Assessment of solvent capsule-based healing for woven fibre-reinforced epoxies

Structural fibre-reinforced composites based on brittle matrices such as epoxy resins may be subject to short-term subcritical matrix damage that develops with time to the point where it may compromise the integrity of the structure. Currently, components that have been damaged at the microscale are usually inspected and, depending on the extent of the damage, repaired with patches or entirely replaced. Self-healing functionality would prevent development of such damage by repairing it as soon as it appears, thereby increasing the service life of the material. Among the many potential methods for introducing self-healing capacity to epoxy resins, incorporation of solvent-filled capsules is particularly promising. Cracks cause the capsules to break, releasing a solvent that is able to swell the crack faces and induce reaction of residual monomer and/or crosslinkable functional groups intentionally left by under-curing the matrix. Thus, a model system based on ethyl phenylacetate (EPA) has been shown to provide up to approximately 90% efficiency in recovering fracture toughness, and to effectively block fatigue crack propagation in pure epoxy resin. The present thesis focuses on the integration of a EPA solvent capsule-based healing system into a glass fibre-reinforced epoxy and the assessment of its effectiveness in repairing static interply damage. EPA-filled capsules with diameters in the range of 90-250 ÎŒm were produced using an oil-in-water emulsion technique and characterized in terms of their thermal stability and mechanical properties. Pure urea-formaldehyde (UF) as well as polyurethane/urea-formaldehyde (PU/UF) double wall capsule shell materials were analysed, showing that the wall thickness of pure UF capsules is independent of their diameter, whereas that of PU/UF capsules is 4 to 6 times greater and scales with the capsule diameter until a saturation value. The modulus of these shell materials was quantified using compression test and an analytical model, and found to range from 1 to 4 GPa depending on the amount of PU content in the shell composition. Bursting forces varied from 5 to 25 mN depending on the shell composition and diameter, thus providing a range of shell compositions to select from in order to tune the capsule behaviour. The integration of the capsules into the composite was carried out by dispersing them manually onto the reinforcement textile prior to processing. During this phase, the optimum capsule size range, position and concentration could be selected to fit in the interstices and survive textile packing. Textile packing properties were tested in compression; the presence of capsules led to a decrease of the reinforcement volume fraction at a given applied load value, in particular at low packing loads, and affected the usual power law behaviour of a textile in compression, up to the rupture of the capsules. Longitudinal permeability was also increased by a factor of 6 to 8 in presence of capsules for a given r