Demand for rare earth permanent magnets (REPMs) has grown drastically the past decades and is expected to increase further due to their use in electronics, electric vehicles and wind turbines. Rare earth supply challenges have increased the urgency to recycle End-of-Life (EoL) REPMs. This paper examined the development of global EoL REPM recycling by applying the Technological Innovation Systems (TIS) framework, assessing temporal development and dynamics between different aspects of the system. The analysis showed an acceleration of recycling innovation activities since 2013, evidenced by e.g. research and development initiatives, (commercial) pilot plants and media and policy attention. Activities were identified globally, with regional concentration of some functions. Innovation in EoL REPM recycling is mainly driven by policies and positive expectations, while entrepreneurial activities also contribute. The EoL REPM recycling TIS holds potential for further growth, if sufficient supplies of recyclable material are secured and a demand for recycled magnets is created. These goals can be achieved by developing the capacity to handle a diversity of waste products, by making recycling cost-effective, or by finding other marketing approaches for recycled magnets. This would enable the emergence of an independent market. Together with other circular economy solutions, EoL REPM recycling can contribute to a more sustainable and resilient magnet supply.

Life cycle assessment (LCA) databases and software evolve. We analyzed to which extent software and evolving life cycle inventory databases affect the comparison of technology alternatives, using a comparative LCA on permanent magnets as a case study, with two selected software tools: CMLCA and Brightway LCA. We migrated the system models from the CMLCA to Brightway LCA software and alternated between the ecoinvent database versions 2.2 and 3.1 to 3.6 in the system background. When using ecoinvent v3.6 instead of v2.2, the change of the indicator results ranged from (Formula presented.) to 283%. The evolution of the ecoinvent database impacted the absolute amounts of the characterized results and the relative performance between alternatives. The impact category with the highest variability was ionizing radiation, which even showed a ranking inversion with ecoinvent v3.4. In contrast, the impact of using CMLCA or Brightway was negligible because the same data and modeling assumptions caused percentage differences below 0.4%. During the semi-automated data migration to Brightway, we identified 23 environmental flows in the CMLCA model that were not paired with their corresponding characterization factors in the published study of reference. This error had led to an underestimation of 63% in the photochemical oxidation indicator of one of the alternatives. This underestimation relates to an interoperability issue regarding the nomenclature of environmental flows in software alternatives and is a matter of data implementation rather than an issue intrinsic to the selected software. Finally, we identified improvement opportunities for the transparency and reusability of LCA models. This article met the requirements for a Gold-Gold JIE data openness badge described at http://jie.click/badges.