Single-Step Identification and Sorting Technology for Recycling of Plastic Packaging and Technical Polymer Products

Abstract This work presents the development and pilot-scale evaluation of Sort4Circle (S4C), a novel single-stage, multi-sensor-sorting approach designed to address emerging challenges in plastic recycling and circular economy systems. S4C singulates and separates mixed waste streams into a continuous sequence of individual items and applies multiple identification techniques prior to their mechanical sorting into highly specific fractions. The S4C approach enables the modular integration of multiple detection technologies, including near-infrared (NIR) spectroscopy, colour detection, AI-based image analysis, medium-infrared (MIR) spectroscopy, fluorescent tracer detection, and optionally laser-induced-breakdown-spectroscopy as well as digital watermark detection, depending on the specific sorting application. The primary objective of S4C is to enhance material identification accuracy while enabling efficient sorting into a large number of fractions. The S4C approach was experimentally evaluated using three pilot setups addressing different aspects of the approach: (1) colour- and fluorescent-marker-based sorting of shredded and small plastic pieces, (2) NIR-, MIR- and colour-based sorting of larger post-production plastic waste and (3) object singulation of rigid lightweight plastic packaging (LWP). Core innovations include the combination of object singulation and handling, application-specific multi-sensor identification of individual objects, and their assignment to multiple output fractions after a single detection stage. High-purity multi-criteria sorting was experimentally demonstrated for shredded and small plastic particles and post-production plastic waste, whereas the LWP experiments focused on singulation performance. In addition to the pilot-scale experiments, the study evaluates market relevance and waste stream characteristics, complemented by a Life Cycle Assessment (LCA). The LCA scenario results indicate a potential reduction in net greenhouse gas balance of approximately − 684 kg CO₂-equivalents per ton of LWP for the modelled S4C scenario, corresponding to a benefit of about 245 kg CO₂-equivalents compared with the modelled state-of-the-art recycling scenario. Additionally, the LCA scenario results indicate an increase in recyclate output by roughly 30%, from about 276 kg to 367 kg per ton of LWP.