Colliding beam facilities: 1974 to the present.

``e⁺e^- annihilations are rather poor producers of meson resonant states and not likely to contribute much new knowledge of their properties" [8]. In April 1974, one could say this without a qualm. By November, nature had overturned it spectacularly, and for over twenty years electron-positron colliders have dominated the field. The bosons produced - the and families and the Z⁰ boson - have been studied with a precision that outstrips virtually all else, and the current standard model is dominated by our knowledge of their properties.

The SPEAR magnetic detector, which was used to make the discovery, is illustrated in figure 1. The SPEAR team could not use a bubble chamber for colliding beams - the target, the duty cycle - everything was wrong. But they attempted to build a ``4 " detector which copied its property of seeing all the things which emerged from each reaction, whatever their directions. No-one had attempted such an ambitious thing before using electronic/spark chamber techniques. But this was what was needed to access colliding beam physics, and to build and run it needed ``a new way to do physics" (W Chinowsky, 1974) - the mega-team, in this case twenty-nine people, was here to stay.

By 1977, I knew that my polarised target experiments were coming to an end at the NIMROD accelerator in England, and I needed to look elsewhere for the future. I did one thing right, in deciding to work at PETRA, the new high-energy electron-positron ring being constructed in DESY, Hamburg, to run from late 1978. The lessons of the November Revolution had been well learned, and this was a new step in complexity and in organisation. The detector was full-solid angle. As at SPEAR, one took all reactions simultaneously, with the same democracy as a bubble chamber, but with a new collision to be inspected every 4 s, rather than every two seconds or so. This placed great importance on the event ``trigger".

One needed to reject unwanted collisions and select wanted candidates at this prodigious rate, and needed to do this in a reliable and understood way, with a known efficiency, known background, and known biases. Electronic logic had developed to the point where complex track-finding triggers could be built to work at this speed. But to make this work in any sensible way, one had to be able to measure the efficiency, to monitor it, and to make improvements. So one needed to reconstruct the results from day one. The lead-time on event reconstruction fell from 18 months to one day, as the trigger experts clamoured for feedback on their performance. The higher energy of the collisions forced every part of the detector to be bigger than at SPEAR.

How to achieve all this? A collaboration of powerful institutes was needed, each with expert teams specialising in certain aspects of the task. In the case of the TASSO collaboration, we began with 87 scientists from nine institutes in three countries: the number of institutes and countries grew a little later. It was a marvellous experience - the thrill of producing new science in all directions, the stimulus of working with enthusiasts from so many places, and the urgency of the competition inside and outside the collaboration, (a spur both to speed and to accuracy,) all combined to raise the level of our work.

A particle physics experiment is a long undertaking, though not perhaps as long as the time scales involved in the design, construction and exploitation of scientific satellites. It starts with a scientific idea, perhaps amongst a group of friends. The development of the detector technique can take over five years - for the quality of the instrument is paramount to the range of the work that can be done. If one takes HERA as an example, meetings on possible electron-proton colliders were held as early as 1975. The subject acquired real momentum (as viewed by this experimenter) in Genoa in 1984, and the first results came in 1992. The full exploitation of the facility will take over ten years.

Particle Physics experiments are like astronomical observatories, in that they permit a range of studies. But by contrast with observatories, each study addresses the entire data sample, accumulated perhaps over several years. Wanted data are selected by triggering and by appropriate cuts in analysis. So different subsystems, designed to measure different aspects of each event, must cohabit without degradation of each others' performance. Thus semiconductor `vertex detectors' which measure track positions close to the production point must have minimal material to reduce both multiple scattering and photon conversion which would degrade the track momentum and multiplicity measurement.

The more sophisticated results are often obtained by combining results from different detector components. An example from the LEP collider at CERN, producing a Z⁰ boson which decays to a quark-antiquark pair, is shown in figure 2. This shows the production of a B_s meson, consisting of a quark pair. The B_s decays after travelling about 3 mm, to . The then decays to , and the to K⁺K^-. The main tracking chamber (`time projection chamber') was used to measure the particle momenta, and to identify K-mesons by their ionisation loss, the vertex detector to separate clearly the production and decay of the B_s, the electromagnetic calorimeter identified the electron, and the penetration of the particles through the hadron calorimeter into the muon chambers identifies the and the .

The ALEPH collaboration, which worked together to construct and exploit this elegant detector, started work in 1982 and now comprises over 400 physicists from 32 institutes in 10 countries.[9] Not all these countries are member states of CERN. People outside the field often ask how such a body is managed. It is done using no documents of greater legal strength than a memorandum of understanding - a document which binds the funding agencies of each country to best endeavours, but is not something on which you could sue. The collaborations operate within a formal structure devised by themselves under the avuncular gaze of the CERN management. In the end, it works because the scientists themselves want it to work. The sanction for failure to deliver results, for poor performance or inadequate maintenance of equipment, or for unethical behaviour is loss of reputation.

The world of Particle Physics is in one sense small. The fear of being drummed out of the club is a powerful sanction, but the very smallness of this world makes it in another sense large. It provided some porosity in the iron curtain long before its rust began to be evident elsewhere, and continues to provide a mechanism for cooperation between both well resourced and poorly resourced countries. Groups from afar that bring little in the way of Swiss Francs can contribute as equal collaborators if they contribute intellectual capital that their partners value.