Structural evolution of layered oxide cathodes for spent Li–ion batteries: Degradation mechanism and repair strategy

Sustainable development has long been recognized as one of the most critical issues in today's energy and environment‐conscious society. It has never been more urgent to recycle and reuse the end‐of‐life cathode materials. Here, this work systematically investigates the structure‐critical degradation mechanism of polycrystalline LiNixCoyMn1−x−yO2 (NCM), combining experimental characterization and DFT simulations. Targeting the key degradation factors, a synergistic repair strategy based on deep mechanochemical activation and heat treatment was successfully proposed to direct regenerate the degraded NCM material. Studies indicate the induction and promotion of synergistic repair technique on the reconstruction of particle morphology, the recovery of the chemical composition and crystal structure, and the favorable transformation of the impurities phase in the failed materials. In particular, the synergistic repair process induces a gradient distribution of LiF and further enables partial fluorine doping into the NCM surface, forming abundant oxygen vacancies and increasing the content of highly reactive Ni2+. Benefiting from the comprehensive treatment for the multi‐scale and multi‐form degradation behaviors, the repaired material exhibits a capacity of 176.8 mA h g−1 at 0.1 C, which is comparable to the corresponding commercial material (172.8 mA h g−1). The satisfactory capacity of the recovered cathode proves that it is an effective direct renovating strategy.


INTRODUCTION
Lithium-ion batteries (LIBs) have been driving the evolution of transportation sector and energy storage systems in recent years owing to their prominent superiorities, and their great potential applications are gradually being exploited, leading to increased demands on the performance. As one of the core components of LIBs, cathode materials have a significant impact on the performance. Among the various commercial cathodes presently available, the layered oxide cathodes, in particular the typical LiNi x Co y Mn 1−x−y O 2 (NCM), are the most promising because of their high average operating potential (3.6 V vs. Li/Li + ), high theoretical capacity (∼270 mA h g −1 ) and excellent rate capability, which are estimated to dominate the global LIBs market in the long run. Due to the rapid adoption of NCM batteries in global mobile electrification systems, large quantities of waste batteries containing high-value metals have been manufactured. The proper handling of end-of-life NCM batteries has become an urgent and promising task, which also implies a substantial opportunity for the recycling industry. 1 Metallurgical process, the most current mature recycling technology, including pyrometallurgy and hydrometallurgy, is growing fast, targeted at recovering valuable metal elements from batteries, such as cobalt, nickel, and lithium. Unfortunately, the metallurgical process is carried out in a way to damage the intrinsic value of the electrode material, potentially enabling recycled products to be less valuable than anticipated as well. Meanwhile, this laborious and complex process is not competitive in future business practices. Accordingly, from a social, economic, and environmental standpoint, there is an urgent appeal for an effective and sustainable recycling strategy to streamline the process, enhance recycling efficiency, lower costs, and widen profit margins. 2 Direct repair of cathode materials is increasingly the focus of researchers' attention as a novel short-range technology for waste LIB disposal, which can accomplish both the recycling of valuable metals and the sustainable reuse of electrode materials. 3 This technique entails the direct replenishment of the lost metal ions into the spent cathode active powder without dissipating the inherent energy. The composition and structural defects of the material particles are then renovated in a manner similar to that of cathode synthesis, ultimately resulting in the recovery of losing capacity. The earliest material to be directly renovated was LiCoO 2 with a layered structure. 4 The lithium source was simply mixed with the recovered cathode powder to adjust the element ratio, and then it was calcined at high temperature. The initial discharge capacity of the recycled LiCoO 2 reached 151 mA h g −1 . Similar methods have also been used to repair other cathode materials, including LiFePO 4 [5][6][7] and LiMn 2 O 4 . 8 Indeed, strenuous efforts have also been made in recent years to innovate the methodology for direct repair of cathode materials. [9][10][11][12][13][14][15][16][17][18][19][20] Some studies have reported the use of solid-phase sintering, 11 hydrothermal, 21 molten salt, 17 electrochemistry, 22 and other novel methods and thoughts 20,23,24 for direct repair cathode, offering innovative insights into cathode materials renovation. However, compared with the above electrode materials, the chemical composition of NCM is more complex. The presence of multiple failure modes in the cycle undoubtedly poses difficulties for direct repair. Therefore, achieving efficient reuse of NCM electrode materials requires targeted remediation. It is essential to first clarify the key behaviors and mechanisms responsible for performance degradation, diagnose and assess the decay properties of structural, and then design targeted repair strategies.
The typical NCM are layered lithium transition metal (TM) oxides belonging to α-NaFeO 2 hexagonal system and R-3m space group. The TMs alternate with lithium occupying in the octahedral sites of the close-packed oxygen framework, whereas the Ni, Co, and Mn ions are statistically distributed according to the material ratios. 25,26 In general, higher levels of Ni, the dominant redox species in the positive electrode, provide more reversible capacity, but they also lead to a severe reduction in structural and thermal stability during cycling, as well as inducing various side reactions. It is not difficult to conclude from previous studies that the failure mode of NCM cathode is synergistic with multiple behaviors, frequently accompanied by multi-scale (crystal, particles, or electrode) and multiform (solid-phase bulk, solid-liquid interface, and oxygen evolution) degradation behavior during the lithiation/delithiation process. 27,28 Generally, the structural decay and capacity loss of NCM cathodes are considered to be inseparable from the Li loss and release of oxygen in cycling. 29 TM ions might migrate, and the layered structures are prone to phase transitions for cation disordered, spinel, and rock salt phases. 30 The hindrance of active Li + embedding triggers the occurrence of multiple side reactions on the electrode-electrolyte interface and produces various ioninsulating by-products, which thicken the interfacial film and increase the internal resistance. 31 Further, the uneven distribution of Li + concentration causes an increase in thermal stress and pressure inside the particles. From this, intergranular microcracks and electrolyte penetration are formed, which accelerates structural damage. [32][33][34] Given the careful consideration, it can be deduced that in order to achieve direct repair and reuse of electrode materials, the failure characteristics of the wastes must be addressed at both macro and micro scales. Currently, there is still a lack of comprehensive and scalable repair methods focusing on detailed diagnostic results of waste NCM materials.
In this work, we systematically investigate the structurecritical degradation mechanism of NCM electrodes, including the evolution of particle morphology, bulk phase structure, Li/Ni translocation, and surface impurities. The effect of various lithium impurities on the degraded NCM surface relative to Li + diffusion kinetics is evaluated in conjunction with DFT simulations. The theoretical results show that LiF has the greatest influence. According to the diagnosis of spent NCM, we report a synergistic repair strategy to directly refurbish the spent cathode powder based on deep mechanochemical activation and heat treatment, and successfully repair the degraded NCM with a greatly boosted capacity. In this technique, reconstruction of particle, restoration of chemical composition and crystal structure, and favorable transformation of impurity phases are achieved simultaneously. It is worth noting that the LiF impurity no longer exists entirely as a coating on regenerated NCM surface but is dispersed into the NCM particles with concentration gradients. Partial fluorine substitutes oxygen on the material surface, which promotes the formation of oxygen defects, and thus improves the electrochemical stability. Our study provides a scalable repair scheme that targets the comprehensive degradation behavior of severely failed electrode materials to maximize resources. It is also instructive to clarify the impurity phase transformation during renovation.

Evaluation of the structural degradation
After excessive de/intercalation of Li + from the matrix structure, the layered NCM material usually undergoes the phase change and structural degradation. To identify degradation mechanisms and develop specific remedial solutions of NCM electrodes, we have analyzed and evaluated the spent LiNi 0.5 Co 0.2 Mn 0.3 O 2 materials (denoted as NCM-S). The morphological characteristics of NCM-S materials were examined using scanning electron microscopy (SEM), and there were clear differences from the commercial LiNi 0.5 Co 0.2 Mn 0.3 O 2 materials (NCM-N). As illustrated in Figure S1, the secondary particles of the NCM-N material have a smooth surface, with some of the micron-sized particles remaining intact in spherical form. In contrast, the surface roughness of NCM-S is altered and cracked, and the boundary of primary particles is blurred ( Figure S1C,D). The cycle-induced differences in the spatial distribution of lithium concentration in the NCM particles frequently cause structural damage and an increase in local micro-stress, which results in the forma-tion and propagation of cracks, and further disrupts the bulk phase structure. 35 Transmission electron microscope (TEM) images confirm the occurrence of phase changes in the surface layer of degraded NCM-S ( Figure S2). Figure 1A summarizes the multiple phase structures derived from the degradation of layered electrodes. Due to the migration of TM ions to the surface during cycling, part of the layered structure gradually transforms into the spinel and rock salt phases. Significantly, in the rock salt phase structure, the migration and diffusion of Li + is intercepted, which is also responsible for the performance degradation. Furthermore, it is obvious that an impurity layer covers the surface of the NCM-S material ( Figure S2A). As the increased instability of the electrode material in contact with the electrolyte over time, a variety of interfacial nonhomogeneous chemical reactions occur, which leads to the generation and deposition of ionic insulators on the cathode interface film, blocking the Li + diffusion path and affecting electrochemical performance. 28,31 The common surface impurities in spent NCM materials mainly consist of the lithium salt, adsorption solvent, carbon, and poly(vinylidene difluoride) (PVDF) binder. NCM-S sample has been pretreated at 600 • C to remove acetylene black and binder, and thermogravimetric (TG) results indicate that they have been cleared ( Figure S3). During this process, the organic and inorganic lithium compounds, which form on the surface of the electrode material due to the side reactions of the interface, will react with the O 2 , H 2 O, and CO 2 in the air, eventually converted into inorganic lithium residues (LiF, Li 2 CO 3 , and LiOH) accumulated on the surface of waste cathode material. 31,35 Therefore, it is speculated that the observed impurity layer is mainly composed of inorganic lithium salt, including LiF, Li 2 CO 3 , and LiOH ( Figure 1B). This is consistent with the results reported in previous studies. Further, the effect of the possible interface impurities (LiF, Li 2 CO 3 , and LiOH) on Li + conduction was quantified by DFT calculations, as shown in Figure 1C. LiF has the highest Li + diffusion energy barrier, rendering the transport of Li + somewhat delayed. Moreover, the Li 2 CO 3 and LiOH impurities on the surface would further react with HF generated from the decomposition of LiPF 6 in the electrolyte, resulting in the LiF accumulation and consequent safety problems. 36 More importantly, the high stability of LiF brings a challenge to achieving reversible recovery of the electrode material structure.
Chemical states of the elements on NCM materials surface were examined by X-ray photoelectron spectra (XPS) to verify the structural changes of failed NCM-S. The peak around 284.8 eV in the C 1s spectrum ( Figure S4B) can be ascribed to carbon in the environment, whereas the peak near 289 eV is assigned to carbonate compounds (Li 2 CO 3 ). In combination with the Li 1s spectrum ( Figure 1D), the peak at 54.25 eV is characteristic of lithium in the lattice of electrode materials, whereas the peaks around 55.25 and 56.2 eV are ascribed to the impurities Li 2 CO 3 /LiOH and LiF, 11 respectively. LiOH is difficult to detect, while it tends to turn into Li 2 CO 3 , but it is indeed one of the impurities. 37 In the F 1s spectrum ( Figure S4C), an Li-F peak (684.8 eV) is also fitted. It is worth noting that the proportions of LiF and Li 2 CO 3 are evidently higher on the NCM-S surface. The O 1s spectra can be fitted with two peaks ( Figure  S4D), correlated to the O Impurity (531.4 eV) and the O Lattice (529.4 eV). 38 The O Impurity is mainly caused by the active oxygen species O − , O 2− , or CO 3 2− of the impurity layer. The O Lattice is formed by the lattice oxygen (O 2− ) of the TM-O bond (TM = Ni, Co, Mn). Of note, in the O 1s region of the NCM-S, the 529.4 eV is largely replaced by a new peak at 529.8 eV, which is attributable to the rock salt NiO phase, originating from the loss of lattice oxygen in the edge region of the particles during long-term cycling. 39,40 As shown in Figure 1E, the presence of NiO is further corroborated by the Gaussian fit to the Ni 2p XPS spectra of NCM-S. 21,40 The above results suggest that the surface/interface structure of the spent cathode has degraded significantly. So, the removal of rock salt phases (NiO) and surface lithium residues (LiF, Li 2 CO 3 ) during NCM renovation is critical to restoring performance.
Additionally, the Li/Ni translocation in the NCM cathode has been proved to have a detrimental effect on the structural stability and kinetics, in terms of the irreversible phase transition, retarded ionic diffusion, and rapid capacity decay. 41 In layered NCM cathode material, the peak intensity ratio of (0 0 3)/(1 0 4) has been well described as reflecting the mixing of Li + and Ni 2+ cations. 42,43 Quantifying the above peak intensity parameters by structural refinement (Table S1), it can be seen that in NCM-S material, the orderliness of layered structure is disrupted, and the Li/Ni antisite is serious (9.79%), compared with NCM-N (0.103%) ( Figure 1F and Figure S5). Research has been carried out to suggest that Li-Ni antisite is a completely spontaneous behavior from either the thermodynamics or the kinetics. 41 The introduction of Li/Ni antisite is influenced by the Li content of the material and becomes easier with increasing Li defects during cycling. Simultaneously, the continuous accumulation of Li/Ni antisites leads to insufficient Li + embedding, thereby causing less and less active lithium in the material. Evidently, the Li + content of the long-term cycled NCM-S is significantly lower than that of the NCM-N ( Figure 1G). As cyclinginduced Li loss is inextricably linked to the structural disintegration and performance degeneration of electrode materials, research into repair technology should first focus on lithium compensation.
Taken together, it can be seen that in the failed NCM, Li defects and Li/Ni antisites are generated within the lattice, and impurity phases are formed externally, directly damaging the grain boundary structure among the primary particles. Consequently, the renovation of the degraded NCM electrode material requires not only restoring the composition but also reviving an ordered crystal structure, removing the impurity phases and restructuring the particle morphology.

Synergistic repair strategy
After identifying the failure mechanism of the cathode material, a corresponding repair strategy is proposed. To maximize the inherent resources of spent LIB cathode materials, we used a solid-phase method to directly refurbish the waste cathode powder, including the synergistic repair process of deep mechanochemical activation and heat treatment.

Reconstruction of morphology
The NCM-S powder was first subjected to primary lithium replenishment with the addition of LiOH and manually ground to obtain the powder as NCM-non-MA. Subsequently, it was deeply re-lithiated using a mechanochemical activation process yielding NCM-MA. It is shown that the mechanical forces refine the LiOH particles, which are caused by the milling or grinding of the material. The morphology of LiOH in the particle swarm was further determined by EDS mapping, clearly visible as a block in the SEM image of NCM-non-MA, and we labeled it in yellow (Figure 2A-D). Contrary to the recognizable morphology of LiOH in NCM-non-MA, it is not readily detectable in NCM-MA. The mechanochemical process allows it to integrate more homogeneously with NCM particles or adsorb on their surface, shortening the actual distance of lithium replenishment. In addition, we found that the mechanochemical activation process also changes the morphology of NCM particles ( Figure 2C). After short high-energy crushing process, the particles of NCM-S are broken along the cracks, and LiOH is changed into tiny particles that are uniformly scattered ( Figure S6A). The shattered particles are gradually refined and spread more uniformly as time passes. After deep grinding, secondary aggregation of the finely crushed NCM-S main particles commences, gradually creating new aggregates ( Figure  S6B). This process arises as a result of the deformation and fracture of solids and involves the changes in internal and surface energy of the solid during milling or grinding, leading to spontaneous aggregation, adsorption, or recrystallization in the activated system. 44 Therefore, we take advantage of the change in particle surface energy induced by this process to make NCM particles undergo fragmentation and then re-agglomeration, thus realizing the remodeling of NCM particles. Following lithium replenishment, the material is heat treated in an oxygen atmosphere and the chemical reactions occur as follows: As shown in Figure S7, the surfaces of both the NCMnon-MA-R and the NCM-MA-R become relatively smooth after heat treatment, and the large particle impurities decrease significantly. Furthermore, NCM-non-MA-R particles, which are only artificially lithium replenished, retain most of the fracture characteristics of its predecessor (NCM-S), presumably with some negative impacts on lithium storage performance. For NCM-MA-R, a large number of secondary particles with surface cracks are broken up due to the mechanochemical activation, and primary particles are exposed. Some of these primary particles undergo re-agglomeration, forming pore-rich aggregates close to be spherical in shape ( Figure S7D,H). Compared to NCM-N, although NCM-MA-R has disintegrated from the secondary spherical aggregates, the primary particles still maintain a similar morphology. The grain size analysis of primary particles shows the average size of 0.633, 0.48, 0.489, and 0.51 μm for NCM-N, NCM-S, NCM-non-MA-R, and NCM-MA-R, respectively ( Figure S7E-H). It can be seen that the statistical results of particle size of NCM-S and NCM-non-MA-R are the smallest, which are still attributed to the existence of fine cracks on the surface. Brunauer-Emmett-Teller (BET) surface area and porosity of the samples were analyzed via nitrogen isothermal adsorption to further explain the variations in microstructure ( Figure S8). The detailed data is listed in Table S2. The BET surface area of NCM-MA-R (23.43 m 2 g −1 ) is significantly higher than that of NCM-N (3.87 m 2 g −1 ), and NCM-MA-R also has a larger total pore volume, indicating a smaller particle size and an abundance of pores. The porous configuration of the NCM-MA-R electrode combined with the high surface area allows for favorable electrolyte infiltration and can possess superior capacity throughout the electrochemical reaction. For NCM-S, however, the extremely high surface area exhibited is presumably due to a large number of cracks in the particles.

2.2.2
Restoration of structure orderliness As the failed electrode materials exhibit significant lithium loss, the lithium content of repaired NCM is one of the key indicators to evaluate the recovery condition of the material composition and structure. Inductively coupled plasma (ICP)-MS was utilized to discern the relative proportion of Li, Ni, Co, and Mn, especially on distinguishing the Li content. Figure 2E depicts the chemical composition of NCM (detail value in Table S3). It can be seen that the Li molar ratios of both the NCM-MA-R and NCM-non-MA-R reach the theoretical level, with NCM-MA-R (1.019) slightly higher than NCM-non-MA-R (0.997). This illustrates that the mechanochemical activation process played a positive role in improving the utilization of the lithium source and promoting the lithium replenishment of the spent NCM.
The structural alterations of the comparative samples were characterized by X-ray diffraction (XRD) to further illustrate the effect of the repair process on the crystal structure of the material. Figure S9 depicts the local distinctive peak region of the collected sample's XRD spectrum. All diffraction peaks can be attributed to a regular layered α-NaFeO 2 -type structure (space-group R-3m, Z = 3). The sample's strong diffraction peak intensity suggests that it has perfect crystallization. Comparing the (0 0 3) peak of samples ( Figure S9), it can be discovered that the peak intensity of NCM-non-MA-R and NCM-MA-R is markedly increased, implying that the repair process in an atmosphere of oxygen restored the material's crystallization to its original level, whereas the NCM-MA-R reflects a more obvious advantage in crystallization. In addition, the characteristic double peaks (0 0 6)/(1 0 2) and (0 1 8)/(1 1 0) of NCM-non-MA-R and NCM-MA-R in the figure are almost identical to those of NCM-N, which indicates a desirable layered structure in restored samples. 45 To future quantify the structural changes before and after renovation, Rietveld refinement was performed on the collected XRD patterns. Rietveld refinement patterns and structure parameters are demonstrated in Figure 2F,G and Table  S4, respectively, in which the lower values of R p are wellmatched with the plots confirming that the refinement findings are logically acceptable. Figure 2H summarizes the evolution of the crystal structure of the NCM layered crystal structure in degradation and renovation. Numerous Li vacancies are generated in the degraded NCM-S, further leading to the occurrence of Li/Ni antisite defects (9.79%). After repairing, the Li/Ni mixing of the NCM-non-MA-R (4.82%) and NCM-MA-R (0.434%) materials decreases to varying degrees, and the fitting results of NCM-MA-R are closer to NCM-N (0.103%), which intuitively and forcefully demonstrates that the synergistic repair as an inductive effect contributes to the recovery of the crystal structure and promotes the ordered arrangement of Li/Ni. This might be attributed to the fact that the mechanical activation process creates a more homogeneous Li-rich environment for the repair reaction, which is more conducive to the reverse conversion of the cation disordered phase and restoration of crystal structure during the heat treatment.

Transformation of impurity phase
In order to explore the removal of impurity phases from the surface/interface of the electrode material, the chemical states of the surface elements were examined using XPS ( Figure 3A-D). The XPS spectra are quite similar between NCM-non-MA-R and NCM-MA-R powders. As shown in the O 1s XPS spectra of the repaired materials ( Figure 3B), it is gratifying to discover that NiO vanishes from the surface of degraded NCM-S ( Figure S4D), and this is consistent with the analysis of the Ni 2p spectrum ( Figure 3D). It means that this process is independent of mechanical activation and mainly dominated by the heat treatment in the oxygen atmosphere. 38 By contrast of NCM-S ( Figure 1D), the proportions of Li 2 CO 3 and LiF are considerably lower in NCM-non-MA-R and NCM-MA-R ( Figure 3A), denoting a reduction in surface impurities after renovation, particularly in NCM-MA-R. For lithium salt impurities present in materials, solidstate nuclear magnetic resonance (NMR) is a very effective means of detection, especially 7 Li. The large sideband pattern on each side of Li resonance centered at 0 ppm (from ca. +1200 to −1200 ppm) represents lithium on the surface, in a diamagnetic environment, where dipole interactions occur with paramagnetic ions located many bond distances away in the bulk material. Typically, diamagnetic lithium salts are by-products of material in the process of synthesis or failure, which can be found as LiOH, Li 2 CO 3 , Li 2 O, LiF, and so on. 46,47 7 Li NMR spectra of the NCM-no-MA-R and NCM-MA-R material are shown in Figure 3K. The diamagnetic surface peak of lithium residues decreases in intensity by approximately 10% upon mechanochemical activation and heat treatment. It is further proved that the synergistic repair strategy has a positive effect on the removal and transformation of residual impurity phase in degraded materials.
However, the mechanism of lithium residues migration and transformation during NCM-S synergistic renovation remain unclear, especially for LiF, which is usually stable and difficult to decompose. Some studies have demonstrated that LiF is unable to be removed from the surface of the regenerated material and remains as a coating. 48 But the decrease in LiF is real in our study. On the one hand, the LiF impurity layer is broken up and dispersed to the inner part of the particles during the mechanochemical activation. On the other hand, it might participate in the material repair reaction. Comparing the XPS spectra of NCM samples before and after ball milling ( Figure S10), it is found that the contents of LiF and Li 2 CO 3 are slightly reduced after ball milling, possibly because the impurities on the surface are integrated into the interior during the process of crushing and aggregation. After heat treatment, there is a significant decrease in the peaks of Li-F in the F 1s spectra of NCM-MA-R powder ( Figure 3C), as well as a disappearance of LiF fitted peak in the Li 1s spectrum ( Figure 3A), which implies that the LiF in NCM-MA-R underwent chemical reactions during the repair process. Among the reasons, we suppose that due to mechanical activation, the LiOH particles are uniformly attached to the NCM-S, forming LiOH-Li 2 CO 3 ( Figure 3I) and LiOH-LiF ( Figure 3J) eutectic molten salt systems thus improving the  7 Li single pulse MAS nuclear magnetic resonance (NMR) spectra of repaired material. feasibility of the reaction. 49 Under an oxygen atmosphere, Li 2 CO 3 can be used as a lithium source to react with NCM as follow: Additionally, LiF is also used as a reactant in the process of preparation and modification of layered oxides cathode to achieve coating cladding or F-doping. 16,50,51 Nevertheless, it is still uncertain in which form of F is present in the repaired material. In our study, we are surprised to find that a small amount of fluorine may be doped in NCM-MA-R materials based on the synergistic regeneration of mechanical activation and heat treatment. Due to the atomic radius and electronegativity characteristics, fluorine is usually considered to substitute for oxygen sites, accompanied by the relaxation of the a and c axes, and the generation of oxygen defects. [52][53][54][55][56] In accordance with the XRD Rietveld results (Table S4), the a and c axes increased from a = 2.8716 Å and c = 14.251 Å (NCM-non-MA-R) to a = 2.8725 and c = 14.259 Å (NCM-MA-R). Furthermore, the repaired NCM fresh electrode sheet was etched at different times and analyzed deeply though highresolution XPS spectra. In the O 1s spectrum ( Figure 3B), we fitted the characteristic peak of the oxygen vacancy at 530.8 eV. 57 The results show that NCM-MA-R has significantly more abundant surface oxygen defects, which are consistent with reports on fluorine doping. 58 The changes of the internal oxidative environment of the repaired material electrode sheets were further investigated after different etching times. As shown in Figure 3G,H, the proportion of O Vacancy peak on the surface of NCM-MA-R is the greatest (30.2%), whereas the intensity of this peak tends to decrease with the increase of detection depth. However, the proportion of the O Vacancy of NCM-non-MA-R is extremely low (5.62%) and decreases internally as well ( Figure 3E,F). It indicates that in the NCM-MA-R, some F may occupy the oxygen site and form defects on the surface, whereas in the NCM-non-MA-R, LiF is retained and exists as an impurity layer. The gradient distribution of oxygen vacancy from outside to inside is very obvious, also suggesting that F-doping is mainly present on the surface. The spectrum of Li 1s proves this speculation more visually. In Figure S11A, the peaks of LiF gradually increase with the etching depth in NCM-MA-R electrode. More importantly, LiF is barely detectable on the surface of NCM-MA-R, which implies that the LiF on the surface is converted to fluorine doping into the lattice. In contrast, the peaks of LiF are always present and in similar proportions after etching in NCM-non-MA-R ( Figure  S11B). This suggests that an LiF impurity coating is still present on the surface of the NCM-non-MA-R, whereas the concentration of LiF in the NCM-MA-R shows an increasing gradient from the outside to the inside. Such gradient distribution of LiF not only protects the internal electrode material but also directly reduces the interfacial resistance.
To further prove the creation of oxygen vacancies, we carried out the electron paramagnetic resonance (EPR) spectroscopy test. As shown in Figure S12, a typical EPR signal centered at g = 2.000 is detected in all samples, suggesting that electrons are trapped on oxygen vacancies. 59 Owing to fluorine doping, the oxygen vacancies of NCM-MA-R are increased, which makes the strongest signal. In contrast, NCM-non-MA-R shows the weakest signal (lower than NCM-S) because of the heat treatment taking place in an oxygen-rich environment. Moreover, the F substitution in NCM triggers a partial reduction of Ni 3+ to Ni 2+ , and a slightly higher Ni 2+ content can improve the structural stability of the Ni-rich cathode material. 60 As shown in Figure 3D, the fitting peak of Ni 2+ in NCM-MA-R (50.9%) is stronger than NCM-non-MA-R (44.9%), further indicating a partial F-doping into the NCM surface. For NCM-non-MA-R, we argue that most of the LiF is still attached to the surface of the particles. The residual layer of lithium salt on the surface is bound to increase the interface impedance of the material, and its influence on the electrochemical performance is described in detail in the following section.

Effectiveness of the repair strategy
Electrochemical tests were performed to assess the effectiveness of the synergistic repair technique and the necessity of mechanochemical activation. First, NCM/Li halfcells were cycled three times at 0.1 C rate to activate the electrode and then tested at a rate of 0.5 C for long cycles between 3 and 4.3 V (vs. Li/Li + ). Figure 4A depicts the initial charge/discharge curves at 0.1 C rate for the electrodes of NCM-N, NCM-S, NCM-non-MA-R, and NCM-MA-R, and the discharge capacities exhibit 172.8, 112.3, 162.9, and 176.8 mA h g −1 , respectively, corresponding to initial coulombic efficiencies (ICEs) of 83.7%, 63.7%, 78.2%, and 84.6%. Evidently, the reversible capacity of both renovated electrodes is significantly improved. The cycling stability of the prepared electrodes is further compared ( Figure 4E). It is promising that the NCM-MA-R still maintains the highest discharge capacity at the beginning of the test, and the capacity remains at 137.2 mA h g −1 after 200 cycles, which closely matches that of the NCM-N (142.9 mA h g −1 ). In contrast, the discharge capacity of the NCM-non-MA-R electrode decreases to 110.9 mA h g −1 after 200 cycles, which is only 77% of its initial. The capacity of NCM-S fails to be maintained in long-term cycling (only 65 mA h g −1 after 100 cycles at 0.5 C rate). In addition, the rate performance of all samples was also assessed at different charge/discharge rates from 0.1 to 5 C between 3.0 and 4.3 V (vs. Li/Li + ) at 25 • C ( Figure 4G). The results suggest that the NCM-MA-R electrode exhibits the least capacity fluctuation. As shown in Figure 4B-D, the discharge capacity of NCM-MA-R decrease by 39 mA h g −1 as the rate increased from 0.1 to 5 C, which is significantly lower than that of NCM-non-MA-R (54 mA h g −1 ) and NCM-S (80 mA h g −1 ). While cycling at 0.1 C again, the NCM-MA-R can still be restored to the capacity of 176 mA h g −1 ( Figure 4G). This is mainly attributed to the smaller particle size and larger specific surface area of NCM-MA-R, which allows a larger contact area between the material and the electrolyte, as well as a shorter diffusion path for Li-ions, thus providing faster ionic diffusion and conduction. Furthermore, due to structural degradation and lithium deficiency, NCM-S materials exhibit continuous fluctuations of ICEs, resulting in poor stability ( Figure 4F) relatively unstable, and the irreversible lithium in the lattice gradually accumulates after a period of cycling, leading to degradation in performance. Additionally, the galvanostatic intermittent titration technique (GITT) was employed to better analyze the kinetics of Li-ion diffusion during the initial charge/discharge operation. The corresponding equation is provided in the following equation: where m B , V M , and M B denote the active mass, molar volume, and molar mass of the active material, respectively, and here, we mainly use LiNi 0.5 Co 0.2 Mn 0.3 O 2 as the basis of calculation. 61 In addition, A is the total contact area between the electrolyte and the electrode, whereas ΔE s and ΔE τ are related to the variation of the voltage, both of which can be derived from the GITT curves ( Figure 4I). As shown in Figure 4H, the GITT curves, obtained in the range of 3.0-4.3 V, were carried out using a current density of 0.1 C to ensure the accuracy of the test. The calculated D Li+ values are shown in Figure 4J,K. Notably, the NCM-MA-R electrode can achieve a higher diffusion coefficient of about 1.0 × 10 −10 cm 2 S −1 , and the values of D Li+ are comparable to NCM-N, as well as significantly higher than those of NCM-non-MA-R, especially during the discharge process. These findings imply that the synergistic repair process can increase the diffusion coefficient of Li-ion even more. To further evaluate the reaction kinetics, the overpotential (OP), incorporating ohmic resistance at various states of the charge, was calculated ( Figure 4J,K). Obviously, we find that with the accumulation of capacity, the OP in both charging and discharging dropped first and then rises, which corresponds to the electrochemical reaction mechanism of the NCM material. More notably, in the (de)lithiation process, the OPs of NCM-MA-R are all much lower than those of NCM-non-MA-R. Moreover, a higher conductivity of NCM-MA-R powders than that of NCM-non-MA-R at different pressures was obtained with the four-point probe technology ( Figure S13). Those imply that the synergistic repair strategy decreases reaction resistance and enhances redox reaction kinetics of the electrode. To further investigate the reason for the difference in lithium storage during cycling, the electrochemical impedance spectroscopies (EIS) of NCM-non-MA-R and NCM-MA-R electrodes were analyzed after various cycles ( Figure 5). All the curves are composed of two semicircles, one small and one great, the first one is related to the surface film resistance (R f ), and the second one can be termed the charge transfer resistance (R ct ). Based on the equivalent circuit ( Figure 5D), we fit the resistance values of NCM-non-MA-R and NCM-MA-R electrodes during cycling, and the results are detailed in Table S5. The change of R f is attributed to the formation of cathode electrolyte interphase. As shown in Figure 5C, during the cycling of the NCM-MA-R electrode, the R f decreases first and then stabilizes around 11 Ω, demonstrating superior interfacial stability. However, the R f of NCM-non-MA-R fluctuates unpredictably, decreasing first and then increasing. What would be unique about NCM-non-MA-R, is that the R f is as high as 62.1 Ω at the beginning of the cycle, presumably caused by the more impurity phases on the surface of electrode material. Combined with the XPS results, it further illustrates that the nonmechanically activated renovation process usually leaves more by-products, such as LiF and Li 2 CO 3 , which have also been proven to be essential in affecting the lithium-ion transport capacity. 37,62 Additionally, the account of R ct for a substantial fraction of overall resistance, especially compared to R s and R f , thus the continual increase in resistance during charge transfer is primarily considered to be the reason for the capacity deterioration. 63 As depicted in Figure 5C, the R ct curve obtained for NCM-MA-R is smooth throughout the cycling, staying around 200 Ω, whereas that of NCM-non-MA-R grows rapidly after 20 cycles and up to 782 Ω after 200 cycles, implying an underpowered electrochemical reaction of NCM-non-MA-R electrode, and corresponding to its capacity decay pattern in cycling. Overall, the NCM-MA-R electrode exhibits superior kinetic properties and favorable interfacial performance during electrochemical cycling. It provides strong evidence to validate the effectiveness of the synergistic electrode repair strategy based on deep mechanochemical activation and heat treatment.
To examine the structural evolution of the electrodes before and after renovation during cycling, the XRD spectrum was acquired in situ for their first two chargedischarge processes at 1 C. Except for the impurity peaks emerging at 44.7 • and 65 • derived from Al foil, and 65.3 • from metal Be, all diffraction patterns can be indexed to The patterns of the regular peaks shifting and splitting reflect a wide range of structural changes caused by Li + de/intercalation. During the whole charge process, the (1 0 1), (1 0 4), and (1 1 0) peaks shift to a higher angle, which suggests the contraction of the a-axis due to the decrease in TM ionic radius. For (0 0 3) and (0 1 8) peaks, the position initially shifts to a lower angle with the removal of Li + and then back to a high region further charging above 4.2 V, directly reflecting the expansion and contraction of the c-axis parameters. 64,65 Moreover, the above peak showed an opposite variation during discharge. As known from the detection results, both renovated samples exhibit the same tendencies in peak evolution ( Figure 6B,C), whereas NCM-S measured weaker peaks owing to its lower capac-ity ( Figure 6A). Visible distinctions are detected through tracking the (0 0 3) peaks. The corresponding peak offsets can be precisely checked after full discharge to 3 V. From the test results, it can be seen that the (0 0 3) diffraction peak of the NCM-S electrode is shifted by 0.08 • from the initial position to a low angle in the fully discharged state of the first cycle, implying a lattice expansion in the c-axis direction, which may be attributed to insufficient reinsertion of Li + from NCM-S electrode during discharge. Moreover, the offset even intensifies by 0.04 • after the second cycle ( Figure 6A,D). Upon long-term de-lithiation and lithiation, the irreversible structural transitions are gradually accumulated, triggering an undesirable voltage and capacity decay. By comparison, NCM-non-MA-R ( Figure 6B) and NCM-MA-R ( Figure 6C) display a more symmetrical phase transition. After two cycles, the (0 0 3) diffraction peak of NCM-non-MA-R shifted 0.04 • toward F I G U R E 7 Mechanistic diagram of the polycrystalline NCM morphology and crystal structure evolution during the degradation and renovation.
the lower angle ( Figure 6E). Fortunately, the 2θ position of NCM-MA-R is hardly shifted to a lower angle, which indicates that the electrode has favorable structural reversibility and stability with less lattice changes along the c-axis after structural reconstruction in cycling ( Figure 6F).
From the above results, it is found that the improvement of the phase composition and crystal structure of the electrode materials directly guide the improvement of the electrochemical activity. The rechargeability of the degraded electrode is efficaciously restored by the synergistic effect of mechanochemical activation and heat treatment. In order to clearly illustrate the superiority and feasibility of the LiNi 0.5 Co 0.2 Mn 0.3 O 2 cathode directly repair strategy in this work, we have summarized previous studies and compared, as shown in Table S6. In terms of technology, our direct solid-phase recycling method involves simpler equipment, less raw material, and lower costs than the molten salt and hydrothermal methods, as well as no secondary contaminants. In the aspect of performance, the positive effect of the mechanochemical activation results in a higher electrochemical lithium storage performance of our renovated porous electrode material than that of similar solid-phase regenerated materials in the past. The performance of the renovated material can reach commercial levels, facilitating the efficient recycling and reusing of resources. This also suggests that the application and regulation of mechanochemical activation has great significance on the road to achieve the goal of short-range direct recycling of spent LIBs.

Repair mechanistic insights
Based on the above discussion, the mechanism of morphology and crystal structure evolution of polycrystalline NCM during degradation and renovation was summarized, as shown in Figure 7. It can be found that within the synergistic repair strategy of NCM electrode materials proposed in this paper, the mechanochemical process plays a crucial role, which mainly includes two aspects. On the one hand, as a method of particle morphology remodeling, mechanochemical activation changes the shape of the particles by inducing changes in their energy, causing fragmentation, aggregation, and homogenization, breaking the original cracked structure and preparing relatively homogeneous porous secondary aggregation. Compared to repair materials without morphology changes, our regenerated secondary aggregate electrodes are remarkably efficient for electrolyte penetration and ions transport.
On the other hand, as an effective means of lithium replenishment, mechanochemical activation creates a more favorable chemical environment for the repair process. Under the action of mechanical energy, the lithium salt is uniformly attached to the NCM particles, which shortens the path of lithium replenishment and facilitates the movement of Li + into the lattice. Meanwhile, the mechanical process breaks the unfavorable impurity phase at the interface and directly reduces the interfacial resistance. It enables the formation of solid solution system of lithium salts during heat treatment, further promoting their removal and transformation. More importantly, a possible favorable reaction environment is created for partial F-doping. Previous studies have also confirmed the feasibility and importance of mechanical activation in the fluorination reaction of electrode materials. 26 The heat-treatment process supplies the required energy to induce repair reactions and reversible phase transitions, thus effectively restoring the crystal structure and lithium storage properties of the degraded electrodes.

CONCLUSION
In this study, we systematically evaluated the structural degradation mechanism of end-of-life NCM electrodes through experimental characterization and theoretical simulation. The failure mode of NCM cathode is synergistic with multiple behaviors. For these three key factors of degradation: cracking of the polycrystalline cathode particles, disorderliness of crystal structure, and accumulation of interfacial insulating phases, we adopted a synergistic repair strategy to repair spent NCM based on mechanochemical activation and heat treatment. It is found that a deep lithiation is carried out by mechanochemical activation, providing a more adequate mixture of lithium salts and facilitating the restoration of the crystal structure. In addition, the morphology of the spent electrode material is reconstructed into porousshaped single crystal aggregates and intergranular cracks are eliminated. Meanwhile, it also allows the insulating layer on the spent NCM surface to be broken and dispersed, which facilitates its removal and transformation during the heat-treatment process, thus reducing the interfacial resistance. For the intractable LiF, the mechanical activation provides a favorable condition for its gradient distribution and partial conversion into F-doping, enabling an increase in the amount of oxygen vacancies and highly reactive Ni 2+ on the surface, which improves stability and safety of the regenerated material. Owing to the restoration of the above degradation behaviors, the repaired NCM-MA-R exhibits remarkable electrochemical lithium storage performance comparable to the fresh NCM material. This work proves the outstanding practical application prospects of the synergistic repair strategy based on mechanochemical activation and heat treatment, and also suggests a feasible, general strategy for sustainable reusing similar spent cathode materials.

Renovation process
All of the spent LIBs used in the study were provided by our partner company, whose capacity was reduced to less than 75% after cycles. The spent LIBs were discharged and then disassembled in a glove box filled with inert argon gas. Manual separation of cathodes and anodes was selected instead of crushing and sieving to reduce impurities and avoid unnecessary negative effects during the renovation process. The electrodes were then immersed in dimethyl carbonate to remove and recover the residual electrolyte. N-methyl-2-pyrrolidone (NMP) was employed to dissolve the binder (PVDF), weakening the bond between the aluminum foil and the cathode materials, thus separating them. After drying, the spent cathode powder and the complete Al foil were obtained. In addition, NMP was recovered and reused several times by reduced pressure distillation. The dried cathodes were calcined at 600 • C to remove conductive carbon and binder impurities, and then the spent cathode powder to be repaired was obtained named NCM-S here. Simultaneously, the generated gases were collected and purified to phase down environmental pollution.
Lithium replenishment was performed according to a stoichiometric ratio of Li and TM elements (Ni, Co, Mn) for 1.05:1. LiOH⋅H 2 O as the lithium source was mixed with NCM-S to achieve the desired stoichiometric ratio, and the mixture was named NCM-non-MA. Subsequently, it was mechanically activated in depth using a ball mill, consisting of a short high-energy crushing process (480 rpm 30 min) and a long deep grinding process (360 rpm 10 h), and the product of the process was defined as NCM-MA. Then, the mixture was calcined at 500 • C for 2 h and 800 • C for 16 h in an atmosphere of oxygen to repair the damage of chemical composition and crystal structure during cycling. The regenerated samples only by grinding without mechanochemical activation were defined as NCM-non-MA-R, and the regenerated samples with mechanochemical activation after grinding were defined as NCM-MA-R. For comparison, the commercial LiNi 0.5 Co 0.2 Mn 0.3 O 2 was denoted as NCM-N.

Materials characterization
ICP-optical emission spectroscopy (iCAP 7400, Thermo Scientific, USA) was used to determine the metal content. The N 2 adsorption-desorption isotherms were measured at −196 • C using an automatic surface area analyzer (Micromeritics ASAP 2020), and the pore size distributions were determined through the Barrett-Joyner-Halenda method. The morphology of NCM material was inspected by an SEM (FEI Quanta SU8010, HITACHI) and TEM (Tecnai G2 F30 FEI Model). XPS were obtained with an ESCALAB 250Xi analyzer, and the binding energy was calibrated by C 1s peak (284.8 eV). Solid state NMR measurements were performed on the as-prepared materials, with natural lithium isotopic abundance using a Bruker AVIII400WB spectrometer. The 7 Li spectra were collected at a spinning rate of 25 kHz, pulse width of 2.5 μs, using a recycle delay of 16 s, 128 transients. Crystallinity and phase analysis were characterized by using a D2 diffractometer (Bruker, Billerica, MA, USA), which operated with Cu K α radiation generated at 10 mA and 30 kV. EPR was carried out by Bruker A300 under the modulation amplitude of 1G. TG analysis and differential scanning calorimetry of the samples were performed on SDTQ600 (TA, USA) at a heating rate of 10 • C min −1 from 30 to 800 • C with air flow.

Electrochemical measurements
The electrochemical performances of NCM materials were evaluated using a CR2025 coin-type cell. The NCM cathodes were prepared by mixing the material with PVDF and conductive agents (acetylene black) at a ratio of 8:1:1 then dried at 100 • C for 12 h. The electrode was punched into cathode discs with 14 mm diameter. The mass loading of the active material was around 1.5 mg/cm 2 . Then, CR2025 coin cells were assembled in an argon-filled glove box with lithium metal as the counter electrode, Celgard-2320 as the diaphragm, and 1 M LiPF 6 in EC:DEC = 1:1 (V:V) with 5% FEC as the electrolyte. Charge/discharge tests were performed in the voltage range of 3.0-4.3 V (vs. Li + /Li) by the Neware Battery Test System (Neware Technology Ltd., China). The GITT test was performed at 0.1 C with a relaxing time of 40 min. EIS analyses were performed using an electrochemical workstation (CHI660E, Chenhua, China). All cells were rested for 12 h at room temperature before testing.

DFT calculations
The spin-polarized DFT calculations were implemented in the Vienna Ab initio simulation package. 66 The ion-electron interactions were expressed by the projector-augmented wave method, 67 and the electron exchange-correlation parametrization was accounted for by the generalized gradient approximation with the Perdew-Burke-Ernzerhof exchange-correlation parametrization. 68 The kinetic energy cutoff was set to 520 eV for all calculations. The total energy difference was less than 10 −5 eV, and the relaxation convergence criterion was set at 0.02 eV Å −1 . For Li 2 CO 3 , LiF, and LiOH, the (2 × 2) supercells were constructed from the primitive cell, calculated using a 3 × 3 × 2 Monkhorst-Pack mesh. Activation energies for the ionic diffusion were calculated using the climbing image nudged-elastic band algorithm. 69 To model the diffusion pathways of Li-ions, one Li atom was removed from these considered systems.

A C K N O W L E D G M E N T S
This work was finally supported by the National Natural Science Foundation of China (NSFC Grants 52074098), the State Grid Heilongjiang Electric Power Co., Ltd., Technology Project Funding (research on sodium vanadium phosphate cathode material and low-temperature performance of sodium ion battery for energy storage, 52243723000C), the Foundation of Key Program of Sci-Tech Innovation in Ningbo (2019B10114), and the Natural Science Foundation of Heilongjiang Province (YQ2021E039).

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.