Neodymium magnets are crucial for many green technologies such as e-mobility and wind energy, resulting in a continuous increase in global demand. Due to low resilience in the supply chains and poor substitutability, various countries have classified the required rare earth metals as critical raw materials. In addition to other measures, the establishment of a circular economy can make a decisive contribution to a secure and sustainable supply. Long-loop recycling based on shredder fractions could play a key role in the future, as it allows for the processing of inhomogeneous and contaminated material streams that are unsuitable for other recovery routes. However, there is currently a lack of applicable processes for extracting the magnet fragments from the ferrous shredder fraction which is typically isolated from the overall waste stream via magnetic separation. Physical sorting is intended to achieve purities that enable efficient treatment using hydrometallurgy for recovery of the rare earth content. Beyond magnetic attraction forces, challenges arise from the low concentration of magnets and the similarities in terms of many physical properties for relevant soft and hard magnetic materials. In this context, the present study investigates the potential of applying magnetic separation to isolate and enrich the neodymium containing fraction from the disperse material mixture. Here two approaches are discussed. First, the study examines whether the transition from ferromagnetic to paramagnetic behavior at temperatures above the material-specific Curie temperature (Curie–Weiss law) can be exploited for sorting. Synergies with the mandatory, prior thermal demagnetization could be utilized for this purpose. Second, the feasibility of fine sorting in an inhomogeneous magnetic field without thermal influence is analyzed. The study follows a two-stage approach. In the first stage, analytical and simulation models are developed to systematically investigate the influencing parameters. These results provide the base for subsequent experimental validation using representative input material.
Ihne et al. (Thu,) studied this question.