

Left to right: Giorgio Bellettini, PaulGrannis, John Peoples, HughMontgomery, and Bill Carithers followingthe seminars announcing the simulta-neous discovery of the top quark atFermilab by the DØ and CDF collabora-tions. Grannis and Montgomery are theco-spokesmen for the DØ Collaboration,John Peoples is the Director of Fermilab,and Carithers and Bellettini are the co-spokesmen for the CDF Collaboration.The collaborations presented theirresults at seminars held at Fermilab onMarch 2, 1995.

14 FALL 1995

for internal use, proliferate and aredisseminated electronically acrossthe globe. Electronic mail is an es-sential component; colleagues work-ing a continent apart on related prob-lems can conduct an almost on-linedialog. Discussion and argument overfine points of interpretation can gen-erate a flurry of productive messagesamong five or six experts. At the oth-er end of the scale, there remains anessential role for formal presenta-tions of work to the full physicsgroup and full collaboration at reg-ular intervals. Repetition is usuallynecessary to bring the non-experts toa reasonable degree of familiaritywith the issues and methods, andto build acceptance of new work.

To what extent did CDF and DØinteract during the time of the topsearch and discovery? At a formallevel, very little. New results fromeach collaboration were periodicallyshown at conferences and public pre-sentations, so the broad outlines ofeach group’s approach were known,but beyond that there were no directinteractions. The grapevine was ofcourse active, with lunch-table con-versations among individuals, butmuch of the information exchangedin this way was the anecdotal ac-count of the day and was often out-moded by the next. There were prob-ably several reasons for this relativeisolation of two groups working to-ward a common goal. At one levelboth groups realized that the inde-pendence had real scientific value.Guarding against making a discov-ery by subconsciously shaping theanalysis is an ever-present worry; bykeeping the walls between CDF andDØ fairly high, this possibility couldbe minimized. Also, the two groups

established on purely technicalgrounds—but often the basis for de-cision is clouded, as the alternativeshave both pros and cons. Adjudicat-ing among these contenders is a dif-ficult and time-consuming problem.In some cases, the choice is madethrough a survival-of-the-fittestprocess—the approach first broughtto full maturity, with documentederror analyses and cross-checks, ischosen. Consensus plays an impor-tant role in rejecting some alter-natives; experts not involved in aparticular question can wield con-siderable influence in rejecting someweaker approaches. When the issueshave sufficiently crystallized, specialadvisory panels drawn from the restof the collaboration can be very use-ful in helping to define the direction.The last resort of turning the issueover to the “managers”—the spokes-persons or physics group conveners—is successful primarily when therivals have agreed that they are un-able to resolve their differences. For-tunately, this rarely happens.

Effective communication onmany levels is an essential ingredi-ent of successful performance in thelarge collaborations. For the topstudies, weekly meetings of a halfdozen subgroups of the top analysisdiscuss problems and status; as manyadditional meetings occur on relat-ed analysis topics such as electronidentification; and a more generalmeeting brings together all thethreads. Video conferencing and elec-tronically posted minutes of themeetings help keep those not at Fer-milab involved, although most ac-tive participants travel there fre-quently. Informal documents pre-pared on sub-topics, intended only

felt the heat of competition; for each,no doubt, the dream of being firstto recognize the top was a powerfulincentive to be rather closed-mouthed about innovations andprogress.

The history of the top search overthe final two years shows evidenceof this competitiveness and friend-ly rivalry at work. As the more se-nior collaboration, CDF had devel-oped its techniques over time to arelatively refined state. It started therace for top with the UA2 experimentin 1988 and had been denied successonly by top’s extraordinary mas-siveness. CDF would have been dis-appointed to lose the race to the new-er DØ experiment. On the otherhand, the Tevatron performance inthe 1992–1994 era improved so muchthat the data in hand was almost allacquired after DØ’s start-up, so thetwo experiments could work on al-most equivalent data sets. DØ, withits strengths of excellent lepton iden-tification and full solid angle cover-age, was able to develop powerfultools for suppressing unwanted back-ground; it was of course determinedto make its mark as the brash new-comer by converging on the top. Butthe competition remained friendlyand supportive throughout; it alsopromoted a quality of work and depthof understanding that would havebeen difficult otherwise. And the ul-timate scientific value of two inde-pendent results was enormous.

How do hundreds of authors agreeon a single paper in a timely fashion?This too is a sociological challenge!The process for writing the final pa-pers began in late January. DØ ap-pointed primary authors—one for ashort letter paper and one for a longer

BEAM LINE 15

paper ultimately not used. The topgroup designated a few individualsas its authorship committee, bring-ing the necessary range of expertiseto the authors. The collaboration re-view was conducted in large partthrough an appointed editorial boardof ten physicists including authors,drawn from widely diverse portionsof the collaboration. The editorialboard spent days carefully reviewingthe backup documentation, probingthe analysis for consistency andcorrectness. It also helped to revisethe paper’s language to make it ac-cessible to the widest possible au-dience. The completed draft was dis-tributed to the full collaboration byelectronic mail, and comments weresolicited from all. Well over a hun-dred written comments were re-ceived (plus many more verbal ones),all of which were addressed by theboard. The revised paper was againcirculated for concurrence of the col-laboration before submission to PRL.

CDF appointed a chief author to beassisted by a group advising on thesections of their expertise. Simulta-neously, a review committee (dubbeda “godparent” committee withinCDF) was appointed. The draft andresponses to questions were alwaysaccessible to the collaboration elec-tronically, both on the World WideWeb (with security) and via ordinaryfiles on the Fermilab computer.

In the final convergence to a re-sult, it was necessary to come to gripswith rather large differences of sci-entific style among the collaborators.Some could find in any result, nomatter how thoroughly cross-checked, the need for further inves-tigation. Some were wary of anyinterpretation beyond a statement of

Happy and relieved physicists who havecompleted PhD theses using the DØexperiment assemble in the DØ main controlroom. Students and postdoctoral candidatesare the lifeblood of the experiment and havemade contributions to every aspect of it.Pictured are Brad Abbott, Purdue; Jeff Bantly,Srini Rajagopalan, Northwestern; DhimanChakraborty, Terry Heuring, Jim Cochran, GregLandsberg, Marc Paterno, Scott Snyder, JoeyThompson, Jaehoon Yu, Stony Brook; ReginaDemina, Northeastern; Daniel Elvira, CeciliaGerber, University of Buenos Aires; Bob Hirosky,Gordon Watts (CDF), Rochester; Jon Kotcher,New York University; Brent May, Arizona; DougNorman, Maryland; Myungyun Pang, Iowa State.

Others who have completed theses onDØ but not pictured are Rich Astur, Bo Pi, SalFahey, Michigan State; Balamurali V., Notre Dame;Ties Behnke, John Jiang, Domenic Pizzuto, PaulRubinov, Stony Brook; John Borders, SarahDurston, Rochester; Geary Eppley, Rice; FabriceFeinstein, Alain Pluquet, University of Paris Sud;Terry Geld, Michigan; Mark Goforth, RobertMadden, Florida State; Ray Hall, Thorsten Huehn,University of California, Riverside; Hossain Johari,Northeastern; Guilherme Lima, LAFEX/CBPF;Andrew Milder, Alex Smith, Arizona; Chris Murphy,Indiana; Haowei Xu, Brown; Qiang Zhu, New YorkUniversity.

the basic experimental numbers.Others favored a more interpretativeapproach. Finalization of the paperitself required the resolution of theseoft-conflicting outlooks on whatmakes good science.

In the end, the chief necessity forthe convergence on the top discov-ery was the willingness of a collab-oration to abide by a majority view.Securing this willingness requires ex-tensive attention to the process—ofbeing sure that all shades of opinion,reservations, and alternate view-points are fully heard and under-stood. It is more important perhapsthat each point of view is carefullylistened to than that it be heeded.A fine line in resolving these view-points must be drawn between autoc-racy and grass-roots democracy. Theprocess must have the confidence ofthe collaboration, or its generaleffectiveness can diminish rapidly.

What did the physicists of CDFand DØ feel on completion of the topdiscovery experiments? Certainly re-lief that the quest was complete. Also

Ferm

ilab

Vis

ual M

edia

Ser

vice

s

16 FALL 1995

pride that such an elusive quarry hadbeen trapped and that the two inde-pendent experiments had drawn sosimilar a profile of the top quark. Vir-tually all in both collaborations couldtake justifiable pride for their per-sonal contributions to the discov-ery—an effective electronics con-troller for the high-voltage systemwas as essential for the discovery asdevising the selection criteria for thetop quark events. Despite the manystrains introduced by the freneticpace, the process of discovering thetop was for most the scientific eventof a lifetime. Graduate students, post-docs, junior and senior faculty, andscientists worked together on equalfooting, with the contributions of theyoungest as significant as those fromthe oldsters. Though for most scien-tists, solving the many small dailyproblems gives sufficient satisfactionto keep them hooked on physics,reaching a milestone touted as a ma-jor discovery was a special thrill notto be forgotten.

After the public announcementsof the discovery, a new challengearose in interacting with the newsmedia, representatives of member in-stitutions, and other scientists to ex-plain and comment on the results.The media blitz caused strains of itsown. As each physicist had con-tributed in some major way to thediscovery, most wanted to share inthe public limelight. In the unfa-miliar world of public exposure,some strains and disappointmentsemerged, but the opportunity to ex-plain the significance of the workto the general public was for most arewarding and educaional experience.

How long did the euphoria of thediscovery last? There were indeed

celebrations, but by a week after thepublic announcements it was backto business, both on the top studiesand on the dozens of other analysesin progress.

DOES THE DISCOVERY of thetop quark close the chapteron our understanding of the

fundamental building blocks of mat-ter? Surely not—it is truly just thebeginning. Though we now havefound the long-awaited sixth quark,with properties as predicted, muchmore accurate measurements areneeded. Dozens of questions abouthow the top is produced and how itdecays are clamoring for answers.Does the extraordinarily large topquark mass hold clues to the natureof symmetry breaking or to the ori-gin of mass? Do the regularities ob-served for the lighter quarks hold forthe top? Are there signs of new par-ticles that decay into the top? Or arethere exotic new objects to be foundin top decays? The very massive topis unique among quarks in that itdoes not live long enough to bebound into hadrons, so it providesa testing ground for the study of a sin-gle “bare” quark. As happens so of-ten in physics, the latest great dis-covery should serve as the crucialtool for the next advance in under-standing the composition of matter.

BEAM LINE 17

THE GHOSTLIEST of elementary particles,neutrinos have fascinated particle physicistsfor half a century and become the objects of

intense research these past few decades. Because they in-teract very weakly with matter, these “little neutral ones”also provide scientists a unique way to peer inside starsand supernovae. Thermonuclear processes occurring with-in the cores of these celestial objects generate profusestreams of energetic neutrinos that easily penetrate theouter layers and travel enormous distances, to be recordedhere on Earth by huge underground detectors.

In Water or Ice?by JOHN LEARNED & MICHAEL RIORDAN

Physicists are building four huge

particle-detector arrays to do high

energy neutrino astronomy.

18 FALL 1995

Now, after twenty years of devel-opment, neutrino astronomy is aboutto forge on to its next great challenge:the study of high energy neutrinosfrom cosmological sources outsidethe Milky Way. In four corners of theglobe, four teams of physicists areracing to build the first full-fledgedhigh energy neutrino telescopes. Andin a remarkable example of interna-tional scientific cooperation, theseteams are collaborating with oneanother and with other interestedgroups on the design of a trulygargantuan neutrino telescope—encompassing a cubic kilometer intotal volume—that they hope tocomplete by early in the followingmillennium.

THE ASTROPHYSICAL neutri-nos observed thus far by un-derground detectors arrived at

the Earth with energies of roughlya million electron volts (1 MeV) ormore. Much lower than this and theybecome essentially impossible to de-tect—although they are perhaps themost plentiful particles in Nature,even more common than photons.Physicists have good reason to be-lieve that every single cubic cen-timeter of the Universe contains oversix hundred neutrinos. With an ex-ceedingly tiny mass equivalent tojust a few electron volts, these ubiq-uitous spooks would constitute thedominant portion of all the matterin the Universe, its much sought-after “dark matter” or “missingmass.”

The observation of MeV neutrinoshas allowed astrophysicists to ex-amine the physical processes respon-sible for stellar burning and super-novae explosions. By determining

the intensities and energies of theneutrinos spewing forth from theseobjects, we have in essence been ableto measure their internal tempera-tures. And discrepancies between thepredicted and observed fluxes of so-lar neutrinos arriving at the Earth’ssurface strongly suggest that we maybe witnessing the results of new andfundamentally different physics. (See“What Have We Learned About SolarNeutrinos?” by John Bahcall, in theFall/Winter 1994 Beam Line, Vol. 24,No. 3.)

For at least 20 years, however,many physicists have recognized thatthe best place to do neutrino astron-omy is at extremely high energiesover a trillion electron volts (1 TeV)—the realm of Fermilab energies andbeyond (see box on next page). Be-cause the probability of a neutrino’sinteracting grows in proportion to itsenergy, so does the ease with whichwe can detect it and determine whereit came from. Thus, for neutrino as-tronomers TeV energies are truly theexperimenters’ dream.

They are also the astrophysicists’and cosmologists’ dream, too, for TeVneutrinos give us the only practicalway to observe physical processes oc-curring at the very center of galax-ies—particularly the active galacticnuclei (AGN’s) that have lately be-come a hot research topic. Nowadaysmany scientists suspect that thevariety of highly energetic cosmicbeasts distinguished by their variouspeculiar radiations are all in fact justdifferent manifestations of a galaxywith an AGN, viewed in various per-spectives and at various times in itsevolution. Most believe that theseAGN’s are the sites of galaxy-classblack holes a million to a billion

times as massive as the Sun; theyswallow up stars, gas and dust whilebelching back perhaps 10 percent oftheir feast in wildly transformedways. Such a tiny, massive object wasrecently shown to exist at the coreof the elliptical galaxy M87, whichis practically in our own cosmolog-ical backyard.

According to one recent scenario,there is a standing shock wave out-side the event horizon of the blackhole, a bit like the standing waves ina river that delight white water en-thusiasts. Infalling matter carriesmagnetic fields, which are drastical-ly compressed and intensified by thetremendous force of gravity when thematter slams into this shock wave.Charged particles are accelerated tohigh energies as they ricochet backand forth between magnetized re-gions, like a ping pong ball betweena pair of converging paddles. Ener-getic pions created in this processrapidly decay into pairs of gammarays or into muons and neutrinos.The gamma rays initiate particleshowers and generate a hellish pho-ton cloud from which only the neu-trinos can easily escape, streamingaway with perhaps half the total en-ergy emitted by this bizarre object.

This process can explain the in-tensities of X rays and ultraviolet ra-diation emitted by some AGN’s, fromwhich the flux of high energy neu-trinos can then be inferred. A num-ber of theorists have now estimatedthis flux; they agree that a significant(meaning measurable) flux exists upto about 1–10 PeV (1,000–10,000TeV). Quite apart from the great in-terest in using such high energy neu-trinos to observe AGN dynamics, theprospect of detecting them has

BEAM LINE 19

naturally spurred particle physiciststo build big neutrino telescopes. Forthe foreseeable future, we have nochance of producing such particlesin earthbound accelerators.

Of course, this is only one partic-ular model that may in fact be wrong.Over the decades we have dreamedof doing high energy neutrino as-tronomy, many promising modelshave come and gone. But the estab-lished existence of extremely highenergy cosmic rays, with energies upto a hundred million TeV, means thatTeV neutrinos must also be plentifulin the Universe. Given what weknow about particle physics, it is in-conceivable that such tremendousenergy could be concentrated intophotons and protons without a sub-stantial amount of it spilling overinto the neutrino sector.

There are a number of more “pro-saic” sources of high energy neutri-nos coming from within the MilkyWay galaxy itself. One source is theshock waves generated by a super-nova, in a mechanism analogous tothat described above for AGN’s. An-other is the lighthouse-like particlebeam thought to be emanating fromits pulsar remnant. X-ray binaries—in which neutron stars blaze away,fueled by feeding off their compan-ion stars—have also been suggestedto be a potent neutrino source. What-ever the case, TeV neutrinos from ourown galaxy are absolutely guaran-teed. The existing flux of high en-ergy cosmic rays, mostly protonswandering about the galaxy, trappedby its magnetic fields, will lead tocollisions with dust and gas, gener-ating neutrinos that come from thegalactic plane. Observing these neu-trinos may help us solve the still

Optimum Energies for Neutrino Astronomy

BECAUSE THEIR DETECTABILITY improves steeply with energy, neutrinos ofTeV energies and above have long been viewed as the natural starting pointfor neutrino astronomy. But since the flux of neutrinos falls off with energy, anoptimum must exist.

Factors favoring higher energies include the increasing neutrino cross sec-tion (their probability of interacting with matter); the increasing range of themuons produced in these interactions (which means the sensitive volume of adetector grows with energy); the decreasing angle between such a muon andits parent neutrino (and thus the point source direction, yielding better resolu-tion); and the increasing signal-to-noise background ratio between astrophysi-cal neutrinos and those produced in Earth’s atmosphere.

The last factor is a bit speculative but has pretty good justification. Thesebackground neutrinos are produced by ordinary charged cosmic rays interact-ing with air molecules. As the energy increases, a smaller and smaller fractionof the secondary particles produced in these collisions can decay into neutri-nos before hitting the Earth. Such local, atmospheric neutrinos constitute thebackground against which point sources of astrophysical neutrinos must be re-solved. Because the atmospheric neutrino spectrum is expected to fall off moresteeply than those of many astrophysical neutrino sources, higher energies areclearly preferred.

While all of the above factors favor higher energies as something like E4,they begin to saturate at a point that is typically in the energy region 1–10 TeV.This turns out to be the optimum range to begin doing neutrino astronomy.

The neutrino cross section is given as a function of energy in the graph. Itspans a range from 0.01 TeV (1010 eV) to 106 TeV (1018 eV, or 1 EeV); the

lower end of this range corre-sponds to the energies of neutri-nos produced at earthbound par-ticle accelerators, while theupper end approaches the high-est energies so far encounteredin cosmic rays. The shadedcurve peaking at 6.4 PeV repre-sents the cross section for theinteraction of electron antineutri-nos with atomic electrons, lead-ing to the production of the W −

boson in what is known as the“Glashow resonance.” Over the

energy range from 4 to10 PeV, this striking process dominates all the otherneutrino interactions. Because of this large cross section, such electron anti-neutrinos will not penetrate through the Earth; instead, they must be absorbedin such relatively short distances (depending on energy) as a few kilometers.

108

106

104

102

1031 106100

Neutrino Energy (TeV)

σ (ν–e + e– → W – )

σTOTνn

Cro

ss S

ecti

on

(10–3

8 cm

2 )

20 FALL 1995

mysterious puzzle of the origin ofcosmic rays.

A much more exotic potentialsource is the gamma-ray bursts re-cently found to populate the sky uni-formly in all directions. If these weirdobjects are indeed occurring outsideour galaxy, as most astronomers nowbelieve, then a large fraction of a so-lar mass is being converted into en-ergy in each event. The neutrino fluxthat some think must accompanythese tremendous bursts of energymight well be seen by undergrounddetectors now being built. Such anobservation would of course helpwith understanding the nature ofthese enigmatic cosmic explosions.

Finally, big neutrino telescopeswill be important aids in searchingfor a key component of dark matter—the vast halo of weakly interactingmassive particles, or WIMPs, thoughtto cluster about the Milky Way andmost other galaxies. Often thoughtto be supersymmetric particles leftover from the Big Bang, these WIMPsshould become concentrated insidethe Earth’s core and that of the Sundue to glancing collisions with atom-ic nuclei there. Occasionally a WIMPwould encounter its antiparticle andannihilate it, however, producing aspasm of other particles. But onlyneutrinos can escape these core re-gions and reach our detectors, whichwould then record fluxes of high en-ergy neutrinos coming from the cen-ters of the Earth and Sun. The bigneutrino telescopes now being builtor contemplated will be sensitive toWIMP masses from 10 GeV to 1 TeV,covering essentially the entire rangeof masses expected for the lightestparticles predicted by supersymmetictheories.

And with a large enough neutrinotelescope, we can study inner spaceas well as outer space. By countingthe flux of neutrinos and antineutri-nos hitting the detector from differ-ent directions, one could measuretheir attenuation in the Earth and do“Earth tomography”—much likecomputer-aided tomography, or CATscans, are done on people usingX rays. With sufficient statistics, neu-trino telescopes will be able to meas-ure the density of the Earth’s core,for example, a feat that is not possi-ble by any other means.

The possibility of completelyunanticipated discoveries is also apotent reason for building big neu-trino telescopes. The history of as-tronomy has shown us time andagain that the most important ob-servations by a new kind of devicewere not even predicted when thefirst such instrument was being built.High energy neutrino astronomy canand almost surely will open a sur-prising new window on the Universe,with implications for distance scalesfrom the largest to the smallest. It

will bring us a grand sense of wit-nessing a new projection of the lifethat exists outside the murky cavein which we find ourselves.

SCIENTISTS WORKING on neu-trino detection have realized fordecades that the Cherenkov ra-

diation generated by neutrino inter-action products in water (and ice) pro-vides an inexpensive way to buildlarge-volume detectors. Charged par-ticles speeding through transparentmedia at velocities near the speed oflight throw a cone of blue light in theforward direction. Sensitive detec-tors can “see” these particles at dis-tances on the order of 100 meters indeep, clear ocean water (which has acharacteristic attenuation length of50 m). Large photomultiplier tubessensitive to this blue light are avail-able for a few thousand dollarsapiece. The typical cost of such aphotosensor module (including elec-tronics, cabling and housing) comesto about $10,000 each. As each mod-ule covers an effective area of 100m2, the cost per square meter forsuch detectors is roughly $100—farcheaper than typical accelerator-based detectors (but at the cost ofpoorer resolution).

Whenever a muon neutrino (or an-tineutrino) interacts with an atomicnucleus by the exchange of a W par-ticle, it produces a muon that carriesoff the bulk of the neutrino’s energy.Because muons have a long lifetimeand do not interact very stronglywith matter, they can carry the newsof neutrino collisions a very longway—kilometers in the case of veryhigh energy neutrinos. A muon pro-duced at PeV energies, for example,can travel 20 kilometers. It is no

High energy neutrino

astronomy can open

a surprising

new window

on the Universe,

with implications

for distance scales

from the largest

to the smallest.

~100 m

BEAM LINE 21

wonder that neutrino detectorscurrently under construction havefocused upon muon detection inorder to attain the large target vol-umes needed.

The precision with which we canmeasure the direction of the muon—and that of the neutrino producingit—will depend largely on how wellwe can determine the exact times atwhich its Cherenkov light strikesthe individual modules. Fortunate-ly, the latest generation of large-areaphotomultipliers has such good timeresolution (a few nanoseconds) thatwe can expect to attain an angularresolution of about a degree for TeVmuons that traverse the entire array.Neutrino telescopes will not be ableto challenge optical or radio tele-scopes in this department, but theirangular resolution improves gradu-ally with energy. And statistical tech-niques can be used to refine our mea-surements still further, so that wemay eventually be able to pinpoint aneutrino source to perhaps a hun-dredth of a degree.

Electron neutrinos and tau neu-trinos do not produce prompt, highenergy muons—and neither do muonneutrinos when they exchange a Zparticle with the struck nucleus. Inthese cases we have to observe thecascades of secondary particles(mostly hadrons, few of which trav-el very far) produced by the recoilingquarks in the struck nuclei. ForCherenkov detection of these largecascades, we must position the mod-ules with typical spacings deter-mined by the mean transparency ofour medium to blue light. Thereforethe sensitive target volume of our ar-ray for detecting cascades grows onlyas the total volume surrounded by

A typical “double-bang” event that mightbe observed in a very large neutrinotelescope, produced by interaction of amulti-PeV tau neutrino. It generates ahuge shower, together with a tau leptonthat travels around 100 meters beforedecaying in a second shower with abouttwice the energy of the first.

it—unlike that for detecting muons,which increases as the area of the ar-ray. At the scale of several kilome-ters, however, these two begin toconverge—a good thing because wecan learn a lot by detecting all typesof neutrinos simultaneously.

A unique and interesting phe-nomenon to watch for is the pro-duction of W− particles by 6.4 PeVelectron antineutrinos striking elec-trons within the target volume. Thecross section for this interaction ismuch greater than for all other weakinteraction processes at this ener-gy; if there happen to be lots of elec-tron antineutrinos in the mix, itshould lead to a nice spike in the dis-tribution of cascade energies. Asidefrom the aesthetic pleasure of find-ing this so-far unobserved funda-mental process, it will give us a goodenergy calibration and a handle onthe fraction of electron neutrinos atthis energy.

With a large enough detector ar-ray we may be able to observe whatwe call “double-bang” events pro-duced by tau neutrinos of a few PeV.Upon striking a nucleus in our tar-get volume, such a neutrino will of-ten produce a tau lepton that travelsabout a hundred meters before de-caying. There will be one cascadeat the point of collision and anoth-er at the point of tau decay, con-nected by a softly ionizing particlebetween the two. The first cascadewill contain about half the energy ofthe second, on average; both shouldbe easily visible, generating about100 billion photons apiece. With asufficient sample of these double-bang events, we would be able tomeasure the percentage of tau neu-trinos at energies of a few PeV; by

22 FALL 1995

comparing this ratio with those ofelectron and muon neutrinos, wewill be able to study neutrino mix-ing with unmatched sensitivity tomass differences. The observation ofsuch events would also be strong ev-idence for neutrino mass.

In addition to sampling theCherenkov light produced by inter-acting high energy neutrinos, physi-cists have considered the possibilityof detecting the radio and soundwaves generated by their violent col-lisions with matter. Small radio puls-es occur because oppositely chargedparticles in a cascade generate op-posing microwave signals that do notcompletely cancel (owing to the factthat its electrons migrate furtherthan its positrons). The best avail-able medium for detecting such puls-es appears to be polar ice, which willtransmit microwaves for kilometersat temperatures below –60°C. Acous-tic pulses occur because cascades de-posit heat energy instantaneously innarrow cylinders up to about 10 me-ters long. Detecting the sound wavesgenerated (with typical frequenciesof 20 kHz) will be difficult, but it maybe possible in the very quiet deepocean for neutrinos with energiesabove a few PeV. Both techniquesshow good promise for extendingneutrino astronomy to ultrahigh

energies and to truly gargantuan vol-umes, but both will probably have toride piggyback first on a Cherenkov-light detector in order to demonstratetheir feasibility.

FOUR LARGE detectors of highenergy neutrinos are now invarious stages of design and

construction. They sport the namesDUMAND (in the Pacific Ocean offHawaii), AMANDA (in ice at theSouth Pole), Baikal (in Lake Baikal,Siberia) and NESTOR (in the Mediter-ranean Sea near Greece). All are atleast ten times larger than their low-energy brethren—such as Japan’sSuperKamiokande detector—in theirenclosed volumes and muon-catching area, and their threshold en-ergy sensitivity is typically a thou-sand times higher.

Having begun life in a series oflate-1970s workshops, DUMAND isthe elderly sibling of the other pro-jects. As currently envisioned, it willconsist of nine strings of photosen-sor modules in an octagonal array(with one string at the center) sub-merged over 4 km deep off the BigIsland of Hawaii. This array will en-close about 2 megatons of water andhave an effective area of 20,000 m2

for the detection of energetic muons,with an angular resolution of about

Size of Array EstimatedDepth Height Width No. No. Year of

Expt Location (km) Medium (m) (m) Strings Modules Operation

AMANDA Antarctica 1–2 ice 390 110 16 478 2000

Baikal Siberia 1.1 lake 70 43 8 96a 1997

DUMAND Hawaii 4.8 ocean 230 106 9 216 1997

NESTOR Greece 3.7 sea 220 32 7b 168 1997a1 module contains 2 photomultiplier tubesb1 tower is equivalent to 7 strings.

a degree. Since the early 1980s, agroup of physicists from Europe,Japan, and the United States has beenworking from the University ofHawaii, doing R&D on this project—acquiring valuable experience in howto apply the techniques of high en-ergy physics to the deep-ocean en-vironment.

Each DUMAND string contains 24photosensor modules spaced 10 mapart along its instrumentedsection—plus floats to provide thenecessary tension, hydrophones,environmental sensors and an elec-tronics package. It is tethered to theocean bottom with a remotely com-mandable release to facilitate re-covery and allow reconfiguration ofthe array if desired. The electronicsunit communicates with all the sen-sors and sends a data stream to shore(including optical, acoustic, envi-ronmental and housekeeping data)via fiber-optic cable through a com-mon junction box.

In December 1993 the collabora-tion successfully laid the shore ca-ble along the ocean bottom from theexperimental site 30 km west of theBig Island to the laboratory at Kea-hole Point. Attached to this junctionbox before deployment, the first de-tector string worked well for a shorttime but its electronics packagefailed due to a leak. Data acquiredbefore this failure confirmedexpectations about the overall sys-tem performance and backgroundcounting rates. After further ship-suspended tests in the winter of1995–96, placement of three com-plete strings will occur in the sum-mer of 1996.

Having begun South Pole opera-tions in 1991, AMANDA is the

BEAM LINE 23

John

Lea

rned

youngest of the four projects but hasmade impressive progress. Its ap-proach is simple: use hot water todrill holes in the Antarctic ice pack,freeze photomultiplier tubes in

place, and use standard electronicstechniques to get a large neutrino de-tector built rapidly with a minimumof new technology development. Theice is very cloudy near the surface

but was expected to be spectacu-larly clear at some depth beyondabout 500 m.

Led by the University of Wiscon-sin, the collaboration installed four

of the planned six strings, each con-taining 20 modules, at depths of800–1000 m during the summer of1993–94. Upon activating thesestrings, however, physicists foundbubble densities in the ice sub-stantially higher than anticipated;scattering of light from these bubblesmade accurate signal timing impos-sible. In effect, AMANDA has blurry

Cable

DUMAND Array

30 km

4.8 km

440 m

Keahole Point Laboratory

Kauai

Oahu

University of Hawaii

“Big Island” Hawaii

Molokai

LanaiMaui

Pacific Ocean

DUMAND Array

Junction Box

3000 m

1000 m 2000 mElectronics

Unit

Photosensor Modules

Photosensor Modules

Floats

Cable

Right: Physicists deploying the first DUMAND string from the deck of theresearch vessel T. G. Thompson in December of 1993. One of the optical

photosensors hangs in midair as the central electronics package is readiedfor submersion.

Bottom: Artist’s conception of the planned DUMAND array and its deep-sealocation off the Big Island of Hawaii. The cable and junction box are now in

position; another attempt to install the first of nine strings of photosensorsand other detectors will occur in 1996.

24 FALL 1995

vision—as if it were looking formuons through translucent glasses.

The big question is whether theice clarity improves enough withdepth to permit a full-scale experi-ment this austral summer (1995–96).Drilling costs grow with depth, andexpensive fiber-optic techniques maybe needed to obtain good timing.AMANDA will run out of ice at about2500 m, although the team may giveup if the clarity is still poor at1500 m.With the observed low light absorp-tion in ice, however, the array mightstill make an excellent calorimeter—and microwave detection techniqueshold great promise in this medium,too.

A group of Russian and Germanphysicists building the Baikal detec-tor includes some of the elders in thisfield, who began work about thesame time as DUMAND. The deep-est (1.4 km) clear lake in the world,Lake Baikal was chosen in the hopesof avoiding the major backgroundflux of light due to radioactivity, butthere appears to be an obscure, unan-ticipated source that varies withtime—making it about as noisy asthe deep ocean. One advantage isBaikal’s ability to place and retrieveinstruments in winter, without useof ships. Its unique cable-layingmethod uses a sled with a huge sawto cut a slot in the ice, through whicha following sled unreels the cable asboth drive shoreward; the slot rapid-ly freezes behind. So far 36 modules(each with two photomultipliers act-ing in coincidence) have been placed,with data being recorded and ana-lyzed at a shore station. This grouphas suffered from the lack of fundingand high-technology materials, but

the recent addition of the DESY lab-oratory in Hamburg, Germany, to thecollaboration brings great hope ofsolving these problems.

Planned for a deep-sea site south-west of Pylos, Greece, NESTOR isclosest in concept to DUMAND. Ata depth of 3.5 km the Mediterraneanis surprisingly clear; a good site hasbeen located and studied a mere 20km from shore. This site is also closeto a line extended from CERNthrough Italy’s Gran Sasso Labora-tory, so that a proposed neutrinobeam might be extended to thedetector to permit a search forneutrino oscillations. The detectorarray will be a hexagonal cluster ofseven towers, each standing 450 mtall and containing 12 umbrella-likefloors with 14 modules per floor. In-stallation of the first tower is ex-pected to occur in 1997. Led by theUniversity of Athens, the NESTORcollaboration includes teams ofphysicists from France, Greece, Italyand Russia—among them a Saclaygroup that provides the support andinfrastructure of a major French na-tional laboratory.

EVER SINCE the late 1970s it hasbeen clear to neutrino as-tronomers that an array filling

an entire cubic kilometer is reallyneeded to begin serious work in thefield. Everyone recognizes that weare really building demonstrationdevices—perhaps big enough to finda few exciting things—as explorato-ry stepping stones to the huge de-tector we really need to finish the job.Many independent calculations haveconverged upon a cubic kilometer asthe appropriate volume. Such a

A well-clad physicist stands beside the hot-water drill used to bore the first four AMANDAholes 1000 meters into the Antarctic ice cap atthe South Pole. This equipment will be used tobore additional holes to depths of about 1500meters during the austral summer of 1995–96.

Bob

Mor

se

BEAM LINE 25

device should function well in detecting neutrinos frompoint sources, the galactic plane, gamma-ray bursters,AGNs and WIMPs.

In 1994 there were several meetings at which physi-cists discussed the idea of forming a world collaborationto design and build such a gargantuan detector. TheAMANDA and DUMAND teams plus a few newcomershave been working with the Jet Propulsion Laboratory,a major national laboratory in Pasadena, California, withthe big-project experience needed in such an effort, whichis clearly beyond the capability of a single university.The Saclay group joined NESTOR partly in order to gainexperience towards building such a facility. Many ofus hope to merge these two efforts into a single inter-national project.

One thing we are all proud of in this emerging fieldis the tremendously supportive interaction among thephysicists involved. Although there is lots of competi-tion, as all of us naturally want to be first with anyimportant discovery, we realize that by helping eachother we also help ourselves in the long run. All fourteams have meanwhile got their work cut out for them.No cubic-kilometer detector will ever begin construc-tion until we can get at least one (and hopefully more)of the current generation up and counting neutrinos.Only then will we have the information we need todecide whether the next step for neutrino astronomyshould occur in water or ice.

Right: One of the Baikal detector modules being loweredthrough the ice together with cabling and electronics. Eachmodule consists of two photosensors in two separatepressure-resistant glass spheres.Bottom: A prototype NESTOR “floor,” which consists of 14photosensors in an umbrella-like array, being lowered into theMediterranean Sea off Greece. Slated for initial deployment in 1997, each tower in this neutrino telescope will consist of 12such floors, whose arms open radially once in the water.

Leo

Res

vani

sB

ob M

orse

26 FALL 1995

Osteoporosis–New Insightsby JOHN KINNEY

Synchrotron radiation is

giving us new insights

into causes and treatments

for osteoporosis.

STEOPOROSIS IS A DISEASE charac-

terized by fragile bones that results in

fractures. Half of women over seventy

will have a fracture as a result of osteoporo-

sis. For many of them, this will lead to a decline in

their quality of life and independence. These frac-

tures, which often occur with little or no injury,

O

BEAM LINE 27

are usually a result of reduced bonemass. Accordingly, research on os-teoporosis has focused on those fac-tors that affect bone mass, such asestrogen deprivation, age, and phys-ical activity. Treatments to controlosteoporosis have been largely basedon trying to maintain or to increaseskeletal mass; however, significantoverlap exists between the bone massof normal individuals and those withosteoporosis. This overlap has re-sulted in investigation into factorsother than low bone mass, such asstructural properties and bone qual-ity, to explain the higher frequencyof fractures among individuals withosteoporosis.

Bone is the hard, marrow-filledcalcified tissue that forms the skele-ton. Near the joints a porous spongy-like bone fills much of the marrowspace as seen in the illustration onthe right. This spongy bone aids intransferring the stresses of motion tothe hard outer shell called the cor-tex. The spongy bone (frequentlyreferred to as trabecular bone) is com-posed of an interconnected structureof curved plates and struts. One ofthe age-related changes that accom-panies bone loss is a decrease in theamount of spongy bone. In normalpeople the volume occupied by thisspongy bone decreases with age. Thisdecrease in bone volume is appar-ently due to a loss in the number ofstruts and a severing of the inter-connections, as opposed to a generaluniform thinning of the elements.Many questions regarding the de-velopment and treatment of osteo-porosis could be answered if the ar-chitecture of the spongy bone—andhow it changes with disease andage—were better understood.

BONE STRUCTURE CHARACTERISTICS

About a hundred years ago, an Eng-lish reverend named Abbott pub-lished a small book titled Flatland. Itwas a story of a two-dimensionalworld occupied by polygons. The pro-tagonist was a square who, one day,was visited by a sphere. The spherecarried the square above Flatland, andallowed him to view his three-dimensional counterpart, the cube.Upon returning to Flatland, thesquare set about to enlighten his two-dimensional world about the exis-tence of a three-dimensional uni-verse. Unfortunately, the square wasnever able to provide convincingproof of the third dimension to hisfellow inhabitants of Flatland andwas eventually labeled a heretic andspent his remaining days in prison.

Though this book was intended asa social satire, it points out the dif-ficulties describing three-dimen-sional structures using only two-dimensional views. The science ofreconstructing three-dimensional in-formation from two-dimensionalviews is called stereology. With stere-ological methods it is possible to es-timate volume fractions and surfaceareas of particles or voids from just afew flat microscopic sections cutfrom an object.

Stereological methods are cur-rently used to estimate three-dimensional properties of bone fromtwo-dimensional or planar sections.These methods allow calculation ofthe volume fraction and surface ar-eas of the spongy bone. It is also pos-sible to quantify other structural fea-tures of the bone required to explainbone formation, resorption, and

An X-ray tomographic microscopeimage of a thin section of leg bone in arat. The thick bone (A) surrounding themarrow cavity is the cortical bone, orcortex. The thin, web-like bone (B) withinthe marrow cavity is the spongy bone.The earliest stages of osteoporosisinvolve the resorption of spongy boneand a breaking of its interconnections.

28 FALL 1995

structure. The application of stere-ological principles has provided im-portant clues as to how a bone’sstructure affects its properties and isaltered by disease and aging. Unfor-tunately, as we understand better thecomplex structures and properties ofbiological systems, we begin to ex-ceed the ability of stereological meth-ods to describe the critical aspects ofthe structure. In particular, conven-tional stereological methods begin tocollapse when questions arise re-garding how the spongy bone net-works are connected. For example,one of the easiest questions to poseabout any structure is to ask howmany objects are contained withinit. Indeed, it is impossible to answerthe basic question “how many?”from a limited number of two-dimensional views or sections. Acomplete three-dimensional visual-ization is necessary.

Questions such as “how many”relate to the topological structureof an object. For any three-dimen-sional structure, the topology canbe described with three variables:i) the number of separate particles(“how many”); ii) the interconnect-edness of the particles (number ofhandles or pathways connecting dif-ferent parts of the object); and iii) thenumber of enclosed surfaces (for ex-ample, bubbles, voids, or entrappedparticles). These variables are ex-tremely valuable for describing theprocess of a disease like osteoporosisand the mechanical behavior of cal-cified tissues.

In the past, quantifying these topo-logical variables required time-consuming, artifact-prone serial sec-tioning of entire specimens. In serialsectioning, a specimen is imbedded

in a rigid holder, and a shallow cut ismade that just exposes the interior.The surface is polished, stained, andphotomicrographed. Then anothercut is made parallel to the originalone. This fresh surface is then pol-ished, stained, and photomicro-graphed. This procedure is repeateduntil the entire specimen has beenexamined. The micrographs fromeach section are then digitized, anda volumetric image is created. In or-der to quantify the interconnected-ness of the spongy bone in a smallanimal, it would be necessary toobtain 50 to 100 sections from eachone. This explains why this methodis rarely used.

Another disadvantage with sec-tioning is that it can only be per-formed on dead animals. This re-quires a large number of animals tostudy at each time point in an ex-periment. For example, in an exper-iment that tests the effectiveness ofa bone-growth drug at a single dose,it is necessary to have control ani-mals, estrogen-deficient animals (in-duced by removing the animal’sovaries), and treated animals. If sixanimals are required at each timepoint for statistical accuracy, and ifwe examine bone loss in the ovariec-tomized animals at three time points,and then examine treated, controlled,and ovariectomized animals at threeadditional time points after treat-ment, a minimum of 78 animalswould be required for this simplestudy. If, on the other hand, it werepossible to examine living animals,then time-point sacrifice would nolonger be necessary. In this case, only18 animals would be required, andthe results would have greater sta-tistical significance because repeated

measures would be performed on thesame animal. The savings inthe number of animals becomes evengreater as the sophistication of theexperiment increases. Hence, thereis a strong motivation for developingimaging methods that providethe same information as serialsectioning—but on living animals.

An alternative method for recon-structing three-dimensional imagesis X-ray computed tomography or CT(see “Positron Emission Tomogra-phy” in the Summer 1993 BeamLine, Vol. 23, No. 2). Because littleor no sample preparation is requiredfor CT, tomographic methods providea cost-effective alternative to serialsectioning, and can, in principle, beused on living animals. CT is mostfrequently used as a medical diag-nostic (higher resolution CT scan-ners have been developed, but theyhave not been approved for use onhumans). Unfortunately, the presentresolution is not high enough toimage spongy bone with the accu-racy required for quantitative mea-surements. To be useful as a substi-tute for conventional methods, theresolution of the CT method must beimproved a hundred-fold over thenewest instruments. This requires anew type of X-ray source and im-proved detectors.

COMPUTED TOMOGRAPHY WITH SYNCHROTRON RADIATION

The attenuation of X rays througha sample is a sensitive measure ofatomic composition and density. Bymeasuring the X-ray attenuation co-efficient as a function of positionwithin a sample with CT, a three-dimensional image of the sample can

BEAM LINE 29

EXPERIMENTAL STUDY OF POSTMENOPAUSAL BONE LOSS

In a recent study, the leg bones(tibias) of living female rats were im-aged with the XTM at SSRL. Imag-ing times with the nearly mono-chromatic X-ray wiggler beam at25 keV were less than 30 minutes peranimal. Shuttering of the direct beamreduced actual exposure times to lessthan two-and-a-half minutes.

The rats were anesthetized, andwhile unconscious they were securedto a rotating platform with their righthind limbs elevated into the X-raybeam. On the day following the ini-tial scans, rats were chosen at ran-dom and their ovaries were removed.Five weeks after removal of theovaries, all animals were imaged afinal time with the same imagingparameters.

Once the data were acquired, thethree-dimensional images of thetibias were reconstructed on a com-puter workstation. The volumetricdata were analyzed by the follow-ing method. Cluster analysis was per-formed on the spongy bone struc-tures in the three-dimensionalimages; it identified all of the spongybone that was continuously inter-connected, and also identified anyisolated structures that were dis-connected from the surrounding cor-tical bone and spongy structure.Cluster analysis provided a directmeasure of the topological variablesthat quantify the number of isolatedbone fragments and the number ofimbedded pores. For the intercon-nected cluster, the connectivity wascalculated from the three-dimen-sional image.

be obtained. Subtle compositionaland structural changes from one po-sition to the next appear as differ-ences in the X-ray attenuation. Aslong as the spatial resolution is smallwith respect to the features of in-terest, the three-dimensional X-rayimages obtained from these mea-surements will provide valuablestructural information.

High spatial resolution dependsupon i) a sufficient X-ray intensity,so that enough X rays reach the de-tector to provide good measurementstatistics; ii) a detector with suffi-cient spatial resolution to discrimi-nate between closely separated X-raypaths; and iii) monochromatic (sin-gle energy) radiation. Synchrotron ra-diation sources, because of their highbrightness and natural collimation,have allowed the development of CTsystems that have spatial resolutionsapproaching 1 µm. For this study wemodified the X-ray tomographicmicroscope (XTM) developed atLawrence Livermore National Lab-oratory in Livermore, California, foruse on beam line 10–2 at the Stan-ford Synchrotron Radiation Labora-tory (SSRL) at Stanford Linear Accel-erator Center in order to image thethree-dimensional structure andmineral density of bone in living an-imals. We wanted to demonstratethat the XTM with synchrotron ra-diation can detect microscopicchanges in the spongy bone structureand connectivity in estrogen-deficient rats, an important animalmodel for osteoporosis. The im-proved resolution available with syn-chrotron radiation over the highestresolution CT scanners is shown inthe two figures on the right.

Conventional X-ray tomographyinstruments do not have the resolvingpower required to detect structuralfeatures within the spongy bone andcortex. Above is the highest resolutionimage of a rat leg bone with specializedcommercial instrumentation. Below isthe same bone imaged with the X-raytomographic microscope andsynchrotron radiation. Variations in thecolor of the bone correspond to smalldifferences in the calcium concentration.

30 FALL 1995

Three-dimensional images of thetibias just prior to removal of ovariesand five weeks after surgery showedthat a 60 percent loss in bone volumeoccurred in the five weeks followingestrogen loss (see “Biological Appli-cations of Synchrotron Radiation” inthe Fall/Winter 1994 Beam Line,Vol. 24, No. 3, page 21). In addition,the three-dimensional images dem-onstrated a significant change froman interconnected plate- and strut-like structure to one that is mostlydisconnected struts. Also, dangling(or dead-end) elements are seen onlyin the estrogen-deficient animals.These dangling elements, althoughstill contributing to the total bonemass, probably do not contribute tothe strength of the bone.

A small region of spongy bone athigher magnification in a rat withovaries removed is shown in the topfigure on the left. Of particular in-terest is the small bone fragment thatis isolated from the surrounding boneand supported only by marrow. Wehave only observed bone fragmentssuch as this in animals with osteo-porosis, where they account for about1.5 percent of the total spongy bonevolume. These isolated bone frag-ments, as well as the more signifi-cant fraction of dangling bone, maybe responsible for the overlap in bonemass between individuals with os-teoporotic fractures and individu-als without fractures.

CAN WE REGROW LOST BONE?

Intermittent parathyroid hormonetherapy has been shown to increasebone mass and improve biomechan-ical strength in osteoporotic rats.

Also, intermittent parathyroidtherapy both alone and in combina-tion with estrogen therapy has beenreported to preserve spongy boneconnectivity. Because the reports ofconnectivity have been based ontwo-dimensional data, we have beenusing synchrotron radiation with X-ray tomographic microscopy to de-termine if intermittent parathyroidtherapy actually does increase bonemass and connectivity. What wehave found is that spongy bone vol-ume and connectivity significantlydecreased after 8 and 12 weeks of es-trogen loss. Parathyroid treatmentappears to increase bone mass bythickening existing bone, not byforming new connections (see mid-dle and bottom figures on the left).From our results, we hypothesizethat to re-establish connectivity inthe spongy bone, treatment wouldhave to be given before significantbone loss has developed, or treatmentwith a bone-forming agent such asparathyroid would be less effective.Further studies need to evaluate crit-ical time points for the administra-tion of bone-forming compounds andfor assessing the effects of these treat-ments on improving bone strength.

New insights gained from thesestudies using synchrotron radiationare allowing us to develop more rapidscreening of new clinical treatmentsfor osteoporosis, a major publichealth problem responsible for overone million fractures a year in theUnited States alone.

Bottom: The spongy bone of an osteo-porotic rat given parathyroid hormoneafter osteoporosis has been established.New bone is formed, and the spongybone is significantly thicker than theoriginal bone; however, the original plateand strut-like structure of the bone is notreestablished.

Top: An X-ray tomographic microscope image of a small region of spongy bone in the leg bone of an osteoporotic rat. The normal plate and strut-like structure of the bone has been greatly eroded, and several isolated bone fragments can beseen suspended in the marrow. Though contributing to the total bone mass, theseisolated fragments do not contribute to the bones’ strength and resistance to fracture.

Middle: The spongy bone in a rat leg at the earliest stages of osteporosis showingthe beginning of perforations to the plate structures.

BEAM LINE 31

HE INTERACTION OF LIGHT with matter is the most funda-

mental process of our physical world. It heats our planet, it supports

life, and it makes objects visible to us. The invention of the laser thirty-five

years ago made possible beams of light of much higher intensity which possess

remarkable directionality and coherence compared to those of ordinary

sources. Lasers of all shapes and sizes have not only revolutionized technology

but have also opened new research possibilities in many fields of science. One

such new avenue is the interaction of photons with high energy electrons,

which lets us study the fundamental electron-photon interaction in the ab-

sence of other matter and in the range of extremely high photon densities.

LASER-ELECTRONINTERACTIONS

by ADRIAN MELISSINOS

T

High power lasers combined

with SLAC’s unique high energy

electron beam are opening up

a new field of research.

32 FALL 1995

High Energy Electron

Laser Photon

Lower Energy Electron

High Energy γ ray

Before

After

That collisions will occur when abeam of electrons crosses through acounter-propagating beam of photonsmay seem obvious to some, but itcould also be relegated to the domainof science fiction by others. The out-come of such an experiment dependson the density of the beams and onthe duration of the crossing, as wellas on the fundamental probability forthe interaction to occur. If we com-pare the energy of an electron in theStanford Linear Accelerator Center(SLAC) beam with that of a laser pho-ton, a great disparity is evident. Theelectron energy is 50 billion electronvolts, whereas the photon energy istypically 1 or 2 electron volts; thecollision can thus be described by theproverbial analogy of a Mack truckcolliding with a ping-pong ball. It isnevertheless true that under appro-priate conditions the electron cantransfer almost all its energy to thephoton, as shown in the illustrationabove left. This process is referred toas Compton (or photon) backscat-tering.

BACKSCATTERING of a laserbeam was first exploited atSLAC in 1969 to create high

energy photons or gamma rays. Thegamma rays were directed into the82-inch bubble chamber, where theirinteractions in liquid hydrogen (onprotons) were recorded and studied.An important aspect of high energyelectron-photon collisions is thatthey depend on the state of polar-ization of the colliding beams. Elec-trons have spin that can point in oneof two directions: along or against thedirection of motion; the same is truefor the photons. By measuring therate of scattered events one can

Right: Backscattering of a photon from ahigh energy electron.Below: The 82-in. bubble chamber atSLAC in 1967. Top to bottom areWolfgang Panofsky, Joseph Ballam,Robert Watt, and Louis Alvarez. Thischamber was a modified version of theoriginal 72-in. chamber built andoperated by the Alvarez group at theLawrence Radiation Laboratory.

BEAM LINE 33

determine the state of polarizationof the electron beam, a techniquethat was pioneered at SLAC on theelectron-positron ring SPEAR in thelate 1970s.

An extension of the polarization-measurement technique, first usedat Novosibirsk in Russia, allows theprecise determination of the energyof the electrons in a storage ring.Three years ago such measurementsrevealed changes in the energy of theLEP electron-positron collider atCERN in Geneva, Switzerland, ow-ing to the gravitational pull of themoon. The tidal force of the moondistorts the lattice of the 27 kilo-meter ring of magnets, resulting in achange of the beam energy at fixedradio-frequency and magnetic fieldvalues. What was first considered apuzzle was eventually resolved bythe late Gerhard Fisher, a long-timeSLAC accelerator physicist.

The energy spread of the photonsin a laser beam is usually very small,and certain lasers can be tuned to

deliver photons of a specific energy.This property in combination withthe polarization dependence of thelaser-electron interaction has madepossible the highly successful polar-ized source that is now in use at theSLAC linac (see “The Heart of a NewMachine” in the Summer 1994 issueof the Beam Line, Vol. 24, No. 2). Inthis case it is the laser photons thattransfer energy to the electrons sothat they can emerge from the semi-conductor cathode and be acceler-ated in the electron gun and then in-jected into the linac. In the strainedgallium arsenide crystal cathode,atomic levels corresponding to dif-ferent polarization states have dif-ferent energies and can be selective-ly excited by the laser. This hasyielded polarizations as high as 85percent. The same principle is usedto achieve polarization in some of thetargets for the electron-scattering ex-periments being carried out in EndStation A at SLAC. Here a titaniumsapphire laser operating at a wave-

Polarized Gun

e–

Left or Right Circularly Polarized Light

Unpolarized Gun

Mirror Box

e–

Ti:Sapphire LasersBunch 1

Bunch 2Combiner

Circular Polarizer

Load Lock

Bunchers

Accelerator

Linearly Polarized Light

The polarized electron source for theSLAC linac.

length of 790 nanometers excites ru-bidium atoms, which in turn collidewith atoms of the helium isotope3He. This causes a significant po-larization of the 3He nuclei, andtherefore of the two protons and neu-tron that form the nucleus.

Incidentally, Compton backscat-tering is not simply restricted to thelaboratory but also plays an impor-tant role within our galaxy, and pre-sumably in all of intergalactic spaceas well. Very high energy cosmic-rayprotons scatter from the photons thatform the microwave background ra-diation that permeates all of space.Although the density of the radiationis about 400 photons/cm3, many or-ders of magnitude lower than thatwithin a laser beam, the distancesover which the interaction can occurare correspondingly enormous com-pared to terrestrial standards, so col-lisions do occur. In fact, it is believedthat the highest energy of the cosmicrays is limited by their collision withthe microwave background photons.

34 FALL 1995

The interaction of electrons withsuch strong electric fields has hith-erto been inaccessible to laboratoryexperiments (similar phenomena arethought to occur in very dense neu-tron stars). Just as atoms are tornapart in fields exceeding 109 V/cm,one can speculate that the electronsthemselves may not withstand thecritical field. What would be themanifestation of such a disruptiveinteraction? Presumably electron-positron pairs would be created outof the vacuum, the energy being sup-plied by the electric field. Experi-ment 144 has indeed observedpositrons produced under such cir-cumstances. Furthermore, the or-dinary electron-photon scattering in-teraction is modified in the presenceof strong electric fields by allowingfor the simultaneous absorption ofseveral laser photons. We will returnto these experimental results.

The laser system for E-144 isbased on the experience gained at theLaboratory for Laser Energetics of theUniversity of Rochester; it usesneodymium embedded in a yttrium-lithium-fluoride crystal matrix(Nd:YLF laser) followed by neo-dymium in a glass matrix (Nd: glasslaser). It has the advantage of beinghighly compact as compared to lasersof similar power and is nicknamedT3, for Table-Top-Terawatt laser.

Neodymium lases in the infrared ata wavelength of about 1060 nano-meters; however, it is possible todouble the frequency of the lightwith good efficiency to obtain beamsin the green wavelength of about 530nanometers. Typically the laser de-livers 1–2 joules of energy in a singlepulse, which is only 1.2 picoseconds(trillionths of a second) long. Thusthe peak power is of the order of 1012

watts, to be contrasted with a com-mon helium-neon laser that operatesat a level of about 10−3 watts. Final-ly the laser pulse must be focused astightly as possible. This implies theuse of large beams and short focallength (to achieve a small f-number)and the absence of aberrations; near-ly diffraction-limited spots of20 mm2 area have been achieved forgreen light using f = 6 optics.

To reach joule energies in a singlepulse, a chain of several laser am-plifiers must be used. Initially amode-locked oscillator produces apulse train of short pulses, typical-ly 50 picoseconds long. This isachieved by placing an acousto-opticswitch that is driven by an externalfrequency with period equal to twicethe time for one round trip of thepulse between the laser mirrors in-side the laser cavity. As a result, thecavity can lase only for a short timewhen the switch is open and in phasewith the external radio-frequency.This frequency has been chosen tobe a sub multiple of the acceleratordrive frequency, and in this way thepulses in the laser train are syn-chronized with the electron pulsesin the linac. A single pulse is selectedout of the pulse train by a Pockelscell, but the energy in the pulse is

AN EXPERIMENT currently inprogress at SLAC (E-144) ex-plores for the first time a dif-

ferent dimension of the electron-photon interaction—what happensif the photon density is extremelyhigh? The experiment became pos-sible because of recent advances inlaser technology that can producedensities as high as 1028 photons/cm3

at the focus of a short laser pulse. Tosuggest the scale of this number, werecall that the density of electrons inordinary matter is in the range of1024/cm3. High photon density cor-responds to high electric field at thelaser focus, approaching E = 1011

volts/centimeter (V/cm). At suchfield strengths, atoms become com-pletely ionized. More importantly,however, when an electron from theSLAC beam passes through the laserfocus, it feels in its own rest framean electric field multiplied by the ra-tio γ = ε/mc2. This effect is a conse-quence of the special theory of rela-tivity, which is applicable when theparticle velocity approaches thespeed of light. For the SLAC beam,the energy ε is about 50 GeV; sincemc2 = 0.5 MeV, the factor γ is of theorder of 105. Thus the electric fieldseen by the electron is about E = 1016

V/cm; this field strength is referredto as the critical field of quantumelectrodynamics.

The T3 laser facility used in SLACExperiment 144.

BEAM LINE 35

of electron (positron) momenta. Thehigh energy gamma rays continuein the forward direction and are de-tected in a separate calorimeter 40meters downstream from the inter-action point.

One of the concerns during theplanning of the experiment waswhether the synchronization of thelaser pulses with the electron beamcould be established and maintained.The length of the laser pulse is300–600 micrometers, and the elec-tron pulse is only slightly longer; fur-thermore, the two beams cross at anangle. Thus the timing and the pathlengths traversed by these two pulsesoriginating from very differentsources must be kept to a toleranceof a fraction of a millimeter. Thiswas successfully accomplished witha phase-locked loop between thelaser mode locker and the linac ra-dio-frequency, as well as by carefulconstruction of the transport line. Acomputer-controlled delay insertedin the laser line was used to set andmaintain optimal timing.

The laser-electron interaction point andthe analyzing spectrometer forExperiment 144 inside the Final FocusTest Beam cave.

only a few nanojoules. The pulse isthen stretched in time and chirpedin frequency before being injectedinto a regenerative amplifier. After100 round trips in the regenerativeamplifier, the pulse energy is at themillijoule level, at which point thepulse is ejected from the regenera-tive amplifier and further amplifiedin two additional stages. The laststage is of special construction usinga slab of lasing material instead ofthe usual rods to permit a repetitionrate of 1 hertz. Finally, the pulse iscompressed in time in a pair of dif-fraction gratings to achieve the de-sired short length.

The experiment is installed in theFinal Focus Test Beam (FFTB) line(See “The Final Focus Test Beam,”in the Spring 1995 issue of the BeamLine, Vol. 25, No. 1), just down-stream of the primary focus. This lo-cation is particularly advantageousbecause the tight focal area and shortduration of the laser pulse must bematched by an electron pulse of cor-responding quality. The laser beamis transported by high performancemirrors to the interaction point,where it crosses the electron-beamline at a 17 degree angle. Immedi-ately after the interaction point, astring of six permanent magnets de-flects the electron beam into a beamdump; these magnets also serve todisperse the electrons that have in-teracted and the positrons that havebeen produced in the interaction sothat their momenta can be measured.Electrons and positrons are detectedin silicon calorimeters placed belowand above the original beam line.These detectors measure the totalenergy deposited in a narrow range

36 FALL 1995

absorptions by a factor of eight, andin general the rate of n-photon ab-sorptions increases as the nth pow-er of the laser energy. That this istrue can be seen in the bottom graphon this page where the observed ratefor two-, three-, and four-photon ab-sorption is plotted as a function oflaser energy. The nonlinear behaviorfor these processes is clearly evident.

The rate for the production ofelectron-positron pairs is relativelylow at present laser intensities; ap-proximately one positron is detectedin 1000 laser shots. The physicalmechanism in this case is the fol-lowing: an electron enters the laserfocus and produces a high energygamma ray; the gamma ray interactssimultaneously with several (at leastfour) laser photons to produce thepair. This result is the first demon-stration of light-by-light scatteringinvolving physical photons. Therewas little doubt that this process,which has been amply verified forvirtual photons, would occur; how-ever, the stage has now been set forfurther precision experiments thatcan probe quantum electrodynamicsin the critical field regime.

IT SHOULD COME as no surprisethat an understanding of the elec-tromagnetic interaction in high

electric fields is essential for the de-sign of the next generation of lin-ear colliders. To achieve the requiredluminosity, the beams at the colli-sion point of these new machinesmust have dimensions of only a fewnanometers. This results in an ex-tremely high electric field inside theelectron (or positron) bunch; the in-coming positron (electron) that

NOT EVERY BEAM electroncrosses through the laser fo-cus, but at high laser den-

sity every electron that does crossthrough the laser focus interacts,giving rise to a backscattered gammaray with energy between 0 and21 GeV (for infrared light and for46 GeV incident electrons). Some5×107 gammas were produced byeach electron pulse, correspondingto conversion of approximately 1 per-cent of the incident electron flux (asexpected from the relative size of thetwo beams). Interactions that involvethe absorption of two or more pho-tons are much less common but stillcopious. In this case it is the recoilelectrons that are being detected; ifa single (infrared) photon is absorbed,the recoil electron energy is alwaysgreater than 25 GeV. Thus electronswith energy below 25 GeV are a mea-sure of interactions where two ormore photons were absorbed (thereis also a finite probability that theelectron scattered twice at differentpositions within the laser focus). Themeasured spectrum of recoil elec-trons for two different laser energiesis shown in the top graph on the left.Recoil electrons in the range between18 and 25 GeV result from the ab-sorption of two photons, between 13and 18 GeV from the absorption ofthree photons, and so on.

The absorption of a single photonis a linear process; this implies thatif the laser intensity is doubled, thescattering rate also doubles. In con-trast, the absorption of several pho-tons is a nonlinear process. When thelaser intensity is doubled the rateof 2-photon absorptions increases bya factor of four, the rate of 3-photon

10–4

10–5

10–6

10–7

101 102

Laser Energy (mJ)103

1/N

γ *dn

/dE

(1

/GeV

)

E = 20.75 GeV17.75

14.75

12.75

n = 2

n = 3

n = 4

10–4

10–5

10–6

10–7

8 12 16

Recoil Electron Energy (GeV)

20 24

1/N

γ* dn

/dE

(1

/GeV

)

400 mJ η = 0.34

27 mJ η = 0.11

Top: The normalized cross section formultiphoton Compton scattering plottedas a function of momentum for laserenergies of 27 and 400 millijoules. Thelaser wavelength was λ = 1054 nano-meters and the electron energy ε = 46 GeV.

Bottom: The normalized cross sectionsfor 2, 3,and 4 photon absorption as afunction of laser energy (λ = 1054 nano-meters and ε = 46 GeV). For linearprocesses the cross section is inde-pendent of laser energy. The opencircles show the predictions of thetheory.

BEAM LINE 37

enters the oppositely moving bunchis affected by the field and can lose asignificant amount of energy in theform of radiation and/or pair pro-duction. This effect is already ob-servable in the collisions at the Stan-ford Linear Collider and has beengiven the name “beamstrahlung.” Itwill be much more important for fu-ture linear colliders, and designersare making every effort to minimizeits impact.

Finally, it is desired to include infuture electron-positron colliders theoption of operating them as gamma-gamma colliders. This has certain ad-vantages in identifying the particlesproduced in the collision, especiallyscalar particles such as the hypoth-esized Higgs boson. The high energyelectron and positron beams can beconverted to gamma-ray beams bybackscattering laser photons from

Experiment 144 group photograph.Pictured left to right are (kneeling) Glenn Horton-Smith, SLAC; TheofilosKotseroglou, Wolfram Ragg, and Steve Boege, University of Rochester;(standing) Kostya Shmakov, Universityof Tennessee; David Meyerhofer andCharles Bamber, University ofRochester; Bill Bugg, University ofTennessee; Uli Haug, University ofRochester; Achim Weidemann,University of Tenessee; Dieter Walz,David Burke, and Jim Spencer, SLAC;Christian Bula and Kirk McDonald,Princeton; Adrian Melissinos, Universityof Rochester. Not pictured are CliveField and Allen Odian, SLAC; SteveBerridge, University of Tennessee; Eric Prebys, Princeton; and ThomasKoffas, University of Rochester.

each of the beams. In this respect E-144 has been the pilot experiment forfuture gamma-gamma colliders bothin terms of the hardware involvedand in measuring the fundamentalprocesses that govern such collisions.There are still many unansweredquestions and much work to be donebefore gamma-gamma colliders canbe designed with certainty.

Investigation and exploitation ofthe electron-photon interaction hashad a long tradition at SLAC. High-er energies, new technology, and newideas have confirmed theoretical pre-dictions that, in some cases weremade decades ago. The recent ex-periments present an opportunity forprobing this fundamental interactionin a new regime. No doubt the an-swers will continue to be important.

Ste

ve B

oeg

e

38 FALL 1995

CONTRIBUTORS

PAUL GRANNIS completed his bach-elor’s degree at Cornell Universityand received a doctorate from theUniversity of California at Berke-ley in 1965. Since 1966 he has beenon the faculty of the State Univer-sity of New York at Stony Brook,with visiting appointments at theRutherford Laboratory, CERN, Uni-versity College London, and FermiNational Accelerator Laboratory.

From its inception in 1983 until1993, he was the spokesman for theDØ experiment at the Fermilab an-tiproton collider; since 1993, he hasserved as co-spokesman for DØ withHugh Montgomery.

Grannis is a Fellow of the Amer-ican Physical Society and vice-chair-man of the Division of Particles andFields of the American PhysicalSociety and a Sloan Fellow from 1969to 1971.

JOHN LEARNED has been at the Uni-versity of Hawaii where he is Pro-fessor of Physics since 1980. He wentthere to get the DUMAND projectstarted and thought he would leavein three or four years. Fifteen yearslater, a little wiser, and much moreexperienced at sea, he and his col-leagues are still pushing to do highenergy neutrino astronomy.

He has also been involved in theIMB nucleon decay experiment, ofwhich he was a co-founder, and hadhis greatest scientific thrill leadingthe discovery of the neutrino burstfrom supernova 1987A. He also hada foray into high energy gamma-rayastronomy in the mid-1980s, build-ing and operating a Cherenkov airshower telescope on Haleakala.

For obvious reasons he prefers tospend his holidays in the mountainsrather than at sea.

BILL CARITHERS, senior staff physi-cist at Lawrence Berkeley NationalLaboratory (LBNL) and visiting sci-entist at Fermi National AcceleratorLaboratory, currently serves, withGiorgio Bellettini, as spokespersonfor CDF, the Collider Detector atFermilab collaboration. He has heldthis position since January 1993.

Carithers did his undergraduatework at the Massachusetts Instituteof Technology and earned his PhD inphysics from Yale University in 1968.Since then he has held positions atColumbia University and RochesterUniversity before making the movewest to LBNL in 1975.

He was a Sloan Fellow from 1972to 1974.

BEAM LINE 39

JOHN H. KINNEY is a research scien-tist in the Chemistry and MaterialsScience Department at LawrenceLivermore National Laboratory. Hereceived his BA degree in mathe-matics and physics from WhitmanCollege and his PhD in materials sci-ence from the University of Califor-nia at Davis in 1983.

In collaboration with researchersfrom the University of California,San Francisco, he has been applyingsynchrotron X-ray methods to thestudy of diseases in calcified tissues.Most recently, this work has beenextended to the study of calcifica-tions in coronary artery disease withresearchers from Stanford University.

He has published over sixty arti-cles in peer-reviewed journals.

ADRIAN MELISSINOS has been on thefaculty of the University of Rochestersince 1958, the year he earned hisPhD from the Massachusetts Insti-tute of Technology. He has served asdepartment chairman and visitingscientist at CERN. His recent re-search efforts are in the applicationof high power lasers to particlephysics and particle accelerators.

Over the years Melissinos hasworked on experiments at the Cos-motron, AGS, Fermilab, SLAC, andCornell. He has been involved in de-veloping techniques for detectinggravitational interactions and insearches for cosmic axions.

He is the author of a widely usedphysics textbook. He is a Fellow ofthe American Physical Societyand a corresponding member ofthe National Academy of Athens.

MICHAEL RIORDAN has been asso-ciated with the Beam Line since hecame to SLAC in 1988 first as editornow as contributing editor. During1995 he was on leave from his posi-tion as Assistant to the Director atSLAC, writing his latest book, CrystalFire, a history of the transistor to bepublished by W. W. Norton in 1997.He is the author of The Hunting ofthe Quark and co-author of TheShadows of Creation in case youcan’t tell from the above photograph.

He did his PhD research at SLACon the famous MIT-SLAC deep in-elastic electron scattering experi-ments.

DATES TO REMEMBER

40 FALL 1995

Jan 7–13 Aspen Winter Conference on Particle Physics: The Third Generation of Quarks and Leptons,Aspen, CO (Aspen Center for Physics, 600 W. Gillespie St., Aspen, CO 81611).

Jan 14–20 Aspen Winter Conference on Gravitational Waves and Their Detection: Data Analysis LIGOResearch Community Space Based Detectors—LISA, Aspen, CO (Aspen Center for Physics,600 W. Gillespie St., Aspen, CO 81611).

Jan 15–26 1996 US Particle Accelerator School (USPAS 96), San Diego, CA (US Particle AcceleratorSchool, c/o Fermilab, MS 125, PO Box 500, Batavia, IL 60510 or USPAS@FNALV.FNAL.GOV).

Mar 18–22 General Meeting of the American Physical Society, St. Louis, MO (ASP Meetings Depart-ment, One Physics Ellipse, College Park, MD 20740-3844 or MEETINGS@APS.ORG).

Apr 1–3 Phenomenology Symposium 1996, Madison, WI (Linda Dolan, LDOLAN@PHENXD.PHYSICS.WISC.EDU).

Apr 9–12 International Magnetics Conference (INTERMAG 96), Seattle, Washington (Courtesy Associ-ates, 655 15th Street, NW, Suite 300, Washington, DC 20005 or MAGNETISM@MCIMAIL.COM).

Apr 22–27 CERN Accelerator School, Course on Synchrotron Radiation and Free-Electron Lasers,Grenoble, France (Mrs. S. von Wartburg, CERN Accelerator School, AC Division, 1211Geneva 23, Switzerland or CASGREN@CERNVM.CERN.CH).

May 26–Jun 9 Aspen Center for Physics Workshop on High Energy Neutrino Astrophysics, Aspen, CO(Aspen Center for Physics, 600 W. Gillespie, Aspen CO 81611).

Jun 10–14 5th European Particle Accelerator Conference (EPAC 96), Sitges, Spain (C. Petit-Jean Genaz,CERN-AC, 1211 Geneva 23, Switzerland or CHRISTIN@CERNVM.CERN.CH).

Jun 24–Jul 12 Division of Particles and Fields/DPB Snowmass Summer Study of Future Directions in USParticle Physics, Snowmass, CO (C. M. Sazama, Fermilab, MS 122, PO Box 500, Batavia, IL60510 or SAZAMA@FNALV.GOV).

Jul 25–31 28th International Conference on High Energy Physics, Warsaw, Poland (A. K. Wroblewski,Institute of Experimental Physics, Warsaw University, ul. Hoza 69, PL-00-681, Warsaw,Poland or ICHEP96@FUW.EDU.PL).