SELF-HEALING COMPOSITE SANDWICH PANELS Hugo R. Williams, Richard S. Trask and Ian P. Bond Advanced Composites Centre for Innovation and Science, Dept. of Aerospace Engineering, University of Bristol, Queen’s Building, University Walk, Bristol, BS8 1TR, UK. KEYWORDS: Self-repair, Sandwich Structure, Biomimetic, Compression-after-impact. INTRODUCTION Sandwich construction offers a competitive design option for advanced structures subject to bending or compression loading. High performing skin materials, especially fibre-reinforced composite, separated by a lightweight core give a structure with high bending and buckling stiffness. Impact damage can have a detrimental effect on the performance of sandwich structures [1,2]. Generally, the primary damage mode for low-velocity blunt impact on a brittle-cored sandwich structure is a cohesive disbond in the core just under the impacted face with consequent loss of skin stability. Residual strengths well below 50% of the undamaged value have been reported in flexure-after-impact of beam specimens [3-5]. The loss of skin support has also been shown to reduce the compressive strength of sandwich panels by over 25% [6-10]. An alternative, bioinspired, approach from the ‘traditional’ damage tolerant design [e.g. 2] is to provide a material with the ability to self-heal [11,12]. This approach has been reported in pure polymers [e.g. 13-15], polymer composites [e.g. 16-18] and has been shown to recover failure mode and load in sandwich beams subject to flexure-after-impact [19]. In the latter case, a simple vascular network was introduced into the foam core of a composite sandwich structure. Rupture of the channels by impact damage allowed the healing agent to infiltrate the damage and cure. This previous study is developed in the work presented here using a compression-after-impact test configuration, improved manufacturing technique and a tailored network layout. METHOD Vascular sandwich cores were manufactured by bonding 1.5mm bore silicone tubing in milled channels between two sheets of 8mm thick Rohacell closed-cell foam sheet using Cytec FM300K film adhesive. The core was cured in a vacuum bag at 175 o C. Holes of 2mm bore were drilled through the foam and tubing at strategic points to connect the silicone supply channels to the skin- core bond region. Pre-impregnated [0,90] S unidirectional Hexcel E-glass/913 epoxy laminates were co-cured onto the vascular core in a vacuum bag at 125 o C. Samples were sectioned into specimens (60x90mm) and the edges ground flat and perpendicular. Each specimen contained five supply vessels feeding a total of 18 risers. Impact damage (3J) was introduced using an Instron Dynatup 9250HV drop tower (5.35kg dropweight, 25mm hemispherical head). Self-healing specimens were infiltrated with Resintech Ltd RT310 epoxy resin and the corresponding hardener in adjacent supply tubes prior to impact and allowed to heal for 48 hours after impact. Similarly, pre-mixed RT310 resin system was used in one group of specimens for comparison. Edgewise compression tests were performed under displacement control according to ASTM C364 using a vertical axis Instron test frame with a 250kN load cell. RESULTS AND CONCLUSIONS Table 1 shows a summary of the results. The impact damage, although leaving virtually no visible damage, reduces the failure load to 70% of the undamaged load. The self-healed specimens recover to 82% of the undamaged load. Destructive sectioning showed the presence of uncured healing agent in the damaged area. When the damage was infiltrated with a pre-mixed resin, not only was the undamaged strength recovered, but the absolute strength increased further. This is because the damaged core and the network have been replaced with a fully cured resin. These data