Vogelgesang, MalteMalteVogelgesangTrojanowski, Bettina KarinBettina KarinTrojanowskiHanstein, StefanStefanHansteinZüge, AndreasAndreasZügeBenner, WladislawWladislawBennerIonescu, EmanuelEmanuelIonescu2025-06-052025-06-052025https://publica.fraunhofer.de/handle/publica/4882882-s2.0-105005059994In a research project on bio-based materials, characteristic near-infrared (NIR) signatures of various inorganic and organic fillers, including wood-based fibers, were identified in polymers. These signatures are strong enough for sensor-based sorting, enabling the removal of these materials from unfilled streams or directing them into specific recycling pathways. This led to the idea of using bio-based materials as markers, offering a sustainable alternative to markers made from rare earth elements like neodymium. The study aimed to develop bio-based markers detectable with conventional NIR sorting equipment. The substances were extracted from apple pomace and cocoa shells, including hemicellulose, theobromine, DEAE glycan, catechin, and others. After initial testing, depolymerized hemicellulose emerged as the most promising candidate for large-scale tests. Bottles and plates containing varying concentrations of markers were produced and cut into flakes for sorting tests. Support vector machines achieved 99.8 % accuracy of classification for inorganic markers and 95.0 % for organic markers. Sorting experiments indicate that bio-based markers can sustainably differentiate polymer materials using existing sorting equipment, but are prone to error from organic dust contamination. Additional research is needed to reduce the required concentrations for robust sorting, making it a viable market product.enfalseNear infrared spectroscopypolymer compoundstracer-based sortingconference paper