Towards the systemic integration of chemical recycling : insights from reverse logistics network design for polystyrene-based building insulation

Transitioning to a circular economy is central to sustainable development. Plastics pose a pressing challenge within this transition, due to their high production volumes, complex material compositions, and environmental persistence. Among plastic applications, those in the built environment, particularly insulation materials such as expanded polystyrene (EPS), are challenging to manage at their end-of-life. In Germany, EPS has been used in external thermal insulation composite systems (ETICS) since the 1970s. As these systems approach the end of their service life, they are increasingly being removed and replaced, resulting in a growing waste stream that is primarily incinerated. This practice contradicts circular economy principles and is challenged by limited treatment capacity, making it not only environmentally unsustainable but also economically inefficient. While mechanical recycling is unsuitable for contaminated EPS waste from ETICS, emerging chemical recycling technologies, particularly depolymerization via pyrolysis and solvent-based processes, offer promising alternatives. However, these technologies encounter barriers for implementation and scale-up, including their high capital intensity, lack of adequate pre-processing infrastructure, and uncertainty regarding the logistically complex, decentralized generation of future waste volumes. To address these obstacles, a systemic planning approach is required, integrating detailed spatio-temporal waste forecasting, techno-economic evaluation, network design, and uncertainty consideration. This dissertation develops an integrated analytical framework to support infrastructure planning for the recycling of plastics via emerging chemical technologies. Focusing on EPS-ETICS waste from the German building sector as a case study, the research combines methodological rigor with applied relevance to address four core challenges: (1) spatial and temporal projection of future waste volumes, (2) comprehensive techno-economic assessment of different recovery options, (3) optimization of reverse logistics networks, and (4) robust planning under uncertainty from multiple sources. The research is structured around three peer-reviewed articles, each corresponding to a distinct stage of the analysis. First, a high-resolution forecasting model estimates future EPS-ETICS waste volumes in Germany over the coming decades, accounting for the building stock composition, probabilistic insulation lifetimes, renovation dynamics, and regional variability. This model supports ex-ante analysis and enables the integration of material flow dynamics into infrastructure design. Second, an optimization model is developed to design reverse logistics networks that include chemical recycling. The model identifies best technology choices, facility locations, processing capacities, and material flows based on a deterministic parameter setting. It incorporates key constraints related to economic costs, material transport, technology requirements, and regulatory compliance. Application of the model to the German case reveals that pyrolysis and solvent-based recycling can outperform incineration in certain scenarios. Solvent-based recycling, however, is more sensitive to regulatory incentives such as recycling quotas. Third, the model is extended into a scenario-based regret-robust optimization framework. This extension explicitly addresses uncertainties in EPS-ETICS waste volume trajectories, driven by regulation changes, economic factors, and technological development. The robust model evaluates trade-offs between system performance and flexibility, ensuring that network designs remain cost-effective and functional across a wide range of plausible futures. The results show that robust infrastructure designs can deliver near-optimal performance even under worst-case conditions. This resilience benefits stakeholders facing long-term investments with large uncertainty, particularly in the context of chemical recycling systems for plastic waste. The dissertation offers several key contributions. Methodologically, it advances the field of reverse logistics network design (RLND) by integrating high-resolution forecasting with robust optimization, offering a flexible framework for infrastructure planning in emerging circular economy applications. Practically, it provides strategic insights for two main stakeholder groups. For policymakers, the research highlights how targeted regulation can unlock the potential of chemical recycling technologies to reduce environmental burdens and enhance resource recovery. For investors and infrastructure planners, the framework offers tools to evaluate investment risks, optimize facility placement, and design systems that remain adaptive under future change. While focused on EPS-ETICS as a representative case, the framework and findings are transferable to other hard-to-recycle plastic streams and contexts characterized by spatially dispersed waste generation and systemic uncertainties. By aligning sustainability goals with decision-support, the research bridges gaps between academic modeling and real-world application. The dissertation also acknowledges some limitations, including the exclusion of EPS-ETICS from non-residential buildings or dynamic interactions between different drivers of uncertainty. Future research is encouraged to broaden the scope, by incorporating additional waste streams and treatment options, and integrating environmental or social sustainability dimensions through multi-criteria decision analysis. In summary, this dissertation presents a comprehensive planning approach for integrating chemical recycling technologies into the circular plastics economy. It offers a novel combination of forecasting, logistics optimization, and robust decision-making under uncertainty, tailored to the realities of infrastructure investment, policy design, and material circularity.