Sustainable upcycling of mixed spent cathodes to a high-voltage polyanionic cathode material

Sustainable battery recycling is essential for achieving resource conservation and alleviating environmental issues. Many open/closed-loop strategies for critical metal recycling or direct recovery aim at a single component, and the reuse of mixed cathode materials is a significant challenge. To address this barrier, here we propose an upcycling strategy for spent LiFePO4 and Mn-rich cathodes by structural design and transition metal replacement, for which uses a green deep eutectic solvent to regenerate a high-voltage polyanionic cathode material. This process ensures the complete recycling of all the elements in mixed cathodes and the deep eutectic solvent can be reused. The regenerated LiFe0.5Mn0.5PO4 has an increased mean voltage (3.68 V versus Li/Li+) and energy density (559 Wh kg–1) compared with a commercial LiFePO4 (3.38 V and 524 Wh kg–1). The proposed upcycling strategy can expand at a gram-grade scale and was also applicable for LiFe0.5Mn0.5PO4 recovery, thus achieving a closed-loop recycling between the mixed spent cathodes and the next generation cathode materials. Techno-economic analysis shows that this strategy has potentially high environmental and economic benefits, while providing a sustainable approach for the value-added utilization of waste battery materials.


Reviewer #2 (Remarks to the Author):
Ji et al. report an upcycling strategy for mixed degraded cathode materials and convert them into a promising polyanionic cathode with an increased working potential.Currently, most battery recycling reports aim at a single component, and tackling the multi-component cathodes is a challenge especially for phosphate cathodes and Mn-rich spinel cathodes.They have relatively lower economic value, thus needing sustainable recycling methods.The authors propose an efficient and interesting method to recover mixed cathodes by using a green and recyclable DES.The authors also provide an estimation based on a technoeconomic analysis, which is helpful for the analysis if such technologies can be applied on large scale.I would recommend the manuscript for publication in Nature Communications after having addressed subsequent issues: 1.The DES (ChCl-OA system) used in this work was green and recyclable, as shown in Figs.1d-f.How is the DES recycled in the process considering oxalates were actually consumed.More discussions are needed in the experimental method part.2. A uniform carbon layer coating can be seen in STEM images (Fig. 2i) and EDS maps (Fig. 2k).Are additional carbon sources added during the recovery process?A brief discussion along these lines is required.3. The choice of Fe and Mn mole ratio seems to be an important factor, which is related to the performance.Please provide a detailed explanation about how to exactly control the ratios in the R-LFMP materials.4. For the TEA analysis in Fig. 4, the upcycle process has a higher cost compared with Pyro, Hydro, Direct processes.What could be the reason?

Reviewer #3 (Remarks to the Author):
This paper explains a new recycling route for the cathode of a lithium ion battery.The novelty is in dissolving 2 types of cathode material (lithium iron phosphate), termed LFP, and lithium manganese oxide (LiMn2O4), termed LMO, to produce lithium iron manganese phosphate, termed LFMP.The use of LFMP as a promising cathode material has been known to the literature for about 1 decade as noted by the authors.The novelty of the paper is the fact that the authors have come up with a processing route that uses a so-called deep eutectic solvent (DES) to manufacture LFMP.The mixture of many chemicals used to react with the 2 cathode materials are given in the methods section without any justification.For example, carbon nanotubes are used without explanation.Response: Thanks for your comments.We have added quantitative measurements to determine the dissolution of S-LFP and S-LMO in the DES.The ICP results (Table R1 and Fig. R1) showed that the leaching rates of Li and P were up to 96.7% and 97.0%, respectively.Lithium and phosphorus remained respectively as Li + and PO4 3-in the DES, which were then recycled as Li3PO4 from the filtrate, as confirmed by the XRD pattern (Supplementary Fig. 3).After the leaching reaction, iron and manganese remained respectively as [FeCl4] 2-and [MnCl4] 2-in the DES, which were precipitated as (Fe, Mn)C2O4•2H2O in the following separation processes.Notably, the DES has greater viscosity than aqueous solutions with majority >100 cP at ambient temperature, making it difficult to directly measure the leaching rates in the DES.Hence, the leaching rates were calculated based the filtrate after removing (Fe, Mn)C2O4•2H2O precipitation.According to the high leaching efficiencies of Li and P (≥ 97%), we inferred that the Fe and Mn can also be leached with the fast reaction kinetics.This is because the structures of LFP and LMO cathodes were broken down and rearranged in the DES system.Therefore, 89.6% of the Fe and 82.2% of the Mn were precipitated as the oxalates after adding water for the separation processes.We attempted to use FTIR spectra to characterize the coordination environment and composition of the dissolved elements in the DES (Fig. R2).Unfortunately, these infrared peaks of these elements were difficult to detect because the DES is an organic solvent, and the peaks of its own organic functional groups were too strong, thus masking the appearance of other peaks.Combined with other characterizations and literature support, we can give a qualitative analysis of the dissolved elements in the DES.precipitation 3,4 .The isolated Li + and PO4 3-were recycled as a lithium salt and the DES was reused for the next loops.

In
We have replaced the original Fig. 1b by the following schematic and added the corresponding description on pages 6-7 in the revised MS, which will be useful for understanding the dissolution and rearrangement of S-LFP and S-LMO cathodes in the DES.Response: Thanks for your suggestions.The control of the Fe/Mn molar ratio is crucial for the regenerated materials in the upcycling processes.As for this important concern, we first re-explained it in the Methods part and the main body part, respectively, in order to show a clear description of this study for the readers.
On pages 9-10 in the revised MS: "Due to the above upcycling strategy, the DES system (ChCl-OA) had high leaching efficiencies (> 97%) for all elements in black mass, and the subsequent precipitation process of Fe/Mn precursors was also demonstrated, enabling that the molar ratios of Fe/Mn can be precisely controlled in the regenerated materials…" On pages 19 and 20 in the revised MS (Methods section): "S-LFP and S-LMO black mass with a specific Fe/Mn molar ratio were added to the DES and the mixture was then heated at 110 ℃ under continuous stirring for 6 h…" "Notably, the Fe/Mn molar ratios in the precursors can be precisely controlled in black mass due to the high leaching efficiencies for all elements, which has been demonstrated in Results part…" "The experimental conditions were the same as for the LFMP materials with different molar ratios of Fe/Mn…" And, we further illustrate the control of the Fe/Mn molar ratio by combining the relevant experimental data, that is, if the Fe/Mn molar ratio is accurate in the black mass, the ratio is ideal in the precursor and LFMP materials.This is because the DES system used for the upcycling strategy has high leaching efficiencies for all the elements in S-LFP and S-LMO cathodes (Fig. R1).In addition, the subsequent precipitation process of Fe/Mn has displayed in Fig. R4.Our strategy also shows a high separation efficiency of all elements.89.6% of the Fe and 82.2% of the Mn were precipitated in the solid solution precursor.While 96.7% of the Li and 97.0% of the P remained in the filtrate, which were then recycled as a lithium salt (Li3PO4).Based on above discussion, the Fe/Mn molar ratio in the upcycling strategy is controllable.Now we believe that the revised MS will provide a clear description on the main results of this study.were still retained at 1, 2, and 5C rates for R-F5M5.In contrast, P-F5M5 only had discharge capacities of 100, 86, and 64 mAh g -1 under the same rates.To further reveal the reasons for the differences in electrochemical performance, we compared and analyzed the dQ/dV curves and in-situ EIS results.R-F5M5 showed sharp redox peaks with mitigated polarization (0.13 V for Fe +2/+3 and 0.15 V for Mn +2/+3 ).
In comparison, P-F5M5 had larger polarizations after 100 cycles (0.13 V for Fe +2/+3 and 0.20 V for Mn +2/+3 ).An extra peak was observed at 3.6 V during discharge, in agreement with the charge and discharge curves.The same phenomenon was seen in the C-LFMP samples (Fig. 3g), ascribed with the structural instability.Response: Thanks for your helpful comments.The potential feasibility for practical application is crucial for evaluating a battery recycling process.Also, cost and environmental issues are also important considerations in developing a sustainable recycling system.We first carried out a scalable experiment in the lab (Ten times of the processes in the MS), as depicted in Fig. R9.Specifically, 56 g choline chloride (ChCl) and 36 g oxalic acid (OA) were mixed in a molar ratio of 1: 1.Then the mixture was heated at 80 ℃ to form a transparent DES, which was used for leaching the black mass.
After which, 1.58 g S-LFP and 0.90 g LMO (Fe: Mn = 1: 1 in a molar ratio) were added into the DES and the mixture was then heated at 110 ℃ under continuous stirring for 6 h.The subsequent experimental procedures were exactly the same as that in the Methods section.As expected, the experimental phenomenon was consistent with the previous processes.Therefore, the upcycling approach is able to expanding at a gramgrade scale.Unfortunately, it is difficult for us to carry out a larger demonstration at the laboratory level, but we believe that this upcycling approach has the potential for practical large-scale applications, and we will continue to expand in this direction in the future.

Fig. R9 | Pictures for a scalable experiment in the lab.
To verify the feasibility of the whole upcycling processes on a large-scale, we further synthesized the R-F5M5-scale sample by using the precursor obtained from a scalable experiment, as shown in Fig. R10.XRD patterns show good phase crystallinity for both (Fe, Mn)C2O4•2H2O precursor and R-F5M5-scale sample.In electrochemical properties tests, it delivered a discharge capacity of 130 mAh g -1 at 1C rate, which was a very satisfying result.In addition, we have expanded the scope of practical applications of the upcycling strategy, that is, a potential recycling approach for LFMP was proposed and verified by experiments.LFMP has the same olivine type crystal structure as LFP and can be dissolved by the DES (ChCl-OA).As shown in Fig. R11, here we used a commercial LFMP (C-LFMP) and R-LFMP as the raw materials, respectively.The same target product was formed, that is, the Fe/Mn oxalate precursor, suggesting that the upcycling approach in this work is also applicative to process LFMP cathodes.Based on above discussion, a sustainable recycling system, involving the upcycling strategy in this work and potential LFMP recycling in the future, is summarized in Fig. R12.LFMP is a promising cathode material with an increased voltage and energy density.Especially in recent years, many battery manufacturers began to layout LFMP batteries.We believe that LFMP will usher in explosive growth in the next few years, so it can be inferred that sustainable recycling of LFMP batteries will also be a hot topic.Finally, we have added an in-depth discussion on the practical applications of this approach, as shown in revised MS and SI.
On page 19 in the revised MS (Discussion section): "Considering that the potential feasibility for practical application is crucial for evaluating a battery recycling process, we carried out a small scale-up experiments demonstration (Supplementary Fig. 31).As expected, the experimental phenomenon was consistent with the previous processes, and the R-LFMP samples synthesized from the scale-up experiments showed satisfying properties (Supplementary Fig. 32), demonstrating that the upcycling approach was able to expanding at a gram-grade scale.Furthermore, we have expanded the scope of practical applications of the upcycling strategy, that is, a potential recycling approach for LFMP was proposed and verified by experiments (Supplementary Fig. 33).A sustainable recycling system, Response: Thanks for your comments.The FTIR results in Fig. 1f showed that the recycled DES had the same infrared peaks compared with the original DES after the first use, suggesting the remained properties of the DES.This is because the liquidsolid ratio of the reaction is roughly 35 and the oxalate is very abundant.Considering that the oxalates were actually consumed after using for several times, the extra oxalate was needed to adjust appropriate proportion of ChCl and OA, based on how many moles were removed in the precipitate.We have added a detailed explanation about the reuse of the DES in the Method section, as shown on page 22 in the revised MS: "Reuse of the DES and recycling of the lithium salt.To regenerate the DES, the filtered liquid was heated to evaporate off the excess water and reused for the next cycle.
After being used several times, the oxalate content can be adjusted through extra addition for reforming the DES, based on how many moles were consumed in the precipitate…" 2. A uniform carbon layer coating can be seen in STEM images (Fig. 2i) and EDS maps (Fig. 2k).Are additional carbon sources added during the recovery process?
A brief discussion along these lines is required.
Response: Thanks for your suggestions.In addition to STEM images and EDS maps results, high-resolution TEM images of R-LFMP showed a uniform carbon layer coating on the particle surface with a thickness of 4-5 nm, as depicted in Fig. R13.It is worthing noting that no extra carbon sources were added during the regeneration process.We attributed it to the following two reasons regarding the carbon layer coating.
First, the S-LFP and S-LMO black mass contained residual conductive carbon (acetylene black or carbon nanotubes) and binder (PVDF), which did not react with the DES and then precipitated along with the oxalate precursor.Subsequently, the conductive carbon and binder were refined and homogenized during the ball milling process.These compounds served as carbon sources in the LFMP synthesis process.
Second, the oxalate precursor itself can also serve as a carbon source, which decomposed into the transition metal oxides and carbon during the heat treatment processes.Therefore, we believe that the above two factors can explain the origin of the carbon coating on the surface of R-LFMP.The utilization of all components in waste batteries is key to realizing sustainable recycling, which has a positive impact on resources conservation and environmental protection.The upcycling strategy of mixed cathodes in this paper is of great significance for approaching this goal.3. The choice of Fe and Mn mole ratio seems to be an important factor, which is related to the performance.Please provide a detailed explanation about how to exactly control the ratios in the R-LFMP materials.
Response: Thanks for your comments.The control of the Fe/Mn molar ratio is crucial for the regenerated materials in the upcycling processes.As for this important concern, we first re-explained it in the Methods part and the main body part, respectively, in order to show a clear description of this study for the readers.
On page 9 in the revised MS: "Due to the above upcycling strategy, the DES system (ChCl-OA) had high leaching efficiencies (> 97%) for all elements in black mass, and the subsequent precipitation process of Fe/Mn precursors was also demonstrated, enabling that the molar ratios of Fe/Mn can be precisely controlled in the regenerated materials…" On pages 19 and 20 in the revised MS (Methods parts): "S-LFP and S-LMO black mass with a specific Fe/Mn molar ratio were added to DES and the mixture was then heated at 110 ℃ under continuous stirring for 6 h…" "Notably, the Fe/Mn molar ratios in the precursors can be precisely controlled in black mass due to the high leaching efficiencies for all elements, which has been demonstrated in Results part…" "The experimental conditions were the same as for the LFMP materials with different molar ratios of Fe/Mn…" We further illustrate the control of the Fe/Mn molar ratio by combining the relevant experimental data, that is, if the Fe/Mn molar ratio is accurate in the black mass, the ratio is ideal in the precursor and LFMP materials.This is because the DES system used for the upcycling strategy has high leaching efficiencies for both S-LFP and S-LMO cathodes, as shown in Fig. R1.In addition, the subsequent precipitation process of Fe/Mn has displayed in Fig. R4.
Our strategy also shows a high separation efficiency of all elements.89.6% of the Fe and 82.2% of the Mn were precipitated in the solid solution precursor.While 96.7% of the Li and 97.0% of the P remained in the filtrate, which were then recycled as a lithium salt (Li3PO4).Based on above discussion, the Fe/Mn molar ratio in the upcycling strategy is controllable.Now we believe that the revised MS will provide a clear description on the main results of this study.

4.
For the TEA analysis in Fig. 4, the upcycle process has a higher cost compared with Pyro, Hydro, Direct processes.What could be the reason?
Response: Thanks for your question.In the cost analysis (Fig. R14), annualized capital cost is a critical factor in all recycling processes.Material cost is relatively lower in the Pyro process because the energy required for smelting is produced by the battery components, such as graphite.In comparison, material cost is a significant contributor to the overall cost in the other processes, such as acid/alkaline reagents in Hydro process and lithium salts in Direct and Upcycle processes.The material cost accounts for 71.1% of the total in the Upcycle process, mainly attributing to the lithium salt used for R-LFMP synthesis.Revenue analysis is directly related to the value of products.In contrast, the regenerated materials (LFMP) are the major contributors to the overall revenue in the Upcycle process.In addition, the recycled lithium salt and Mn salts also contribute to the revenue in the Upcycle process.Therefore, our proposed Upcycle process has the highest profit (profit = revenue -cost), that is 4.43 $ per kg of feedstock.

Figure 1 ,
there should be a quantitative measurement of the dissolution of S-LFP and S-LMO in the deep eutectic solvent (DES).Such quantitative data is essential to support the author's claims and to provide a clearer understanding of the LFMP preparation process, including the composition of elements and the concentration in the leaching solution of the black mass with DES.While the authors have presented indirect evidence using SEM, XRD, and XPS, it would be more comprehensive to include scientific evidence regarding the elements in the leaching solution.

Fig. R1 |
Fig. R1 | The separation rate of Li/P/Fe/Mn based on ICP results.

Fig. R2 |
Fig. R2 | FTIR spectra of pristine DES and the black mass dissolved in the DES at different times.

Fig. R4 | d
Fig. R4 | d Demonstration of the formation of DES. e Demonstration of the formation of the precursor.f FT-IR spectra of original and recycled DES.g The separation efficiencies of Li/Fe/Mn/P based on ICP results.

3 .
To ascertain the superiority of the prepared samples, it is crucial to analyze the electrochemical performance of these samples in comparison to LFMP synthesized with highly purified and well-defined raw materials.I recommend including results from LFMP prepared with such refined raw materials and expanding upon the discussion in this context.Response: Thanks for your comments.We have synthesized the comparative samples with highly purified raw materials (Chemicals involve FeC2O4, MnC2O4•2H2O, which are analytical grade and purchased from Macklin), as shown in Fig.R5.All the experimental conditions were the same as for the R-LFMP materials except for the commercial Fe/Mn precursors.Here we take pristine LiFe0.5Mn0.5PO4(P-F5M5) as a comparison sample for discussion.As shown in Fig.R6, P-F5M5 had an initial discharge capacity of 130 mAh g -1 with a low initial Coulombic efficiency (ICE) of 68%.As contrast, the initial discharge capacity of R-F5M5 was 152 mAh g -1 with ICE of 94%, indicating the superior performance and cyclic reversibility.Regarding the rate performance (0.1-0.5-1-2-5-10C), the discharge capacities of 128, 110 and 90 mAh g-1

Fig. R8 |c
Fig. R8 | Impedance spectra collected during the first cycle.a R-F5M5, b P-F5M5.c Rct value change trends.

Fig. R10 |
Fig. R10 | a XRD patterns for the precursor and R-F5M5 synthesized by the scale-up experiments.b Electrochemical performance of R-F5M5-scale sample.

Fig. R11 | a
Fig. R11 | a Picture for C-LFMP and R-LFMP dissolved in the DES.b XRD patterns for the obtained precursors.

Fig. R12 |
Fig. R12 | Schematic for the upcycling strategy in this work and the potential LFMP recycling in the future.

Reviewer # 2 ( 1 .
the upcycling strategy in this work and potential LFMP recycling in the future, is summarized in Supplementary Fig. 34.LFMP, regarded as the structural upgrade product of LFP, has attracted much attention in both academic and industrial fields in recent years.The DES system in this work was also applicable to process LFMP, suggesting that the proposed upcycling strategy can achieve a closed-loop recycling between the mixed spent cathode chemistries and the next-generation batteries."Revision marked with red in the revised MS and SI) Ji et al. report an upcycling strategy for mixed degraded cathode materials and convert them into a promising polyanionic cathode with an increased working potential.Currently, most battery recycling reports aim at a single component, and tackling the multi-component cathodes is a challenge especially for phosphate cathodes and Mn-rich spinel cathodes.They have relatively lower economic value, thus needing sustainable recycling methods.The authors propose an efficient and interesting method to recover mixed cathodes by using a green and recyclable DES.The authors also provide an estimation based on a techno-economic analysis, which is helpful for the analysis if such technologies can be applied on large scale.I would recommend the manuscript for publication in Nature Communications after having addressed subsequent issues: Response: Thanks for your comments on this manuscript, which are useful for us to revise the paper.We have done a full investigation and added some experiments to address your concerns.All the changes were highlighted with red in the revised manuscript (MS) and supplementary information (SI).Below are the point-to-point responses to your comments.The DES (ChCl-OA system) used in this work was green and recyclable, as shown in Figs.1d-f.How is the DES recycled in the process considering oxalates were actually consumed.More discussions are needed in the experimental method part.

Fig. R13 |
Fig. R13 | a TEM image, b-e HRTEM images, f SAED pattern, g Enlarged regions in Ⅰ, h Line profiles in Ⅴ of R-F5M5.

Fig. R1 |
Fig. R1 | The leaching rates of Li/P/Fe/Mn based on ICP results.

Fig. R4 | d
Fig. R4 | d Demonstration of the formation of DES. e Demonstration of the formation of the precursor.f FT-IR spectra of original and recycled DES.g The separation efficiencies of Li/Fe/Mn/P based on ICP results.

Fig. R14 |
Fig. R14 | Techno-economic analysis results of the Pyro, Hydro, Direct, and Upcycle processes.b Cost analysis.c Cost percentages of the Upcycle process.

Table R1 |
The leaching rates of Li/P/Fe/Mn based on ICP results.
Note: 1 mmol S-LFP and 0.5 mmol S-LMO were used to calculate the leaching rate.Three experiments were used to get the average value.The leaching rates were calculated based the filtrate after removing (Fe, Mn)C2O4•2H2O precipitation.