ADVANCED MATERIALS & PROCESSES | JULY/AUGUST 2023 22 toxic waste, and the risk of making a product that is not competitive with current state-of-the-art cathode material. In particular, the separation and binder removal processes may fail to eliminate contaminants such as fluorine, carbon, or metals from current collectors. These impurities can then be carried through the rest of the direct recycling process and have negative effects on the performance of end materials. Metallic contaminants are particularly problematic in direct recycling: Even trace quantities (~1%) of impurity metals such as Cu can substantially disrupt battery performance[17]. An example of black mass before and after purification is shown in Fig. 4. Methods to remove metallic contamination are already used in hydrometallurgical recycling. In this case, all metals are chemically leached to ionic form, and the undesired impurities can then be removed through solvent extraction or precipitation. However, most hydrometallurgical leaching conditions have been designed to also break down the cathode metals, which prevents subsequent direct recycling. New approaches that selectively remove metallic contaminants from black mass while keeping cathode material intact are gaining traction. These strategies include high-precision optical sorting to remove larger metallic impurities and tailored alkaline chemical treatment to remove metallic contaminants from finely shredded material where optical sorting is unfeasible[18]. Alternatively, solvent- based delamination of electrode films is being evaluated as an upstream means Fig. 5 – (a) Comparison of cost and GHG emissions for direct recycling vs. hydrometallurgical recycling; (b) cost; and (c) GHG emission breakdown for a redox mediator-based direct recycling process. (a) (b) (c) of avoiding shredding altogether[19]. This could substantially reduce the amount of contamination introduced into the direct recycling process. There is a pressing need to use physics-based electro- chemical models and semi-empirical/machine-learning-based life models to help accelerate development of direct recycling. These models have been key to many recent advancements in Li-ion battery development, including better energy density, improved rate capability, increased lifetime, and reduced cost. These models can help answer some of the questions and concerns that remain with direct recycling. For instance, extensively validating the lifetime of cells that use directly recycled materials is a monumental task. Life models can be employed to reduce the length and number of use-cases that must be tested to validate recycled material lifetimes. Additionally, there is an urgent need for research-backed standards defining the properties recycled materials must have to be effective and safe in batteries. Often, the assumption is that recycled materials must be the same as pristine materials. This may or may not be realistic for direct recycling. Achieving complete separation of active materials, removing contaminants to match virgin material purity, and completely restoring all physiochemical properties may not be feasible—or may not make economic sense. Electrochemical models and process-cost models should be used to rationally determine requirements that balance process cost with performance gains and losses. For example, if allowing up to 1% of active material mixing decreases separation cost by 50% and leads to no measurable change in lifetime and rate capability, then it likely makes sense to modify the present standards. Finally, technoeconomic models and life cycle analysis can be used to scan the large parameter space of requirements for a given recycling process and determine optimal tradeoffs. Fig. 4 – Industrial black mass powders before (right) and after (left) black mass purification treatment.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==