Recovery of LiCoO2 and graphite from spent lithium-ion batteries by molten-salt electrolysis

Summary The recovery of spent lithium-ion batteries has not only economic value but also ecological benefits. In this paper, molten-salt electrolysis was employed to recover spent LiCoO2 batteries, in which NaCl-Na2CO3 melts were used as the electrolyte, the graphite rod and the mixtures of the spent LiCoO2 cathode and anode were used as the anode and cathode, respectively. During the electrolysis, the LiCoO2 was electrochemically reduced to Co, and Li+ and O2− entered into the molten salt. The O2− was discharged at the anode to generate CO2 and formed Li2CO3. After electrolysis, the cathodic products were separated by magnetic separation to obtain Co and graphite, and Li2CO3 was recovered by water leaching. The recovery efficiencies of Li, Co, and graphite reached 99.3%, 98.1%, and 83.6%, respectively. Overall, this paper provides a simple and efficient electrochemical method for the simultaneous recovery of the cathode and the anode of spent LiCoO2 batteries.


INTRODUCTION
6][7] The global battery markets are expected to exceed $100 billion by 2025. 8With the rapid growth of power battery production, the demand for the Co and Li will also continue to rise, and the supply and demand will be more tense.If properly recycled, spent LIBs will have huge economic value and become a veritable ''urban mine''.
The LIBs' recovery methods mainly include pyrometallurgy, hydrometallurgy, electrometallurgy, and biometallurgy.0][11][12][13] This method has the advantages of high chemical reaction rates, large processing capacities, relatively flexible raw materials, and simple operation.However, it would cause environmental pollution and consume a huge amount of energy. 14Hydrometallurgy includes pretreatment, leaching, separation, and recovery of precious metals from the leachate.The purpose of leaching is to convert the metals of the cathode material into ions in the solution.The leaching process is usually carried out using inorganic acids, organic acids, and bases as leaching media.0][31][32] S.Y.Zhou et al. used circular a spent LiCoO 2 electrode plate as the cathode, a platinum plate as the anode, and DL-malic acid as the electrolyte to recover spent LiCoO 2 .The leaching efficiencies of 97.25% for Li and 96.21% for Co could be obtained by electrolytic leaching. 32During the electrochemical recovery process in molten salts, electrons can be manipulated to drive the reduction, and thus the reduction of different substances can be controlled by adjusting the applied electrode potential to improve the efficiency of the reaction. 33,34At the same time, molten-salt systems have high ionic conductivities and wide electrochemical windows, making them excellent electrolytes for extraction and refining metals/alloys, energy storage, and materials synthesis. 35Zhang et al. used molten-salt electrolysis to reduce LiCoO 2 to Li 2 O and Co/CoO in molten Na 2 CO 3 -K 2 CO 3 at 750 C and the recovery efficiencies of Li and Co were 85% and 99%, respectively. 337][38] The existing recycling methods of LiCoO 2 batteries using molten-salt electrochemistry do not consider the simultaneous recovery of the cathode and anode materials.In molten-salt electrochemical recovery of LiCoO 2 , the chemical bonds of LiCoO 2 are broken electrochemically to separate Li and Co. Due to the insolubility of Co in the molten salt, the resulting solid product can be separated from the molten salt.In this way, the impurities in the anode material can be further removed, so the anode material can be recycled.
In this paper, a molten-salt electrochemical method was used to recover the cathode and anode of spent LiCoO 2 batteries in NaCl-Na 2 CO 3 molten salt.Correspondingly, the effects of various electrolysis parameters, such as voltages, time, intensity of pelletizing pressure, and other parameters on the products were investigated.The mechanism of recovery was analyzed and finally, the recovered graphite was reused.Using molten-salt electrolysis to co-recover the cathode and anode materials of spent LiCoO 2 batteries can shorten the battery recycling process and save the recycling cost.The method also realizes the recycling of spent LiCoO 2 batteries without using strong acids and bases.

Thermodynamic analysis and electrochemical measurement
The standard equilibrium potential (DE) of an electrochemical reaction was obtained from the Gibbs free energy change (DG) of its reaction, as shown in Equation 1.
where n is the number of molars of electrons transferred by the reaction and F is the Faraday constant (96485 C/mol).Thermodynamically, the inorganic salts that constitute the molten-salt electrolyte must be more stable than LiCoO 2 .This phenomenon assures the LiCoO 2 reduction if the decomposition voltage of LiCoO 2 was lower than that of the molten salts (NaCl and Na 2 CO 3 ).From Figure 1A, it could be seen that the reductive potentials of both LiCoO 2 and cobalt oxides were more positive than that of NaCl and Na 2 CO 3 .Therefore, LiCoO 2 can be converted to Co by electrolysis, whose process contains the preferentially production of cobalt oxide, and then the Co metal.Since the voltage difference of Co and Na generation is about 0.6 V, the generation of side reactions could be suppressed by controlling the voltages.
As shown in Figure 1B, a pair of redox peaks (c0/a0) correspond to the precipitation and dissolution of alkali metal ions on the Mo electrode in the molten NaCl-Na 2 CO 3 , respectively.When the mixture of LiCoO 2 and graphite powders were loaded onto the Mo rod as the working electrode, two reductive peaks c1 and c2 appear at À0.9 V (vs.Ag/Ag + ) and À1.3 V (vs.Ag/Ag + ) corresponding to the reduction of LiCoO 2 and cobalt oxide (Figure 1C), respectively.The peaks of c1 and c2 were the proofs of the stepwise reductive process of LiCoO 2 , i.e., LiCoO 2 was reduced to cobalt oxide first, followed by the reduction of cobalt oxide to Co. Therefore, the type of electrolysis products could be controlled by changing the voltages, i.e., the target product Co could be obtained at a specific potential.

Electrolysis
From Figure 2A, when intensity of pelletizing pressure of the cathode material was 10 MPa and the electrolysis time was 5 h, the recovery efficiencies of Li and Co were the highest at 99.0% and 98.8% by constant-voltage electrolysis at 1.5 V.As the voltage increased, the electron transfer rate increased with a limited rate.When the electrolysis voltage was increased from 0.9 to 1.5 V, the recovery efficiency of Li did not increase significantly due to the limited reaction rate.In other words, the mass transfer rate in the molten salt could not keep up with the electron transfer rate provided by the voltage.However, the recovery efficiency of Li decreased significantly when the voltage reached 1.8 V.This may be attributed to the fact that the higher voltage could cause side reactions that disturb the target reaction.According to the product information, the impurities contained Mg and Al.At 680 C and an electrolysis voltage of 1.62 V, the electrochemical reaction is as follows: 2MgCl = 2Mg + Cl 2 (g).At 680 C and an electrolysis voltage of 1.78 V, the electrochemical reaction is as follows: 2AlCl 3 = 2Al + 3Cl 2 (g).
The highest recovery efficiencies of Li and Co of 99.3% and 98.1% were achieved when intensity of pelletizing pressure on the cathode raw material was 5 MPa (seeing in Figure 2B).While intensity of pelletizing pressure at 20 MPa resulted in the lowest recovery efficiencies of Li and graphite at 93.1% and 72.9%.This distinction was ascribed to the penetration extent of molten salt into the cathode sheet, which largely affected the mass transfer efficiency.Generally, the electrolysis reaction on the cathode required sufficient contact between the cathode and the molten salt to ensure the transfer of electrons.The excessive intensity of pelletizing pressure not only affected the electrolysis efficiency but also delayed the Li + entrance to the molten salt.The latter affected the mass transfer efficiency and further reduced the efficiency of the electrolysis reaction.
When the intensity of pelletizing pressure of the cathode material was 5 MPa at a voltage of 1.5 V, the recovery efficiencies of Li and Co gradually increased with the increase of electrolysis time until it stabilized (Figure 2C).When electrolysis occurred in the beginning, the reactants were very sufficient and the reactants on the surface layer of the cathode were more easily in contact with the molten salt.It was beneficial for mass transfer, so the recovery efficiency of Li increased rapidly.With the reaction proceeding, the reactants gradually decreased and the concentration of Li in the molten salt gradually increased, and it was more difficult for the cathode to contact the molten salt internally.This phenomenon caused the sluggish reaction rate corresponding to the flat of Li recovery efficiency.Therefore, 5 h was the optimal constant electrolysis time, in which the recovery efficiencies of Li, Co, and graphite were 99.0%, 98.8%, and 80.6%, respectively.
The molar ratio of the cathode and anode materials obtained after the disassembly of the spent LiCoO 2 batteries is 1:2.87.From Figure 2D, it could be seen the lowest recovery efficiencies of Li and Co (48.5% and 34.0%) were obtained after electrolysis at a molar ratio of LiCoO 2 to graphite of 1:0.At 680 C, the recovery efficiencies of Li obtained after electrolysis were 99.34% and 68.47% when the molar ratio of LiCoO 2 to graphite were 1:2.87 and 1:1, respectively.The cathode mixture materials made of LiCoO 2 and graphite reacted a partial carbothermal reduction reaction at 680 C, which was more favorable to the reduction of LiCoO 2 under the condition of sufficient C. Compared with only LiCoO 2 powder, the mixtures of graphite and LiCoO 2 as the cathode could not only realize the simultaneous recovery of spent LiCoO 2 cathode and anode materials but also improve the conductivity of the cathode in favor of faster electrolysis reaction efficiency.As the carbothermal reduction reaction of LiCoO 2 generated CO 2 gas, a fine pore structure was formed inside the cathode material, which was conducive to the full contact between the molten salt and LiCoO 2 on the cathode.This could improve the electrolysis reaction efficiency.In this experiment, the recovery efficiencies of Li and Co reached 99.3% and 98.1%, respectively, and the recovery efficiency of graphite was 80.6% when the mixed spent electrode materials with a molar ratio of 1:2.87 was used as the electrolytic cathode.The recovery of spent cathode and anode electrodes together not only eliminated the step of separating the black powder again when the battery was disassembled in the industry but also realized the recovery of the spent cathode and anode electrodes.
The magnetically separated products obtained by electrolysis at 0.9,1.2,1.5 V (5, 10, and 20 MPa), and 1.8 V for 1-6 h were Co, and no cobalt oxides were present (seeing in Figures 2E-2H).There was residual cobalt oxide in the products from magnetic separation at 1.5 V for 0.1 h with the intensity of pelletizing pressure 5 MPa (Figure 2G).Cobalt oxide exhibits para-magnetism and anti-magnetism at different temperatures, while Co has strong ferromagnetism.It is easily separated Co from graphite by magnetic separation.The product of electrolysis under suitable conditions was Co which facilitated the magnetic separation of the metal from the graphite.
The remaining product of magnetic separation after electrolysis at 1.5 V-0.1 h is cobalt oxide and graphite (Figure 3A).The remaining electrolytic products after magnetic separation at 0.9, 1.2, 1.5 (5, 10, and 20 MPa), and 1.8V for 5 h and 1.5V for 1 $ 5h were pure graphite (Figures 3A-3D).In summary, under suitable electrolytic conditions, pure recycled graphite material could be obtained by magnetic separation.As could be seen from Figures 3E-3H, the sizes of the recycled graphite particles did not differ from those of the spent anode.0][41][42] As the lithium embedding and de-lithiation process continues, the graphite layer spacing will continue to expand until it reaches a certain limit, so the capacity will gradually increase until it stabilizes.As the battery is charged and discharged over time, the graphite negative electrode expands to such an extent that its structure is permanently damaged and cannot be recovered, causing the lattice structure to collapse.The lattice collapse leads to structural rupture and the nanoparticles detached from the graphite negative electrode become attached to the surface of the micron-sized graphite particles together with the binder, electrolyte, and conductive agent.Impurities are no longer observed on the regenerated graphite surface, and the surface tends to be smooth with a perfect graphite morphology.These results can be attributed to the release of internal stresses in the spent graphite during the recycling process and the lattice tends to be fully restored.The recycled graphite became pure and maintained the laminar structure, which was suitable for the LIBs' anode.

Mechanism analysis
Observing Figure 4A, the current plateau was much smaller at 0.9 V compared to 1.8 V, which may be related to the slower reduction ratio at lower voltages.The product after electrolysis, leaching, and filtrating of the molten salt was Li 2 CO 3 .At constant voltage electrolysis, the first current plateau of each voltage was caused by the reduction of LiCoO 2 to cobalt oxide.The processes remained for 39 min, 18.5 min, and 9 min at 1.2V, 1.5V, and 1.8V, respectively.Then, the reduction entered another rate-limiting step.The second current plateau of each voltage was generated by the reduction of cobalt oxide to Co.The subsequent slow decrease of current was due to the fact that fewer reactants are in contact with the molten salt, and the determinant of the reduction reaction rate gradually changed from the conduction of electrons to ion diffusion.The electrolysis curve shows a slow decrease in current until it finally stabilized at a certain current value.As shown in the reaction mechanism diagram in Figures 4B and 4C, during the electrolysis experiment, LiCoO 2 on the cathode got electrons to be reduced to cobalt oxide or Co. Correspondingly, the resulting O 2À entered the molten salt and then generate CO 2 by losing electrons on the graphite anode.

Electrochemical performance of recovered graphite in LIBs
The recovered graphite could be reused as LIBs' anode, and the electrochemical performances were evaluated.Figure 5A shows the CV curves of the recovered graphite in a graphite||Li half-cell.The peak R1 at around 1.3 V corresponded to the formation of the solid electrolyte interface (SEI) and disappeared in subsequent cycles, which means that SEI had become stable.The peaks of R2 and R3 were related to the lithiation process of recovered graphite.The reduction peak R3 appeared near 0.01 V in the first cycle and moved toward 0.2 V (peak R2) in the second and third cycles, indicating the embedding of Li + in the graphite layer in the second and third cycles.The oxidation peak O1 was observed near 0.38 V in the first cycle, corresponding to the de-embedding of Li + in the graphite layer, and moved to 0.3 V in the next two cycles, which can be attributed to the delayed penetration of the electrolyte into the electrode sheet.The peak current gradually increased from the first to the third cycle, indicating that the electrode was gradually activated and the reaction rate accelerated with charge/discharge cycles.
Figure 5B shows the cycling performance of the recovered graphite.The charge-discharge curve of recycled graphite at 1 C (assuming 1 C = 372 mA h À1 ) showed its first discharge-charge specific capacity of 266.1 mA h g À1 .The first cycle resulted in an irreversible capacity of 58.0 mA h g À1 due to the formation of the SEI, so the second cycle showed a rapid decrease in capacity compared to the first cycle.For the first three cycles, the battery was activated by charging and discharging at 0.1C.And from the fourth cycle onwards, the battery was tested at 1C, so it appears that the battery capacity attenuated rapidly in the first few cycles.As the lithium embedding and de-lithiation process continues, the graphite layer spacing will continue to expand until it reaches a certain limit, so then the capacity will gradually increase until it stabilizes.As shown in Figure 5C, the recovered graphite exhibited excellent rate performance.The discharge-charge capacities at 0.1, 0.2, 0.5, 1, and 2C were 321.2, 321, 294.7, 243.3, and 140.4 mA h g À1 , respectively.The discharge capacity remained at about 316.2 mA h g À1 at 0.5 C after the rate capability test, indicating that the revered graphite had good stability after high-rate cycling.Figure 5D shows the charge/discharge curves (0.5 $ 5C) of recovered graphite, which shows that recovered graphite exhibited good reversibility in the graphite||Li half-cell.From the above results, the recovered graphite showed excellent battery performance for LIBs.

Conclusions
The experimental apparatus and flow chart for this work were shown in Figure 6.Molten-salt electrolysis has been demonstrated as an efficient way to co-recover the cathode and anode of spent LiCoO 2 batteries in molten NaCl-Na 2 CO 3 .At the optimal electrolysis conditions (1.5 V for 5 h and at 680 C), the recovery efficiencies of Li, Co, and graphite reached 99.3%, 98.1%, and 83.6%, respectively.The use of electrons instead  of chemical reagents to destroy the crystal structure of LiCoO 2 achieved the separation of Li, Co, and graphite.This electrochemical process greatly reduced the secondary waste.In addition, the relatively low operational temperature avoided the direct carbothermic reduction and co-recycled the Co and graphite.The recovered graphite showed similar electrochemical performance as the original commercial graphite.Further, the molten-salt electrochemical method can be used for the recovery of other types of spent LIBs with the aim to reduce secondary waste by using renewable electricity as the clean agent and driving force.

Limitations of the study
There are some limitations to this study.We collected the cathode material of spent LiCoO 2 batteries and performed X-ray diffraction spectroscopy (XRD) inspection and XRD Riedveld refinement.As shown in Figure 7 and Table 1, the wR of the refined result was 2.232%, indicating that the spent cathode powder primarily consists of LiCoO 2 . 43The element content of cathode of the spent LiCoO 2 batteries was tested by inductively coupled plasma optical emission spectrometer (ICP-OES) and X-ray fluorescence spectrometer (XRF).The test results are shown in Table 2.The spent cathode powder consists predominantly of LiCoO 2 , representing 98.8813wt %.The primary impurity within the spent cathode powder is Al, which originates from a small quantity of aluminum foil scraped along with the spent cathode powder.In this study we have only discussed the LiCoO 2 recovery process and have not discussed the effect of other impurities on the reaction process.
There is a limitation that the experiment was not amplified.As only 0.26 g of graphite was used in each experiment, a very small amount of weight can cause a large change in recovery.In our experiments, we separated the graphite and placed it in a beaker to dry, then scraped it out of the beaker and weighed it again.If the inner wall of the beaker is no longer smooth, some graphite will be lost, introducing an error due to handling problems, and if the beaker is very smooth, this error is greatly reduced.As speculated, the total graphite recovery in the manuscript is slightly lower than the actual recovery.So, there is some error in the graphite recovery ratio in Figure 2, but the data still have some general reference value.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
We greatly thank the financial support from the Fundamental Research Funds for the Central Universities (N2025034, N2025035), Xingliao Project (XLYC 1807042), and the 111 Project (B16009).

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS
There are no experimental model and study participant details to be reported.

Materials and pretreatment
Sodium carbonate (Na 2 CO 3 , AR, 99%) and sodium chloride (NaCl, AR, 99.8%) were purchased from Tianjin Damao Chemical Reagent Factory.The spent LiCoO 2 batteries were obtained from the local electronic market in Shenyang, Liaoning Province, China.HNO 3 (65% $ 68%) was purchased from Sinopharm Chemical Reagent Co., Ltd.Argon (Ar) gas, 99.99%, was purchased from Shuntai, Co., Ltd.The spent LiCoO 2 batteries were placed in saturated NaCl solution for 24 h to be thoroughly discharged and then dried under vacuum at 60 C for 12 h.The dried batteries were disassembled in a fuming hood, and the obtained cathode electrodes were pyrolyzed at 450 C for 1.5 h to remove the electrolyte and organic binder.Then the obtained cathode was roasted in the air at 800 C for 2 h to remove the acetylene black. 44,45The degraded LiCoO 2 powder was obtained by scraping off the Al foil and removing the residual Al foil fragments using an 800-mesh sieve.We collected the obtained the cathode material of spent LiCoO 2 batteries and performed X-ray diffraction spectroscopy (XRD) inspection and XRD Rietveld refinement.As shown in Figure 7 and Table 1, the wR of the refined result was 2.232%, indicating that the spent cathode powder primarily consists of LiCoO 2 . 43he disassembled anode electrodes were placed in deionized water and soaked for 12 h.After the anode electrode material was separated from the copper foil, the collected anode material was washed and filtered several times using deionized water, and then placed in a vacuum environment and dried under vacuum at 60 C for 12 h.Cyclic voltammetry of LiCoO 2 powder in molten NaCl-Na 2 CO 3 salt Firstly, a mixture of NaCl-Na 2 CO 3 powder (NaCl/Na 2 CO 3 molar ratio of 0.577:0.423)was dried under vacuum at 200 C for 12 h to remove moisture.The Al 2 O 3 crucible (F=100 mm, H=100 mm) was filled with anhydrous NaCl-Na 2 CO

Figure 1 .
Figure 1.Thermodynamic data and cyclic voltammograms data with a scan rate of 100mV s -1 at 680℃ in molten NaCl-Na 2 CO 3 Potential versus temperature curves in molten NaCl-Na 2 CO 3 (all thermodynamic data were obtained from HSC Chemistry 6.0) (A), cyclic voltammograms of Moblank electrode (B), and Mo-LiCoO 2 electrode (C) with a scan rate of 100 mV s À1 at 680 C in molten NaCl-Na 2 CO 3 .

Figure 4 .
Figure 4. I-t curves, experimental reaction mechanism diagram, and structure chart of raw materials and products I-t curves and XRD image of the product after leaching and filtration of the molten-salt after electrolysis (molar ratio of LiCoO 2 :C = 1:2.87)(A), experimental reaction mechanism diagram (B), and structure chart of raw materials and products (C) at 680 C in molten NaCl-Na 2 CO 3 .

Figure 5 .
Figure 5. Cyclic voltammetry curves of graphite cell, cycling stability, rate performance, and potential-capacity plots at different rates Cyclic voltammetry curves of graphite cells at a sweep rate of 0.1 mV s À1 (A), cycling stability for 300 cycles at 1 C (B), rate performance (C), and potential-capacity plots at different rates (D) of the regenerated graphite button.

Figure 6 .
Figure 6.Experimental apparatus and flow chart Experimental apparatus (A) and flow chart (B) for the recovery of Co, Li, and graphite from spent LiCoO 2 batteries.

Figure 7 .
Figure 7. XRD Rietveld refinement of the cathode material of spent LiCoO 2 batteries

TABLE
Cyclic voltammetry of LiCoO 2 powder in molten NaCl-Na 2 CO 3 salt B Constant voltage electrolysis of mixed LiCoO 2 and graphite d RESOURCE AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS d METHOD DETAILS B Materials and pretreatment B B Recovery of Co, Li, and graphite B Electrochemical performance of recovered graphite in LIBs B Material characterization d ADDITIONAL RESOURCES

Table 2 .
Composition of spent cathode powder

Table 1 .
Cell parameters of sample for the cathode material of spent LiCoO 2 batteries 3 mixtures as the electrolyte, and the Al 2 O 3