,r ( ~~~~~~~~9r~~~~Aolr A~r /1 40, Ue-

6.9./ D~~~gy A.~:..

A PERIODICAL OF PARTICLE PHYSICS

SPRING 1991 VOL. 21, NUMBER 1

EditorsRENE DONALDSON, BILL KIRK

Contributing EditorMICHAEL RIORDAN

Editorial Advisory BoardJAMES BJORKEN, JOHN MATTHEWS, MARTIN PERL

JOHN REES, RONALD RUTH

MARVIN WEINSTEIN

Photographic ServicesTOM NAKASHIMA

BETTE REED

IllustrationsTERRY ANDERSON, KEVIN JOHNSTON

SYLVIA MACBRIDE, JIM WAHL

DistributionCRYSTAL TILGHMAN

CONTENTS

FEATURES

1LOOKING AHEAD: The Next 15 Yearsin U.S. High Energy Physics

The challenge is to construct the new SSCLaboratory on a timely schedule whilemaintaining the vitality of theU.S. high-energy physics program.

Stanley G. Wojcicki

7 EGS A TECHNOLOGY SPINOFFTO MEDICINE

The EGS code has been adopted

by medical physicists who are makingimprovements that are important

to high-energy physics.

Ralph Nelson and Alex Bielajew

DEPARTMENTS

The Beam Line is published quarterly by theStanford Linear Accelerator Center,P.O. Box 4349, Stanford, CA 94309.Telephone: (415) 926-2282.BITNET: BEAMLINE@SLACVMSLAC is operated by Stanford University undercontract with the U.S. Department of Energy.The opinions of the authors do not necessarilyreflect the policy of the Stanford LinearAccelerator Center.

Cover: A Monte-Carlo simulation of an electromagnetic showerin a liquid-hydrogen bubble chamber. Unaffected by theapplied magnetic field, one of the (yellow) photons travelingupwards within the "bremsstrahlung core" interacts to pro-duce a (red) positron and a (green) electron. A more ex-panded view showing all of the shower radiation is providedon page 9.

12 TOWARD THE NEXT LINEAR COLLIDER:ACCELERATOR STRUCTURE DEVELOPMENT

Gregory Loew, Harry Hoag, and Juwen Wang

16 LETTERS TO THE EDITORS

17 DATES TO REMEMBER

18 CONTRIBUTORS

NJG AHEAD:ext 15 Years

Li Energy Physics

N'EY G. WOJCICKI*

C : ; --00 idii000 S C - Stti:i tD i;X00 t gy

[ T0·G .t0S:;20iE000008000j0020;0i TC400000;4000000:?08040 I 0: ;T * .i$0

Fhe challenge is to construct the new SSC Laboratoryon a timely schedule while maintaining the vitality

of the U. S. high-energy physics programand planning for a proper transition

of the existing laboratories.

HE U. S. HIGH ENERGY PHYSICS program is

in the throes of a major transition. A chainof events has begun with the approval ofthe Superconducting Super Collider Labora-

tory (SSCL) that will have a great impact on our field in

this country. This important new initiative will offer

unparalleled new opportunities to extend the energyregion that can be explored and hopefully will provide

answers to some key questions in high-energy physics

(HEP) today. But at the same time, as with any major

change, it presents some important new challenges.Those challenges are the focus of this article. I cautionthe reader that the opinions expressed here representmy own views and do not necessarily reflect either theresults of High Energy Physics Advisory Panel (HEPAP)deliberations or formal policies of the u.S. fundingagencies.

*Stanley Wojcicki is presently chairman of the the U. S. Departmentof Energy's High Energy Physics Advisory Panel (HEPAP).

BEAM LINE 1

r

At the September 1990 meetingof HEPAP I gave my view that inthinking and planning for the futurewe should focus on three majorchallenges:

i) The need to construct the SSCon a timely schedule and to insure atthe same time that in spite of itslarge scale the traditional goals of HEPlaboratories, namely free and un-encumbered pursuit of basic knowl-edge by all qualified scientists, willnot be compromised.

ii) Maintaining the required vi-tality in the U.S. HEP program duringthe intervening years to insure re-search opportunities for young peopleand significant U.S. participation inaddressing the current questions inthe field.

iii) Planning for a proper transi-tion for the existing U.S. HEP labo-ratories into a future when they willno longer find themselves at the en-ergy frontier. This is especially rel-evant for Fermilab and SLAC, our twolabs that for many years have pro-vided the highest energies attainablein the proton and electron realm,respectively.

I elaborate on these challenges inthe remainder of this article.

SC IS UNDOUBTEDLY the larg-est pure basic research con-

struction project this country hasever undertaken. Its size is frequentlyused as an argument that one needsto have a major change in the con-struction procedures that will haveto be followed to assure its successfulcompletion. It might be thus usefulto compare the essential differences(in scale and other features) betweenSSCL and our previous largest

construction project, i.e., Fermilab,which was also a new "green-field"laboratory.

Initial Fermilab construction costwas $250 million, in then-year cur-rency. Subsequent additions and im-provements probably double theoverall investment. The accepted,DOE-endorsed SSCL cost is $8.2 bil-lion. Clearly these numbers are notdirectly comparable because the SSCcollider will be completed about 30years after completion of theFermilab 400 GeV accelerator. Howmuch the purchasing power of the"scientific" dollar will have de-creased during that time lapse is asubject of mild controversy. It is cer-tainly not less than a factor of 4,which is the approximate consumerprice index increase during that in-terval, and probably not more than afactor of 8. But in addition, the ac-counting methods have changedsomewhat in the intervening years.Certain costs included in the SSCLestimate (detectors, some R&D costs)were not a part of the Fermilab offi-cial cost. I would thus guess that theSSCL probably represents about afactor of 4 +1 real increase over theinitial Fermilab investment.

What about increases in scale ortechnical complexity over existingcapabilities? Fermilab's 400 GeVnominal energy represented a factorof 13 increase over BNL's AGS orCERN's PS. SSC will give us a factor of20 increase over the Tevatron'seventual energy of 1 TeV. SSC iscertainly more complex than theinitial Fermilab accelerator, but thatincrease has to be viewed in the prop-er context. During the last 20 yearsthere has been enormous progress inunderstanding beam dynamics, in

available instrumentation, and ingeneral experience with high energyaccelerators and colliders.

I would thus argue that from thepoint of view of scale, complexityand even cost (after all U.S. GNP hasalso increased significantly duringthe intervening period) the SSCL doesnot present a radical departure fromour past experience. There are, how-ever, several other factors that docontribute to the difference and mayinfluence significantly its develop-ment and the progress of u.S. HEP. Iwould like to elaborate on my viewson this issue.

ROBABLY THE MOST impor-tant factor is the significantly

greater size and complexity of theSSC detectors. The general-purposeSSC detectors will effectivelybecomelaboratories within a laboratory. Theold procedures of selecting them andtheir leaders, their design, construc-tion and associated management, andthe old pattern of detector fundingare no longer applicable. New wayshave to be found. I believe that theconstruction and management areespecially difficult issues, since SSCLwill need to play an important partin the management even though, atthis time at least, it is very muchshort of the required expertise to doso.

The required heavy dependenceon industry is another major newchange. The Fermilab magnets (bothmain ring and Tevatron) were con-structed in house; a large fraction ofSLAC klystrons were also builtwithin the laboratory. On the otherhand, the SSCL will be entirelydependent on outside suppliers for

2 SPRING 1991

its magnets, with the resultant expos-ure to potential cost overruns andschedule slippage. This new modusoperandi will also require significantadministrative overhead within thelaboratory.

The increased oversight by thegovernment agencies is another im-portant new development. Over thelast thirty years, we have seen asteady growth in the requiredamount of reporting, in the num-ber of reviews, and in the com-plexity of the approval system forlaboratory decisions. This growth,if left unchecked, has the potentialof diverting a significant fractionof laboratory resources into ef-forts related simply to satisfyingformal government procedures.

Finally, I wish to point out thatfor the first time in our history thenew facility is replacing a single-purpose HEP lab as the frontier en-ergy institution. The Fermilab 400-GeV machine took over from AGS atBNL (a multi-purpose laboratory) asthe highest energy machine; the AGSreplaced the Bevatron at LBL (anothermulti-purpose institution) about adecade earlier. I believe that this dif-ference is important for socio-politi-cal reasons because it creates an is-sue about Fermilab's future that ap-pears more difficult to solve. I shallreturn to that point later.

N THE ABOVE DISCUSSION I triedto describe the circumstances in

which we find ourselves today andwithin which we have to work to ac-complish our goal: the timely con-struction of a first-rate laboratory inthe true tradition of other U.S. HEPinstitutions. Let me now address two

; The constructiontime" for the n:ew, ;SSC3

La boratr is longM

previous HE:: proj:ec:: : :do- i:::: ::: _t: I; : ::: : : 0:: 7 t .

and very long whencompared with :almost-anyother scientific

e a eavor

additional issues of extreme impor-tance to this first challenge that alsohave a close connection to the secondone-i.e., the vitality of the U.S. HEPprogram in the next decade-namelythe issues of timeliness and man-power.

The SSCL was recommended byHEPAP in 1983, received its firstconstruction funds in 1988, and isscheduled for completion in 1999.Thus the "construction time" forthis new laboratory is somewherebetween 11 and 16 years. This is atime scale that is long compared withprevious HEP projects, and very longwhen compared with almost anyother scientific endeavor. It is alsovery long compared with "natural"personal time scales in HEP: agraduate student's research career(-3-4 years), a post-doctoral ap-pointment (-3 years), or the length ofservice of assistant professors beforebeing considered for tenure (-6 years).Thus every effort must be made toinsure that this schedule will beadhered to. There will undoubtedlybe pressures to delay the completiontime as a way to respond to future

fiscal crises, but to give in to thesepressures would be self-defeating.Not only would the total cost risesignificantly but one might alsoexperience significant personnellosses as quality people becamediscouraged by the ever-lengtheningtimescale.

This brings me to an issue that Iconsider most crucial in the wholeSSCL picture, that of manpower andrecruitment. The SSCL is goingthrough a crucial design phase rightnow, needing to finalize some of thefeatures of the booster complexwithin a matter of months. Thus it isessential that the relevant and as yetstill vacant senior positions in theaccelerator area be filled as soon aspossible. This will require participa-tion by the U.S. HEP community inthe SSCL to an extent greater thanhas been the case up to now, and itwill mean that some of the person-nel from the existing labs will haveto help out significantly, either byjoining the SSCL or by having theirhome labs assume some responsibil-ity for part of the SSC design and/orconstruction. This undoubtedly willhave an impact on at least some ofthe programs at the existing labs.

T HE NEXT CHALLENGE ISthat of maintaining a vital U.S.

program in the intervening decade.Clearly that program will be basedprincipally on the existing U.S. ac-celerator laboratories, with a smal-ler but nevertheless significant ef-fort at non-U.S. facilities or centeredaround non-accelerator experiments.I have already noted some of thereasons why the current program hasto thrive, mainly in connection with

BEAM LINE 3

__

the long SSC time scale. It is onlythrough the vitality and excitementof ongoing research that we, the u.S.HEP community, will be able to at-tract and train the next generation ofparticle physicists and keep the cur-rent practitioners in the field. If weaccept the goal of a strong U.S. HEPprogram for decades to come, wehave to insure that there is no lack ofimportant research opportunities inthe intervening time before SSCcompletion.

But there are many other reasonswhy we have to insure that ourexisting program remains strong.First of all, some of our currentfacilities are today optimally poisedto explore some of the most im-portant questions in the field. Second,the history of the field has over andover again demonstrated theimportance of maintaining sufficientdiversity in the available facilities.Different questions frequently re-quire different lines of attack, andcomplementary approaches, based ondifferent kinds of accelerators orcolliders, are often necessary. Thusit is very likely that the presentlyoperating facilities will still be rele-vant and essential in the SSC era.Finally, it is to the currently operatinglabs that one must look for the testbeams, operating experience, and realexperimental environment, all ofwhich will be essential in designing,building, and testing SSC instru-mentation.

But how do we deal with the un-avoidable conflict between currentprograms and SSCL needs in terms oftheir demands on financial and per-sonnel resources? It is the second ofthese that I am mainly concernedwith here, because it is rapidly

becoming clear that this may be themost criticalissue in assuring timelySSC success. No easy solutions exist,but I believe that we must make ourcommunity more aware of the needto make judicious compromises here.One possible general guideline wouldbe that we should not embark on anymajor new initiatives (in the experi-mental or accelerator area) if theirscientific payoffs will not come untilafter the SSC startup and if they sig-nificantly affect potential SSCLpersonnel needs.

E FINALLY COME to the thirdmajor challenge, namely

defining the proper future of the fourexisting accelerator laboratories inthe SSCL era. The role of the existinglaboratories will change in the future,just as has always happened in thepast whenever a new, higher energyfacility became operational. Ourchallenge is to optimize thistransition and to preserve the uniquecapabilities of the presently operatinglaboratories.

Even though SSC and its highenergy p-p collider will be the

cornerstone of the U.S. high energyprogram in the future, the need fordiversity of the program requiresstrong complementary activities atother laboratories. There is amplehistorical evidence documentingunique contributions of both fixedtarget and collider experiments. Evenmore dramatic is the record of thevery fruitful complementarity ofelectron and hadron machines. Overtime the pendulum of success hasswung from one type of machine toanother. The late 50s and early 60sprobably belonged to the hadronmachines; then the electronmachines came to the fore duringthe succeeding 10 years, to besomewhat overshadowed again bythe hadron colliders in the 80s.

The present emphasis on hadroncolliders is at least partly attributableto our current lack of technicalcapability to build an electron linacthat can achieve the same energyscale as the SSC. However, recentadvances in the technology ofelectron machines can provide im-portant new tools in the electronsector through the means of highluminosity e+e - colliders optimizedfor the 10-GeV energy region (B fac-tories). The physics goals of such amachine have been stronglyendorsed by the recent HEPAPSubpanel on the U.S. High EnergyPhysics Research Program for the1990s (Sciulli Subpanel) as anexcellent way to study the questionof CP violation, one of the mostfundamental problems in particlephysics today. Thus if resources couldbe found to build such a machine,the particle physics program wouldbe greatly enriched.

4 SPRING 1991

T URNING NOW to a more spe-cific discussion about the fu-

ture of our four existing acceleratorlaboratories, one can be relativelysanguine about Brookhaven andCornell. BNL's future direction israther well-defined and envisages agradual transition from its presentprincipal emphasis of an experimen-tal program based on the high-currentcapability of the AGS to a programbased on the new Relativistic HeavyIon Collider (RHIC), a machine that

stands at the interface between thetraditional fields of particle andnuclear physics.

The B physics program at Cornell,building on the present capabilitiesof CESR and the upgraded CLEO de-tector, and presumably strengthenedin the future by additional machineimprovements, will most likely pro-vide exciting frontier physics for therest of this decade. Even if a new Bfactory were to be built somewhereelse in the world during the nextseveral years, the 10 GeV e+e-regionis rich enough so that there will be

plenty of opportunities for excitingphysics left to the Cornell machine.Of course, eventual upgrade of thepresent machine to a B factory is themost obvious step towards main-taining frontier capability of thatlaboratory into the next century. Butwhether or not this will come topass, it seems likely that the ingenu-ity and inventiveness of the Cornellstaff and their relatively small scaleoperation will somehow allow themto maintain an important niche inthe U.S. HEP program.

The Fermilab and SLAC situationsappear more difficult, both becauseof their single-purpose nature andbecause of their larger scale of

operation. If the Main Injector isconstructed at Fermilab, with a com-pletion date around 1995, one canlook forward to at least 10 moreyears of frontier physics at Fermilabbased on gradual luminosity increaseof the Tevatron; specialized fixed-target work based on 1-TeV protons;and high-intensity, medium-energyfixed-target work at the Main Injec-tor. It seems that even in the SSC eraFermilab will have a unique niche,albeit somewhat reduced in scope,based on its multi-faceted fixed-targetprogram and collider experimentalcapabilities that will probably offermore flexibility and a lower pressureenvironment than will be availableat the SSCL. But there is no doubtthat in parallel a significant fractionof the Fermilab effort will have to bedirected towards SSCL experimentsalong the lines already initiated.

LAC'S SITUATION appears to besignificantly more difficult. There

is no doubt that SLAC's past contri-butions to the field of high energyphysics have been immense-notonly in the key experimental break-throughs that have shaped our cur-rent understanding of the basic con-stituents of matter and the forcesthat govern them, but also in thevery important area of acceleratortechnology. And yet today the labo-ratory finds itself in a difficult situa-tion. At least three major factors canbe identified as contributing here:the community's commitment to theSSC, which results in pressures onthe resources going to the existinglaboratories and tends to hamper anymajor new initiatives; greater thananticipated difficulty of bringing on

line the high-luminosity linear col-lider based on the retrofitted SLAClinac; and the major commitment byCERN to Z° physics and the result-ant success of LEP and its associateddetectors. Thus for the present, SLACfaces the challenge of generating suf-ficient frontier scientific activity atthe laboratory. How to achieve thisis the subject of intense discussionwithin SLAC, with the currently pre-ferred solution relying on SLC and a Bfactory. For the future the main chal-lenge is how to preserve the scien-tific vitality of the laboratory and theexpertise that is located there, andhow to ensure that there is an or-derly progression towards true linearcolliders, where SLAC can makemajor contributions and where it isbound to play a major role. Suchmachines will be essential if we areto maintain in the future the dualcapability of experiments with bothhadron and electron machines.

I would like to express next somepersonal views on one possible coursefor SLAC in the next decade. I use adecade as the time of reference be-cause I doubt that a very large linearcollider could be initiated much be-fore then. When it happens, I have nodoubt that SLAC will be a majorplayer in that enterprise, regardlessof where it is built. But it seems tome that until that time the naturalfocus for SLAC's main activities liesin three general areas: acceleratorresearch based around the final focustest beam facility, exploitation oflocal facilities and work at otherlaboratories. The last two points needa few explanatory comments. Thelocal facilities include: SLC, 50 GeVelectron beams, possible restart ofPEP, and B physics if a B factory is

BEAM LINE 5

built at SLAC. Even though the Bfactory, as discussed above, wouldaddress several key questions in HEPtoday, it would probably not have agreat impact on physics in this de-cade because such a facility wouldstart producing significant resultsonly towards the end of this century.The extent to which SLC shoulddominate the local SLAC programwill presumably be determined inconnection with the project reviewin early 1992. Regarding the outsideactivities, the proper role of SLACwould be to participate in those ef-forts where the laboratory could pro-vide unique expertise or facilitiesrather than act as a competitor touniversity groups. A major institu-tional involvement by SLAC in theSSCL experiments would undoubt-edly help both laboratories in thelong run. In addition, SLAC's exper-tise in many areas of acceleratortechnology, if applied to the SSC ac-celerator construction effort, wouldmake an important contribution tothe task of creating that new facility.Furthermore, some diversification ofSLAC into areas outside of particlephysics, but having some intellec-tual connection to it, might be agood thing for the overall long-termhealth of the laboratory.

would like to end this essay bymaking some brief comments

about the role of the u.S. HighEnergy community in the futureevolution of our program and aboutthe outlook for internationalcollaboration in the future. The HEPcommunity is viewed from the

outside as being immenselysuccessful in bringing their views tothe attention of our policy makersand having them act frequently alongthe lines we desire. The initiation ofthe SSCL is pointed to as one majorexample of this pattern. HEPAP isviewed by outsiders as a majorcontributor to this success, a successthat is mainly attributable to the factthat it allows us to speak with onevoice vis a vis the outside world. Ibelieve that our ability to unify afterspirited internal discussions is veryimportant and must be maintainedif we are to be successful in thefuture.

The other point to make here hasto do with our increased visibility onthe national scene. As our field be-comes more visible and as the fundsallocated to it are viewed as compet-ing with other discretionary expen-ditures, it is only natural that thepublic at large and its elected repre-sentatives should insist on under-standing what they are getting fortheir money. Thus it will becomeeven more important for members ofour community to take time to ex-plain to the nation at large the rea-sons why investment in basic re-search is important, and to repay thecountry for the support provided toour field. We have to do this not onlyby speaking regularly to our electedrepresentatives but also by writingto local newspapers, by being will-ing to speak to interested civic groups,and perhaps most importantly byparticipating actively in improvingscience education in this country,and thus helping to improve the sci-entific literacy of our population.

T URNING TO the internationalscene, I see that some of the

same forces and resultant tensionsthat are being experienced in ourcountry exist also on the world scene.The trend towards higher energiesinexorably brings increased costs perfacility and hence greater centraliza-tion and fewer frontier laboratories.It would thus argue for closer col-laboration and more cooperativeventures. The desire, however, toprotect the investment in the exist-ing labs is undoubtedly one of thecentrifugal forces that acts againstmore collaborative planning. Ofcourse our own system of year-to-year appropriations without firmlong-range commitments is also asignificant negative force in that di-rection. In addition, the scale of theresources required for meaningfulcollaborative efforts is such thatgovernment bodies at the highestlevel will have to be thoroughly in-volved. All of these are factors thattend to dampen any optimism that Imight have about a significantlyhigher level of international col-laboration in the future.

In conclusion, let me stress onceagain that the above views about thefuture represent my own crystal- (orcloudy-) ball gazing and should notbe interpreted as official views ofHEPAP, DOE, NSF, or any laboratory.But these issues will and should bediscussed in the future by HEPAP,and thus their serious considerationat this time by the community as awhole would be quite useful. Thefuture of our field requires that wecome up with thoughtful and opti-mized solutions.

0

6 SPRING 1991

!l:gg iii:i::.i . :. .. .

i:: 0 W;S X E CE iER-dE:ghE

..... i'~!iii i ~iSi. ii * . . . ..

.. ..... ::p S..1.

A Teinolc'. ::iii iiii" i

,f

LU 1VI.:CUIic1'1

.e GS code has been adopted by medicalhsi:Cists who are making impiovements that

are important to high-energyphysics.

O VER THE YEARS, medicine»been the beneficiary ofa:^M

mayi technologies derived

from high-energy :sic s . Electronaccelerators, for examp"l, are now "routinely employed in ext3ernal-b eam a-

radiotherapy. Particle detectors have,: ~---:, :::ii;i: spawned ideas for Positrn mision

=!= =Tomography (PET) and Comp iEad:-;- .iAxial Tomography (CAT). Theorie"-sif : :

...... y. of electron multiple scattering and :slowing-down are increasingly usedin present-day treatment-planningalgorithms to calculate the amountof energy deposited in the humanbody during the course of radio-therapy.

The EGS4 computer program, de-veloped originally for detector designand shielding analysis, is yet anothermedical application derived fromhigh-energy physics. It has been wide-ly adopted by medical physicists whonot only are sophisticated users but :ii;s!:ialso provide feedback for further de- Jvelopments of the code.f i

BEAM LINE 7

'' iTiEENTEEi:EER

GEL'dELaiE'

EG S, WHICH STANDS FORElectron-Gamma Shower, is a

general-purpose package for theMonte Carlo simulation of thecoupled transport of electrons andphotons in an arbitrary geometry; itcan be used for particles with ener-gies above a few keV up to severalTeV. Stated more simply, EGS pro-vides a way of calculating the flow ofradiation energy carried by electronsand photons as they travel randomly("walk") through matter. Althoughthe individual interactions made bythe particles are fundamental andwell understood, the transport pro-cess taken as a whole leads to amathematical problem prohibitivelydifficult to solve in a way that is ofsufficient generality to be useful. Butthe problem can be solved fairly eas-ily on a computer using statisticalgame-playing techniques (hence thename Monte Carlo), which is whatthe EGS Code System accomplishes.The "seed" for EGS was a com-puter program brought to SLAC byHans-Hellmut Nagel of Bonn Uni-versity around 1965. There were sev-eral other programs receiving wideattention during the early 1960s, oneof which was developed at the OakRidge National Laboratory for use inthe design of beam stoppers, colli-mators, and targets at SLAC.Althoughthe results from the Oak Ridge pro-gram were extremely useful duringthe construction of the Two-MileAccelerator and its beam lines, thecomputer code itself was never re-leased to the scientific community.Consequently this program, likemany others of its genre, eventuallyfaded from the scene.

The EGS Code System, on theother hand, has become a standard

tool in the design of calorimetry sys-tems in high-energy physics, as wellas a benchmarking tool in radiationdosimetry (i.e., a standard by whichother calculations are measured).This success can be attributed di-rectly to the continuous efforts andsupport provided by SLAC, its col-laborators, and a user communitythat is well over 1000 strong.

It was clear from the beginningthat Nagel's program was simply toorestrictive to be useful in solvingcurrent problems of interest to high-energy physics. The code needed tobe completely rewritten in order toachieve the generalization required.Working independently at first,Ralph Nelson at SLAC and RichardFord at the Hanson High EnergyPhysics Laboratory on the Stanfordcampus decided to combine theirprogramming efforts and producedthe first version of EGS in 1978.

T HE EGS3 CODE SYSTEM, as itwas known at that time, was

designed to simulate the flow of elec-trons and photons in arbitrary geom-etries at energies ranging from wellbelow an MeV to several thousandGeV. A vast bookkeeping effort is ac-complished by the program as it keepstrack of the variety of interactions

and steps involved in the radiation-transport process. As demonstratedin the picture on the facing page, EGSallows one to visualize the myriadparticle tracks produced by anelectromagnetic cascade shower.

This example, a color-graphicportion of which is shown on thecover, represents an EGS simulationof a photon-initiated shower devel-oped within a 40-inch bubble cham-ber similar to the one used at SLACduring the 1970s.

But it takes more than pretty pic-tures to make a computer programcredible. The documentation re-leased with EGS3 in 1978 containedimportant comparisons with bench-mark experiments, both at high andlow energies. The most importanttests, however, were performed bythe growing number of EGS usersthemselves.

Because EGS was well docu-mented, as well as versatile, user-friendly, and upward-compatible-buzz words common in the lexiconof today's software industry-a largeuser community soon developed. Thefact that anyone could get the EGSCode System free, together with sup-port for its use, was a significantfactor in its growing popularity.

In retrospect, probably the singlemost important event that made EGSan everyday word in high-energyphysics was the discovery of the J/yparticle in the fall of 1974. This dis-covery led to a dramatic increase inthe use of storage rings and a need forsophisticated calorimetry surround-ing the interaction regions of thesebeams. EGS has played an importantrole in the design of many, if notmost, of the electromagnetic showercounters built since then.

8 SPRING 1991

HOW HAS EGS been influencedby medical physics? In the lat-

ter part of 1982 SLAC's RadiationPhysics group, together with itscounterpart at KEK in Japan, starteda collaboration to extend the flex-ibility of EGS in a general way, butwith a specific class of problems inmind-those involving the design offuture high-energy accelerators. Fromthe beginning, however, there wasalso a growing awareness of the needto extend EGS downwards to muchlower energies. The program was be-coming increasingly popular as a low-energy tool in a variety of problemsoutside the field of high-energyphysics, and various people hadmentioned problems and limitationsin the existing code. It was antici-pated that these low-energy effectswould eventually show up in de-signs of future detectors for high-energy physics.

Serendipitously, a detailed low-energy benchmarking effort wasbeing conducted at this time by agroup of EGS users at the NationalResearch Council of Canada (NRCC)in Ottawa. Initiated by David W.O.Rogers, the NRCC work involvedadopting EGS for use as a theoreticaltool in ionizing radiation standardswork to serve the low-energyradiation protection community (e.g.,diagnostic x rays) as well as theradiotherapy community (energiesless than 50 MeV). There was alsostrong need in the medical physicscommunity for theoretical tools withgood predictive capability, given thatthe major mode of radio-therapytreatment was changing fromrelatively low-energy Cobalt-60gamma rays (about 1.25 MeV) toradiation produced by higher energy

electron linear accelerators (4-50MeV). This conversion had as muchimpact on the use of EGS as the T/lydiscovery in high-energy physics.Analytical tools dominant at Cobalt-60 energies were being rendered in-effective at higher energies becauseof the complications of electrontransport.

With its elegant treatment ofelectron transport, EGS was ideallysuited to these problems. Hence, acollaboration was formed betweenSLAC, KEK, and NRCC groups re-sulting in the EGS4 Code System,which was introduced in the fall of1985. Since then, the interest shownin the program by medical physicistshas been overwhelming. As the mapindicates, this interest is worldwide.

EGS simulation of a hydrogen bubblechamber. A single 1-GeV photon strikesa 3-mm lead slab from the bottom andproduces radiation. Positrons andelectrons (solid lines) bend in a 10-kGmagnetic field, whereas photons (dots)do not. A close-up color rendition of thissame event is shown on the cover.

T ODAY OVER 60 PERCENT ofthe requests for EGS come from

scientists working in medically re-lated disciplines. Based on currenttrends, roughly one in eight of uswill find ourselves undergoing radio-therapy during our lifetime, barringmiracle cures for cancer. A completegeometry for electron acceleratortreatment is shown in the followingsketch, which serves to illustratethe extraordinary complexity of theclinical situation.

BEAM LINE 9

nt dos'imetrv--that is the

Top: EGS user sites throughout theworld. Bottom: A typical geometry forradiotherapy treatment.

accurate determination of the radia-tion dose given during a treatmentprocedure--must account for scat-tering from components inside themachine as well as structures withinthe human body itself. Bones andlungs, for example, produce "inter-face effects" because of differencesin the transport of electrons set inmotion by photons in adjacent ma-terial. An experiment was per-formed to determine how accuratelyEGS could model features within thehuman body. Using a 20-MeV ac-celerator, scientists at the NRCCplaced small cylinders of aluminumand air within a large tank of waterand irradiated them with electrons,as shown on the facing page.

Measurements were made atvarious locations in the water, par-ticularly near the surface of the cyl-inder, and computer simulationswere performed using EGS. In theresults presented in the figures, thesmooth curves represent the mea-sured data and the histograms areEGS calculations.

A typical depth-dose curve in ahomogeneous "phantom" of water-i.e., containing no voids or solidmaterials--is shown in the top rightfigure on page 11. The other twocurves demonstrate how an alumi-num cylinder attenuates, and an aircylinder enhances, the dose alongthe central axis within the phantom.

The radial dose profile at variouslocations downbeam from the aircylinder was also measured and theresults are shown in the final set offigures. Clearly the dose perturba-tions caused by discontinuities arewell predicted by EGS, lending con-siderable confidence to the ability ofthe program to simulate the passageof electrons through the human body.

INCE 1985, the new low-energyfocus has led to improvements

to EGS that directly benefit the high-energy community. Calorimetry, forexample, requires measuring elec-tron-photon showers until the par-ticles are greatly degraded in energy.By means of an EGS option calledPRESTA, developed recently for low-energy dosimetry research, electrontransport can now be simulated veryaccurately in the design of calorim-eters and other detectors.

Feedback from medical physicshas also come in the form of basicinstruction. Medical physicists haveso far organized five "hands-on"teaching courses on EGS4. Completewith computer laboratories, thesefour-day courses attract not onlymedical physicists but also studentsfrom high-energy physics, the nuclearpower industry and military research.Efforts along these lines have ledrecently to the publication of a booktitled Monte Carlo Transport of

10 SPRING 1991

/

20 MeVElectrons

C)o0

L Ei 4o

o -

IC2 2

0 2 4 6 8 10 12Depth (cm)

0

0 2 4 6 8 1

Depth (cm)

Electrons and Photons (reviewed in the Spring 1990issue of the Beam Line). Also, the new reference stan-dard, The Dosimetry of Ionizing Radiation, devotesalmost two hundred -ages to EGS-related calculationsand discussion.

6

W x HAT ABOUT THE FUTURE OF EGS? Future re-leases of the code will reflect more strongly the 4

newly formed symbiotic relationship between high-energy physics and medical physics. In addition, thereare many interesting applications in cosmic-ray phys- 2ics, space science, nuclear power, radiation processing,and even such a diverse field as the paper industry.Low-energy electron beams can be used as quality as- H 0surance tools in the measurement of paper thickness!

In 10-15 years, based upon current initiatives, it is o 6likely that most external-beam radiotherapy planningwill be accomplished using EGS Monte Carlo methods.With computers rapidly becoming faster and cheaper, X 4the Monte Carlo technique, popularized by vonNeumann, Ulam, and Fermi in the 1940s, is no longerthe tool of a select few with access to high-power 2 2supercomputers. EGS has become the standard tool forthe hospital physicist; it has been run on all computerarchitectures, from PC's to Cray's. The number of Monte § 0Carlo papers in the journals Medical Physics andPhysics in Medicine and Biology increased fivefoldfrom 1983 to 1988, and this trend continues today. EGSis expected to stand at the forefront of this surge.

0

2

Top left: Schematic diagram of an experiment designed toverify the EGS code for medical physics applications. Top

rit:. - ., V- .; -,U., L W.itin.. i r- , - ,/,,, ,- , D,-L -,--.r DU-,-i, ;,- I 0yriLg: Uose Vs. oepin wI irn a waheir LpnanLUIII. DULLUIIL:. clRUll -2 -1 0 1dose profiles down beam of the air cylinder. Radial Position (cm)

0 12

2

BEAM LINE 11

Air or aluminum cylinder 6

TOWARD THE NEXT LINEAR COLLIDER

Accelerator Structure DevelopmentGREGORY LOEW, HARRY HOAG, AND JUWEN WANG

1 HE NEXT LINEAR COLLIDER (NLC)will probably consist of two seven-kilometer longtunnels pointing at each other: one will house theelectron linac, the other the positron linac. Of thesefourteen kilometers, about twelve will be occupied bythe microwave structures that accelerate the beams.What kind of accelerator structures will we need forthe NLC?

The accelerator structure is the heart of the entirelinear collider. It must be efficient, reliable, easy toalign, pump and cool. It must accelerate not just onebut a train of high current bunches in each pulse tomeet the luminosity requirements of the collider. Itmust do this while preserving the tiny size and energyspread of the beams as they are injected from specialdamping rings. Operating at high field gradients, prob-ably between 50 and 100 MV/m, it must not generateparasitic electrons (called "dark current") produced byfield emission, which can absorb energy, get acceler-ated and produce detrimental x-ray radiation and un-desirable steering effects. Finally, the structure mustnot be too expensive to fabricate and install.

Building on a long tradition of work on linac struc-tures at SLAC, much progress has been made towardthese goals. Considerable advances have taken place bythe use of specialized computer codes for structuredesign and for studies of the effects of higher-ordermodes on the bunches as they are accelerated. Experi-mental work has included studies of field emission andrf breakdown at high gradients, use of materials otherthan pure copper, design of prototype structures andfuture fabrication techniques.

12 SPRING 1991

The choice of frequency has been moved upwardsabove SLAC's frequency of 2.856 GHz in order to achievethe required energy in a machine of reasonable peak rfpower, stored energy and length. An upper bound to theoperating frequency is imposed by the minimum sizeof an accelerating structure that can be built to thenecessary cross-sectional tolerances, by the iris holediameter which, as it decreases, rapidly increases thedeflecting fields left behind a bunch, and by the klys-tron amplifiers which can be built to supply the drivingpower (see Fall/Winter 1990 issue of the Beam Line). Asa practical compromise, a frequency of 11.424 GHz (fourtimes SLAG) has been chosen.

The new structure, however, cannot be a simplescaled-down version of the SLAG disk-loaded waveguide.This is because the very high-intensity electron andpositron bunches leave behind energy in the long-rangehigher-order mode (HOM) wakefields, which can causeunacceptable energy spread and bunch-to-bunchmisalignment.

WO WAYS OF STOPPING the multibunch inter-action are being investigated. One, in line with

suggestions by Bob Palmer, is to provide channels forthe power in these modes to propagate radially outwards,thus lowering the field levels on the beam axis belowthe tolerable levels. Several methods of incorporatingthese channels have been investigated. One methodwas referred to in Ron Ruth's article, "The Next LinearCollider," which appeared in the Summer 1990 issue ofthe Beam Line. A related approach, which is currentlybeing examined because it looks simpler to build, isillustrated in the drawing above. The left-hand side ofthis figure shows an exploded view of the simple disk-loaded waveguide in which the beam-inducedwakefields are trapped and left to damage the beamproperties. The right-hand side shows the new structure,in which three rectangular waveguides divergesymmetrically and radially from each central acceleratorcavity. The size of the waveguides is carefully chosen

Shown schematically at the left is a conventional disk-loadedwaveguide accelerator structure made of stacked cups. Insuch a structure the beam and the fundamental acceleratingwave propagate together along the axis, but higher ordermodes are trapped and disrupt the beam. At the right is oneversion of a new accelerator structure in which the higherorder modes are allowed to escape through short radialwaveguides and are absorbed in microwave terminations,while the beam and the fundamental mode propagate alongthe axis as before.

so that the fundamental travelling mode whichaccelerates the beam (and carries up to hundreds ofmegawatts of peak rf power) cannot escape. It "thinks"the structure is the same as the one on the left of thefigure. However, all the HOMs have higher frequenciesand can propagate out of the radial waveguides, wherethey are absorbed in microwave terminations.Experimental tests at SLAG have shown that structureswhich rapidly damp the undesirable modes can indeedbe designed and constructed.

A second way of dealing with the HOMs is to pro-gressively alter the geometry of successive cells insuch a way that the resonant frequencies of the HOMsare changed without altering the propagation charac-teristics of the fundamental accelerating mode. Then,the HOM waves "decohere" (i.e., they become all mixedup), and cumulative interaction with the beam can beminimized. An experiment to test this idea has re-cently been carried out with the help of Jim Simpson atthe Argonne National Laboratory, using two struc-tures assembled at SLAG. In the Argonne facility, two

BEAM LINE 13

"n-,4- 1 n ; ..

electron bunches, a driving bunch and a witness bunchtrailing behind at adjustable distance (up to about30 cm), can be made to traverse a test structure. Thefacility is equipped with a spectrometer that can mea-sure the longitudinal and transverse forces created bythe driving bunch on the witness bunch. The effect isillustrated below. The measurement showed a rapiddecrease of the forces behind the leading bunch for the"detuned" structure tested and was in good agreementwith theoretical predictions.

One concern arising from the work of KathyThompson at SLAC is that the waves may "recohere"at large distances behind the leading bunch and stillcause beam disruption. Thus, the first method of HOMdamping cannot be entirely discarded, and there areadvantages to judiciously combining the two methods.For example, an adequate solution may be to build astructure in which the dipole mode frequency is detunedby 10% over the length of the structure and its qualityfactor Q is loaded down to between 20 and 100.

Another problem that must be considered in high-gradient structures is field emission causing the un-desirable "dark current" mentioned above. This is asubject that has been studied at great length at severalfrequencies in standing-wave cavities. The notion isthat electron field emission is enhanced by a

combination of metallic protrusions (mountains) anddielectric impurities partially covering these mountains(snow). RF processing can gradually remove some ofthis "snow" but, when pushed to surface fieldsexceeding 300 MV/m, can cause a mountain to eruptinto a volcanic flare, vaporizing the metal and leavingbehind it a broken-down landscape of molten copperand "lunar craters," as shown in the photograph at thetop of the next page. Interestingly enough, it appearsthat the presence of these craters does not seem to bedetrimental to the operation of the structure atsomewhat lower fields. Earl Hoyt and others at SLACare helping to analyze the surfaces before and afterhigh-gradient operation.

The drawings on the right-hand side of the next pageshow two experimental cavities that are presentlybeing used to study these effects. The top one is a

The force exerted by a leading bunch on a trailing bunch isindicated by the length of the vertical arrows. In the ArgonneNational Laboratory experiment, the force on a trailing bunchwas negligible after a few oscillations of the "wake field"shown below.

-So= Design Distance Between

Il I II~g~'e~Boat

TrailingBunch

Deflecting Force on a Traiat a Distance S Behind the

14 SPRING 1991

--

---. A

i

An electron microscope "photograph" of damage causedby surface fields in excess of 300 million volts per meter. Thecraters are a few microns across. The planned accelerationfield for an NLC is well below the value that is expected tocause this type of damage.

seven-cavity X-band structure that will be tested forfield emission, using a new high power klystron. Thebottom one is a demountable S-band cavity designed toexamine various geometries and metallic surfaces.

NE OF THE GREAT CHALLENGES ahead is to de-termine how to fabricate the future structures,

once the design is optimized. The total structure lengthis estimated to be about 12 km. The length of eachcavity has to be one-third of a wavelength at 11.424 GHz,or 0.8 75 cm. This means that we need about 1.4 millioncavities! If we want to avoid tuning every individualcavity, the critical dimensions have to be held towithin about one micron. Cavities such as those shownon page 13 must be aligned, water-cooled and sup-ported in an outer vacuum envelope. A lot of fun andhard work to make these structures affordable liesahead! 0

The first structure shown in the upper right is a 7-cavity X-bandtest structure. In the lower right is an S-band demountable testcavity.

BEAM LINE 15

LETTERS TO THE EDITORS

Dear SLAC Beam Line Editors:I very much appreciated your article by Rocky Kolb, "Particle Astrophysics

and the Origin of Structure." I found it a very coherent and good review. I waspleased to see the COBE data in the article and proud to see the COBE DMR mapin color on the cover. I was disappointed that the article did not mention thatthe map was produced by a team headed by my group at LBL. COBE is a col-laboration of about 20 scientists; however, the leadership and prototypeinstruments for mapping the Cosmic Microwave Background Radiation (CMBR)were produced by Lawrence Berkeley Laboratory. More than a decade ago werealized that the large angular scale CMB anisotropy provided information onphysics at very high energies. I think as Kolb does that the primordialfluctuations were created by quantum fluctuations during the inflationaryepoch at an energy of about 1015 to 1016 GeV and remain as perturbations in theCMBR isotropy. Measurement of the fluctuation spectrum will provide excitingnew information joining particle physics and cosmology together at birth.

George Smoot*COBE DMR Principal Investigator

From the Editors:We are very pleased to acknowledge here the seminal contributions of

George Smoot, his colleagues at the Lawrence Berkeley Laboratory, and theother members of the COBE collaboration. We were also reminded in anotherletter from Bob Birge of LBL that "Earlier similar pictures were produced from

U-2 flights some years ago (again funded by LBL) when the asymmetry was firstdiscovered." The significance of these pioneering studies was recently recog-nized in the award to George Smoot of the NASA Medal for Exceptional Sci-

entific Achievement (see below).We welcome letters to the Editors about the articles that we publish. We

also renew our invitation to prospective writers to contact us about possiblearticles in particle physics or related (even marginally related) fields.

*Editors' Note: George Smoot of the Astrophysics Group at Lawrence BerkeleyLaboratory was recently awarded the NASA Medal for Exceptional Scientific Achieve-ment in a ceremony at Goddard Space Flight Center in Greenbelt, Maryland, "inrecognition of his outstanding scientific achievements in observational cosmologythrough conception and leadership on the Cosmic Background Explorer (COBE)Mission."

16 SPRING 1991

DATES TO REMEMBER

Jun 10-14

Jun 17-Aug 9

Jul 25-Aug 1

Aug 5-16

Workshop on the Design of a Detector for a High-LuminosityAsymmetric B Factory: Summer Session at SLAC, Palo Alto, CA(Anamaria Pacheco, SLAC, Bin 95, P. 0. Box 4349, Stanford, CA94309, BITNET ANAMARIA@SLACVM).

Summer School in High Energy Physics and Cosmology, Trieste, Italy(ICTP, P. O. Box 586, 1-34100 Trieste, Italy, BITNETVARNIER@VX1 CP2.INFN.IT).

15th International Lepton Photon Symposium at High Energies,Geneva, Switzerland (by invitation)(L. Griffiths, LP-HEP91 ConferenceSecretariat, CERN, 1211 Geneva 23, Switzerland).

SLAC Summer Institute on Particle Physics, Lepton-Hadron Scattering,Palo Alto, California [Jane Hawthorne, SLAC, Bin 62, P. 0. Box 4349,Stanford, CA 94309, BITNET SSI@SLACVM or (415) 926-2877].

Aug 18-22 Particles and Fields '91, American Physical Society Division of Particlesand Fields and Division of Particle Physics, Canadian Association ofPhysicists, Vancouver, Canada [PF91 Secretariat, TRIUMF, 4004Wesbrook Mall, Vancouver, BC, Canada V6T 2A3, (604) 222-1047,BITNET PF91@TRIUMFCL].

Aug 23-Sep 2 1991 CERN School of Computing, Ystad, Sweden. (Ingrid Barnett,CERN CH Division, 1211 Geneva 23, Switzerland, BITNETBARNETT@CERNVM).

Aug 27-30 International Conference on Neutron Scattering, Oxford, England.(ICNS '91 Secretariat, Rutherford Appleton' Laboratory, Chilton,Didcot, Oxon OX 1 OQX, England)

Sep 16-27 CERN Accelerator School: Advanced Accelerator Physics,Noordwijkerhout, Netherlands (S. von Wartburg, CERN AcceleratorSchool, SL Division, 1211 Geneva 23, Switzerland BITNETCASNIK@CERNVM).

BEAM LINE 17

CONTRIB UTORS

STANLEY G. WOJCICKI is chairman of the High EnergyPhysics Advisory Panel (HEPAP) and a member of theExecutive Board of the American Physical Society. Hechaired the HEPAP Subpanel in 1983 that recommendedinitiation of the Superconducting Super Collider (SSC)project and subsequently served for four years as DeputyDirector of the SSC Central Design Group, the organi-zation responsible for initial design and cost estimatesfor the SSC. He came to Stanford University in 1966 andwas honored with the Dean's Award for DistinguishedTeaching in 1979. His field of research is elementaryparticle physics, and at the present time he is workingon an experiment to search for rare K decays atBrookhaven National Laboratory.

RALPH NELSON joined SLAC in 1964 as one of theoriginal members of the Radiation Physics Group. Hisinterests have always revolved around the science ofradiation transport, of which the EGS computer code isbut one part. His work ranges from electron-photondosimetry (he co-authored a book on the subject in1974) to the theory and measurement of muon shielding.He has been a consultant to numerous organizationsincluding the SSC Laboratory, Varian Associates, andSincrotrone Trieste. He is an Adjunct Professor withthe Nuclear Science Facility at San Jose State Univer-sity where he teaches shielding and dosimetry to bothmedical and health physicists.

18 SPRING 1991

ALEX BIELAJEW is a research scientist with the Institutefor National Measurement Standards at the NationalResearch Council of Canada. Since graduating fromStanford University in 1982 he has pioneered the use ofMonte Carlo techniques as a theoretical tool in ioniz-ing radiation dosimetry and developed theory relatedto radiation standards. Concentrating his efforts atimproving accuracy for low energy electron transport,e.g., the PRESTA algorithm, Alex devotes considerabletime to lecturing and organizing courses extolling thevirtues of EGS, supporting the burgeoning UNIX-basedEGS community, and serving as a center of softwaredistribution and expertise.

GREGORY A. LOEW is Deputy Director of SLAC'sTechnical Division and a member of the SLAC Faculty.In 1958 he joined Project M which was later to becomeSLAC. Starting in the 1960s he became Head of theAccelerator Physics Department, a job which he heldfor about 20 years before assuming his present position.His first assignment when hired was to design theconstant-gradient structure for the two-mile accelera-tor. Throughout the years, he has maintained an activeinterest in the field of linac structures, the subject of hisarticle in this issue of the Beam Line.

0

rT

BEAM LINE 19

HARRY HOAG is a physicist-turned-microwave engineer who came toSLAC in 1964 to work on phasing systems, beam position monitors, andother microwave devices used on the linear accelerator. He was acontributor to and an editor of the SLAC "Blue Book." Over the years hehas worked on many projects, including rf superconductivity, SLED, andthe development of very sensitive beam position monitors which were akey part of polarized electron scattering experiments. At present, he isinvolved in the development of new structures for the NLC and is alsoworking on a number of microwave projects connected with SLC.

JUWEN WANG, physicist in the Accelerator Theory andSpecial Projects Department, came to SLAC in 1980 asa visiting scientist. He has a long history of expertiseand proficiency in the field of linear accelerators thatdates back to the 1960s when he was trained at Tsinghua University and subsequently at Stanford Univer-sity. He has actively participated in the research ofaccelerating structures for future linear colliders. Hisactivities span theoretical calculations of rf parameters,wakefields, and beam loading as well as the mechanicaldesign, microwave measurement, and high-powertesting of various linac structures.

20 SPRING 1991