Radiation-induced electrical degradation of ceramic materials : an artefact ?

Recently three experimental groups around Hodgson [1,2], Pells [3], and Shikama [4] have reported a phenomenon in irradiated ceramics, called radiationinduced electrical degradation (RIED). Their experimental results suggest that ceramic materials suffer a permanent loss of their electrically insulating ability after exposure to low doses of irradiation. Since the use of electrically insulating ceramic materials is essential for various com~nents of a fusion reactor, such as rf windows, neutral beam injector insulators, first wall current breaks, magnetic coil insulators and diagnostic toois, the phenomenon has immediately become of serious concern for fusion reactor designers. RIED has been found after electron [2], proton [3], and neutron [4] irradiation in a temperature range between 300 and 600°C in Al,O, and MgAl,O, and at doses of only 10p5, 10m3, and lo-’ dpa for electrons, protons, and neutrons, respectively. It occured when the specimens were exposed to an electric field during the irradiation. The presence of both a displacive and ionizing effect of the irradiation has been shown to be essential for its occurrence [Z]. The microst~ctural origin is yet unresolved, but it was suggested that the formation of metallic colloids [2,4], as they have previously been observed in irradiated alkaline halides and HVEM irradiated a-alumina, is responsible for the permament rise in electrical conductance. A review by Zinkle and Hodgson [S] has recently been published in this field. To contribute to the understanding of this phenomenon, two years ago we started a study on the influence of 10 MeV proton and 28 MeV alpha particle irradiations on the electrical resistivity of AlaO,, MgAl,O,, Si,N,, and AlN. The objectives of these experiments were (a) to further establish the correlation of RIED with the physical parameters, temperature, dose, dose rate, and electrical field strength, and (b) to explore possible differences between ionic bonding (Al,O, and MgAI,O,) and covalent bonding (Si,N, and AlN) in ceramics. Our prime objective, however, was the goal of clarifying the microstructural origin of this phenomenon (hopefully as a first step towards a possible solution of the problem it poses). The irradiations reported in this Letter were performed by 28 MeV alpha particles in vacuum on 140160 pm thick polycrystalline specimens. The specimens were solid state bonded to a nickel block which was kept at 500°C by a controlled resistance heater. A dc voltage of 50 V (corresponding to about 350 kV/m) was applied across the specimens during irradiation. The standard irradiation configuration consisted of four specimens of 3 mm diameter mounted as a square array, with a 0.5 cm2 square beam aperture. The resistance was measured during irradiation and during intermissions, when the beam was turned off, by means of electrical current reading. For transport of the electrical signals from the target chamber to the remote measuring equipment triaxial cables were used with the inner shield of the cable held at the same potential as the signal carrying lead. A more detailed description of the irradiation chamber, measuring equipment and specimen preparation technique will be given in a forthcoming paper. The present investigation used a standard threeterminal electrical guard technique for measuring the electrical resistances in the insulating ceramics. Its principle is shown in fig. 1. Platinum electrodes were positioned on the specimens by sputter deposition. Contact to the platinum electrodes was made by spring loading tungsten wires. Between the central electrode