Mechanochemical upcycling of spent LiCoO₂ to new LiNi_0.80Co_0.15Al_0.05O₂ battery: An atom economy strategy

Significance

The rapid expansion of the electric vehicle market generates a large volume of spent lithium-ion batteries (LIBs). The employment of harsh chemicals and lack of atom economy in the recycling method leads to significantly more environmental pollution and CO₂ emissions. Herein, Al foil wastes inherent in spent LIBs are used as the reducing agent and purifying agent in situ to induce the reduction of transition metal oxides and the targeted separation of impurity metals, and ultimately incorporated into the upcycled product at 100% atom efficiency.

Abstract

The use of strong acids and low atom efficiency in conventional hydrometallurgical recycling of spent lithium-ion batteries (LIBs) results in significant secondary wastes and CO₂ emissions. Herein, we utilize the waste metal current collectors in spent LIBs to promote atom economy and reduce chemicals consumption in a conversion process of spent Li_1-xCoO₂ (LCO) → new LiNi_0.80Co_0.15Al_0.05O₂ (NCA) cathode. Mechanochemical activation is employed to achieve moderate valence reduction of transition metal oxides (Co³⁺→Co^2+,3+) and efficient oxidation of current collector fragments (Al⁰→Al³⁺, Cu⁰→Cu^1+,2+), and then due to stored internal energy from ball-milling, the leaching rates of Li, Co, Al, and Cu in the ≤4 mm crushed products uniformly approach 100% with just weak acetic acid. Instead of corrosive precipitation reagents, larger Al fragments (≥4 mm) are used to control the oxidation/reduction potential (ORP) in the aqueous leachate and induce the targeted removal of impurity ions (Cu, Fe). After the upcycling of NCA precursor solution to NCA cathode powders, we demonstrate excellent electrochemical performance of the regenerated NCA cathode and improved environmental impact. Through life cycle assessments, the profit margin of this green upcycling path reaches about 18%, while reducing greenhouse gas emissions by 45%.

While lithium-ion batteries (LIBs) have revolutionalized consumer electronics, transportation, and energy markets, millions of tons of important metal resources are embedded in municipal waste streams after the LIBs fail (1–3). Ill disposal of spent LIBs causes heavy pollution and fire accidents (4). As a typical anthropogenic mineral, the concentration of key metals in spent LIBs is 3 to 10 times that of natural mineral concentrates, and thus, its reuse will improve the supply security of lithium and cobalt resources (5, 6). Furthermore, the recycling of spent LIBs can reduce CO₂ emissions coming from the mining and smelting of primary mineral resources, thereby contributing to the goal of carbon neutrality.

The recycling and upcycling of spent LIBs are under heavy development (7–9). Researchers are no longer just satisfied with a metal recovery rate greater than 98%, but are also beginning to care about how to reduce the consumption of chemicals in the recovery process in order to reduce the environmental footprint while improving the profitability (10–12). Thus, in situ recycling methods have emerged, namely, using physicochemical means to stimulate the redox potential of the inherent components of spent LIBs to enhance the dissolution efficiency of transition metal (13, 14). Spent LIBs are mainly composed of transition metal oxide cathode, graphite anode, current collectors (Cu/Al foils), polymer separators, and steel shells (15). Among them, zero-valent metals such as Fe shells, Al foils, and Cu foils, are natural reducing agents that have been used to develop in situ recycling technology (16, 17). Joulie et al. proposed that the current collectors can act as reducing agents to enhance the leaching of cathode materials of spent ternary LIBs, and thermodynamic analysis was performed (18). Peng et al. also found that a dual reducing agent system composed of Cu foils and Fe²⁺/Fe³⁺ impurities can significantly increase the leaching rate of Li and Co, from 80% to about 100% (19). Fe²⁺ ion has a lower reduction potential and can reduce the Li_1-xCoO₂ (LCO) cathodes, while the oxidized Fe³⁺ iron is reduced by Cu foils, so Fe²⁺ ion in the system can play a cyclic role of soluble redox mediator (20). Although Al⁰ has a stronger reducing ability to directly reduce the cathode materials without the participation of iron ions, it is easy to react with acid to generate H₂, which not only increases the consumption of harsh H₂SO₄, but also poses the risk of fire and explosion. On the other hand, Al⁰ production is energy and CO₂ intensive. If Al foils can be used as a reducing agent to replace hydrogen peroxide, the additional environmental impact of the subsequent regeneration of waste foils and the production process of traditional reducing agent can be reduced by 13.8 ton of CO₂ per ton Al (20, 21). Therefore, the in situ utilization of Al⁰ is very important for carbon emission reduction in the recycling of spent LIBs. It is necessary to develop green innovations that improve the reduction efficiency of Al foils while reducing the risk of hydrogen generation.

The key green chemistry principle of atom economy implies that all raw materials should be converted into the final product as much as possible (22). In previous in situ recycling, raw materials (current collectors) are employed as reducing agents and then separated in the form of hydroxide precipitation, but they do not enter the final products at all, so the final atom efficiency is 0%. Waste will still be generated after reuse, which not only does not reduce the amount of waste, but may actually make subsequent high-value utilization more difficult. For example, a unit mass of Al foils will generate 2.88 times as much Al(OH)₃ waste after reductive leaching. Al(OH)₃ exists in floc form with amorphous gel-like properties, which is difficult to be separated from the liquid. Therefore, based on the atom economy principle, using Al⁰ as a functional reagent to enhance metal recovery and then entering the final electrode product may be a promising path for upcycling spent LIBs with the least economic and environmental costs.

In this work, the waste Al foils are fully exploited to its redox potential to replace multiple chemical reagents, and finally integrated into new LiNi_0.80Co_0.15Al_0.05O₂ (NCA) high-performance cathode materials. Firstly, mechanical activation pretreatment is employed to stimulate an in situ reduction reaction between the Cu/Al foils and spent cathode materials, followed by efficient CH₃COOH leaching without generating H₂ gas. Then, an Al foil-induced self-purification mechanism for targeted cation (Cu/Fe) precipitation in the liquid system has been developed to support the selective removal of Cu/Fe impurities. After all, Al foils enter the leaching solution as Al³⁺ ions, they are used as an aluminum source to prepare NCA precursor solution together with the Ni and Co elements. Based on a detailed life cycle assessment (LCA), we evaluated the new technology in terms of its environmental impact and economic benefits and found the new process to be highly beneficial. The mechanochemical upcycling of spent LCO batteries based on an atom economy strategy can be a model for the technological innovation of recycling other electronic wastes.

Results and Discussion

Mechanochemical Reduction: Lattice Evolution and Element Migration.

The phase analysis of ≤4 mm crushed product of spent LIBs is shown in Fig. 1A. It is a mixed powder composed of LCO, graphite, zero-valent Al and zero-valent Cu. The X-ray diffraction patterns of the mechanically activated products after different grinding times are shown in SI Appendix, Fig. S1. Although the characteristic peaks of graphite nearly disappear, the degree of attenuation of the main LCO electrode peak is small with increasing grinding time, which indicates that long-term low-speed ball milling at 400 rpm will not induce the lattice collapse of LCO. In contrast, the mechanochemical force from high-speed ball milling can cause the LCO lattice to change significantly, as shown in Fig. 1B. With increasing grinding speed, the peak intensity of LCO decreases rapidly, and the characteristic peaks of the Cu foils and Al foils nearly disappear. Specifically, the drifting of the (002) peak of LCO in Fig. 1C suggests that when ball milling exceeds 600 rpm, the mechanochemical force can destroy the order of the LCO lattice in the c-axis direction to the maximum extent, promoting the transformation of the crystal to the amorphous state. Fig. 1D shows that as the grinding speed is increased, the (111) peak of Al foils and the (111) peak of Cu foils attenuate significantly until they disappear. Additionally, the (200) peak of Al foils decreases and gradually merges with the (104) peak of LCO, which indicates that Al should react with LCO to form a mixed metal oxide. On the other hand, the surface morphology of the raw crushed products in Fig. 1E shows that LCO is a smooth-surfaced blocky particle. After mechanochemical activation, the diameter of LCO particles decreased greatly to ~2 μm, and fine scaly particles appeared on the surface, as shown in Fig. 1 F–H. The detailed micromorphology and element distribution of the mechanically activated products in Fig. 1 I–L show that the Cu and Al are uniformly distributed on the LCO particles. This indicates that LCO, Cu foils, and Al foils, which will not react spontaneously, undergo oxidation reaction under the induction of mechanochemistry. Since the face-centered cubic phases of Al and Cu disappear after the use of mechanical force, along with the Al and Cu being evenly distributed on the surface of the LCO final particles, it can be inferred that the Al⁰ and Cu⁰ lose electrons, convert into ionic states (Al³⁺, Cu^1+/2+) and adsorb on the LCO particles. Therefore, it can be speculated that the LCO particles undergo a reduction reaction, resulting in the collapse of the layered structure.

Fig. 1.

Lattice evolution, element migration and chemical environment around atoms of spent LCO cathodes induced by mechanochemical action [(A) lattice evolution of raw crushed products; (B) lattice evolution process at different grinding speeds; (C) local magnification of (002) peak of LCO; (D) local magnification of the diffraction angle of 36°≤2θ≤47°; (E) surface morphology of raw crushed products; (F) surface morphology of the mechanochemical activation products; (G) and (H) are partial enlarged views of (F); (I–L) are surface morphology and element distribution of optimized grinding product].

Microsurface Analysis and Atomic-Level Evidence.

The valence state changes of zero-valent metals and high-valent transition metal elements in spent LIBs are direct evidence of mechanochemical interactions. Here, we focus on exploring the functional groups or valence components of metal elements within 5 nm of the surface before and after mechanical activation by XPS analysis. The Al 2p XPS spectra of the crushed materials are shown in Fig. 2 A and B, and the peak of Al₂O₃ appears at 75.1 eV (23). It is worth noting that due to the high activity of Al element, a dense oxide film tends to form on the surface of raw Al foils. However, the intensity of the Al₂O₃ peak increased five times after high-speed ball milling, which verified that a large amount of Al³⁺ ions were generated by mechanochemical action and adhered to the sample. The Cu 2p XPS spectra in Fig. 2 C and D give that Cu foils in the raw crushed materials are still mainly in the form of Cu⁰ [located at 932.7 eV (24)], and a small amount of Cu²⁺ [located at 935.5 eV (25)] is also attached to its surface. The latter may be caused by hydrofluoric acid (HF) etching generated by electrolyte hydrolysis (26). More importantly, we detected 68.47% Cu²⁺ and 31.53% Cu¹⁺ [located at 933.3 eV (27)] on the surface of the mechanochemical products, which verified the above conjecture (Al⁰→Al³⁺, Cu⁰→Cu^1+,2+). The Co 2p XPS spectra in Fig. 2 E and F suggest that the content of high-valent Co³⁺ [located at 780.4 eV and 795.9 eV (28)] on the surface of the mechanochemical products decreased from 26.43 to 13.22%, down by about 50%, while the content of low-valent Co²⁺ [located at 781.8 eV and797.3 eV (28)] increased from 73.57 to 86.78%.

Fig. 2.

Microsurface functional group and chemical environment around atoms of spent LCO cathodes induced by mechanochemical action [(A–F) XPS spectrum of Al 2p, Cu 2p and Co 2p, where A, C, and E are raw materials and B, D, and F are mechanochemical products; (G) Co K-edge XANES of Co foils, Co-RM, Co-MA; (H) valence fitting analysis of Co atoms in Co-MA; (I) fourier transform of EXAFS spectrum in R space and the best fitting result for Co-MA; (J) fourier transform of EXAFS spectrum in K space and the best fitting result for Co-MA; (K–M) WT for the k³-weighted EXAFS signals of Co-MA, Co-RM and Co₂O₃].

The Co K-edge X-ray absorption near-edge structure (XANES) patterns of raw materials (Co-RM), mechanical activation products (Co-MA), and standard cobalt metal oxides are shown in Fig. 2G and SI Appendix, Fig. S2A. The absorption edges of Co-RM and Co-MA are located between the standard curves of CoO and Co₃O₄, which means that the average valence state of Co atoms is located between Co²⁺ and Co³⁺, and there is no Co⁰ at all before and after mechanochemical action. Further fitting analysis results of valence compositions of Co atoms, as shown in Fig. 2H and SI Appendix, Fig. S2B, indicate that the content of Co²⁺ in the cathode materials from spent LIBs after mechanical activation increased from 0.12 ± 0.05 to 0.72 ± 0.77, an increase of more than six times. It proves that the mechanochemical reaction after the introduction of zero-valent metals is a mild and well-controlled redox reaction. It does not directly reduce high-valent Co³⁺ to Co⁰, but only converts a large amount of high-valent Co³⁺ into low-valent Co^{(2+, 3+)}. The Fourier transform of the Co R-edge EXAFS of Co-MA shows two strong peaks around 1.75 Å and 2.25 Å in Fig. 2I, which are mainly attributed to Co-O and Co-Co coordination (29, 30). It should be noted that the two consecutive peaks at 3 Å correspond to the Co-Co signals with different spacings in LiCoO₂. Through the least-square analysis and quantification of extended X-ray absorption fine structure (EXAFS) fitting curve, as shown in Fig. 2J, we determined the chemical bond lengths of the cobalt atomic centers and their coordination numbers of the raw materials and mechanical activation products in SI Appendix, Table S1. As the valence of cobalt decreases, the bond length of Co-O increases, while the bond length of Co-Co decreases. After the mechanochemical action, the bond length of Co-O in the waste cathode materials increases from 1.90 ± 0.01 Å to 1.94 ± 0.02 Å, and the coordination number increased from 3.0 ± 0.6 to 3.2 ± 1.0; the bond length of Co-Co decreases from 2.80 ± 0.02 Å to 2.43 ± 0.04 Å, and the coordination number increases greatly from 3.2 ± 0.9 to 7.0 ± 0.2. All these prove that the mechanochemical reaction induces the efficient reduction of waste cathode materials, and the low-valence cobalt increase significantly. Furthermore, the growth of chemical bonds of Co-O verifies the occurrence of lattice collapse and cell distortion in spent LCO cathodes, and the large increase in the coordination number also implies the transformation of the LCO cell from an ordered arrangement to an amorphous state, which all show a surge in the internal energy of the sample after the mechanochemical reaction. These deductions can be confirmed by the wavelet transform, as shown in Fig. 2 K–M and SI Appendix, Fig. S3. The Co K-edge contour plots of Co-RM show two maximum intensities, and their positions are consistent with Co₂O₃ scattering, while Co-MA only shows one intensity maximum, and its position is between Co foil and CoO. Only an increase in the internal energy can induce such an irregular directional shift in the local coordination environment of Co centers. In summary, the enhanced mechanochemical action by introducing zero-valent metals promotes the controllable redox reaction of waste cathode materials, and meanwhile, a large amount of internal energy [i.e., Wigner energy (31)] is stored inside, causing its crystal structure to collapse, which will effectively promote the subsequent leaching.

Verification and Optimization of Leaching Effect of In Situ Reduction.

The stored internal energy from ball-milling promotes dissolution. We use acetic acid (CH₃COOH) to verify the leaching effect of weak organic acid on the different samples, as shown in SI Appendix, Fig. S4A. The direct leaching of 2 mol/L CH₃COOH of the ≤4 mm crushed products from spent LIBs shows that only 57.17% Li and 33.4% Co are dissolved into the liquid phase, and the leaching rates of Cu and Al are 17.96% and 50.22%, respectively. After the addition of 6 vol.% H₂O₂, the leaching rates of Li and Co increase to 100% and 73.21%, respectively, while that of Al decreases to 15.14%. Nevertheless, CH₃COOH leaching of the mechanical activated products results in the leaching rates of Li and Co reaching 92.37% and 92.51%, respectively, while the leaching rates of Cu and Al also increase significantly to 67.51% and 65.43%. Since it does not introduce impurity ions, H₂O₂ is recognized as one of the best reducing agents and its redox reactions with LiCoO₂ at pH ~1 are shown below (32):

$${\mathrm{H}}_2{\mathrm{O}}_2={\mathrm{O}}_2\left(\mathrm{gas}\right)+2{\mathrm{H}}^{+}+2{\mathrm{e}}^{-}$$ [1]

$${\mathrm{LiCoO}}_2+4{\mathrm{H}}^{+}+{\mathrm{e}}^{-}={\mathrm{Li}}^{+}+{\mathrm{Co}}^{2+}+2{\mathrm{H}}_2\mathrm{O}$$ [2]

However, the thermal stability of H₂O₂ is poor, and the reaction rate of reaction (1) is significantly higher than that of reaction (2), resulting in excessive consumption of H₂O₂. Under mechanochemically enhanced leaching, the waste cathode materials undergoes a direct redox reaction with Al foils and Cu foils as follows:

$$3{\mathrm{LiCoO}}_2+\mathrm{Al}+12{\mathrm{H}}^{+}=3{\mathrm{Li}}^{+}+3{\mathrm{Co}}^{2+}+{\mathrm{Al}}^{3+}+6{\mathrm{H}}_2\mathrm{O}$$ [3]

$$2{\mathrm{LiCoO}}_2+\mathrm{Cu}+8{\mathrm{H}}^{+}=2{\mathrm{Li}}^{+}+2{\mathrm{Co}}^{2+}+{\mathrm{Cu}}^{2+}+4{\mathrm{H}}_2\mathrm{O}$$ [4]

During the mechanochemical process, the Al foils and Cu foils, which are originally insoluble in weak acetic acid, lose their electrons to form soluble Al³⁺ and Cu²⁺, while the acid-insoluble Co³⁺ in LCO is reduced to soluble Co²⁺ by obligatory absorption of electrons. This charge rebalancing simultaneously enhances the leaching of Li, Co, Al, and Cu in weak acids. In addition, SI Appendix, Fig. S4 B–D shows the optimization of the leaching conditions of the 600 rpm-activated products. With a liquid-to-solid ratio of 100 mL/g, an acetic acid concentration of 2 mol/L and a reaction temperature of 60 °C, the leaching rates of Li, Co, Al and Cu in the ≤4 mm crushed products are close to 100%. Therefore, the mechanical activation pretreatment can promote the charge neutralization of insoluble metals in the crushed products and realize the in situ reduction in the spent LCO cathode. Synchronous leaching of multiple metals is thus achieved.

Self-Purification by ORP-Control via the Inherent Al Foils in the Leaching Liquid System.

In conventional hydrometallurgical recovery, differences in the solubility product constants (K_sp) can be used to achieve the removal of metal impurity precipitates; for instance, Fe, Al, and Cu ions can be separated from Ni, Co, and Mn ions by adjusting the temperature, pH, and initial concentration. We consider Cu cation as a typical impurity and use the E-pH Pourbaix diagram of Cu-Co-H₂O system to simulate the evolution of the Cu-ion phase boundary during the pH adjustment process, as shown in Fig. 3A. When the pH of the solution is adjusted from 3.0 to 6.5, the concentration of Cu ions decreases from 1 mol/L to 1 × 10⁻⁶ mol/L, demonstrating the removal of soluble Cu ions; however, the theoretical removal boundary of Cu ions coincides with the precipitation boundary of Co ions, which is the initial concentration of Co-ion precipitates. Therefore, adjusting the pH is likely to cause the coprecipitation of Cu and Co ions, which would result in the loss of precious Co ions needed to prepare NCA precursor. Temperature is also an important parameter in the conventional separation method. The E-pH phase diagrams of typical metal impurity ions at different temperatures are shown in SI Appendix, Fig. S5. The precipitation boundary lines of the Cu and Al ions shift in the direction of decreasing pH with increasing temperature; however, the difference in pH between the two fluctuates by only 0.6, which also makes it difficult to realize the complete separation of the Al and Cu ions.

Fig. 3.

Mechanism and effect of self-purification based on E-pH diagrams [(A) E-pH diagram of Co and Cu with different concentrations; (B) phase migration of Cu with different concentrations; (C) mechanism of Al⁰-induced directional transfer of Cu ions; (D) selective purification effect of Cu ions in solution].

CH₃COOH is used as the leaching agent, and the pH of the leaching solution fluctuates between 2.0 and 4.0. Thus, the phase distribution of impurity ions in the liquid system is shown in the orange background area in Fig. 3B. The oxidation/reduction potential (ORP) E, which is the y-axis of the Pourbaix diagram, gives us a systematic handle to separate out Cu cation impurities from the Li/Co in the aqueous solution. The ORP corresponds to effective O₂(gas) and H₂(gas) partial pressures inside the water (ORP↑ pO₂↑ pH₂↓), and zero-valent Al, being highly reactive with pO₂ once its oxide passivation layer is removed by the acid, can greatly reduce the ORP of the aqueous solution. When the solution pH = 2.0 to 4.0, the electrode potential for the conversion of Cu²⁺ to Cu is approximately 0.4 V, while the electrode potential for the conversion of H⁺ to H₂ is 0 to 0.2 V; thus, the potential of Cu ions is always higher than that of H ions. Moreover, the potentials of the Co ions will not exceed that of the H ions until the solution pH is >5. Therefore, when Al foils are added to the CH₃COOH leaching solution, Cu²⁺ will preferentially absorb electrons and cross the curve corresponding to the reduction reaction and enter the zero-valent Cu phase region; thus, the directional precipitation of Cu is realized, while H⁺, Li⁺and Co²⁺ still exist in the liquid phase, as shown in Fig. 3C. More importantly, the precipitated zero-valent Cu will not redissolve into the dilute acid solution, avoiding secondary pollution problems. To test the actual effect of the targeted purification mechanism, we conducted verification experiments on leaching solution systems, as shown in Fig. 3D. The Cu ion residue in the solution after the targeted precipitation tends to be near zero, and the loss of precious metal ions can be almost ignored; these results prove the success of the self-purification strategy. Notably, the pH at which Fe³⁺ begins to precipitate in the solution is 1.149, and when the solution pH reaches 2.813, Fe³⁺ can be completely precipitated; thus, the solution pH range is limited by the CH₃COOH leachate, and the ORP influenced by the added larger Al foils not only removes soluble Cu ions but also thoroughly remove the Fe ions from the solution (33). Actual ion concentration in different batches of self-purification residual liquid has been shown in SI Appendix, Table S2, which demonstrates the feasibility of the synergistic removal of copper and iron impurity ions. Therefore, the self-purification of the Cu and Fe ions in the leaching solution can be effectively realized by reusing the large Al foils pieces from the crushed products of spent LIBs.

Crystal Structure and Electrochemical Properties of the Regenerated High Nickel Lithium-Ion Batteries.

The Al³⁺ ions in the self-purified leachate are no longer removed, but instead further used directly as an Al source for the preparation of cathode materials; thus, Al self-supply method is realized. The morphology and elemental distribution of the precursor and the regenerated cathode materials are shown in SI Appendix, Figs. S6 and S7, all exhibiting similar particle size, plump shape and uniform elemental distribution. To check the crystalline state of the regenerated LiNi_0.80Co_0.15Al_0.05O₂ (NCA) cathode, we used GSAS software with the EXPGUI program to refine the structure of its X-ray diffraction (XRD) pattern (34); the results are shown in SI Appendix, Table S3 and Fig. 4A. According to the Rietveld refinement, the fitting parameters (R_wp and R_exp) of the differential peaks are 2.84% and 1.69%, respectively, and χ² is 3.93, which indicates that the lattice structure of the regenerated NCA cathode is very close to the theoretical crystalline state. The Rietveld refinement results show that the regenerated NCA cathode has a perfect layered structure, 100% belonging to the R3M space group, and does not contain nonlamellar oxide phase such as NiO. The reference value (c/a) of the ideal densely stacked hexagonal-laminated lattice is 4.899, while the c/a value of the regenerated NCA cathode is 4.937, which shows good stability of the layered structure (35). The spatial parameter I₀₀₃ / I₁₀₄>1.2 means a lower degree of cation mixing, while the I₀₀₃ / I₁₀₄ of the regenerated NCA cathode is 2.412, which indicates that the cations in the crystal structure are in a highly ordered arrangement (36, 37). Further fitting results also show that only 0.916% of the divalent Ni ions occupy the position of the Li ions in the LiNiO₂ phase, which is lower than 1.5% in similar literature (38). The actual electrochemical cycling performance of the regenerated NCA cathode and commercial NCA materials are shown in Fig. 4 B–E and SI Appendix, Fig. S8, respectively. The results suggest that the charge transfer resistance of our regenerated NCA cathode at room temperature is 229 Ω, which is higher than that of commercial NCA materials. The rate performance of the regenerated NCA cathode at different charge and discharge rates are as follows: 211.78 mAh/g at 0.1C, 190.08 mAh/g at 0.2C, 184.92 mAh/g at 0.3C, 177.69 mAh/g at 0.4C, 170.45 mAh/g at 0.5C, and 155.99 mAh/g at 1C. The first discharge capacity of the regenerated NCA cathode is 215.28 mAh/g, and the capacity retention rate after 180 cycles is approximately 80%; in contrast, the first discharge capacity of commercial NCA materials is 194.56 mAh/g, and their capacity retention rate after 180 cycles is approximately 75% under similar physicochemical conditions. The NCA cathode materials prepared by the direct upcycling of spent LIBs exhibit excellent crystallinity and a low degree of cation mixing; thus, their electrochemical cycling performance is up to or even slightly better than that of commercial products.

Fig. 4.

Lattice analysis and electrochemical performance of newborn NCA cathodes prepared from spent LCO batteries (A is structural refinement analysis of NCA cathode materials; B is electrochemical impedance spectroscopy; C is cyclic voltammetry curve; D is battery cycle performance at 0.2 C; E is battery rate performance).

Technoeconomic Analysis and Environmental Impact Assess ment.

In this study, the coupling of materials flow and life cycle cost analysis is used to complete the technoeconomic analysis of the mechanochemical upcycling method. The economic values and quantitative parameters of each step are given in the SI Appendix, and the results of technoeconomic analysis are shown in Fig. 5A. The price of waste raw materials is quoted from China National Resources Recycling Association, the prices of chemical additives and recycling products are from the industry quotation of Shanghai Nonferrous Metals Network on Sep. 9, 2022, and the prices of industrial water and industrial electricity are Chinese average price over the same period. The technoeconomic analysis shows that the total cost of raw materials for spent LIBs, and their disposal process is 3644.37 CNY per kilogram, the sales revenue of the recycling product is 4300.68 CNY per kilogram, and the comprehensive profit rate reaches 18%. The coprecipitation step is the most intensive step of chemical additives, and the cost of the corresponding chemical additives is 1590.86 CNY, accounting for about 43.65% of the total input. Due to shipping obstacles and ore supply difficulties caused by the COVID-19 pandemic, lithium carbonate is the most expensive additive, and its procurement cost accounts for about 52.42% of the total input. Since NCA batteries have been used as key components in advanced electric vehicles such as Tesla Model S, 3, X and Y, the regenerated NCA cathode products are a main source of economic income, and its economic benefit accounts for about 98.95% of the total output.

Fig. 5.

Technoeconomic analysis and environmental impact assessment of mechanochemical upcycling process for spent lithium-ion batteries [(A) technoeconomic analysis; (B) global warming potential assessment; (C) human toxicity potential assessment].

The environmental impact comparison between our new green upcycling process and the conventional hydrometallurgical process for recycling spent LIBs is analyzed by the life cycle assessment (LCA) method, and data lists are shown in the SI Appendix. The LCA results of greenhouse gas emissions, water acidification, and water eutrophication in Fig. 5B and SI Appendix, Fig. S9 show that acid leaching and coprecipitation are the main steps that increase the environmental impact. In the conventional acid leaching step, to neutralize the alkaline leaching residue, a large amount of strong inorganic acid (e.g., H₂SO₄) is consumed, and an additional reducing agent is necessary, which increases the environmental footprint. Thus, all environmental impact indicators of conventional technologies are higher than those of the new technology. In the coprecipitation step, the new technology is supplemented with appropriate amounts of (CH₃COO)₂Ni and (CH₃COO)₂Co. Although the environmental impact of the preparation process of these organic additives is higher than that of inorganic additives, acetate waste has high vapor pressure; thus, it is easy to volatilize and recover, resulting in a smaller environmental impact. As a consequence, the new technology has lower global warming potential, acidification potential and eutrophication potential than the traditional technology, with decreases of 45.83%, 9.16%, and 15.02%, respectively. Fig. 5C shows that the new technology is far lower than the traditional technology in terms of human toxicity potential, and the final potential value drops by 95.09%. This reason is that due to the high hole-blocking rate in the deep sieve-screening process in the traditional technology, a large amount of cobalt is not effectively recovered, causing them to be abandoned in the ecosystem, resulting in significant heavy metal pollution. In contrast, the new technology adopts moderate sieve-screening, and the recovery rate of the cathode materials increased to more than 98%. This in situ green upcycling therefore illustrates how one can significantly reduce environmental pollution and greenhouse gas emissions while supporting profitability of LIBs recycling industry and other circular economy industries.

Conclusion

In this work, we use the mechano-chemical upcycling of spent LiCoO₂ to new LiNi_0.80Co_0.15Al_0.05O₂ battery to illustrate the unique advantages of atom economy strategy. All components in the waste should be reasonably disposed of and enter the recycling products as much as possible. In contrast to waste reuse concept in the past, waste Al foils are used as a supply source for the regenerated NCA cathode materials after being employed as substitute reducing agents and purifying agents. Despite the lower consumption of chemical additives, the metal recovery rate from spent LIBs still reaches nearly 100%, and the electrochemical performance of the regenerated NCA cathode materials has reached the commercial application standard. More beneficially, compared with the conventional recycling process of spent LIBs, the carbon emission, acidification potential, and eutrophication potential of this new technology decrease by 45.83%, 9.16%, and 15.02%, respectively, while the human toxicity potential drops greatly by 95.09%. Such a direct upcycling paradigm can be useful in a broader field of waste management and carbon neutralization, and assist the decarbonization and sustainable development of new industries in the emergent global energy transition. As a final note, the pollution of organic volatiles is another difficult problem that hinders the industrial-scale recovery of spent LIBs. We believe that the mechanochemical reaction process will also promote the degradation of long-chain alkanes, and it will be parsed in detail in another work.

Materials and Methods

Raw Materials and Reagents.

The spent LCO batteries of the year-2015 model Motorola mobile phones were dismantled and used as experimental materials. The reagents used in the experiments were analytically pure and purchased from the unified reagent management platform of Tsinghua University. The experimental water was deionized water that was uniformly prepared by the School of Environment, Tsinghua University. It should be noted here that this study demonstrates the whole process of preparing NCA batteries from spent LCO batteries. Since ternary LIBs have similar crystal structure and material composition to LCO batteries, this method can theoretically be extended to the preparation of spent ternary LIBs into LiNi_xCo_yMn_zAl_mO₂ batteries.

Mechanochemical Upcycling Process of Spent LCO Batteries.

The data quality flow chart of the green upcycling of spent LCO batteries is shown in SI Appendix, Fig. S10. First of all, all spent LIBs need to be fully discharged in a 5% NaCl solution for 24 h to induce the lithium ions in the graphite anode to return to the cathode materials (39). After complete crushing and screening, undersized products with particle sizes of less than 4 mm (≤4 mm crushed products) and oversized products with particle sizes of greater than 4 mm (≥4 mm crushed products) were obtained, accounting for 75.44 wt.% and 24.56 wt.% of the total mass, respectively. This sieve aperture can ensure that the materials under the sieve undergo sufficient redox reaction (33). The ≥4 mm crushed products were sorted in sequence by a pulsating airflow to recover the plastic diaphragm, magnetic separation to recover the steel flakes, color separation to recover the Cu foils, until finally, a small amount of larger-sized Al foil fragments were obtained (40). After mechanical activation, ≤4 mm crushed products were leached with acetic acid and filtered to obtain a leaching solution and graphite residue. On the other hand, the large Al foils obtained from the ≥4 mm crushed products were thrown into the CH₃COOH leachate to precipitate and separate the copper sludge and a small amount of iron sludge, in order to obtain a purified leaching solution.

Preparation of New Ternary Cathode Materials from Purified Leachate.

Firstly, the purified leachate needs to be characterized for its specific elemental concentrations to calculate the subsequent additions of lithium, nickel, and cobalt. Next, the purified leachate should be stirred uniformly at 500 to 600 rpm at moderate concentration and temperature, while coprecipitation and aging are performed after slow dropwise addition of diluted nickel acetate, cobalt acetate, or sodium meta-aluminate according to the molar ratio of Ni: Co: Al = 16:3:1. Under low-speed grinding and stirring, the NCA precursor obtained from coprecipitation should be thoroughly mixed with lithium carbonate at a molar ratio of 1:1.05. Finally, the NCA precursor mixtures can be sent to a high-temperature furnace for roasting and crystallization in a liquid oxygen environment to achieve the regenerated NCA cathode materials. The oxygen-rich calcination process requires a slow ramp-up to 800 °C at 5 °C/min, holding for 24 h, followed by a slow ramp-down at 5 °C/min. Excessive temperature fluctuations and insufficient oxygen content are prone to result in the failure of the NCA cathode.

Electrochemical Performance Test.

First, the NCA cathode materials, carbon black, and polyvinylidene fluoride were evenly mixed at a mass ratio of 8:1:1 and stirred evenly with N-methylpyrrolidone as the solvent. Next, the fresh slurry was evenly coated on an Al foil, dried in a vacuum drying oven at 70 °C for 12 h, and punched into a circular disk with a diameter of 8 mm to obtain a serviceable cathode. The assembly of the battery was carried out in a glove box filled with argon gas. Finally, the CR2032 button cell was assembled with the above cathode, lithium metal (anode electrode), a Celgard 2,400 microporous polypropylene film (separator), and 6 μL of electrolyte composed of 1.0 mol/L LiPF₆ in EC:DEC:EMC = 1:1:1 for electrochemical performance testing. The mass loading of NCA cathode materials is 1.7 mg/cm².

Material Characterization and Analytical Calculations.

The analytical instruments used in this research were provided by the testing platform at Tsinghua University, and all instruments had obtained China metrology accreditation. Under the test conditions of a copper target at 40 kV and 30 mA, X-ray diffraction (XRD, Philips PW 1700, USA) was employed to characterize the crystal structures of different samples. Field-emission scanning electron microscopy equipped with energy dispersive spectroscopy (Carl Zeiss MERLIN Compact, Germany) was performed to capture the surface morphology and the elemental distribution, respectively. Synchrotron X-ray absorption near-edge structure (XANES) spectroscopy was supported by Shanghai Light Source. Inductively coupled plasma optical emission spectrometry (ICP-OES, PerkinElmer, OPTIMA 2000, USA) was used to analyze the concentration of metal ions in the leaching solution. Based on Gibbs free energy, Faraday's law and Nernst equation, E-pH Pourbaix diagram can be obtained for phase transfer predictions (41, 42). The calculation methods and process parameters of E-pH curve has been shown in SI Appendix. On the other hand, the data list and soft parameters of life cycle assessment method and technoeconomic analysis have also been shown in SI Appendix.

Supplementary Material

Appendix (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.