4.1. Recycling Processes
In this section, a critical discussion on the suitability of the recycling processes and their viability to support a CE model is presented. Table 2
presents a summary of processing stages, final products and their target use, for all technologies discussed in Section 3
As seen in Table 2
, only two of the state-of-the-art recycling processes are specialized for treating LIB, i.e., Sumitomo–Sony and Akkuser. Of the emerging processes, only the OnTo Process is capable of treating primary and secondary LIBs simultaneously. The rest of the processes are specialized for one or the other. Nevertheless, while Umicore and Recupyl processes were not initially designed for the processing of LIB, the increasing presence of these types of battery in waste streams pushed for a redesign to accept LIBs as part of their feed. At the same time, the Umicore and Sumitomo–Sony processes claim that their products can be mixed with virgin materials and used as raw materials for batteries. In this manner, the requirement for pristine material is reduced without sacrificing quality requirements.
In the new process developed by Retriev Technologies (Figure 3
), the material is ground to a smaller particle size (105 μm vs. 707 μm) compared to the original process (Figure 2
). This change in particle size improves the separation of anode and cathode by froth flotation while removing carbon and binder, the presence of which deteriorates the performance of remanufactured cathode [104
]. Nevertheless, employing a water-intensive technology such as froth flotation involves environmental risks, as it is well known that some hazardous battery components are water-soluble [51
]. On the other hand, the wet-crushing system of the original Retriev Process, Figure 2
, is at a disadvantage for carbon–binder separation [106
]. This drawback is tackled by thermal treatment at 500 °C, where the binder is decomposed.
The Recupyl process is capable of recovering Co-containing cathode powder and LiFePO4
whenever it is present in the feed. In addition, processing of the electrolyte LiPF6
is possible, recovering PF6
and an ammonium salt during a hydrolysis phase (Figure 4
The high levels of recycling efficiency of the Akkuser process (i.e., >90%) and its low energy consumption (0.3 kWh/kgmaterial
) set this process in a privileged position compared to the rest. It is, however, only possible to reach this cost/efficient value because this process is based only on mechanical processing steps, and aims to obtain a black mass for cathode precursor manufacturing by a third party [107
]. Thus, the stages for refinement are not directly related to Akkuser, and have been left out of this process [75
]. Moreover, the Akkuser process reports losses of plastics only during processing. The process was included in this review because it is a real industrial-scale operation [75
There has been a general shift regarding the final aim for materials recovery from the SoA to emerging technologies. While the former focus on metals recovery (or alloys), the latter also aim to recover cathode or cathode precursor material. In addition, the goal of diminishing material losses is more explicitly pursued by the emerging processes. The extensive processing in emerging processes is directly related to the quality of the final recovery. In processes dominated by pyrometallurgy, the only expected product is a metallic alloy. More complex processes entailing hydro- and pyrometallurgy and mechanical processing are able obtain a wider variety of materials. Additionally, processes using pyrometallurgical operations report larger losses, e.g., Umicore Valéas™ and Sumitomo–Sony processes do not recover electrolyte, plastics, organic material, metals, and graphite. In the case of the Umicore Valéas™, the slag is not listed as a loss, but as a by-product, because this process explicitly pursues an afterlife for the slag phase as additive for the construction industry [34
]. However, using a material in applications with low economic value can be considered a downgrading operation. Nevertheless, as seen in Table 2
, methods presenting an initial pyrometallurgical step do not require a discharging step, as is the case for mechanical and hydrometallurgical processes. Discharge in a controlled environment is bound to the mechanical and hydrometallurgical stages, due to the risk of ignition [51
]. For example, the Akkuser process requires environmental control carried out by a cyclone air system. In the Retriev process, the concentration of O2
in the environment needs to be adjusted, as it is highly reactive with the pure Li found in primary LIBs.
Other significant differences between SoA technologies and emerging processes is that the main products of the former are not specifically recovered for LIB manufacturing. In comparison, the emerging technologies aim to recover cathode and/or anode precursor materials to satisfy the specific needs of the LIB industry.
Consequently, the emerging technologies require a larger number of mechanical processing stages, in addition to a combination of hydro- and pyrometallurgical stages. Evidently, the higher complexity of the emerging technologies raises questions of their viability compared to simpler process recovering a limited variety of products. For instance, the Battery Resources process requires the consumption of various chemical reagents (e.g., MnSO4
, and CoSO4
) but obtains the most suitable product for use as a cathode material, at least out of the processes analyzed herein. This brings up the need for balance between the total number of stages, the overall process complexity, and the quality of the products [45
]. It is well known that a more complex process, of any nature, is associated with a higher possibility of failure, and requires larger quantities of energy and chemical reagent input while not fully eliminating the generation of waste. For instance, the Aalto University process presents a high quality of products, but also demands a large number of reagents in the hydrometallurgical stages and high energy in the pyrometallurgical step, in addition to efficient mechanical pre-processing stages.
Comparing the different operations present in the processes, there is a clear advantage to the use of mechanical processing coupled hydrometallurgical operations from a CE perspective, i.e., an increase in the variety and usability of the recoveries [73
]. This can especially be seen in the products of the Toxco and Recupyl processes, where the losses are diminished considerably in comparison to Umicore Valeas™ or Sumitomo–Sony. Even if the main recoveries of the Umicore Valeas™, Toxco, and Recupyl processes are intended to be used for cathode production, only the latter two recover Li2
in the same operation, which are considered cathode precursors. If a cathode precursor is defined as a material that do not require further processing to be used in cathode manufacture, the CoCl2
obtained in the Umicore Valeas™ process cannot considered a cathode precursor, as it requires further chemical treatment for cathode material synthesis [70
]. Thus, it is possible to say that the Toxco and Recupyl are more in line with the idea of circular economy, in comparison to the rest of the processes.
As seen in Table 2
, how close the recovered Li components are from being re-used as cathode material is directly related to the complexity of the process. For example, the Recupyl process recovers a cathode precursor, while Sumitomo–Sony produces CoO. Evidently, there are advantages and drawbacks for recycling processes based on either pyro- or hydrometallurgy. In general, pyrometallurgical processes are more energy intensive and lead to larger material losses, but the metallic components are easily upgraded to a commercially valuable state. On the other hand, processes based on hydrometallurgy can recover materials with chemical characteristics suitable for LIB manufacturing, thus “closing the loop” in a more efficient manner. However, hydrometallurgical processes are bound to mechanical processing operations in order to liberate the valuable elements confined in the LIBs. In addition, hydrometallurgical operations require vast amounts of reagents and strictly controlled chemical environments to perform dilution or precipitation of the target materials. In summary, the technologies presented in Table 2
suggest that processes with a higher degree of complexity are necessary to close the materials cycle for all components of LIBs.
From a CE perspective, the preferred processes should entail multiple layers of closed material loops using the fewest possible external inputs to obtain, for instance, cathode precursors. As described in previous sections, both emerging and state-of-the-art processes achieve to some extent the aims of CE for LIBs, bringing valuable materials back into the economic cycle. However, the recovery of materials is only one aspect of the CE. In a closed loop process, resource inputs should also be minimized. Emerging processes require the consumption of additional reagents to achieve the target recoveries. The internal loops in the Battery Resources and LithoRec processes support reduced consumption of external resources for the pyrometallurgical refining process for cathode precursors, e.g., Li2
. In addition, the OnTo process reports the possibility of recovering battery components suitable for LIB manufacturing [100
Even though emerging processes appear to be closer to the ideal of circular economy, they are not without their drawbacks. For instance, they are mostly focused on one type of battery chemistry, reducing the flexibility of the feed compared to SoA technologies. Some of these technologies require chemical reagents considered hazardous, e.g., the bromoform used in the OnTo process to separate anode and cathode during dense media separation has been associated with kidney and liver injuries [109
]. Evidently, there is still room for improvement. For example, the Accurec and LithoRec processes do not recover electrolyte, while the available information on the Battery Resources process does not mention electrolyte recovery whatsoever. The OnTo process is unique in this sense as it allegedly recovers the electrolyte components, i.e., solvent and Li salt [51
]. It has been expressed that the OnTo process is able to recover about 80% of the LIB components, making it an attractive possibility for development at the industrial scale [99
]. Interestingly, during a test conducted by Rothermel et al. [111
] graphite recovered from the OnTo process outperformed new synthetic graphite in discharge capacity, although it presented a lower discharge capacity compared with non-synthetic graphite. This drawback was attributed to the environment used during the extraction of electrolyte with supercritical CO2
, as it may have damaged the crystalline structure of the graphite [111
]. As the emerging processes are still in early stages of development, additional risks may become evident during the scale-up phase [27