Just as in the atom, where the electrons fill different energy levels or "shells", the nucleons (neutrons and protons) in an atomic nucleus are also arranged in similar shells. Each time a shell has the maximum number of particles it can accommodate, the nucleus, like the atom, is particularly stable.

These "magic numbers" (2, 8, 20, 28, 50, 82 and 126) were discovered in the 1940s and soon explained by the nuclear shell model. Unlike the atom, the atomic nucleus consists of two different types of particle - the protons and the neutrons. A nucleus with completely filled shells for protons and for neutrons is called "doubly magic".

Of the roughly 2500 different nuclear isotopes known to date, only nine had a doubly magic shell structure. Nickel-48, with 28 protons and 20 neutrons, becomes number 10 in this list, and probably, at least for quite a while, the last one.

Beyond the importance of nickel-48, owing to its doubly magic properties, this nucleus is also of particular interest because it is at the extreme limit of nuclear stability, where the nuclear forces are no longer able to bind all protons and neutrons together.

At the "drip lines", nuclei decay by the emission of excess protons or neutrons. All commonly used models for atomic nuclei predict that nickel-48 is already beyond this proton drip line and is thus unstable with respect to the strong interaction, which means that this nucleus is only held together briefly owing to electrical forces between the protons.

Therefore, a possible decay mode of nickel-48 is the emission of two protons forming a helium-2 nucleus, analogous to alpha decay, where a helium-4 nucleus is emitted. This former type of radioactivity has never been observed. In addition, nickel-48 is the only

doubly magic nucleus with a bound mirror nucleus, which will allow for interesting mirror symmetry studies.

In September 1999 a collaboration of French, Polish and Romanian physicists began an experiment at the Grand Accélérateur National d'Ions Lourds (GANIL) in Caen, France, to search for nickel-48, the last doubly magic nucleus accessible with present methods.

A primary beam of nickel-58 with an average intensity of 1012 ions per second and an energy of 95 MeV per nucleon hit a natural nickel target in the superconducting solenoids of the SISSI device.

The proton-rich projectile fragments were selected by the LISE3 separator and finally identified by their time of flight, their energy loss and their total energy in a detection set-up consisting of a microchannel plate detector and a stack of five silicon detectors. This allowed the measurement of 10 independent parameters to identify each fragment arriving at the focal plane.

Features of GANIL

The success of the present experiment is a result of the combination of specific and powerful features available at GANIL:
* a primary beam intensity never reached before was achieved through an intense ion-source development programme: a new technique allowed nickel to be treated as a gas in the ion source, yielding a gain of a factor of 20 compared with past experiments;
* the transmission of the GANIL cyclotrons was optimized to accelerate a high-intensity primary beam;
* the efficient production and collection of projectile fragments by the SISSI superconducting device;
* the powerful separation and identification by the LISE3 separator with its velocity filter and an efficient detection set-up.

The experiment ran for about 10 days, revealing for the first time four production "events" of this new nucleus. Although optimized for the transmission of nickel-48, it also produced other exotic proton-rich nuclei in the vicinity - about 100 events of nickel-49, 50 of iron-45 and 290 of chromium-42. This confirms a similar experiment conducted about three years ago at the GSI laboratory, Darmstadt, where 5, 3 and 12 events, respectively, of these latter isotopes were reported for the first time.

The new observation gives a lower limit for the half-life of nickel-48 of about 0.5 ms. This contradicts a number of models that predicted nickel-48 to be highly unstable, with half-lives of far less than 1 ms, the typical flight time of the projectile fragments between the production target in the SISSI device and the detection set-up at the end of the LISE3 separator.

However, the few events observed at GANIL do not allow a detailed comparison with nuclear models. To do this requires higher-statistics experiments to determine, for example, the exact half-life of this nucleus. Such experiments should be possible in the near future at GSI as well as at GANIL, where continuous improvements in source development and the acceleration process should yield even higher production rates.