For future electron-positron linear colliders, high-intensity electron and positron beams are needed. These must be sufficiently well defined (low-emittance) in order to reach a high luminosity at the collision point. While intense electron beams can be produced without major difficulty, the production of intense positron beams is more of a problem.

A significant R&D effort is under way in many laboratories to find a positron source satisfying the requirements of intensity and emittance and being reliable over long periods of time. Recently an experiment on a special kind of positron source carried out at CERN yielded promising results.

The basic route to creating positrons is via the production of hard photons (gamma rays) by electron-positron pairs in a material. In conventional sources, a powerful electron beam hits an amorphous target (without any particular crystal orientation). In such a target the electrons are attracted by the nuclei and radiate photons (bremsstrahlung). These in turn produce electron-positron pairs in the target. The rates for these two successive steps increase by the square of the atomic number of the target, so that heavy nuclear materials, such as tungsten, are preferred.

Another approach is to use a crystal, the atomic rows of which are aligned with the incident electron beam, instead of an amorphous target. Here the electrons will be attracted not only by the individual nuclei but also by many successive nuclei of a same row, as though the atomic mass were multiplied by the number of successive nuclei. This gives more intense radiation (coherent bremsstrahlung).

An electron can even revolve around the atomic row many times. It is then "channelled" and radiates as though it were a helicoidal undulator, the period of which would be in the order of 1 µm, with a field equivalent to thousands of Teslas. This channelling radiation, which is even more intense than coherent bremsstrahlung, gives more electron-positron pairs.

Crystal targets are therefore thinner than amorphous ones that give the same number of positrons. This is useful for limiting the energy deposited in the target and, hence, the heating.

The aim of the WA103 experiment, carried out at CERN in 2000 and 2001 after similar experiments at Orsay (France) and KEK (Japan), was to observe and measure the enhancement of positron production by a crystalline source and to measure the energy and angular distributions of the emitted positrons.

An electron beam of 5-40 GeV was used in the West Hall of CERN's SPS synchrotron. Two incident energies were selected - 6 GeV and 10 GeV - the former corresponding to the choice of the Next Linear Collider; the latter to the Japanese Linear Collider.

The experimental set-up accommodated different kinds of target - all crystal (8 mm) or compound (4 mm crystal followed by 4 mm of amorphous target). The emitted positrons are detected in a drift chamber partially in a magnetic field. The part of the chamber outside the magnetic field provides the positron emission angle, while that in the magnetic field provides the positron momentum. The emitted photons are monitored in a preshower detector and in a "spaghetti" calorimeter (scintillating fibres in lead).

Different laboratories took part: LAL-Orsay (acquisition electronics and goniometer); IPN-Lyon (simulations and goniometer control); the Max-Planck Institute, Stuttgart (tungsten crystals); the Budker Institute, Novosibirsk (conception and realization of the drift chamber, track reconstruction programme and simulations); and the institutes of Kharkov and Tomsk (photon detectors). The spaghetti calorimeter was provided by LNF-Frascati. The tests on the detectors were carried out at LAL-Orsay with the participation of the Budker Institute and IPN-Lyon physicists. Data-taking was done by Franco-Russian teams in which the physicists from the Budker Institute played an essential role. In this long collaboration, physicists from the College de France-Paris participated in the initial simulations.

The first results showed the channelling enhancement, which was close to that predicted by simulations. Comparison of the positron energy spectra obtained for a tungsten crystal oriented on its <111> axis and for an amorphous target of the same thickness showed a positron yield boosted by a factor of 3-4 for a 4 mm target and a 10 GeV electron beam.

A boost slightly larger than 2 is seen for the 8 mm crystal target. A large number of soft positrons were also seen, which is very interesting for the accelerator acceptance downstream of the target. The observations made on the photon detectors confirmed the enhancement of the number of photons and of the radiated energy.

This appears to be a very promising avenue for future linear colliders.