Carlo BernardiniDepartment of Physics

Rome University “La Sapienza”carlo.bernardini@roma1.infn.it

“I dedicate this paper to some friends, now regrettably deceased,who carried out the most brilliant part of this work: BrunoTouschek, the leader of the group, Giorgio Ghigo, and PierreMarin. Their versions of the events would have been moreenlightening than mine”.

(Physics in Perspective 6 - June 2004)

In 1958 the ItalianNational Institutefor Nuclear Physics(INFN) was close tocompleting theconstruction of alarge particleaccelerator, a 1100-MeV (million-electron-volt)electronsynchrotron, at theFrascati NationalLaboratories(LaboratoriNazionali diFrascati, LNF),about 25 kilometersSouth of Rome.

The electron synchrotron group with Giorgio Salvini

The idea (particularly by Gilberto Bernardini) was to build anational accelerator laboratory at a central site that would bestaffed by physicists and engineers who were dedicated to themachine. The staff was greatly assisted – in some cases trained –by outstanding physicists from Italy and abroad, especially byEnrico Persico (1900–1969), a friend of the young Enrico Fermi,who was in charge of accelerator theory; Mario Ageno(1915–1992), Fermi’s last student in Rome, who was in charge ofthe design of the injector; Italo Federico Quercia (1921–1987), whowas in charge of organizing the many services required,particularly the electronics, Matthew Sands (b. 1919) and RobertL.Walker (b. 1919) from the California Institute of Technology, andlater by Boyce McDaniel (1917–2002), Albert Silverman (b. 1919),and Robert Wilson (1914–2000) from Cornell University.

Gilberto Bernardini (left)and

Enrico Persico

Edoardo Amaldi and Gilberto Bernardini (circa 1960)

There were some theorists at Frascati, since Salvini recognized theimportance of having theorists working at the same place whereexperiments were being performed. At first, Giacomo Morpurgoworked there but did not communicate very much with theproblems with the experimentalists and distributing calculations tohis students.

The planning,building, andexploitation of theFrascati NationalLaboratories wasunder the Direction ofGiorgio Salvini (b.1920).Giorgio Salvini withGrace Kelly andPrince Rainier ofMonaco during a visitat FrascatiLaboratories

Edoardo Amaldi was president ofINFN at that time; his great skillswere reminiscent of those ofOrso Mario Corbino, Fermi’ssponsor in Rome.

Giuseppe Fidecaro, Edoardo Amaldiand Werner Heisenberg in Geneva.

O. M. Corbino(1876–1937)

Enrico Fermi

Bruno Touschekand

Edoardo Amaldi

Meanwhile, Edoardo Amaldi, a true “talent scout,” hadremembered someone who had visited Rome in 1938: BrunoTouschek (1920–1978), a brilliant Austrian theoretician who hadsurvived a difficult time under the Nazis. Amaldi offered him aposition in Rome, which Touschek accepted. At first, Brunocollaborated with Luigi Radicati and Giacomo Morpurgo onfundamental problems, particularly on time reversal and weakinteractions, then with Marcello Cini on weak interactions.

Theory was greatly influenced by the development of quantumelectrodynamics(QED), a highly successful theory (exploiting second quantizationand Feynman diagrams) that had served as a prototype for Fermi’stheory of weak interactions and worked in the lowest-orderperturbative approximation because of the smallness of therelevant coupling constant, the fine-structure constant.

Bruno Touschek did not like muchof the theoretical machinery of thattime (dispersion relations, Reggepoles, evasion and conspiration andso on), but he regarded as quiteimportant the problem of analyticityof form factors and their analyticalcontinuation to time-like values ofsquared-momentum transfer.

Some people speculated on the possible breakdown of QED andlooked for it in high-precision measurements, such as those of thegyromagnetic ratio g-2, and in electron-collision experiments. Inthe latter case, physicists generally believed that the breakdownmight occur at some very high energy (or better, momentum-squared transfer) characterized by a length (or mass) cutoffoccurring in the modified electron-photon vertex or in the electronor photon propagators.

The most naïve proposal was that of the so called “heavy-electron,” e*, which was supposed to decay into an electron andgamma ray, a decay mode that already had been shown byGiuseppe Fidecaro and his collaborators to be forbidden for themu meson.

The idea of exploring collisions inthe center-of-mass system to fullyexploit the energy of theaccelerated particles had beengiven serious consideration by theNorwegian engineer and inventorRolf Wideröe, who had constructeda 15-MeV betatron in Oslo and hadpatented the idea in 1943 afterconsidering the kinematicadvantage of keeping the center ofmass at rest to produce largermomentum transfers.

Rolf Wideröe(1902–1992)

This idea was also taken seriously by aPrinceton-Stanford group that included WilliamC. Barber, Bernard Gittelman, Gerry O’Neill, andBurton Richter, who in 1959, following asuggestion of Gerry O’Neill in 1956, proposed tobuild a couple of tangent rings to study Møllerscattering.

Andrei Mihailovich Budker (1918–1977) initiated asomewhat similar project at Novosibirsk, wherethe VEP-1 electron-electron collider was underconstruction in 1961.Donald W. Kerst (1911–1993), who hadconstructed the first successful betatron at theUniversity of Illinois in 1940, also wasconsidering colliders, particularly for protons, in1959 using the Fixed-Field Alternating-Gradient(FFAG) concept; his contributions to acceleratorphysics and technology as technical director ofthe Midwest Universities Research Association(MURA) while at the University of Wisconsin inMadison during this period were of much interestto the accelerator community.

A. M.Budker

O’Neill’s rings drawn by

Bruno Touschek

The attention of these people was apparently focused more on thekinematic advantage of colliding beams than on the new physicsto be learned from them. To achieve head-on collisions betweenaccelerated particles in flight required storing them in magneticdevices (storage rings) to allow them to collide repeatedly as theycrossed at various points in their circular orbits.

That was precisely Bruno Touschek’s starting point at Frascati. Heconsidered the kinematics as rather obvious; to him the possiblephysics to be learned from colliding particles was far moresignificant. He had a very strong picture of the microscopic worldin his mind. He conceived the vacuum as a reactive dielectricresonating at frequencies ν = mc2/h, where c is the speed of light, his Planck’s constant, and m in this case is the mass of a bosonhomologous to the photon, that is, a neutral vector meson.

Some people at this time were speculating on the existence ofsuch mesons. I was corresponding, for instance, with YoichiroNambu in Chicago, who had suggested searching for suchparticles as plausible intermediaries of the strong interactions; theconcept of “vector dominance” was close at hand.

Bruno’s view, as he put it, was that a physical system can becharacterized appropriately by investigating its “geometry” andits “dynamics.” Its geometry, its size and shape, is observable byemploying space-like photons as in electron-proton scatteringexperiments (“diffraction of electron waves”); this was preciselywhat Robert Hofstadter was doing at Stanford using the Mark IIILinac to measure form factors of nuclear particles.

No one, however, had as yet observed the dynamics; for this oneneeded to produce time-like photons of sufficiently large energy toexcite resonant modes of the vacuum corresponding to themasses of the vector mesons. Thus, Touschek concluded, weshould make electrons and positrons collide and annihilate in thecenter-of-mass system to produce time-like photons (in thedominant one-photon channel where the small value of the fine-structure constant helps).

Bruno Touschek gave a seminar on March 7, 1960, at Frascati inwhich he guaranteed that an electron and a positron necessarilymeet in a single orbit because QED is CP (charge-parity) invariant.His skeptical colleagues did not have the courage to doubt him.His seminar was attended by many people, among them Salvini,Fernando Amman (who a couple of years later would be in chargeof the Adone collider), and Raul Gatto, who with Nicola Cabibbo

many or few hadrons be produced?

B.Touschek R.Gatto

N. Cabibbo

immediately began to investigate all possible electron-positron reaction cross sectionsand derived formulas describing the relevant parameters,particularly in hadronic physics. The central question wasasked: Will

Touschek’s seminar appears to have been thoroughly preparedduring the whole period going from February 18 to March 7, as canbe inferred from his notebook preserved in the archives of thePhysics Department of La Sapienza University in Rome.

N = number of particles accepted per pulseν = repetition rate of the Synch (ν = 20)τ = lifetime of the beam,q = effective x-section of the circulating beamσ = x-section for the process to be observedc = velocity of lightπR = half circumference of the storage magnet

Touschek’s first note-book(February 18, 1960)

Touschek, in his peculiar style, tried to convince Salvini toimmediately convert the Frascati electron synchrotron into acollider ring (as actually was done ten years later for theCambridge Electron Synchrotron). Salvini wisely refused : TheFrascati electron synchrotron was unfit for this purpose, andmany experiments had already been performed or scheduled forit. Salvini however warmly

agreed with theproposal to prepare anew machine. Wetherefore immediatelyconstituted a smallgroup of people toinvestigate the mostpressing problems thatwould have to beaddressed to build anelectron-positroncollider ring ex novo.

The original group consisted of GianfrancoCorazza (b. 1924), Giorgio Ghigo (1929–1968),Bruno Touschek, and myself.We agreed that the energy of the electron andpositron beam should be 250 MeV, which wasa reasonable amount higher than the thresholdfor producing a positive and negative pionpair.

Giorgio Ghigo

Carlo Bernardini

Gianfranco Corazza working on the electron synchrotron donut

Bruno Touschek

On The Storage Ring.

The following is a very sketchy proposal forthe construction of a storage ring inFrascati…No literature has been consulted inits preparation, since this invariably slowsdown progress in the first stage…I shallpresent you here all I have thought about itand much, which others have suggested to me andto anticipate the question: No, I have notproperly read O’Neill, but I hope that somebodywill[…] I prefer to think of it as an experimentrather than as a machine […]Talking of it as an experiment I propose tostudy the reactions 2γ (A)

µ+ + µ− (B) π+ + π− (C)

e+ + e−

And I admit that I think that there isnothing else of importance, which can bestudied with the same set up. The first of the processes listed is twoquantum annihilation. The cross section forthis process is

σ(A)=6.3.10−30cm2

At 250 MeV[…]I propose to use (1A) as amonitoring process[…]

3 Injection of electronsand positrons.4 Design of the magnetto achieve compactnessand to leave enoughroom to allow access tothe electron and positronbeams, and design of theRF cavity to compensatefor the synchrotron-radiation losses.

Problems:1 The evaluation of the “source factor,” which from then on wascalled the “luminosity.”2 Analysis of the beam lifetime and of the processes that might influence it.

(1) Electron beam from the electronsyncrotron

(2) External target: production ofbremsstrahlung gamma-rays

(3) Donut walls(4) Internal target: tantalum converter(5) RF Cavity

With the magnetic field (B) directeddownward, electrons are injected intothe ring when it is in the positionshown at the right.The ring is then translated and rotated180° as shown above.Positrons are then produced and orbitclockwise.

We did a lot of calculations trying to estimatethe injection efficiency; most of them, however,were quite unreliable because of theirsensitivity to imperfectly-known magneticparameters.Touschek kept himself fully informed on alldetails of these technical aspects of our work-in-progress with AdA, but his main concern, asalways, was the physics.

To get physical results, the oppositely orbiting electron andpositron beams had to meet and overlap completely. Touschekthus was fascinated with the luminosity formula, which actuallyfollowed from a classical calculation of the “quality” of theoperating ring, something like the duty cycle of an engine.To the question: “How can you be sure that electrons andpositrons will meet?” he answered: “Obviously, TCP [time-charge-parity] theorem! Actually, CP is enough!” Another questionsometimes was: “Will electromagnetic interactions with the wallsof the donut separate the beams?” Bruno’s answer: “Scheisse!”And so on.

To our surprise even a singleelectron was visible to thenaked eye through one of theportholes. A common joke wasto store a few electrons andastonish distinguishedvisitors, among whom wereEdoardo Amaldi, Philip IvorDee from Glasgow (a formerstudent of Ernest Rutherfordand a good friend ofBruno),Wolfgang Paul fromBonn, Guy von Dardel fromLund, Boyce McDaniel andAlbert Silverman from Cornell,and Matthew Sands andRobert Walker from Caltech.

The phototube record showing steps that correspond tosingle electrons entering or leaving AdA.

On February 27, 1961, just less than a year afterTouschek’s seminar, we got the first stored electronsand/or positrons.

The injection trials werecarried out first byirradiating the internalconverter target when thering was installed far awayfrom the electronsyncrotron on a tower(tripod).Then we moved the ringnear the syncrotronmounting the magnet on asupport that rotatedaround a horizontal axis.This made possible toinvert the magnetic field toswitch from electron topositrons, and vice versa.

Many discussions occurred as to which were the electrons and whichwere the positrons; they ended when Bruno drew a famous cartoonunderlining the inconclusive debate.

The difficulties in using a well-collimated bremsstrahlunggamma-ray beam in Frascati toproduce electrons and positronsin AdA was the subject ofwidespread bad humor, but theyfavored the rapid acceptance ofPierre Marin’s proposal, on behalfof André Blanc Lapierre, theDirector of the Orsay Laboratory,to move Ada to Orsay, where theirelectron linac beam could bepositioned very close to AdA’sevacuated donut and internalconverter. Bruno and I agreedwith Marin’s proposal, and weeasily convinced Salvini andAmaldi to accept it. We thereforeimmediately organized thetransportation of AdA to Orsay ona truck in the first half of June1962.

Jacques Haïssinskiand

Pierre Marin

AdA on the rotating and translating platform atOrsay. The injector beam channel is visible on theleft.

We eventually were able to inject a non-negligibleelectron/positron current, which was extremely satisfactory to us.

I already had reconsidered the calculation of the transverse size ofthe beam – its horizontal and vertical dimensions – and I realizedthat its vertical dimension was much smaller than we had believed.

We calculated that at an energy of 200 MeV and a pressure of 10-9torr the root-meansquare dimensions of the electron and positronbunches were 1.8 millimeters in the horizontal direction, 1.5microns in the vertical direction, and 255 millimeters in thelongitudinal direction, where only the transverse (horizontal andvertical) dimensions matter for the beam’s luminosity. Ourenthusiasm, however, was short-lived.That night in March of 1963 we were steadily filling the collider ringas usual; the vacuum was excellent, as was the electron-positroninjection current. Everything gave the impression that we werereaching higher stored currents than ever before. Then, at acertain point, we noticed that the injection rate was decreasing,and sometime later the stored current increased no further – it hadreached saturation.

The vacuum pressure gauge showed no change; a single beamwas in. Touschek went crazy. It was about 2 or 3 o’clock in themorning – a time when we usually were at work at Orsay.Touschek left the laboratory and went to the Café de la Gare,which was open to serve passengers leaving and boarding thenight trains.We continued to try to inject.Suddenly, Bruno reappeared (I cannot claim that he was exactlysober at the moment) announcing: “I got it! It is Møller scatteringin the bunch!” He then exhibited a formula, explaining that he hadcalculated that saturation should occur at the beam intensities wehad reached because electron-electron scattering in the beam’sbunches was transferring energy from the betatron oscillations inthe traverse directions into the longitudinal stability zone, whichwas limited in the amount of energy it could accept.Bruno, however, was desperate. In thinking about the completecalculation, we understood that this disadvantage appliedespecially to small colliders. AdA thus looked like a flop. I began toreconsider the situation, trying to be optimistic. For someunknown reason, I pictured the beam in my mind as a strapbecause of the “different” mechanisms that resulted in forming itshorizontal and vertical sizes. “Different” means “uncoupled.”

Suppose, then, that you introduce acoupling. I telephoned Frascati:“Please prepare a small quadrupolemagnet, of such and suchdimensions, to be inserted into thequasi-straight section of AdA. I willfly tonight, come to the lab tomorrowmorning, and bring the machinedmagnet back to Orsay in 48 hours.Thank you, guys.”But we had won the battle and notthe war. When the coupling wasturned off, the beam lifetimedecreased drastically for the firsthour, so that the number ofannihilation events recorded by ourCerenkov counter was low. Brunowas disappointed, but not entirely: Athigher energies, which for instancewe later had with the Adone collider,we knew that this saturation effect,known as the “Touschek effect,” wasnot a disaster.

Coefficient of the term linear in thestored particle number of inverselifetime as a function of energy(original drawing by Bruno Touschek)

Δ December run

σ January run

λ February run

µ April run

Bruno noticed that therate of gamma raysobserved in thedirection of beam 1 isproportional to thenumber of particles N1in it, while for beam-beam events the rate isproportional to thenumber of particlesN1N2 in both beams 1and 2. Thus, theobserved gamma-rayrate divided by N1depends linearly on N2,and the slope of the lineis a measure of the raterecorded by thedetector monitoring thereaction, that is, of theluminosity of the beam.

Counting rate per particle in beam n° 1 asa function of particle number in beam n° 2

Bruno and I took charge of the data analysis. We used hisformula to predict the beam size and calculate its luminosity,finding a luminosity of 1025 particles per square centimeter persecond – small but not negligible.

“The paper describes a series ofexperiments carried out with thepurpose of observing the γ-raysproduced in the collision betweenstored beams of electrons andpositrons. The interaction rate … wasfound to be in good agreement withthe hypothesis that there is acomplete overlap between the twobeams…”C. Bernardini et al., “Measurementsof the Rate of Interaction betweenStored Electrons and Positrons”,Il Nuovo Cimento 34, December 16,1964 - Received July 16

1960Bruno TouschekCarlo BernardiniGianfranco CorazzaGiorgio Ghigo(Giancarlo Sacerdoti: magnet;Antonio Massarotti andMario Puglisi: RF Cavity)

1961(joined in:)(Ubaldo Bizzarri)Giuseppe Di GiugnoRuggero Querzoli

1962(joined in:)(François Lacoste)Pierre MarinJacques Haïssinski

1963(final group:)Carlo BernardiniGianfranco CorazzaGiuseppe Di Giugno(Giorgio Ghigo)Jacques HaïssinskiPierre MarinRuggero QuerzoliBruno Touschek

The adventure that was AdA thus came to its happy end. I particularly want toemphasize not only our scientific achievements with it, but also the exceptional– I would say unique – atmosphere of collaboration and friendship that weexperienced during those four years, 1960 to 1964.

ADONE - A Draft Proposal for aColliding Beam ExperimentBruno Touschek,Rome, November 9, 1960“It is assumed that experiments inwhich there are only two particlesin the final state are most easy tointerpret.There are 16such reactions…”

Touschekhesitated to getinvolved with Adone: there weretoo many problems of a“nonphysical” nature (placingorders, carrying out other duties,drawing up plans, lack ofimprovisation, and the like).

Adone, a 3000-MeVelectron-positroncollider at Frascati(each beam had anenergy of 1500 MeV),was FernandoAmman’smasterpiece. Ammanconceived its ring in1961–1963 using themost advancedconcepts, a powerfulelectron linac to usewith the electron-positron converter,and a hall suited forexperiments.

Adone was a unique new tool,and prominent Italianphysicists wanted to measuresomething with it. My personalfeeling (which I still maintainwas right, after so many years)was that one should firstexplore, with unsophisticatedmulti-purpose experimentaldevices (counters and sparkchambers covering a widesolid angle), if and howhadrons are produced atsignificant energies,particularly in the form ofnarrow resonances, to profitfrom the very precise energydefinition of the collidingbeams. We misseddiscovering the J/Ψ particleonly because it was found at50 MeV above Adone’smaximum beam energy!

Summary of the hadronic totalproduction at various energiescompared to µ-pairs

Bruno Touschek

and

his dog Lola

(1965)

UnfortunatelyBruno Touschekdid not liveenough to seethat essentiallyall high energyphysics comesfrom colliders...He died inAustria in 1978,at 57.

AdA under its glass showcase - Frascati, Open Air Museum.

1961 AdA, Frascati1964 VEPP 2, Novosibirsk, URSS1965 ACO, Orsay, France1969 ADONE, Frascati, Italy1971 CEA, Cambridge, USA1972 SPEAR, Stanford, USA1974 DORIS, Hamburg, Germany1975 VEPP-2M, Novosibirsk, URSS1977 VEPP-3, Novosibirsk, URSS1978 VEPP-4, Novosibirsk, URSS1978 PETRA, Hamburg, Germany1979 CESR, Cornell, USA1980 PEP, Stanford, USA1981 Sp-pbarS, CERN, Switzerland1982 Fermilab p-pbar, USA1987 TEVATRON, Fermilab, USA1989 SLC, Stanford, USA1989 BEPC, Peking, China1989 LEP, CERN, Switzerland1992 HERA, Hamburg, Germany1994 VEPP-4M, Novosibirsk, Russia1999 DAΦNE, Frascati, Italy1999 KEKB, Tsukuba, Japan1999 PEP-II, Stanford, USA2003 VEPP-2000, Novosibirsk, Russia ? LHC, CERN, Switzerland