Integrated Local‐Microstructure Engineering Toward Mechanochemically Robust Ultra‐High Nickel Cathodes

ABSTRACT

Ni‐rich layered cathodes show exceptional promise for next‐generation high‐energy‐density lithium‐ion batteries. However, higher nickel utilization in layered cathodes readily exacerbates both bulk mechanical failure and interfacial chemistry instability, significantly hindering their practical application. Herein, a comprehensive local lattice regulation strategy was proposed via Mg/Nb co‐doping, which constructed a chemically and mechanically co‐robust LiNi_0.95Co_0.03Mn_0.02O₂ cathode from surface to bulk. The surface‐reconstructed ultrathin disordered rock‐salt phase effectively stabilized the electrochemical interface. Meanwhile, the bulk cation‐disordered structure integrated with a coherent spinel‐like phase mitigated lattice strain and enhanced structural integrity. Therefore, the modified ultra‐high nickel cathode exhibited excellent long‐term cycling stability, high rate capability, and thermal stability. It exhibited high initial coulombic efficiency of 93.24% and a discharge specific capacity of 240.11 mAh·g⁻¹ at 0.1C. It could also deliver a high initial capacity of 210.44 mAh·g⁻¹ at 1C and retain 97.37% of its capacity after 100 cycles. Moreover, it exhibited outstanding performance during cycling at 3C, delivering a superior capacity retention of 81.65% after 500 cycles. It also delivered a remarkable specific capacity of 147.43 mAh·g⁻¹ even at the rate of 15C. This integrated microstructure regulation strategy would pave the way for commercializing ultra‐high‐nickel cathodes in next‐generation high‐energy‐density batteries.

This manuscript presents an integrated microstructure regulation strategy, wherein the surface disordered layer coupled with the coherent spinel‐like phase and local cation‐disordered structure in the bulk, synergistically stabilizes the ultra‐high nickel cathodes. This integrated microstructure suppresses interfacial side reactions and lattice strain while enabling fast Li⁺ transport, delivering exceptional structural stability and electrochemical performance.

1. Introduction

The rapid evolution of the electric vehicle market creates an urgent need for cathode materials with higher energy density, longer cycle life, and lower cost to power the next generation of batteries [1]. Developing Ni‐rich cathodes with higher nickel content can deliver higher energy density and lower‐cost batteries, but often requires compromising cycling life and thermal stability [2, 3, 4, 5]. Under high voltage, the highly reactive Ni⁴⁺ formed on the surface would trigger interfacial side reactions between the cathode and electrolyte, leading to electrolyte decomposition and non‐uniform cathode‐electrolyte interphase (CEI) layer formation [6]. Simultaneously, oxygen release from the lattice would reduce Ni³⁺/⁴⁺ to Ni²⁺, resulting in irreversible transformation from layered to inert rock‐salt phase [7]. The generated thick CEI and NiO rock‐salt barrier collectively hinder interfacial charge transport, accelerating capacity fade and increasing thermal runaway risk [8, 9]. Additionally, the anisotropic volume changes within the bulk, particularly during the H2‐H3 phase transition, could progressively accumulate lattice strain and induce microcracks [10, 11]. As these cracks propagate from the interior to the surface, electrolyte would penetrate into the secondary particles, further triggering side reactions, thickening the NiO rock‐salt layer, and accelerating capacity degradation [12].

To enhance the surface chemical stability of Ni‐rich cathodes, protective coatings have been widely established to impede the direct contact with electrolyte, thus reducing side reactions [13]. For example, the electrochemically passive compounds, such as metal oxides, fluorides, and phosphates, have been explored as physical barriers [14, 15, 16]. However, precisely controlling the lattice match between the coating species and the layered structure of the Ni‐rich cathode presents a significant challenge for practical application. Furthermore, achieving tunable coating thickness with uniformity remains a hurdle in the manufacturing process [17]. Besides, strategy relying solely on surface modification is inadequate to mitigate both intragranular and intergranular cracking in polycrystalline particles, as these issues are inherent to localized lattice distortions in the bulk. These cracks further facilitate interfacial side reactions and ultimately compromise the effectiveness of surface coatings [18]. Thus far, the practical application of ultra‐high‐nickel polycrystalline materials continues to be hindered by the critical challenge of synergistically stabilizing their bulk and interfacial lattices.

Recently, to enhance the intrinsic mechano‐chemical stability of cathode materials, local structural engineering for the inherent particle architecture is regarded as an effective approach. Regarding surface microstructure engineering, gradient surface architectures that feature a controlled transition from ordered bulk to disordered surface have been adopted. This gradient structure can be fabricated through surface‐enriched doping of elements, such as Mg²⁺, Ta⁵⁺, and Mo⁶⁺, thus forming a thin and compatible disordered surface layer to mitigate interfacial side reactions [19, 20, 21]. To address anisotropic lattice variations and the subsequent formation of microcracks, bulk structural reinforcement has been strategically employed [22, 23, 24]. For instance, constructing an appropriate cation‑disordered structure in the bulk can effectively suppress the anisotropic volume change and improve the Li⁺ diffusion of Ni‑rich cathodes [25, 26]. On the one hand, within the cation‐disordered structure, Ni_Li sites act as pillars to mitigate c‐axis lattice deformation [23]. Meanwhile, Li_Ni sites transform the Li⁺ diffusion path from the original 2D into 3D mode, thereby promoting the Li⁺ diffusion [27]. On the other hand, Ni_Li sites also increase the formation energy of oxygen vacancy, thus improving the reversibility of lattice oxygen [28]. Besides, constructing locally coherent heterogeneous phases within the bulk lattice via heteroatom doping (e.g., Nb⁵⁺, W⁶⁺) can restrain lattice strain and microcracks through cross‐fixation effect. This heterogeneous architecture also provides an expressway for Li⁺ diffusion by reducing the diffusion energy barrier [29, 30]. Although targeting either the bulk or surface can resolve localized instability, such isolated strategies often fail to reconcile the coupled mechanochemical failure driven by the interdependent exacerbation between bulk structural degradation and interfacial parasitic reactions. Therefore, constructing an integrated surface‐to‐bulk microstructure to cooperatively enhance both the surface chemical stability and bulk mechanical stability of Ni‐rich cathodes remains a significant challenge.

In this work, we achieved an integrated surface‐to‐bulk local microstructure regulation by Mg/Nb co‐doping strategy. The modified NCM95 cathode (NCM95‐MN) exhibited multi‐regional local structural reconstruction. First, the surface gradient rock‐salt structure induced by Mg‐enriched doping helped suppress side reaction between cathode and electrolyte. Second, bulk local cation‐disordered structure caused by lithium deficiency promoted Ni_Li occupation to restrain structure collapse. Third, intralattice‐bonded spinel‐like phase induced by Nb doping contributed to maintaining robust structure while promoting Li⁺ diffusion. Therefore, benefiting from an integrated local structures engineering strategy, NCM95‐MN exhibited significantly suppressed anisotropic lattice strain, microcracks and interfacial side reactions. Consequently, it delivered higher initial discharge capacity of 210.44 mAh·g⁻¹ and better capacity retention of 97.37% after 100 cycles at 1C, as well as 81.65% capacity retention after 500 cycles at 3C. This work demonstrated that a rationally designed local microstructure regulation strategy could synergistically enhance the surface and bulk stability of ultra‐high‐nickel cathodes while maintaining favorable thermal stability.

2. Results and Discussion

Figure 1a demonstrates the integrated local microstructure design strategy, which introduced a disordered phase on the surface and integrated intralattice‐bonded spinel‐like phase coupled with local cation‐disordered structure into bulk. To facilitate the construction of surface disordered phase and intralattice‐bonded spinel‐like phase, Mg and Nb were selected based on the following considerations. First, Mg tended to segregate on the surface to form a disordered rock salt phase, and Nb could be doped into the bulk to induce spinel‐like phase [19, 29]. Second, low‐valence Mg²⁺ at Li site could restrain the collapse of Li layer, meanwhile, high‐valence Nb⁵⁺ at TM site helped to reduce the lattice strain [31, 32]. Nb also regulated grain refinement and introduced bonding phases within the lattice, thereby homogenizing the lattice strain [33]. Third, Mg/Nb co‐doping could also modulate the stoichiometric ratio between the high‐nickel precursor and LiOH, thereby inducing a cation‐disordered structure within the bulk phase and further reinforcing the layered framework.

FIGURE 1.

(a) The schematic diagram of the local microstructure regulation strategy for NCM95‐MN. (b1–b4) EPMA images of morphology and elements (Ni, Mg, and Nb) distribution in the NCM95‐MN. The HAADF‐STEM images for (c) surface gradient structure, (d) coherent spinel‐like phase, and (e) cation‐disordered structure in the NCM95‐MN. The ND patterns of (f) NCM95 and (g) NCM95‐MN.

The morphology of NCM95 and NCM95‐MN was first confirmed by scanning electron microscopy (SEM) coupled with high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM). Figure S1a–d showed that secondary particles of both samples displayed spherical morphology. By comparison, primary particles of NCM95‐MN had been refined with fewer voids existing in the bulk, due to the growth regulatory effect of Nb. Furthermore, the morphology of the single‐element doped samples was also obtained by SEM in Figure S1e,f. The images revealed that while Mg incorporation had no appreciable effect on primary particle size, Nb doping facilitated substantial grain refinement, as reflected in the reduced particle size of NCM95‐Nb and NCM95‐MN relative to NCM95 and NCM95‐Mg. Electron probe microanalyzer (EPMA) in Figure 1b and Figure S2a–c verified the distribution of elements in NCM95‐MN, in which the signal of Mg was stronger on the edge of the particle and its concentration gradually decreased from surface to interior. In contrast, the signal and concentration variation trend of Nb was the same as transition metals (Ni, Co and Mn), showing uniform concentration distribution in the interior. Besides, HAADF‐STEM complemented by corresponding fast Fourier transform (FFT) analysis, was used to characterize the local crystal structure within the particles. Figure 1c revealed a gradual transition from a disordered rock‐salt phase to a layered phase in the surface region of NCM95‐MN. Additionally, as shown in Figure S3, the low‐magnification HAADF‐STEM image, which encompassed a wider range of the particle, confirmed that the cation‐disordered surface layer was widely distributed on the surface of NCM95‐MN. Furthermore, the results of electron energy loss spectroscopy (EELS) showed that Ni L‐edge pre‐peak shifted toward lower energy loss in Figure S4a–d further confirmed the altered local chemical states of Ni toward the particle surface. Meanwhile, the cation‐disordered layer exhibited a consistent thickness of approximately 5 nm along the particle periphery. These results indicated that a local disordered phase with a lower valence of Ni had been successfully induced by surface‐enriched doping of Mg, which would act as a physical barrier to reduce side reactions between the cathode and electrolyte.

Besides, the spinel‐like phase that consisted of two coherent planes was also observed in the NCM95‐MN. In Figure 1d, the intersect angle between these two planes was measured as 110°, which corresponded to the intrinsic angle between (003) plane and (104) plane [22]. Similarly, as shown in the low‐magnification STEM images in Figure S6a,b, the coherent spinel‐like phase was also widely distributed within the particles. The crossing fixation effect of this special structure further helped to maintain the structure integrity of NCM95‐MN. Furthermore, the geometric phase analysis (GPA) of this spinel‐like region showed homogeneous color distribution, revealing the coherent growth between the twin planes did not induce excessive lattice strain. This suggests that the spinel‐like region was well integrated within the bulk.

Moreover, the bright dots observed in the bulk Li layer in Figure 1e demonstrated that partial Ni²⁺ occupied within the Li layer. This local cation‐disordered structure could restrain the lattice collapse in the direction of c axis, thereby reducing microcracks during cycling [23]. By contrast, the HAADF‐STEM images and EELS spectrum in Figure S5a–c showed that NCM95 without Mg/Nb doping remained ordered structure from bulk to surface. The locally disordered atomic arrangement was further verified by neutron diffraction (ND, Figure 1f,g) and X‐ray diffraction (XRD, Figure S7a,b). The XRD patterns confirmed that both samples crystallized in the α‐NaFeO₂ structure with good crystallinity. Moreover, it should be noted that the I(003)/I(104) intensity ratio in XRD patterns was governed not only by the degree of Li⁺/Ni²⁺ cation mixing but also by the preferential crystallographic orientation of primary particles. A higher ratio of I(003)/I(104) was generally associated with a lower degree of antisite defects or a more pronounced preferential grain growth along the direction parallel to the (003) planes [34, 35, 36, 37, 38]. In the present case, the I(003)/I(104) intensity ratio increased from 1.41 for NCM95 to 1.57 for NCM95‐MN. However, this increase did not necessarily indicate reduced cation mixing. Cross‐sectional HAADF‐STEM images in Figure S1c,d revealed that the primary particles in NCM95‐MN exhibited a more evident radially oriented elongation along the (003) planes, confirming the existence of preferential crystallographic orientation [39]. This morphological feature, rather than a lower Li⁺/Ni²⁺ disorder, accounted for the elevated intensity ratio. Rietveld refinement of the XRD patterns yielded a Li⁺/Ni²⁺ cation mixing degree of 3.1% for NCM95‐MN, which was indeed higher than that of NCM95 (2.6%). Given the inherent low sensitivity of XRD to light elements such as lithium, ND measurements were further performed to provide complementary quantification. The ND refinement yielded Li⁺/Ni²⁺ cation mixing degrees of 1.9% for NCM95 and 3.1% for NCM95‐MN, in agreement with the XRD refinement trend and further confirming the elevated antisite defect concentration in NCM95‐MN. (Detailed refinement parameters were provided in Tables S1–S4)

Furthermore, the XRD pattern of the single‐element doped samples (NCM95‐Mg and NCM95‐Nb) in Figure S8 showed both of them exhibited increased I(003)/I(104) intensity ratio relative to NCM95, which indicated a lower degree of Li⁺/Ni²⁺ disorder for them compared to NCM95‐MN. Regarding the (003) peak, NCM95‐Mg displayed a diffraction angle comparable to that of NCM95, whereas NCM95‐Nb and NCM95‐MN exhibited a discernible shift toward a lower angle. In conjunction with the Rietveld refinement results for NCM95‐MN, this phenomenon indicated that the incorporation of Nb was responsible for the expansion of the c‐axis lattice parameter. These results demonstrated that multiple local structures had been successfully integrated within both the bulk and surface of ultra‐high nickel particles. Moreover, to investigate the effect of Mg/Nb co‐doping and the lithium‐deficient design on surface residual lithium species, XPS depth profiling was performed on the pristine NCM95 and NCM95‐MN electrodes before cycling in Figure S9. The O 1s spectra revealed that NCM95‐MN exhibited consistently lower Li₂CO₃ peak intensity and correspondingly higher lattice oxygen peak intensity than NCM95 at all probed depths. Additionally, with increasing etching depth, the attenuation of the Li₂CO₃ signal was more pronounced, and the enhancement of the lattice oxygen signal was more rapid for NCM95‐MN. These results demonstrated that Mg/Nb co‐doping combined with the lithium deficiency design effectively reduced the accumulation of residual lithium species on the particle surface, allowing the clean lattice oxygen framework to be exposed at shallower depths.

To evaluate the effect of local microstructure regulation, the electrochemical performance of NCM95 and NCM95‐MN cathodes was investigated by testing coin‐type cells within 2.8–4.3 V. The galvanostatic charge and discharge curves at the first cycle in Figure 2a revealed that NCM95‐MN achieved a higher initial Coulombic efficiency than NCM95, leading to an enhanced discharge capacity of 240.11 mAh·g⁻¹ at 0.1 C. Besides, although the potential difference between the H2‐H3 oxidation and reduction peaks of both samples was 0.03 V, the H2‐H3 phase transition peak in NCM95‐MN (Figure 2b) appeared broadened and less intense, indicating a more gradual structural evolution. Given that this transition typically triggers detrimental effects, such as microcracking, gas release, and cation leaching, thereby its mitigation is crucial for enhancing cycling stability [40]. Benefiting from improved structural stability, NCM95‐MN exhibited a high initial discharge specific capacity of 210.44 mAh·g⁻¹ at 1 C and better capacity retention of 97.37% after 100 cycles in Figure 2c, reflecting its excellent cycling stability. Moreover, different Mg/Nb ratios were screened during optimization. In Figure S10a,b, a lower loading (0.5 mol% Mg and Nb) yielded insufficient stability enhancement, whereas a higher loading (2 mol% Mg and Nb) caused significant capacity sacrifice despite improved retention. Hence, the selected 1 mol% Mg and Nb composition was identified as the optimal trade‐off point. The differential capacity (dQ/dV) profiles were further analyzed to investigate the phase transition dynamics during cycling. After 100 cycles, NCM95‐MN exhibited less H2‐H3 peak intensity reduction and minor peak shift compared to NCM95 in Figure 2d,e, further indicating its remarkable structure stability. Moreover, the integrated spinel‐like phase acted as a fast ionic conductor, effectively boosting Li⁺ diffusion kinetics. Consequently, NCM95‐MN exhibited enhanced rate capability across various C‐rates (Figure 2f), delivering a remarkable specific capacity of 147.43 mAh·g⁻¹ even at a high rate of 15C. Notably, NCM95‐MN also showed enhanced high‐temperature cycling stability (80.33% retention after 100 cycles at 60°C), significantly outperforming the 63.56% of pristine NCM95 in Figure 2g. These findings demonstrated that a synergistic microstructural strategy from surface to bulk was instrumental in bolstering the thermal runaway resistance of ultra‐high nickel cathodes. Besides, the temperature‐dependent electrochemical impedance spectroscopy (EIS) measurement was performed on both NCM95 and NCM95‐MN, and the charge‐transfer activation energy (Ea) was extracted using the Arrhenius equation in Figure S11a–c. The refined activation energy for NCM95‐MN was 57.95 kJ/mol, which was slightly lower than that of NCM95 (62.05 kJ/mol). This result provided direct thermodynamic evidence that the cation‐disordered structure in NCM95‐MN did not impede Li⁺ transport as conventionally expected. However, the modest reduction in Ea indicated that the specific local microstructure configuration, which was characterized by Mg/Nb co‐doping and the formation of coherent spinel‐like domains, created a lower‐energy migration pathway for Li⁺. Benefiting from the dual enhancement of kinetics and structural stability, NCM95‐MN exhibited outstanding performance during extended cycling at a high rate of 3C, delivering a superior capacity retention of 81.65% after 500 cycles at 30°C (Figure 2h). Furthermore, to verify the synergistic advantage of the Mg/Nb co‐doping strategy, the electrochemical performance of single‐element doped samples NCM95‐Mg and NCM95‐Nb (with a doping ratio of 1 mol%) was also evaluated. As shown in Figure S12a,b, it can be concluded that although single‐element doping provided a certain enhancement in electrochemical performance compared to that of NCM95, yet their electrochemical performance remained inferior to that of NCM95‐MN. This performance gap clearly demonstrated that individual Mg or Nb doping alone was insufficient to achieve the comprehensive enhancement delivered by the co‐doping strategy.

FIGURE 2.

(a) The galvanostatic charge–discharge curves and (b) the dQ/dV curves of NCM95 and NCM95‐MN at the first cycle within 2.8–4.3 V at 0.1C at 30°C. (c) The cycling performance of two samples within 2.8–4.3 V at 1C at 30°C. The dQ/dV curves of (d) NCM95 and (e) NCM95‐MN at different cycles within 2.8–4.3 V at 1C at 30°C. (f) The rate performance of two samples within 2.8–4.3 V at 30°C. (g) The high temperature cycling performance of two samples within 2.8–4.3 V at 1C at 60°C. (h) The rapid charge performance of NCM95‐MN within 2.8–4.3 V at 3C at 30°C.

To elucidate the structural reversibility enhancement of NCM95‐MN via integrated microstructure regulation, in situ XRD was used to track the structural evolution during the first two cycles at 0.2 C (2.8–4.3 V). Upon charging to 4.3V, clear differences of (003) peaks and lattice parameter c for two samples appeared in Figure 3a,b. NCM95 showed rapidly broadened (003) peak and larger c‐axis contraction at the end of charge. In contrast, NCM95‐MN exhibited narrower (003) peak shift and more moderate c‐axis shrinkage. The c‐axis contraction during H2‐H3 phase transition originated from the alleviative electrostatic repulsion, which was induced by the enhanced Ni─O orbital hybridization and substantial Li vacancies generated at the end of charge [41]. Correspondingly, NCM95‐MN exhibited smaller volume changes (6.26%) than NCM95 (8.08%). In addition, NCM95‐MN also showed narrower (101) shift toward a higher angle and a slighter a‐axis contraction compared to NCM95 in Figure S13a,b, suggesting its better crystal lattice stability at the planar level. Further analysis of the c/a lattice parameter ratio could reflect the characteristics of anisotropic stress within the crystal structure. As shown in Figure 3b, NCM95‐MN exhibited a 75.6% lower cumulative velocity of shear stress c/a against NCM95, indicating effective restraint of lattice strain during the H2‐H3 phase transition.

FIGURE 3.

(a) The variation of (003) peak for NCM95 and NCM95‐MN obtained from in situ XRD. (b) The variation of c‐axis lattice parameters, volume, and sheer stress c/a in the NCM95 and NCM95‐MN. (c) The R space curves and (d) the wavelet transform of NCM95 and NCM95‐MN obtained by the EXAFS spectrum before and after cycling.

To further confirm the reversibility of Ni chemical states in NCM95‐MN, an XAFS (X‐ray absorption fine structure) spectrum was performed. As shown in Figure S14a,b, NCM95‐MN exhibited a slighter Ni K‐edge displacement than NCM95 after extended cycling, suggesting that the tailored microstructure effectively suppressed the formation of irreversible Ni⁴⁺, which is highly reactive and detrimental to stability [42]. Besides, the shortened Ni─O and Ni─TM interatomic distance also could be observed in the extended X‐ray absorption fine structure (EXAFS) spectrum in Figure 3c. By comparison, the intensity fluctuations of the Ni─O and Ni─TM peaks were minimal in NCM95‐MN, indicating that the lattice distortion within the local coordination environment was significantly alleviated. Noticeably, the intensity of the Ni─O peak increased significantly in NCM95, pointing to local coordination structure of Ni changed obviously, which was attributed to the generation of lithium vacancies during irreversible lithium intercalation and deintercalation [43]. Besides, the Ni─TM peak intensity of NCM95 also decreased more, demonstrating reduced order and more defects [44]. Furthermore, the wavelet transform analysis in Figure 3d revealed that the distribution of the Ni─O peak in the NCM95 significantly contracted after cycling, signifying poor irreversibility of the Ni redox reaction. In contrast, NCM95‐MN exhibited a highly reversible Ni redox process, as evidenced by its well‐preserved wavelet transform peak intensity and distribution.

To gain deeper insights into the benefits of local microstructure modulation, multi‐scale electron microscopies via SEM and TEM were employed to reveal the internal‐to‐external structural changes. As shown in Figure 4a,b, obvious intergranular cracking occured in NCM95 after cycling. The propagation of these severe microcracks could facilitate the exposure of nascent surfaces to the electrolyte, thereby exacerbating parasitic side reactions concomitant with irreversible phase transitions [45, 46]. Furthermore, cross‐sectional SEM images on multiple secondary particles for both samples were also captured. The cycled NCM95 exhibited more intergranular microcracks and a substantial population of fragmented particles, with cracking observed in the majority of examined cross‐sections in Figure S15a,b. In contrast, the incidence of microcracking in cycled NCM95‐MN particles was substantially reduced in Figure S15c,d. As a result, NCM95 exhibited severe surface lattice degradation into a thick rock‐salt phase, along with the formation of a dense CEI film (Figure 4c,d). The rock‐salt phase generated after cycling was an inert phase with low electronic and ionic conductivity, which could obstruct lithium‐ion diffusion channels and result in capacity loss [47]. In contrast, fewer intergranular microcracks could be observed in NCM95‐MN (Figure 4e,f). Besides, owing to the local microstructure regulation, the cycled NCM95‐MN not only retained its intrinsic disordered rock‐salt phase with a maintained thickness of approximately 5 nm at the surface, but also preserved a mechanically robust layered structure from the bulk to the surface (Figure 4g). Additionally, the formation of a thin and even CEI film in Figure 4h further demonstrated that interfacial side reactions were effectively inhibited in NCM95‐MN. GPA analysis derived from Figure 4c1,c2,g1,g2 also revealed that the strain within the layered domains of the NCM95‐MN bulk was significantly suppressed. Moreover, the conclusion that lattice strain was suppressed in NCM95‐MN remained valid when the geometric phase analysis was extended to encompass a larger field of view in Figure S16a,b.

FIGURE 4.

The SEM images of cycled (a, b) NCM95 and (e, f) NCM95‐MN. (c) The TEM image of NCM95, in which region I and region II were analyzed by fast Fourier transform (FFT) maps, and the strain distribution in the region of (c1) was indicated by GPA analysis in (c2). (d) The CEI generated on the surface of cycled NCM95. (g) The TEM image of NCM95‐MN, in which region I was analyzed by fast Fourier transform (FFT) map, and the strain distribution in the region of (g1) was indicated by GPA analysis in (g2). (h) The CEI generated on the surface of cycled NCM95‐MN. (i) The load‐depth curve and (j) the elastic modulus and hardness of the materials obtained from the nanoindentation test of cycled NCM95 and NCM95‐MN.

Additionally, to investigate the mechanical strength of the samples, nanoindentation tests at the electrode level after cycling were also conducted. As shown in the load‐depth curve in Figure 4i, during the loading stage, the indentation depth of NCM95‐MN was smaller than that of NCM95 under the same load, indicating that NCM95‐MN had higher hardness and thus stronger resistance to local plastic deformation. During the unloading stage, the slope of the curve measured for NCM95‐MN was also larger than that of NCM95, indicating that NCM95‐MN has a faster elastic recovery, thereby proving its stronger ability to resist elastic deformation. At the same time, the residual depth of NCM95‐MN after unloading was also smaller than that of NCM95, further verifying its stronger ability to resist plastic deformation. Specifically, the elastic modulus and hardness values were calculated from the load‐depth curve in Figure 4j. The elastic modulus and hardness of NCM95‐MN measured were 6.43 and 0.0865 GPa, respectively, while those of NCM95 were 5.20 GPa and 0.0556 GPa. These results further quantitatively verified that the NCM95‐MN after cycling had stronger resistance to plastic and elastic deformation, thereby possessing higher mechanical strength. These suggest that the integrated local microstructures could effectively mitigate stress accumulation, thereby substantially lowering the risk of microcrack formation and parasitic interfacial side reactions.

Furthermore, XRD patterns in Figure S17 showed that the (003) reflection of NCM95 exhibited a pronounced shift toward lower angles and significant broadening after cycling. In contrast, NCM95‐MN exhibited negligible variation, underscoring its exceptional structural reversibility. Subsequently, further Rietveld refinement was performed on the XRD patterns in Figure S18. Specifically, after 100 cycles, the degree of Li⁺/Ni²⁺ cation mixing in NCM95 increased from 2.6% to 3.08%, whereas the degree of Li⁺/Ni²⁺ cation mixing of NCM95‐MN remained nearly unchanged. Furthermore, the relative change in the c‐axis lattice parameter was significantly smaller for NCM95‐MN (0.0072%) compared to NCM95 (0.25%) (more detailed refinement parameters provided in Tables S5 and S6). These quantitative results confirmed that the Mg/Nb co‐doping strategy effectively suppressed the accumulation of antisite defects and mitigated anisotropic lattice distortion during prolonged cycling, thereby providing statistical validation for the enhanced bulk structural stability inferred from the TEM observations.

The stability of lattice oxygen is critically linked to the thermal stability and safety of Ni‐rich cathode materials [48, 49]. To evaluate the beneficial effects of the proposed integrated local microstructure modulation on ultra‐high Ni cathodes, we systematically investigated the evolution of lattice‐oxygen covalency, the formation of active oxygen species, and the changes in thermal stability in NCM95‐MN. As shown in the in situ Raman spectra in Figure 5a, the characteristic E_g and A_1g vibration observed at about 450 and 550 cm⁻¹ correspond to O─TM─O in‐plane bending and TM─O stretching, respectively [50]. Upon charging to 4.3 V, the intensities of the E_g and A_1g vibration peaks in NCM95 rapidly decreased, whereas the covalent bonding variation within NCM95‐MN exhibited a gradually weakening trend. When discharged to open circuit voltage (OCV: 3.3 V), the E_g and A_1g vibration of NCM95 failed to recover to their pristine intensity in Figure 5b. In contrast, discharged NCM95‐MN exhibited comparable intensity of both A_1g and E_g vibration to the original state of OCV, indicating good lattice oxygen stability. Moreover, electron paramagnetic resonance spectroscopy (EPR) demonstrated that NCM95‐MN showed lower oxidized oxygen intensity than NCM95 at 4.3 V (Figure 5c), confirming that the detrimental over‐oxidation of lattice oxygen in NCM95‐MN had been effectively mitigated.

FIGURE 5.

(a) The 2D contour maps of in situ Raman spectra for NCM95 and NCM95‐MN during cycling within OCV‐4.3 V at 0.2C. (b) The comparison of Raman spectra at different stages (OCV, charge to 4.3 V and discharged to OCV) in the NCM95 and NCM95‐MN. (c) The EPR spectroscopy of NCM95 and NCM95‐MN. (d) The TG‐MS curves of NCM95 and NCM95‐MN. (e) The DSC curves of NCM95 and NCM95‐MN. (f) The schematic diagram of the relationship between oxygen stability and thermal stability.

To evaluate the oxygen evolution behavior of both samples at high states of charge (SOC), thermogravimetric analysis coupled with mass spectrometry (TGA‐MS) was further employed in Figure 5d. As can be seen, the oxygen release temperature of NCM95‐MN delayed by approximately 20°C, and the oxygen release intensity was much lower than that of NCM95. Meanwhile, the results of differential scanning calorimetry (DSC) in Figure 5e revealed that the exothermal peak temperature for NCM95 was about 225°C, whereas it shifted to a higher temperature of 235°C for NCM95‐MN. Besides, the total heat generated in NCM95‐MN (22.5 J/g) was obviously less than that in NCM95 (119.8 J/g), indicating enhanced thermal resistance in NCM95‐MN. Noticeably, NCM95‐MN even exhibited higher exothermic peak temperature and less total heat release compared to commercial NMC811 (232°C, 103.4 J/g) in the Figure S19 [51]. Figure 5f clarified the relationship between oxygen stability and thermal stability. Benefiting from the integrated local microstructure induced by Mg and Nb co‐doping, the Ni─O covalency in NCM95‐MN was significantly enhanced. This effectively suppressed lattice oxygen over‐oxidation and subsequent oxygen loss, thereby substantially mitigating the risk of thermal runaway.

Eventually, long‐term post‐cycling analysis of NCM95‐MN was further conducted to investigate the stabilization mechanisms enabled by its integrated microstructure. X‐ray photoelectron spectroscopy (XPS) was first performed to map the surface chemical landscapes and component distribution of the two cathodes. Compared to NCM95, NCM95‐MN exhibited substantially lower peak intensities for C─O, C═O and O─C─O₂ in the C 1s spectra, along with attenuated signals for Li₂CO₃ and ROCO₂Li in the O 1s spectra (Figure 6a), underscoring its significant reduction in interfacial side reactions and byproduct accumulation [52]. Besides, decomposition of LiPF₆ salt was evaluated by F 1s spectra in Figure 6a and P 2p spectra in Figure S20. The lower intensity of LiF and Li_xPO_yF_z was observed on the surface of NCM95‐MN, reflecting less LiPF₆ salt decomposed. To further investigate the spatial distribution of CEI degradation products in cycled electrodes, we performed an in‐depth TOF‐SIMS analysis of three types of fragments (carbon‐, fluorine‐, and phosphorus‐containing fragments) to evaluate the CEI. As illustrated in Figure 6b and Figure S21, NCM95‐MN exhibited thinner C₂HO‐ and PO₂‐ fragment signals compared to NCM95, indicating suppressed carbonate solvent degradation and reduced LiPF₆ salt decomposition [53]. Additionally, the deeper penetration of Li₂F₃‐ fragments in NCM95 was attributed to the presence of more intergranular cracks, which facilitated electrolyte infiltration and lead to the formation of a thicker CEI. In contrast, benefiting from the dense and homogeneous CEI formed on the surface of NCM95‐MN (Figure 6b), a significantly weaker Li₂F₃‐ signal was observed in its electrode [54]. The above results suggested that the NCM95‐MN could effectively suppresses CEI growth during long‐term cycling.

FIGURE 6.

(a) C 1s, O 1s, and F 1s XPS spectra of the cycled NCM95 and NCM95‐MN. (b) 3D rendering TOF‐SIMS fragments of C₂HO, Li₂F₃ and PO₂ in the cycled NCM95 and NCM95‐MN. The 2D contour maps of DRT for the cycled (c) NCM95 and NCM95‐MN. (d) The GITT curves of cycled NCM95 and NCM95‐MN at high voltage.

Moreover, the bulk and surface degradation of NCM95 and NCM95‐MN after high‐temperature cycling were also characterized. The XPS spectra (C 1s, F 1s, and P 2p) revealed distinct differences in the CEI chemistry between the two samples in Figure S22a–c. NCM95‐MN developed a fundamentally more stable and passivating CEI compared with NCM95 under elevated temperature conditions. Specifically, the C 1s spectrum of NCM95‐MN exhibited both a reduced proportion and a lower absolute intensity of oxygenated carbon species (C─O, C═O, and O─C─O₂), indicating substantially suppressed solvent decomposition and carbonate formation at 60°C. In the F 1s spectrum, NCM95‐MN displayed a decreased intensity of fluorine‐containing byproducts arising from LiPF₆ degradation, while simultaneously showing a higher area fraction of LiF. This enrichment of LiF was widely recognized as beneficial for forming a stable and thin passivation layer that effectively blocked continuous electrolyte attack without significantly impeding Li⁺ transport. Correspondingly, the P 2p spectrum of NCM95‐MN showed a decreased proportion and diminished intensity of phosphorus‐containing organic species. This trend further corroborated the mitigated decomposition of the LiPF₆ salt. Concurrently, post‐cycling TEM imaging confirmed that NCM95‐MN maintained its structural integrity with the preserved surface disordered phase and bulk layered phase in Figure S23b, whereas NCM95 exhibited pronounced surface and bulk degradation in Figure S23a. Collectively, these results demonstrated that Mg/Nb co‐doping remained highly effective under elevated temperature conditions: the integrated surface‐to‐bulk local microstructure synergistically cooperated with the stable CEI to suppress continuous electrolyte decomposition, mitigate LiPF₆ degradation, and preserve both the interfacial chemical stability and the bulk structural integrity of NCM95‐MN during prolonged high‐temperature cycling.

Furthermore, due to the occurrence of interface side reactions, the transition metal ions on the surface of Ni‐rich cathodes would be chemically eroded by the acidic substances (such as HF) produced by the electrolyte decomposition, thereby leading to transition metal dissolution in the electrolyte. Therefore, investigating the transition metal dissolution situation of two samples after cycling could also reflect their interface stability. The ICP‐MS analysis results in Figure S24 showed the Ni, Co, and Mn dissolution concentrations of NCM95 were 1218, 90, and 183 µg/mg respectively, while the Ni, Co, and Mn dissolution concentrations of NCM95‐MN were significantly reduced to 487, 45, and 118 µg/mg. This indicated that the transition metal dissolution of NCM95‐MN was significantly inhibited, which was mainly due to the pre‐induced surface disordered phase that acted as a passivation layer on the surface of NCM95‐MN, thereby inhibiting the interface side reactions and reducing the transition metal dissolution after cycling.

To investigate the influence of the improved interfacial stability on the electrochemical kinetics of the NCM95‐MN cathode, the impedance spectrum of cycled cathodes was further analyzed by combining the distribution of relaxation time (DRT) in Figure 6c. As displayed in the 2D contour maps of DRT, the τ1 peak in the high‐frequency region corresponded to the resistance of the ohmic contact resistance (R_s). The τ2 peak was the resistance of the cathode/electrolyte interface (R_sf), while the τ3 and τ4 peaks in the mid‐frequency region were assigned to charge transfer resistance (R_ct1 and R_ct2). The τ5 peak in the low‐frequency was ascribed to Li⁺ diffusion inside the material [55]. It is noteworthy that the τ3 and τ4 peaks shifted markedly to lower frequencies and intensify around 4.2 V, aligning with the occurrence of the irreversible H2‐H3 phase transformation. Such behavior indicated an augmented charge‐transfer impedance, which consequently extended the time required for Li⁺ to traverse the electrode/electrolyte interface. In comparison with NCM95, NCM95‐MN exhibited notably smaller peak shifts and intensity variations for τ3 and τ4, indicating reduced charge‐transfer impedance. Furthermore, the τ5 peak intensity in NCM95‑MN was significantly lower, which could be attributed to the coherent spinel‑like phase that reduced the Li⁺ migration energy barrier and enabled rapid Li⁺ diffusion pathways [56]. The enhanced Li⁺ diffusion in the NCM95‐MN was also evaluated by the galvanostatic intermittent titration technique (GITT) in Figure 6d. The results showed that NCM95‐MN exhibited a higher Li⁺ diffusion coefficient, particularly above 4.2 V.

3. Conclusions

In summary, an integrated local microstructure regulation strategy was developed to enhance the mechanochemical stability of the NCM95‐MN cathode from surface to bulk. The surface‐near ultra‐thin disordered rock‐salt layer served as a passivation barrier, effectively mitigating catastrophic side reactions at the cathode‐electrolyte interface. Within the bulk phase of the particles, the collapse along the c‐axis and anisotropic lattice strain were effectively suppressed through stabilization induced by locally disordered cation mixing and the presence of a coherent spinel‐like phase, thereby significantly mitigating particle cracking. Consequently, NCM95‐MN could deliver a high initial capacity of 210.44 mAh·g⁻¹ at 1C, along with excellent cycling stability, retaining 97.37% of its capacity after 100 cycles. It also demonstrated excellent fast‐charge stability, retaining 81.65% of its capacity after 500 cycles at 3C. More importantly, NCM95‐MN also exhibited excellent performance even under elevated temperatures, maintaining over 80% capacity retention after 100 cycles at 1 C and 60°C. This integrated local microstructure regulation strategy provided critical guidance for advancing ultra‐high‐nickel cathode materials, representing a pivotal direction for their practical implementation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma73673‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.