Dielectric strength of ion irradiated polyimide and epoxy/fiber composites

During long-term operation of the new FAIR facility, parts of the superconducting magnets will be exposed to high radiation levels, cryogenic temperatures, and dynamic mechanical loads (Lorentzian forces during pulsed operation). Depending on the position of the different components, the radiation due to beam losses consists of a cocktail of gammas, neutrons, protons, and heavier particles [1]. Although the number of heavy fragments of the initial projectiles is small compared to neutrons, protons, or light fragments (e.g. α particles), their large energy deposition can induce extensive damage at rather low fluences (dose calculations show that the contribution of heavy ions to the total accumulated dose can reach 80% [2]). In the MeV to GeV energy regime, beam-induced radiation damage strongly depends on the specific sensitivity of the material and scales with fluence and electronic energy loss of the ions. In particular, organic polymers to be used e.g., as cable insulation for the superconducting FAIR magnets, may undergo severe degradation accompanied by outgassing of small volatile radiolysis products [3,4]. This study tackles the dielectric strength of polyimide (Kapton) as electrical insulation and G11-type epoxy/glassfiber composites as structural support material. Kapton foils of thickness 12, 25, and 50 μm were irradiated with 21 and 800 MeV protons (ITEP) and with various heavy ions of MeV-GeV energy (UNILAC, GSI). In addition, three types of 1-mm thick epoxy/glassfiber sheets were exposed to 180-MeV/u Xe ions (SIS, GSI). The irradiation experiments with protons and Xe ions took place in air, while the UNILAC irradiations were performed in vacuum. To test degradation of the insulating properties, breakdown voltage measurements were carried out using a current-limited 20-kV high voltage tester available at CERN. The ramping speed of the DC voltage was 1.3 kV/s. The location of breakdown events was inspected by means of optical microscopy and typically occurred inside the Rogowski-type stainless steal electrodes having a diameter of 10 mm. Any significant geometric influence on the electric field is therefore excluded. The tests took place in air, at room temperature, and at a humidity of 24-30%. No systematic errors due to temperature and/or humidity fluctuations were found. For Kapton, the measurements show an overall decrease of the breakdown voltage with increasing dose (Fig. 1). For light projectiles, such as protons and C ions of rather small electronic energy loss (dE/dx between 0.03 and 0.5 keV/nm), the decrease of the breakdown voltage becomes significant at doses above 1 MGy. In the case of heavy ions (dE/dx .> 16 keV/nm), the breakdown voltage changes at a much lower dose (note the semi-log presentation of Fig. 1). The expected maximum voltage in the superconducting coils of the FAIR magnets is about 3 kV. In the tested dose regime up to ~80 MGy, the degradation due to light ions is insignificant for the operation voltage. The situation is much more crucial for heavy ions, where already a dose of a few kGy results in a severe decrease of the breakdown voltage. At around 0.1 MGy, the values are close to the voltage requirement for the FAIR magnets. These results give a first indication that individual tracks completely passing through the Kapton insulation may represent a serious security risk for the insulation of the FAIR magnet coils.