Experimental and numerical study of water diffusion in thermoset polymers for advanced recycling

The increased use of fibre reinforced polymer composites in the industry, coupled with increasingly stringent European regulations, has resulted in a need to develop effective recycling technologies for composites. Such sociotechnical pressure will continue to grow in the coming years, driving the development of advanced End-of-Life solutions. Recycled carbon fibres, if properly treated, can indeed have a reduced global warming potential and require less energy compared to the production of virgin fibres. This thesis investigates supercritical solvolysis, an innovative chemical recycling technique, within the framework of the European EuReComp project, with a specific focus on water diffusion in thermoset polymers. A multiscale modelling strategy is adopted to provide insights into transport phenomena across different length scales, combining continuum-based Finite Element analysis with atomistic Molecular Dynamics simulations. At the microscale, a comprehensive series of numerical analysis was conducted to quantify the effect of voids on diffusion efficiency varying void size, void content and temperature. The results highlight that temperature is the primary driver of diffusion, whereas microstructural variations play a more subtle role. At the nanoscale, Molecular Dynamics simulations are implemented to evaluate the diffusion coefficient in a series of different cross-linked epoxy systems, with careful attention to data extraction and numerical post-processing. In particular, the Crank-Nicolson method is applied to overcome the limitations of classical analytical solutions of Fick’s diffusion laws. The comparative assessment of diffusion coefficients derived from different evaluation strategies is expected to clarify the interplay between molecular structure, cross-linking degree and temperature. By bridging nanoscale and microscale perspectives, this work contributes to understanding diffusion processes with insights into local structure-property correlations and molecular motion mechanisms that are inaccessible through experimental activity alone. The results are meant to reinforce the link between computational modeling and experimental validation in the sustainable development of composites and to aid in the design of more effective recycling methods.