Multi-dimensional correlation of layered Li-rich Mn-based cathode materials

Li-O-Li configuration

LRMO cathode materials consist of LiMO₂ (M = Ni, Co, Mn and so on) and Li₂MnO₃ components, where Li and TM atoms are located at the center of octahedra surrounded by six oxygen atoms and the atom layer alternately arranges as O/Li/O/TM along the c axis. The occupation of extra Li atoms at the TM sites results in a √3a × 3a LiMn₆ superstructure, which gives rise to a change in the coordination environment of oxygen, i.e., three Li-O-TM configurations transform into two Li-O-TM configurations and a special Li-O-Li configuration^[25] (shown in Figure 1A). The intense interaction of Li-O ionic bonds leaves the distribution of coordination electrons between Li and O closer to the oxygen element, forming an isolated unhybridized O 2p orbital. The Li-O-Li configurations stimulate a higher energy band, enabling the charge compensation of oxygen during charging (the downward shift of the Fermi level). TM-O bonds rotate along the other directions, resulting in the coplanar arrangement of adjacent unhybridized O 2p orbitals, which interact with each other to ultimately form O-O dimers. The unhybridized oxygen orbitals caused by the Li-O-Li configurations can also be triggered by other configurations^[26] (such as Li-O-vacancy and Li-O-Na configurations).

Figure 1. (A) Local coordination of oxygen with one Li-O-Li and two Li-O-M configurations and a schematic of the band structure for Li-rich layered oxides^[25]. (B) Schematic of π-type hybridization of the Li-O-Li axis with neighboring TM and the band structure for the π-type hybridized state of the Li-O-Li axis interacted with M t_2g^[27]. (c) Local coordination around Mn³⁺ and Mn⁴⁺ in LiMn₂O₄ (left) and λ-MnO₂ (right) obtained from theoretical calculations. The gray and red spheres represent Mn and oxygen atoms, respectively^[33].

TM-O hybridization

The charge compensation of anions is more closely associated with TMs coordinated by oxygen atoms^[27], which is attributed to the two Li-O-TM configurations and one Li-O-Li configuration. For the typical Mn-based material Li₂MnO₃, Mn bonds with oxygen in a form where the non-bonding oxygen energy band with high activity is separated from the Mn t_2g energy band and electrons directly escape from it, contributing to the severe irreversible release of oxygen^[28]. The introduction of Fe into the inert structure of Li₂TiO₃ successfully motivates the anion redox activity^[13]. This can be attributed to the fact that electron holes tend to be selectively generated on oxygen ions bonded to Fe atoms instead of Ti atoms. The mechanism can also be extended to the stimulation of anions in Co-doped Li₂TiS₃.

In the model Li₂RuO₃ material, the strong covalent interaction between Ru and oxygen causes overlap of the Ru 4d and O 2p non-bonding bands, exhibiting a competitive anion-cation redox coupling reaction with highly reversible anion redox^[10]. The redox activity of Li₂RuO₃ can be regulated by the weaker covalency of the Sn-O bond. Furthermore, the intensely bonding Ir-O immensely suppresses the charge compensation ability of oxygen for Li₂IrO₃ and all the capacity derives from the redox reaction of Ir hybridized with oxygen^[29]. When Ir is partially replaced by Sn, the lower number of valence electrons at the end of charging significantly promotes the ligand-to-metal charge transfer, resulting in short Ir-O π bonds and O-O dimers with a bond length of 1.4 Å. In Li₂RhO₃, with an intense π-type interaction [Figure 1B], the weaker difference in electronegativity between Rh and oxygen causes a more intense TM-O interaction and increases the contribution of O 2p to the high-energy π* orbital. Oxygen exhibits high reactivity under the occupied anti-bonding state as the π electron donor of the central Rh, displaying a novel oxygen redox reaction at low voltages^[30] (which is different from the redox of non-bonding oxygen), where oxygen participates in the reaction prior to Rh. The functional mechanism is demonstrated to be triggered by the oxygen-dominated π* hybrid states.

As a result, the non-bonding O energy band is subtly regulated by the varieties and numbers of adjacent TM elements, owing to the π-type interaction between Li-O-Li configuration and TM t_2g bands. The hybridization state of TM-O is key for adjusting the anion charge compensation ability. The ability to resist oxygen loss and structural deterioration is promoted with the increase in M-O covalent bonds^[31]. The redox mechanism derived from the 3d metals and more heavy metals with a d⁰ electron configuration in Li₂MnO₃ is related with the lone-pair electron consumption of oxygen (locally non-bonding O 2p state), which leads to irreversible anion redox. Electrode materials formed by the coordination hybridization between oxygen and 4d and 5d metals with intermediate dⁿ electron configurations are able to induce reversible cation and anion redox activity to realize reversible extra capacity.

The anion and cation redox ability can be visibly judged according to a Zaanen-Sawatzky-Allen diagram via the comparison of U (Coulombic interaction, depending on the d metal) and D (energy difference between (M-O) and (M-O)*, known as the charge transfer parameter)^[32]. When U << D, the d band established in the (MO)* state locates first on the Fermi level, resulting in pure cation redox. When U > > D, the O 2p state on the p band lies at the Fermi level, thereby motivating anion redox that leads to the generation of highly unstable electron holes and highly activated free radicals in localized states. O-O dimers then form and escape from the lattice, giving rise to the serious irreversible release of oxygen. When U ≈ D/2, the d bands in the antibonding (MO)* states overlap with the O 2p non-bonding band, which stimulates the anion/cation redox mechanism and stabilizes the Fermi level via the decrease in local symmetry in the case of electron loss. Therefore, it is necessary to choose appropriate TM-O hybridization, local coordination and a combination to improve the possibility of reversible anion redox in the actual electrochemical window.

Jahn-Teller effect

The LiMn₆ superstructure formed by Mn plays an important role in motivating anion redox in LRMO cathode materials. Simultaneously, the anions in turn activate Mn to participate more in the redox reaction. The coupling of Mn and oxygen is supposed to have significant synergistic effects on stabilizing (O₂)^n- species and activating more Mn ions to participate in the electrochemical process. However, the valence of Mn undergoes a deep evolution in a wide voltage window of 2.0-4.8 V. The layer structure of LRMOs gradually transforms into the Li₂MnO₄ spinel phase with cycling, which induces Jahn-Teller distortion^[33][Figure 1C] that seriously influences the structural stability and electrochemical performance.

The Jahn-Teller effect refers to the asymmetric occupation of electrons in degenerate orbitals that causes a distortion of the geometric configuration, thus reducing the symmetry of the molecule and the orbital degeneracy, and further decreasing the energy of the system. In a Mn³⁺O₆ octahedron, there is only one electron in the doubly degenerate e_g orbitals (d_x²_-y² and d_z² orbitals) for high-spin Mn^3+[34], which leads to asymmetry of the electron distribution. To reduce the energy of the system, the bond length of Mn-O along the z-axis is elongated and four Mn-O bonds in the horizontal direction are shortened, causing a structural distortion that is not observed in the Mn⁴⁺O₆ octahedron. It is generally considered that the critical mean valence of Mn is 3.5+ when Jahn-Teller distortion occurs. Moreover, unstable Mn³⁺ is prone to undergo a disproportionation reaction of Mn³⁺→Mn⁴⁺ + Mn²⁺. Mn²⁺ dissolves into the electrolyte and deposits on the anode, resulting in a loss of active material and an increase in the electrochemical impedance on the electrode interface. Therefore, it is necessary to regulate the variation of Mn valence during the electrochemical process to decrease the content of Mn³⁺.

Crystal structure and nanodomains

Owing to the composition of the R$$\bar 3$$m LiMO₂ phase and C2/m Li₂MnO₃ component, LRMO materials display complex structural characteristics. The (001) crystal plane of the C2/m phase just overlaps with the (003) crystal plane of the R$$\bar 3$$m phase and both the interplanar spacings are close to 4.7 Å, which causes the two components to appear dissolved at the atomic scale and mixed at the nanometer scale [Figure 2A]. Interestingly, weak peaks at 20-25° appear in the X-ray diffraction (XRD) patterns, owing to the loss of the systematic absence for the LiMn₆ superstructure. Jarvis et al.^[35] demonstrated that Li[Li_0.2Ni_0.2Mn_0.6]O₂ was composed of a solid solution with C2/m monoclinic symmetry and multiplanar defects via the combination of atomic-scale high-angle annular dark-field aberration-corrected scanning transmission electron microscopy (HAADF-STEM) and XRD. In particular, there are two-phase regions, owing to the loss of ordering between Li⁺ and Mn⁴⁺ in the TM layer with less lithium. Yu et al.^[36] also clarified the coexistence of LiMO₂ and Li₂MnO₃ with sizes of 2-4 nm in Li_1.2Ni_0.15Co_0.1Mn_0.55O₂. Yu et al.^[37] revealed the symbiotic heterogeneous interface of rhombohedral LiMO₂/monoclinic Li₂MnO₃ at the atomic scale in Li_1.2Mn_0.567Ni_0.166Co_0.067O₂ via aberration-corrected STEM and computer simulations. The two structures were separated by the (003)_rh and (001)_mon crystal planes along the c-axis [Figure 2B].

Figure 2. (A) Crystal structure of rhombohedral LiMO₂ (left) and monoclinic Li₂MnO₃ (right)^[37]. (B) HAADF images (i) and intensity profiles (ii) of the two-phase intergrowth and heterointerface along the [001]_rh zone axis direction^[37]. (C) Increasing stacking disorder results in asymmetric peak broadening with no decrease in the peak area ratio between the superstructure peaks and the 001 direction (inset)^[41]. (D) Morphological and structural designs for a delocalized excess Li system for enabling high electrode density and reversibility^[20]. (E) 2D-ordered cation layered structure with coplanar unhybridized O 2p orbitals (left) and 3D-disordered cation framework with a random spatial distribution of unhybridized O 2p orbitals (right)^[49]. (F) Mixing energy predicted from bonding model for the solid-solution LiMO₂ (M = Ni, Co and Mn) phase^[51]. HAADF: High-angle annular dark-field.

In addition, stacking faults^[38,39] along the c-axis exist both in the two-phase symbiotic heterogeneous interface and simple structure, which is the root cause of broadened superstructure diffraction peaks. Shunmugasundaram et al.^[40] discovered that the location of the superstructure diffraction peaks varied monotonously with the Ni content, while the probability of stacking faults increased monotonously with the content of Ni according to the XRD patterns fitted by FAULTS. Therefore, the presence of heterovalent ions, such as Ni²⁺, disturbs the stacking order along the c-axis and broadens the superstructure peaks of the material [Figure 2C]. Similarly, House et al.^[41] determined 20% stacking faults in a pristine Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ material. These widespread stacking faults result in local segregation of Li/Mn, expansion of the interplanar spacing and local heterodomains^[42], which influence the extraction/insertion of Li⁺, the structure size in the direction vertical to the TM/Li layer, the coordination environment of lattice oxygen and the distribution of 2p electrons, thereby changing the reaction behavior and activity of LRMOs.

Cation arrangement

The LiMn₆ superstructure is the key factor to induce anion redox and O-O dimers in the TM layer and its distribution deeply affects the whole structural evolution and electrochemical performance of the material^[43]. Hwang et al.^[20] discovered that Li_1.11Mn_0.49Ni_0.29Co_0.11O₂ with a relatively delocalized distribution of the superstructure exhibited more stable lattice oxygen [Figure 2D], which was attributed to the more dispersed distribution of highly active oxygen in an oxidized state that effectively suppressed the polymerization of O-O dimers and the release of O₂ during the extraction of Li⁺. In addition, Song et al.^[44] found that a disordered distribution of Li/Mn in the TM layer improved the symmetry of LRMOs and substantially adjusted the redox activity of oxygen, which helped to realize the reversibility of the structure, thereby alleviating the Jahn-Teller effect, decreasing the band gap and expanding the two-dimensional (2D) diffusion channel of Li⁺. The Li vacancy in the superstructure inhibited the dimerization of O-O in anion redox. This disordered arrangement is also universally applicable to other Li-rich systems. For Li₂RuO₃, the local symmetry around the oxygen ions is adjusted by the disordered arrangement in the Li/TM layer, so that the response structure to anion redox of the material transforms from the original O-O dimers to the retractable O-Ru-O configurations^[45], thereby inhibiting the significant release of oxygen and promoting the cycle stability.

The cation arrangement between the TM and Li layers profoundly influences the coordination environment of oxygen and the whole structural stability of the material^[46-48]. In LRMO materials with 2D cation arrays, the oxygen ions in Li-O-Li configurations interact with adjacent oxygen ions to reduce the O-O distance and their energy in the oxidized state. The interaction of different oxygen ions leads to a decrease in the bond angle of O-TM-O and the overlap of TM e_g and O 2p orbitals^[49], which weakens the bonding of TM-O and damages the stability of the structural framework. Two unhybridized O 2p orbitals are non-coplanar in the structure with disordered three-dimensional (3D) cation arrays and the lack of interaction and bonding between the two orbitals inhibits the dimerization of O-O. This results in a relatively stable lattice structure of oxygen [Figure 2E] that hinders the diffusion and migration paths of Li⁺. Liu et al.^[47] synthesized Ti-doped Li_1.2Ti_0.26Ni_0.18Co_0.18Mn_0.18O₂, where Li⁺ was partially replaced by Ti⁴⁺, and the Li/Ti cation disorder structure strengthened the tolerance of deformation and restricted the migration of TM ions in the TM layer during the delithiation process. This contributed to lowering the electrochemical reaction potential and improving the reversibility of anion redox. Lee et al.^[50] certified that the structural stability of the bulk and surface was enhanced with an increasing degree of cation disorder in Li_1.2Ni_0.2Mn_0.6O₂ during the initial activation process. This ultimately inhibited the structural deterioration with subsequent cycling and maintained the electrochemical activity and rate performance even in the case of cation disorder.

Distribution of elements

The distribution of elements in pristine particles affects the structure and electrochemical properties of LRMOs. Liang et al.^[51] proposed a TM bond model and predicted the order of the strength of TM-TM bonds (Mn⁴⁺Mn⁴⁺ > Ni²⁺Mn⁴⁺ > Ni³⁺Mn⁴⁺ > Co³⁺Mn⁴⁺ > Co²⁺Mn⁴⁺ > Ni²⁺Ni⁴⁺) by studying the interaction and distribution of TM elements [Figure 2F]. It was found that Co and Mn could easily segregate to form clusters in the Ni environment, which was a significant reason for the deterioration of the structure and the degradation of the electrochemical performance during cycling. Adjusting the interaction of TM elements and promoting the uniformity of elements can prominently promote the structural stability of layered NCM materials. This mechanism is also applicable to LRMO cathode materials, with the segregation of Ni on the surface impeding the migration channel of Li⁺ and weakening the interaction of Ni-Mn, thereby causing the reduction of manganese ions and the fast degradation of voltage and capacity. In contrast, the uniform distribution of elements at the atomic scale enhances the interaction of TM elements, stabilizes the crystal structure and contributes to excellent capacity retention and more smaller voltage degradation^[52,53].

For LRMOs, the distribution of elements can be locally adjusted to produce outstanding electrochemical performance^[54,55]. The gradient structure that Mn gradually decreases and Ni and Co gradually increase from the center to the surface with a certain slope inhibits the transformation to the spinel phase and the Jahn-Teller effect, thereby markedly reducing the degradation of voltage, which promotes the cycle performance. In Ni-rich layered materials, reducing the content of Ni in the surface layer and increasing the concentration of Mn^[56] not only maintains the high capacity of the center of the material but also improves the stability of the surface layer. This seems to imply that a similar content of TM components can improve the structural tolerance of the material. The inhomogeneous distribution of Li also significantly influences local anion redox. Zhu et al.^[57] synthesized an LRMO material with a Li-poor shell, which inhibited the generation of O-O dimers on the surface and the release of O₂.

C2/m:R$$\bar 3$$m

The Li₂MnO₃ phase with the C2/m space group as the sole component of an LRMO endows the material with the “Li-excess” characteristic and profoundly affects its structure and electrochemical performance^[9]. Li is extracted from the lattice in the form of “Li₂O” and leaves the MnO₂ phase to stabilize the structure, enabling the composite LRMO layered material to maintain good tolerance to high voltage. Yang and Xia^[58] doped NCM811 cathode materials with a certain Li₂MnO₃ component to sufficiently inhibit the phase transition between the H2 and H3 hexagonal phases with distinctly different cell parameters, which dramatically decreased the variation of the structure volume and alleviated the drastic structural deterioration and capacity degradation. More importantly, the non-activated Li₂MnO₃ can store the excess lithium, so that Li⁺ in the TM layer migrates to the adjacent Li-deficient layer during charging to stabilize the structure and realize the sufficient release of carriers and capacity.

The pristine layered structure of Li₂MnO₃ transforms to the spinel phase during cycling and the capacity source changes to maintain the capacity. However, the difference in the voltage plateau between the two phases leads to voltage degradation, which is a unique phenomenon and problem of LRMOs. The intrinsic atomic structure and arrays of the Li₂MnO₃ component significantly influence the electrochemical performance. Song et al.^[59] found that the structural defects in different orientations and the imperfect atomic arrays of the Li₂MnO₃ component caused an increase in atomic defects and the concentration of vacancies, ultimately leading to the breaking of grains and performance degradation. The sluggish electrochemical kinetics of the Li₂MnO₃ phase restrict the rate performance and also limit the performance release of the R$$\bar 3$$m component. However, the control of the current density can regulate the activation of Li₂MnO₃ to improve the electrochemical performance.

The introduction of the R$$\bar 3$$m component into Li₂MnO₃ (via the doping of Ni or Co) can significantly improve the irreversible oxygen loss. It has been shown that the intrinsic charge compensation of anions for Li₂MnO₃ leads to the completely irreversible release of O₂, but the irreversible oxygen loss accounts for 25% in Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ with a Li₂MnO₃ ratio of 50%. This means that Ni and Co play an important role in stabilizing lattice oxygen in the bulk material. Liu et al.^[60] studied the electrochemical properties of xLi₂MnO₃·(1-x)0.5LiNi_1/3Co_1/3Mn_1/3O₂ (x = 0.3, 0.5 or 0.7) and found that materials with a medium Li₂MnO₃ content had higher initial specific capacity. The materials with higher Li₂MnO₃ contents showed better cycling performance and more significant voltage attenuation due to the stable electrode/electrolyte interface and structural rearrangement at high potentials [Figure 3]. In addition, Shi et al.^[61] replaced the R3m component in 0.5Li₂MnO₃·0.5LiNi_1/3Co_1/3Mn_1/3O₂ with LiNi_0.8Co_0.1Mn_0.1O₂, which increased the average discharge voltage from 3.5 to 3.8 V and the specific energy from 912 to 1033 Wh kg^-1. Thus, it is necessary to appropriately adjust the component and ratio of the C2/m and R$$\bar 3$$m phases to enhance the degree of uniformity and arrangement rules at the atomic level, improve the stability of Mn octahedra and balance the coordination effect of Mn valence, reversible capacity and the Jahn-Teller effect.

Figure 3. Typical charge-discharge curves of (A) initial and (B-D) cycling process for xLi₂MnO₃·(1-x)LiNi_1/3Co_1/3Mn_1/3O₂ (x = 0.3, 0.5 or 0.7) cathode materials^[60].

Ratio of Ni, Co, Mn and Li/O

Every element and stoichiometric ratio play a critical role in determining the whole structure and electrochemical activity^[62]. Hua et al.^[63] systematically studied the relationship between Li content and the structure of Li_xNi_0.2Mn_0.6O_y by matching the Li content with the precursor. When x < 0.4, the material was indexed to the spinel phase (Fd$$\bar 3$$m), while at 0.4 < x < 1.2, the material was a complex of the spinel (Fd$$\bar 3$$m) and rock salt (Fm$$\bar 3$$m) phases, including Li and the layered monoclinic phase (C2/m). At 1.20 < x < 1.52, the material was indexed to the pure monoclinic phase (C2/m). The ratio of Li/O represents the content of Li-O-Li configurations in the LRMO material, thereby affecting the anion redox and the structure of the energy bands^[64]. The lower ratio of Li/O means the TM-O hybridization and non-bonding O 2p bands move to lower energy levels, which contributes to a higher potential of redox reaction that promotes the electrode reaction kinetics, reversibility and energy efficiency [Figure 4A]. The higher ratio of Li/O accelerates the anion redox reaction, inevitably introducing an excess disordered structure that is regarded as the buffer regulating the conflict between oxygen evolution and the crystal structure and making the LRMO more stable and flexible^[65].

Figure 4. (A) Initial charge-discharge plots with Coulombic efficiency (i) and cycle performance (ii) of the as-prepared Li_xMn_0.56Ni_0.16Co_0.08O₂ samples^[64]. (B) Initial charge-discharge curves (i) and average discharge voltage during cycling with the differential capacity curve (dQ/dV) (ii) of the Li[Li_0.2Ni_0.2-x/2Mn_0.6-x/2Co_x]O₂ (0 ≤ x ≤ 0.24) samples^[70]. (C) Capacity (i) and average voltage retention (ii) of as-prepared Li_0.7Mn_0.78Co_0.22O₂ (LMCO), Li_0.75Mn_0.78Co_0.11Ni_0.11O₂ (LMCNO) and Li_0.74Mn_0.78Ni_0.22O₂ (LMNO) during cycling^[69].

The regulation of lithium and oxygen is also applied to other cathode materials. Xiao et al.^[38] designed different cooling technologies and synthesized Li₂MnO₃ with various nonstoichiometric oxygen defects and stacking faults. These defects decreased the valence of Mn and the increase in Mn³⁺ content accelerated the activation of Li₂MnO₃, but this simultaneously led to a more rapid structure deterioration to the spinel and capacity degradation. Wu et al.^[66] introduced an appropriate excess of lithium into NCM811, resulting in the reduction of partial Ni³⁺ to Ni²⁺ and the decrease of Li/Ni disorder, which improved the electrochemical stability and structural robustness during cycling. Ji et al.^[67] designed a novel cathode material with excellent rate performance and reversible redox of oxygen via the introduction of excess Li into spinel LiMn₂O₄. The cations over stoichiometric ratio and the derived local disorder eliminated the typical first-order phase transition reaction of the ordered spinel material and realized higher practical capacity.

The oxidation of oxygen in Li₂MnO₃ during charging generally causes the irreversible loss of lattice oxygen. However, the introduction of Ni and/or Co can dramatically inhibit the loss of oxygen, which results in the formation of Li[Li_0.2Ni_0.2Mn_0.6]O₂ and Li[Li_0.2Ni_0.13Co_0.13Mn_0.54]O₂. The introduced Ni is enriched on the particle surface to form Li-deficient and Ni-rich rock salt shells as protection layers to inhibit oxygen loss^[68]. Though the addition of Co cannot cause the formation of the rock salt phase, it still restricts the oxygen loss overall. In addition, the overlap of the Co^3+/4+ t_2g and O 2p bands and the intense covalency of Co-O in LRMO can promote the coupling between Co and O and the anion redox^[69]. Nevertheless, this leads to an increase in Mn³⁺ content, a decrease in electrochemical reaction potential, the generation of oxygen vacancies and the migration of TM ions to the Li layer^[70], which accelerates the capacity and voltage degradation [Figure 4B]. Accordingly, the appropriate increase in Ni content and a lower ratio of Co/Ni can efficiently adjust the coordination environment and local electron structure of oxygen to reduce the proportion of Mn³⁺/Mn⁴⁺ redox couples at low voltages and enhance the operation voltage^[19]. This also enables the TM-O and non-bonding O 2p bands to move down to lower energy levels, thereby inhibiting anion redox and the corresponding phase transition and efficiently improving the structural stability and capacity/voltage retention. Enhancing the Ni content not only changes the center of redox in LRMO cathode materials but also promotes the reversibility of Mn migration^[71] by a shielding effect, where the random transition of Mn is restricted in a certain region by the electrostatic shielding effect of Ni ions [Figure 4C].

Charge compensation mechanism

The LRMO cathode material xLi₂MnO₃·(1-x)LiMO₂ possesses a superior discharge capacity of over 250 mAh g^-1 and a complex crystal structure, so that the charge-discharge mechanism is difficult to reasonably explain through the escaping/embedding mechanism of typical LIB layered materials. The initial charge-discharge curves of the typical lithium-rich material Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ mainly include two parts: a slope below 4.4 V indexed to the solid solution reaction and a voltage plateau at 4.5 V. The discharge curve acts as a slope with a large voltage span. The material exhibits completely inconsistent charge/discharge behavior with a Coulombic efficiency of ~80%, which is different from other cathode materials with good reversibility (e.g., LiFePO₄, LiCoO₂, LiMn₂O₄, LiNi_xCo_yMn_1-x-yO₂ and so on). The dramatic irreversible phase transition of Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ inevitably occurs in the initial cycle, which strongly influences the subsequent charge and discharge cycles. Therefore, the initial charge/discharge behavior is an important scientific problem that cannot be ignored in the study of LRMO materials.

According to the absorption edge analysis of each element at different states of charge [Figure 6A], it is found that the slope below 4.4 V during charging corresponds to the charge compensation of Ni²⁺/Ni⁴⁺ and Co³⁺/Co⁴⁺ without the variation of the Mn⁴⁺ valence, which is attributed to the charge compensation mechanism of the LiMO₂ component^[9]. The absorption edge of the TM at 4.5 V shows that there is no change in the K edge of Ni and Co^[81]. The Mn K-edge X-ray absorption near-edge structure spectrum indicates that the environment of Mn changes at 4.5 V according to the continuous variation of the absorption edge, which results from the distortion of MnO₆ octahedra. Hence, the lattice oxygen coordinated with Mn can more easily participate in charge compensation and generate electron holes. The unique anion charge compensation of LRMOs occurs in the plateau region at 4.5 V (the triggering condition has been discussed above), which is ascribed to the Li-O-Li configurations and the hybridization conditions of TM-O.

Figure 6. (A) 1s-4p peak energy in the initial cycle showing almost no change in the environment around Ni and Co on charging across the 4.5 V plateau^[83]. (B) Compositional phase diagram showing the electrochemical reaction pathways for a xLi₂MnO₃·(1-x)LiMO₂ electrode^[9]. (C) Operando mass spectrometry of ¹⁸O-labelled Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ cathode during the initial cycle^[83]. (D) O K-edge mRIXS of a series of Li₂RuO₃ electrodes at different electrochemical states. The solid red arrows indicate the striking, sharp features of oxidization at the high voltage state. (i-vii) represent the mRIXS of Li₂RuO₃ at the state of the pristine, charged to 4.1 V, charged to 4.2 V, charged to 4.6 V, discharged to 3.8 V, discharged to 3.35 V, and discharged to 2.0 V, respectively. (viii) Summary of Ru and O redox during the initial cycling profile with blue for lattice Ru^4+/5+ and yellow for lattice oxygen redox^[28]. (E) Charge-discharge profiles of Li_2-xIr_1-ySn_yO₃ for a full cycle (black) and for ~1.5 e^- in Ir per cycle (pink)^[29]. (F) Enlarged ABF-STEM image. The O-O pairs arise from twisting the opposite triangular faces of the IrO₆ octahedra (shown in yellow)^[11]. (G) O K-edge mRIXS of Li₂Ir_0.75Sn_0.25O₂ and Li₂Ir_0.75Sn_0.25O₂ charged to 4.60 V showing a localized RIXS feature at 530.7 eV excitation energy and 522.8 eV emission energy^[29]. mRIXS: Mapping of resonant inelastic X-ray scattering.

In the initial cycle, oxygen evolution occurs in the Li₂MnO₃ component with the extraction of Li⁺ manifesting as the first-order phase transition reaction in the form of “Li₂O” net loss [Figure 6B], which involves the nanodomains and dislocations formed by localized uneven delithiation^[82]. It is clarified that CO₂ gas forms at the beginning of the voltage plateau by operando mass spectrometry and oxygen isotope labeling^[83], which can be attributed to the oxidic lattice oxygen with high activity reacting with the electrolyte (possibly including the decomposition of residual Li₂CO₃ on the surface). Subsequently, O₂ escapes from the lattice at the end of the voltage plateau^[84], which does not derive from the oxidation of the electrolyte [Figure 6C]. However, the amount of oxygen release is considerably lower than the estimated value generated by the activation of the Li₂MnO₃ component. Considering that O₂ cannot migrate from the bulk to the surface, the emission of O₂ occurs only at the near surface. According to calculations, the equivalent thickness of the surface where oxygen is released is ~2-3 nm^[85].

The anion charge compensation mechanism is distinct for different lithium-rich materials^[28]. For Li₂RuO₃, the cation Ru⁴⁺/Ru⁵⁺ and anion O^2-/O₂^n- couples reversibly participate in charge compensation [Figure 6D], owing to the d-sp hybridization related to the redox coupling without irreversible oxygen evolution. For Li₂IrO₃, only the Ir⁴⁺/Ir⁵⁺ and Ir⁵⁺/Ir⁶⁺ couples participate in the charge compensation without anion redox via the high-efficiency mapping of resonant inelastic X-ray scattering (mRIXS) results. Though it was demonstrated that the IrO₆ octahedron is distorted with the generation of analogous O-O dimers, the distance of O-O is much larger than ~1.5 Å for the generation of peroxide and superoxide groups^[11] [Figure 6F]. However, the substitution of Sn⁴⁺ can promote the reaction activity of anions [Figure 6E and G] by the local coordination environments.

Energy band reconstruction

The inconsistency of the initial charge/discharge curves certainly means the reconstruction of the bulk structure and the corresponding energy band. According to the above discussion, the redox of Ni and Co is the same as the charge compensation of LiMO₂ with good reversibility, which cannot influence the structure and energy band. Therefore, it is the key to defining the anion charge compensation mechanism. It is found that the valence of Ni and Co remains unchanged at the beginning and the end of discharge and the variation of valence during discharging is reversible via the comparison of the absorption edge for Ni, Co and Mn. However, the valence of Mn decreases slightly, which may be ascribed to the activation of Mn in the Li₂MnO₃ component^[86]. Based on the test and calculation results, more than half the discharge capacity (~280 mAh g^-1) is still unidentified. If we exclude the side reaction with the electrolyte, this part of the capacity can be regarded as the contribution from the reduction of oxygen.

In order to verify the reversibility of anion redox, Li et al.^[87] determined the reversible evolution of O-O dimers during charging for the first time via in-situ surface enhanced Raman spectroscopy. The results seemed to suggest the reversibility of the redox behavior of oxygen but failed to explain why the reduction plateau of oxygen disappeared during discharging. Yabuuchi et al.^[84] considered that oxygen was electrochemically reduced on the electrode surface below 3 V during discharging. Sun et al.^[64] demonstrated that there was larger voltage hysteresis (≈ 0.5 V) during the anion redox via adjusting the cut-off voltage. Gent et al.^[88] revealed that the strong couple effect between anion redox and the partially reversible cation migration contributed to the decrease over 1 V in the reduction potential of the bulk oxygen, causing the rearrangement of the anion and cation redox potentials.

However, this does not seem to explain intuitively and thoroughly the root cause of the energy band structure rearrangement. Considering the irreversible release of O₂ in the near-surface region of the lithium-rich material during the first charging process, it should be questioned whether O^2- will also dimerize to form O₂ molecules after being oxidized inside the particle body, which are restricted in the crystal lattice and cannot escape. This assumption seems more reasonable. Recently, House et al.^[41] used high-resolution RIXS and ¹⁷O magic angle spinning nuclear magnetic resonance (NMR) spectroscopy to successfully confirm the existence of O₂ molecules in the crystal lattice instead of O₂^2-, longer peroxide-like O-O bonds or electron holes on O^2-. In addition, they calculated and simulated the intralayer migration and disordered arrangement of TM elements in the fully charged state, which destroyed the LiMn₆ honeycomb ordered superstructure and formed vacant clusters that can accommodate O₂ [Figure 7]. The migration in the TM layer is irreversible. During the discharge process, the coordination environment of active oxygen changes from O-Li₄Mn₂ to O-Li₆ with the insertion of Li⁺, which increases the number of oxygen non-bonded orbitals and increases the energy level^[14]. Therefore, the reduction of oxygen occurs in the low voltage stage, which leads to a larger voltage hysteresis. The above perspective is currently the most satisfactory conclusion of the reconstruction of the structure and the rearrangement of the energy band during the first cycle of LRMO materials, but it still requires the unremitting efforts of researchers to verify and improve it.

Figure 7. (A) Operando XRD of Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ at different stages of the initial cycle, where P is pristine, BOP is the beginning of the plateau, FC is full charge and FD is full discharge. The peak area ratio of the superstructure peaks relative to the 001 peak decreases across the plateau, which indicates the loss of the honeycomb structure^[41]. (B) Macroscale changes of the cathode particles during the first cycle^[41]. (C) Structures of atomic-scale ordering changes within the TM layer modeled by DFT^[41]. (D) Local coordination environment around oxygen in pristine honeycomb-ordered O-Li₄Mn₂^[14]. (E) Local coordination around oxygen in disordered Li-rich cluster O-Li₆ after the first cycle. O 2p states belonging to O and that have six coordinating Li are higher in energy than those that have four coordinating Li, which results in the initial voltage hysteresis^[14]. XRD: X-ray diffraction.

Element migration

The topological reaction occurs before 4.4 V, where Li⁺ ions are extracted from the LiMO₂ component with no variation in the frameworks^[89]. During the electrochemical process, the Li⁺ ions in the LiMO₂ component are gradually consumed, so the Li⁺ ions in Li₂MnO₃ migrate from octahedral sites in the Mn layer to tetrahedral sites in the Li layer to participate in charge compensation, which provides extra binding energy to stabilize the structure. Li⁺ ions continue to be extracted from the Li₂MnO₃ component with a series of complicated element migrations and phase transition processes at the 4.5 V voltage plateau^[90,91]. The charge compensation of oxygen ions at the 4.5 V plateau leads to TMO₆ octahedral distortion^[92] and even the formation of oxygen vacancies. Taking the Li_1.5MnO₃ model containing Li vacancies, Yu et al.^[93] found that the migration energy of Mn from the 4g site to the 2b Li site was ~3.07 eV through theoretical calculations. With the introduction of oxygen vacancies, the migration energy was rapidly reduced to 1.86 eV, which promotes the migration of Mn ions.

It is generally considered that partial Mn ions migrate from octahedral sites to the coplanar tetrahedral sites in the Li layer and ultimately from tetrahedral sites to Li sites at this stage. Radin et al.^[94] claimed that the Mn migration to tetrahedral sites was actually indexed to the oxidation of Mn⁴⁺/Mn⁷⁺ simultaneously triggering the generation of O₂ molecules and peroxide ions [Figure 8A] according to first-principles calculations based on the Li-Mn-O phase diagram. Mn⁷⁺ was then reduced and returned to the original octahedral sites and there were also Mn⁷⁺ ions in tetrahedral sites at the end of charging. Kleiner et al.^[95] quantified the migration track of TMs via synchrotron radiation XRD differential Fourier fitting. It was found that ~8% TMs occupied the tetrahedral sites in the Li layer during charging and ~2%-5% TMs occupied Li octahedral sites without the occupation of tetrahedral sites after discharge. Their results strongly suggested that 3%-6% TMs can reversibly migrate from the tetrahedral sites to the original TM octahedral sites during delithiation [Figure 8B].

Figure 8. (A) Alternative charge mechanisms in LRMO cathode materials. Oxidation of Mn⁴⁺ to Mn⁷⁺, followed by the spontaneous formation of peroxide or trapped oxygen molecules accompanied by the reduction of Mn^[94]. (B) TM migration in an LRMO from the octahedral TM sites via tetrahedral sites into octahedral Li sites^[95]. (C) Comparison of crystal structures. Schematic illustrations of crystal structures of O3-type (left) and O2-type (right) LRMOs with TM migration paths. Although the TM ions in the O3 structure can readily occupy Li sites that share only edges with neighboring cations, the TM ions in the O2 structure are subject to strong repulsion when they occupy Li sites that face share with neighboring cations^[98]. LRMOs: Lithium-rich manganese-based layered oxides.

The TM ions partially migrate to Li sites, leading to the formation of localized spinel-like phases. Therefore, the electrochemical activity of Mn is activated and Mn will be excessively reduced to Mn³⁺ below 3 V. This tends to stimulate the Jahn-Teller effect, as well as the disproportionation of Mn³⁺, resulting in structural deterioration and element dissolution. In particular, in the near-surface layer of LRMOs, TM migration and structural reorganization are more serious due to the O₂ release and the destruction of the oxygen framework, resulting in the formation of disordered rock-salt phases and surface densification^[91]. Therefore, irreversible element migration is a vital reason for the phase transition and the degradation of electrochemical performance. It is particularly important to suppress the irreversible migration of TMs, especially the migration process from the transition-state tetrahedral site to the Li octahedral site. The O2-type cathode material^[96,97] then emerges and the special oxygen stack (ABCBA) enables TM ions to undergo intense electrostatic repulsion when occupying the octahedral Li sites coplanar with adjacent TM cations. This effectively suppresses the irreversible TM migration and phase transition^[98] [Figure 8C]. Furthermore, if the TMs can realize reversible migration between the TM and Li layers during the charge-discharge process, it is also worthy of affirmation. Ma et al.^[99] clarified that Mo can reversibly migrate between the Li and TM layers in Li₂MoO₃, with the material maintaining good structural and electrochemical stability with the “unit cell breathing”^[12] mechanism.

Oxygen vacancy transfer and element segregation

As LRMOs are cycled under a wide voltage range of 2-4.8 V and high voltage conditions, the lattice oxygen redox and complex structural evolution can continue to exist. From the above discussion, the TMs in the TM layer not only migrate to the Li layer after the first activation but will also migrate in the layer to form localized nanopores to restrain O₂ molecules. Moreover, the oxygen ions around the nanopores become more active due to the change in the coordination environment and the increase in the number of non-bonded 2p orbitals, which in turn makes it easier to form oxygen vacancies and promote the migration of TMs^[100]. The nanopores gradually expand and merge [Figure 9A], and around the pores, the continuous transfer of TMs leads to segregation and densification of the structure. This phenomenon continues with charge and discharge cycling. In this case, the active lattice oxygen gradually forms O₂ molecules and gathers in large pores, which cannot provide capacity. The densified rock salt phase structure around the pores is also electrochemically inert, leading to continuous capacity attenuation, which also hinders Li⁺ transport and charge transfer and increases electrochemical impedance and voltage hysteresis. Grenier et al.^[100] used small-angle X-ray scattering to observe the formation of characteristic nanopores during the initial oxygen oxidation of lithium-rich materials (rather than other traditional layered cathode materials) and believed that cation migration and disorder may promote the formation of nanopores. The 3D electron tomography reconstruction method was employed by Hu et al.^[15] and it was detected that Li_1.2Ni_0.15Co_0.1Mn_0.55O₂ produces obvious holes inside after 15 cycles. The decrease in oxygen content and the segregation of Mn were found in the host material around the pores [Figure 9B].

Figure 9. (A) HAADF-STEM images of Li_1.2Mn_0.6Ni_0.2O₂ cathode showing the gradual propagation of the nanovoid-populated zone from the particle surface towards the interior of the particle with increasing cycles^[101]. (i-iii) Represent the HAADF-STEM images of the material after different cycles (5 cycles, 45 cycles, 200 cycles, respectively). (B) Annular dark-field STEM image (i) of an opened pore. Mn (ii) and oxygen (iii) composition extracted from the quantification of the electron energy-loss spectroscopy map^[15]. (C) Charge-discharge curves and phase structural evolution of LRMO during electrochemical cycling^[105]. (D) Mean discharge/charge voltage of the two cells cycled simultaneously (cell 1/2: 22 °C, cycles 1-49; cell 3/4: 25 °C, cycles 1-49; 22 °C, cycles 49-105), with the error bars representing the standard deviation between the two cells^[95]. (E) Contribution towards the discharge capacity from each element at various cycles (left). An illustration of the Fermi level being lifted up as a result of the electronic structure change (right)^[15]. HAADF-STEM: High-angle annular dark-field aberration-corrected scanning transmission electron microscopy.

In addition, there are differences in oxygen release on the surface and inside of LRMO particles. Csernica et al.^[80] used transmission-based X-ray absorption spectroscopy and ptychography on a mechanically cross-sectioned Li_1.18xNi_0.21Mn_0.53Co_0.08O_2-δ electrode to quantitatively analyze its oxygen deficiency during cycling. It was found that the degree of oxygen release was more serious on the surface of the material particles and under the conditions of the same size of the primary particles, larger secondary particles can alleviate the problem of oxygen release and maintain the electrochemical capacity. This seems reasonable because oxygen released at the material interface will diffuse into the electrolyte, leading to a series of side reactions and HF erosion, thereby aggravating element dissolution, O₂ release and structural deterioration near the surface of the cathode. Yan et al.^[101] found that the pores of the material are more inclined to the particle surface distribution and exhibit a larger pore size. They then gradually diffuse and expand into the material and the diffusion ability of oxidized O^n- into the material is significantly higher than that of O^2- by theoretical calculations. Similarly, Lin et al.^[102] also found abnormal segregation of Ru on the surface in Li₂Ru_0.5Mn_0.5O₃ after cycling. The surface layered structure was replaced by 1-2 nm metal clusters and the overall structure also became loose and porous. In other materials, such as spinel LiMn₂O₄ and layered LiCoO₂, the original oxygen vacancies and TMs continuously migrate during non-equilibrium charging and discharging. Along the direction of the close-packed oxygen layer, stacking faults are particularly prone to large lattice strains and stress concentrations^[103], which also cause the generation and evolution of cracks and gaps.

Local phase structural transformation

During the cycling of LRMO materials, the layered structure gradually becomes unstable with the generation of oxygen vacancies and the migration of TMs. When the TM, especially Mn, occupies the octahedral site of the Li layer, the surrounding re-embedded sites of Li⁺ are transferred to the tetrahedral vacancy, forming a spinel-like LiMn₂O₄ phase of a sub-nanodomain. Although a considerable amount of Mn will reversibly return to the original position after it migrates from its own octahedral position to the tetrahedral position of the Li layer in each charge and discharge process and the recurring spinel nanodomain proposed by Xiao et al.^[104] appears, there is still some Mn irreversibly migrated to the Li layer. This process continues to accumulate, from the first charge and discharge Li site octahedra are occupied by TMs (mainly Mn) at ~2%, which gradually increases to ~5% after 100 cycles^[95]. In the wide voltage range of 2-4.8 V, the Mn ions undergo a profound valence state change. When discharged to ~2.8 V, Li⁺ ions are excessively embedded in the 16c octahedral position of the cubic spinel LiMn₂O₄ phase, forming a tetragonal spinel Li₂Mn₂O₄ phase, and the valence state of Mn decreases from +3.5 to +3, resulting in Jahn-Teller distortion and the dissolution of Mn²⁺ caused by the disproportionation reaction of Mn³⁺. This leads to further deterioration of the material structure and eventually an inactive disordered rock salt structure (Fm$$\bar 3$$m space group). On the surface of the material, due to the loss of Mn and the migration of Ni, a large number of maligned NiO rock salt phases will be formed^[105] [Figure 9C].

The series of phase structural transformations mentioned above lead to a decrease in electrochemically active materials that can accommodate the reversible deintercalation of Li⁺. In addition, the local disordered rock-salt phase of cations, especially on the surface, seriously hinders the migration of Li⁺ and charge transfer and increases the electrochemical impedance and voltage hysteresis. Although the production of the active spinel phase during cycling compensates to a certain extent for the capacity decline caused by the gradually weakened anion charge compensation, this leads to another unfavorable factor where the O^2-/O^(2-n)- electrochemical couple with the high-voltage reaction is replaced by the Mn³⁺/Mn⁴⁺ couple with the low-voltage reaction [Figure 9E]. The continuous TM valence decreases as the oxygen vacancy increases, which also leads to a decline in the redox potential of Co and Ni. Although there is a small energy level difference between the Ni²⁺/Ni³⁺ and Ni³⁺/Ni⁴⁺ redox couples, the difference between Co²⁺/Co³⁺ and Co³⁺/Co⁴⁺ is quite large. Therefore, the reduction of the electrochemical reaction potential of the redox couples during cycling and the aforementioned kinetic factors are the direct causes of the voltage decay^[15]. It is noteworthy that the charge/discharge average voltage changes during the cycle are inconsistent and the charge average voltage changes very little^[95] compared with the serious attenuation of the discharge average voltage [Figure 9D], which can also be attributed to the combined effects of kinetics and thermodynamics. Because electrochemical polarization increases the potential during the charging process, the thermodynamically driven redox couples change will lower the electrode potential. In addition, TMs mainly migrate to the tetrahedral position during the charging process and partially irreversibly migrates to the octahedral Li position during the discharge process, which may also cause a difference in the thermodynamic electrode potential during charging and discharging.

Li⁺ migration

In layered cathode materials, Li⁺ mainly carries out 2D interlayer migration in the Li layer. There are two migration paths for Li⁺ from one octahedral position to the next octahedral position^[106] [Figure 10A]. The first is that Li⁺ directly passes through the center of O-O on the common edge of two octahedrons, which is known as oxygen dumbbell hopping (ODH). The second is to migrate to the tetrahedral site coplanar with the two octahedral by relying on the vacancy left by other Li⁺ diffusion. It then moves to another octahedral position to complete a migration process known as tetrahedral site hopping (TSH). The diffusion barrier of the TSH path is much lower than that of the ODH path through theoretical calculations. Therefore, when the material is charged in the early stage, Li⁺ migrates in the ODH path first because there are not many vacancies. With the continuous emergence of Li⁺ and the increase of vacancies, the migration path of Li⁺ gradually changes from ODH to TSH and the migration rate is gradually accelerated. Liu et al.^[107] pre-introduced Li vacancies in Li-rich materials, which not only reduced the energy barrier of Li⁺ in the initial ODH diffusion path but also accelerated the transition from difficult ODH to facile TSH.

Figure 10. (A) Illustration of ODH and TSH types of Li⁺ diffusion pathways^[106]. (B) Schematic diagram of Li⁺ distributions around tetrahedral sites in the cation-disordered oxide: 0-TM; 1-TM; 2-TM; 3-TM^[108]. (C) Illustration of a TMO₆ octahedron with two face-sharing tetrahedral sites. Migration of the TM from the octahedron transforms a 1-TM channel into a 0-TM channel^[109]. (D) D_Li⁺ of a LRMO material under different states during the initial charge^[110]. (E) Nyquist plots of Li_1.2Mn_0.56Ni_0.16Co_0.08O₂ electrodes at various potentials^[111]. (F) dQ/dV curve of O2-LRMO measured at a current density of 5 mA g^-1, which proves the reaction mechanism of anions at high potential^[98]. (G) Variation of discharge capacity as a function of current density is estimated for the two classified voltage ranges of 2.0-3.4 and 3.4-4.8 V^[98]. (H) Schematic diagrams of density of states at different states of charge (top) and anionic oxygen oxidation coupled with cationic reduction (bottom), showing the effect of different activation energy in the intermediate states on anionic kinetics in Li_1.2Ru_0.4Co_0.4O₂^[112]. ODH: Oxygen dumbbell hopping; TSH: tetrahedral site hopping.

When a TM occupies the Li site in the crystal structure of the material (especially after cycling and near the surface of the material), the migration channel of Li⁺ gradually deteriorates from 0-TM to 3-TM^[108] [Figure 10B] and the diffusion path of Li⁺ will not only be blocked, but also suffer from the electrostatic repulsion of the high-valence TM, which greatly increases the migration energy barrier of Li⁺. Huang et al.^[109] found that Cr can realize the non-topotactic reaction of reversible octahedral to tetrahedral migration in the process of delithiation in lithium-rich disordered rock salt materials. Octahedral TMs are the main obstacle to Li⁺ diffusion. When Cr migrates to the tetrahedral position, it will correspondingly provide a tetrahedral vacancy for Li⁺, which improves the migration conditions of Li⁺ [Figure 10C]. This principle can be used for reference to alleviate the deterioration of Li⁺ diffusion kinetics caused by the deterioration of the layered oxide structure. In addition to Cr, other electronic configurations, such as d⁵ and d¹⁰, are also conducive to the migration to the tetrahedral position, but the matching between the TM size and tetrahedral pores needs to be considered simultaneously.

Anion reaction kinetics

In the first charging activation process, our previous work^[110] found that the diffusion coefficient of Li⁺ (D_Li⁺) increases slightly at the initial charging stage, which corresponds to the release of Li⁺ from the LiMO₂ component. However, D_Li⁺ drops sharply at the 4.5 V plateau [Figure 10D], which corresponds to the charge compensation process of oxygen in the Li₂MnO₃ component. This anomalous behavior cannot be explained from the perspective of the Li⁺ migration channel. Although a small amount of TM migration occurred during the process, which hindered the migration of Li⁺, the generation of Li vacancies and the expansion of c-axis layer spacing were still favorable for Li⁺ diffusion. Considering the electrochemical redox reaction with both electron and ion conduction, the above apparent decrease in D_Li⁺ only represents a decrease in the electrochemical reaction rate rather than a decrease in the diffusion capacity of Li⁺. Therefore, the electron transfer of anions is likely to be a limiting factor.

Nayak et al.^[111] used electrochemical impedance spectroscopy (EIS) to test the reaction kinetics of a LRMO at various reaction steps at different potentials and found that the charge transfer resistance increased significantly at voltages of > 4.4 V [Figure 10E], indicating that the electrode reaction kinetics at high potentials were controlled by anionic charge transfer rather than Li⁺ diffusion. Yu et al.^[36] used time-resolved X-ray absorption spectroscopy (XAS) to determine the characteristic reaction kinetics of each element. Compared with Ni and Co, the reaction kinetics of the Mn site before and after the initial activation of Li₂MnO₃ was much worse. These results suggest that Li₂MnO₃ (anionic charge transfer) is a key component in limiting the rate capacity of LRMOs. Moreover, the anion reaction kinetics in the O2-type LRMO cathode material are also slower than that of the cations. Eum et al.^[98] analyzed the dQ/dV of different potential ranges and found that the charge compensation of the anion in O2-type materials during charge and discharge occurs at a high potential [Figure 10F], which is different from the anion reduction in the O3-type material at a low potential. This creates the conditions for simply examining the reaction kinetics of anions and cations by controlling the applied current and analyzing the capacity evolution at different potential ranges [Figure 10G]. In order to improve the reaction kinetics of the anions, Li et al.^[112] found that the strong covalent property of Co and O enables the charge compensation of oxygen to occur in advance before the complete oxidation of Co³⁺ to Co⁴⁺. This strong coupling effect makes the rapid redox of Co^(3+δ)+/Co⁴⁺ a medium to improve the kinetics of anion redox in lithium-rich materials [Figure 10H]. Other anions, such as S^2-, exhibit faster reaction kinetics and lower voltage hysteresis than O^2- but also lower the redox potential^[113] and affect the energy density due to the weaker electronegativity.

Co-precipitation method

Generally, various metal salts are prepared into a mixed solution according to the target stoichiometric ratio and the metal salt solution and the precipitating agent (e.g., OH^-, CO₃^2-, C₂O₄^2- and so on) are added dropwise to deionized water simultaneously. An appropriate amount of chelating agent (e.g., ammonia) is added to coordinate with metal ions to buffer the precipitation reaction. Meanwhile, the pH and stirring speed are controlled at a certain temperature to allow metal cations to co-precipitate. The precursor and lithium source (e.g., LiOH, Li₂CO₃ and so on) are then uniformly ground and then calcined at a high temperature to obtain the final lithium-rich oxide. This method can realize ion mixing on the nanometer scale and the subsequent formation of the phase structure at high temperature does not require long-range diffusion of TM ions and avoids problems such as uneven cation distribution and lattice defects. This simple and quick synthesis process is suitable for large-scale promotion and application.

Among them, hydroxide precipitation, especially Mn(OH)₂, is easily oxidized in an alkaline environment and an inert protective gas is required to isolate oxygen. However, compared to other carbonates and oxalates, hydroxides have higher reactivity and promote a better crystal structure. During the precipitation process, the nucleation and growth of crystals should be controlled by adjusting the conditions, such as pH and chelating agent. The morphology of the primary and secondary particles formed by aggregation is controlled by adjusting the temperature and stirring speed^[79]. During the preparation of the precursor, the sheet morphology can be controlled by adjusting the ammonia concentration, thereby synthesizing particles with specific {010} crystal planes dominantly exposed to promote the rapid diffusion of Li⁺. Furthermore, primary particles with a radially ordered arrangement can be generated in the final secondary spherical particles by optimizing the ammonia concentration and stirring speed during co-precipitation, so that the radial primary particles of the material undergo a coordinated volume change during circulation to reduce stress, strain and the formation of microcracks. It is also possible to prepare precursor materials with a concentration gradient (or core-shell structure)^[114,115] by adding pre-designed metal salt solutions of different concentrations in batches [Figure 11A], thereby preparing gradient materials with special element distributions and better performance. However, it is necessary to pay attention to the the phenomenon that the gradient distribution would be weakened in the high-temperature calcination process.

Figure 11. Schematic diagram of (A) materials preparation with gradient distribution of elements by step-by-step co-precipitation method^[114]. (B) Controlling nucleation growth to prepare special exposed crystal faces and morphologies by solvothermal method^[123]. (C) One-step preparation of O2-type Mn-based cathode material by electrochemical ion exchange^[127]. (D) Synthesis process of one-dimensional core-shell structure by electrospinning method^[139].

The calcination temperature and time have an important influence on the crystallinity and morphology of a material. The higher the calcination temperature and the longer the hold time, the larger the particle size of the primary particles^[116-118], and even the transformation from polycrystalline secondary particles to large single crystal particles occurs, with the structure with a special morphology being more susceptible to damage during the calcination process. It may also cause excessive oxidation of TMs and a loss of Li/O, leading to increased cation mixing and phase transformation^[119]. In contrast, a lower calcination temperature and time also reduce the crystallinity of the material. Considering the special morphology and high crystallinity, Chen et al.^[18] heated the precursor with the special morphology at a low temperature in advance to pre-oxidize the TM, thereby reducing the reaction temperature and time, while maintaining the crystallinity of the material during the calcination process. The pre-oxidation process can also remove part of CO₂ or H₂O from the precursor in advance to form a particle morphology with a multilayer hollow core-shell structure. The rate of heating and cooling also affects the morphology and crystal structure of the material. A lower heating rate is more conducive to the maintenance of the particle morphology and a faster cooling method causes more stacking faults, cation disorder and other lattice defects^[38].

Solvothermal method

The solvothermal method refers to the rearrangement/recrystallization of ions in the liquid phase at a certain temperature and pressure, which can realize some reactions that are difficult to complete in solid or normal liquid phase reactions. Some new materials with special structures or morphologies can be prepared^[120-122]. Specifically, the corresponding molar ratio of metal salt and lithium salt is dissolved in a single or mixed solvent^[120] (e.g., deionized water, alcohol, dimethyl formamide and so on) with a certain additive, such as acetic acid^[75] and hexamethylenetetramine (HMT)^[76]. Subsequently, the prepared solution is transferred to a stainless-steel reactor. The precursor can be collected after reacting at a certain temperature and then calcined at a certain calcination condition to obtain the final material. Among them, the Li source can also be introduced after the solvothermal reaction. The morphology of the materials prepared by the solvothermal method is complex and diverse, including spherical secondary particles, single crystal nanosheets, microrods and 3D layered morphology, formed by orthogonal stacking of nanosheets. Microwave assistance can accelerate the reaction rate in the solvothermal process and increase the pores between the spherical secondary particles for promoting the migration of Li^+[77]. The microwave heating method can also be applied to the high-temperature calcination process, which greatly shortens the reaction time.

During the nucleation process, some additives or solvent molecules have the lowest adsorption energy on the {010} crystal plane and tend to be adsorbed on the surface of the crystal plane^[123], which results in slower crystal growth in the direction of the crystal plane and promotes the continuous accumulation and growth of atoms along the c-axis direction [Figure 11B]. This promotes the final particle to exhibit more exposed {010} crystal planes. Liu et al.^[76] synthesized a series of precursors with a controllable morphology by a solvothermal method, where HMT acts as the precipitant and a mixture of dimethyl formamide and deionized water acts as the solvent. The prepared 3D cube-maze-like Li_1.2Mn_0.54Co_0.13Ni_0.13O₂ material exhibited excellent Li⁺ diffusion ability and cycle stability by virtue of its unique 3D structure.

Ion-exchange method

The ion-exchange method means that after the material is placed in the designed solution, a certain ion in the material is spontaneously exchanged with the corresponding ion in the solution and the basic framework of the material is maintained to achieve the preparation of new materials. Among them, the number of ions in the solution must exceed the number of ions in the material to be replaced several times to be completely exchanged. At present, the Li/Na ion exchange reaction is usually used to prepare LIB cathode materials, which can obtain unique structural characteristics based on the precursor with a layered structure^[124]. However, the large difference between the Li⁺ and Na⁺ radii causes drastic changes in the interlayer spacing and structure of the material, which may be accompanied by partial lattice distortion and particle breakage^[125].

The conspicuous O2-type LRMO materials are prepared by molten salt ion exchange based on the precursor of P2-type Na-ion battery cathode materials^[126]. Specifically, the P2-type precursor is mixed with an excessive amount of composite lithium salt (LiNO₃:LiCl = 88:12 w/w) with a lower eutectic point and then heated to melt the composite lithium salt and cause the Li/Na ion exchange to occur. After a period of the reaction, it is cooled to room temperature, washed clean to remove impurities and the O2-type material is finally obtained. This kind of O2-type structure cannot be obtained with the traditional high-temperature calcination process and can only be obtained using the characteristics of the P2-type material, where the oxygen array slips easily at high potential to generate the O2-type structure.

Furthermore, we have developed an electrochemical ion exchange method^[127] based on the working principle of alkaline-ion batteries [Figure 11C]. By assembling the precursor into a hetero-type battery, the ions in the electrolyte and bulk material can be exchanged through the electrochemical driving force by continuous charge and discharge. This method generally needs to circulate the material a certain number of times to achieve a relatively complete exchange. A relatively complete ion exchange will be achieved after certain electrochemical cycles by this method, which is simple to realize one-stop ion exchange and enhanced performance. Ion exchange can also occur when the precursor material is immersed in an organic or inorganic solution^[128,129] containing a relatively high concentration and excess of Li⁺ at room temperature.

Other preparation methods

The sol-gel method^[130-132] refers to the stoichiometric ratio of metal salts dissolved in water to prepare a solution, with a certain proportion of complexing agents (e.g., glucose, citric acid and so on) added and heating and stirring to evaporate the solvent. During this process, the metal salt is gradually hydrolyzed or complexed to form a sol and finally dehydrated to form a gel. The gel is dried and calcined to obtain a metal oxide precursor. It is then calcined at a high temperature to obtain the final cathode material after the precursor is uniformly ground. This method can realize the mixing of target ions at the ion/nanoscale, lower the subsequent calcination temperature and is beneficial to realizing the elemental distribution of the cathode material with high uniformity.

The electrospinning method is used to jet-spin the target metal salt polymer solution in a strong electric field to produce a polymer filament precursor with a nanometer diameter and then perform high-temperature sintering to prepare the cathode with one-dimensional nanowires^[133-137]. Other petal-shaped nanosheets and nano-block particles can also be prepared by adjusting the sintering temperature and time^[138]. In addition, using different preparation processes, such as changing the nozzle structure and controlling the experimental conditions, the precursors with solid, hollow or core-shell structures can be obtained, and finally the preparation of materials with special morphologies or an in-situ coating can be realized [Figure 11D]^[139].

The spray drying method is used to spray the mixed solution, sol or suspension and other fluid materials that need to be dried under high pressure and dispersed into mist-like droplets. The droplets are sprayed into the drying chamber with fluid hot air, which can instantly remove the moisture and a dry powder particle with uniform mixing of elements will be obtained. The precursor material is then sintered and finally the layered cathode material is prepared. The nanoscale primary particles dispersed in the solution can be spray-dried to produce secondary spherical particles and a more uniform coating can also be achieved by spray drying the coating material^[140], such as graphene, and the electrode material. Herein, the properties of different preparation methods and their advantages and disadvantages are prepared in Table 1.

Regulating activation

When an LRMO is charged to ~4.5 V for the first time, the charge compensation of oxygen occurs, accompanied by the in-plane/out-of-plane migration of TMs and the reorganization of the energy band structure. The irreversible structural changes that partly occur during this process to the subsequent electrochemical behavior and structural evolution are affected profoundly. Watanabe et al.^[141] and Ito et al.^[142,143] carried out voltage regulation, reducing the upper voltage to 4.5 V during the initial activation and gradually increasing the upper cut-off voltage to 4.8 V, which can significantly improve the cycle performance of the material. After thorough research, this stepwise pre-cycling treatment can inhibit the microcracks on the particle surface and the periodic distortion of the crystal structure, significantly improving the cycle resistance of the material. In our previous work [Figure 12A], we also found that pre-cycling can alleviate the severe lattice distortion and the formation of spinel-like phases during the activation process and rationalize the intercalation/deintercalation interval of Li^+[110].

Figure 12. (A) Charge-discharge curves and electrochemical performance of LRMO materials under different voltage systems. (i) Initial two cycles of charge-discharge curves under the condition of Vr-1 (tested between 2.0 and 4.8 V). (ii) Initial four cycles of charge-discharge curves under the condition of Vr-8 (pre-cycled at 2.0 to 4.5 V for two cycles and tested between 2.0 and 4.6 V). (iii and iv) Cycle performance of materials under Vr-1 and Vr-8, respectively^[110]. (B) Initial charge-discharge curves (i), the evolution of TM occupancy in Li layer (ii) and two-phase content by Rietveld refinement (iii) for Li_1.2Mn_0.567Ni_0.167Co_0.078O₂ at 25 and 55 °C^[93]. LRMO: Lithium-rich manganese-based layered oxide.

The range of the voltage test is also related to the evolution of electrochemical performance and structure^[144]. TMs migrate from the TM layer to the vacancy in the Li layer at a high charging voltage (> 4.5 V), tending to form a more favorable spinel structure. The Mn will be excessively reduced to a charge below +3.5 at low discharge voltages (< 2.8 V), which promotes the Jahn-Teller effect and the disproportionation reaction of Mn³⁺. For this purpose, Yang et al.^[145] used an LRMO to perform charge and discharge cycles in the range of 2.8-4.4 V, which effectively inhibited the phase transition at high and low voltages and showed excellent cycle stability. Nayak et al.^[146] studied the effect of electrochemical activation on the performance of the material. It was found that the materials that have not undergone activation show the equivalent capacity of LiMO₂ at 2.3-4.3 V but still show a voltage decay of > 0.1 V after 100 cycles. The activated material exhibits a capacity exceeding the non-activated conditions by 50 mAh g^-1 with a lower and more stable discharge average voltage. Pradon et al.^[147] evaluated the influence of the constant voltage method adopted in industry to eliminate dynamic polarization of the materials. It was found that the more time spent at high potentials, the more structural defects that were formed. Yin et al.^[148] also studied the structural evolution of the material during a constant voltage of 5 h after being fully charged to 4.8 V and found a new phase produced under continuous oxidation at this stage, which aggravated the migration of Mn and the deterioration of capacity and voltage.

Kaewmala et al.^[149] studied the influence of current density on the structural changes and cycle stability of an LRMO and found that the activation of the Li₂MnO₃ components was affected by the current density. High-current cycling can effectively reduce the activation of Li₂MnO₃ and the formation of the spinel phase to obtain better cycling performance and faster Li⁺ diffusion. Yu et al.^[93] studied the influence of activation temperature on the phase structural transformation [Figure 12B] and found that the remaining proportion of the Li₂MnO₃ component after activation at 25 °C is 17%, while the remaining Li₂MnO₃ component after activation at 55 °C only remains 7%. This shows that the activation process of LRMOs is very sensitive to temperature and will be aggravated with increasing temperature. Therefore, the activation process, voltage range, current density and other factors are very important for LRMOs. Appropriate working conditions need to be designed to make the electrochemical performance of the material reasonable.

Surface modification

In the process of electrochemical charge-discharge, there are serious problems, such as oxygen release, holes, microcracks, particle powdering, TM dissolution and irreversible phase change of the surface structure, which are the main reasons for the decay of electrochemical properties, so it is necessary to focus on improving the stability of the surface of the material.

Coating is a common means to modify the surface of a material. Coating materials are generally divided into oxides^[150-152], phosphates^[153], fluorides^[154,155] and conductive polymers^[156], and the surface treatment object can be an electrode material or precursor. Inert oxides, such as Al₂O₃^[157] and SnO₂^[158], can only build an inactive isolation layer on the surface of the materials to avoid the side reactions caused by direct contact with the electrolyte, but they may reduce the electron/ion conductivity and affect the rate performance. Other functional oxides enriched with oxygen vacancies, such as Ce_0.8Sn_0.2O_2-σ^[159], MoO_x^[160] and so on, not only avoid direct contact between the host material and the electrolyte, but also provide oxygen vacancies to capture oxygen radicals, a fast ionic conductor reacting with surface Li to improve rate performance^[161,162], additional charge compensation to improve the capacity and other functions^[163] [Figure 13A]. Fluoride coatings provide F-O bonds on the surface, thus reducing oxygen loss and the decomposition of the electrolyte. Phosphates can also stabilize the surface lattice oxygen to reduce oxygen loss and PO₄^3- can form covalent bonds with metal ions to inhibit TM migration. Conductive polymer coating layers can significantly improve the electrochemical performance due to its high electronic conductivity and good coating uniformity. In addition, the binding force between the surface coating layer and the host material cannot be ignored. It is a good choice to choose a coating that is structurally compatible with a bonding effect^[164].

Figure 13. (A) Schematic illustration of oxygen vacancies in Ce_0.2Sn_0.8O_2-σ and its protective effect on an LRMO material^[159]. (B) HAADF image near the surface (i) and the EDS mapping of Mn and Nb (ii) of the LRMO particle by Nb surface doping^[170]. (C) Schematic of gas-solid interface reaction between Li-rich layered oxides and carbon dioxide^[173]. (D) Optimized structure of BO₄^5--doped LRMO (left), in which B atoms occupy the tetrahedral interstitial sites. The simulated local structure (right) around the boracic polyanion, showing that the M-O bonds are lengthened due to the strong B-O bond^[182]. (E) Schematic illustration of Li-rich layered material with composited LiMn₆ and SbNi₆ superstructures^[186]. (F) (i) Length of TM-O (Cl) bonds of Li_1.2Mn_0.53Ni_0.27O₂ (LMNO) and Li_1.2Mn_0.53Ni_0.27O_1.976Cl_0.024 (LMNOC), (ii) energy barriers of TM migrating from the TM layer to the Li layer and (iii) operando differential electrochemical mass spectrometry curves of O₂ for LMNO and LMNOC^[187]. LRMO: Lithium-rich manganese-based layered oxide; HAADF: high-angle annular dark-field.

Surface doping is based on surface coating. By increasing the calcination temperature, the coated elements spontaneously diffuse into the surface layer lattice of the material to form localized interface doping^[165-167]. Surface doping can enhance the interaction between the doped elements and the host material and stabilize the surface interface without affecting the bulk structure^[168,169]. Liu et al.^[170] modified the surface of Li_1.2Mn_0.54Co_0.13Ni_0.13O₂ by Nb⁵⁺ surface doping. The doped Nb⁵⁺ is located in the Li layer near the surface of the material, thereby inhibiting lattice distortion and deactivating the surface oxygen, which is beneficial to preventing cation mixing and improving the structural stability [Figure 13B]. Zhong et al.^[171] doped PO₄^3- on the surface of Li_1.2Mn_0.54Co_0.13Ni_0.13O₂, where the PO₄^3- inhibited the irreversible release of lattice oxygen on the surface of the material. Moreover, PO₄^3--M with a stronger binding energy effectively reduced the migration of TMs to the Li layer and a fast ionic conductor, Li₃PO₄, was also formed on the interface of the host material by the interdiffusion of ions. Thus, the ionic conductivity of the modified material was improved and the generation of the CEI film was reduced.

In addition, other surface treatments, such as mild acid treatments^[172], can achieve H⁺/Li⁺ exchange and remove Li⁺ and O^2- in the Li₂MnO₃ component in advance to reduce oxygen release and irreversible capacity loss during the initial cycle. Simultaneously, the residual Li₂CO₃, LiOH and other impurity components on the surface of the host material were removed, which improves the reaction activity of the cathode interface. Gas phase^[73,173,174] (NH₃, CO₂, CO and so on) treatment can produce certain element doping or oxygen vacancy generation on the surface of the material, inhibit the oxygen release and promote the Li⁺ diffusion [Figure 13C]. After molten salts or other highly reactive solutions are used to act on LRMOs^[175-177], epitaxial heterostructures^[150] (spinel and cation disordered rock-salt phases) are inclined to form on the surface. These structures formed by the phase transformation of layered materials exhibit higher thermodynamic stability, can inhibit the activity of surface oxygen and stabilize the interface to avoid the migration of structural defects from the surface to the bulk.

Bulk doping

Compared with the surface layer, the structural stability and electrochemical reversibility of the bulk phase of the material are relatively good. However, due to its intrinsic anion redox reaction, there are still problems such as TM migration, oxygen vacancy transfer and phase structural transformation. Therefore, bulk doping of the material is expected to improve the electrochemical performance and structural stability of the material radically. Generally, bulk doping can be carried out at three positions: Li sites^[178]; TM sites^[107]; oxygen active sites^[96,179,180], which can produce different effects and results.

Liu et al.^[181] prepared Li_1.167-xK_xMn_0.583Ni_0.25O₂ by ball milling and solid-phase sintering, using K⁺ to partially replace Li⁺ in the material. The rate performance and cycle performance of the pristine material are significantly improved. Studies have found that the doped K⁺ is pinned at the Li site, which hinders the migration path of part of the Mn and inhibits the formation of the spinel phase. Zhang et al.^[178] successfully incorporated Na⁺ into the Li_1.23[Ni_0.2464Mn_0.462Co_0.0616]O₂ crystal through a solid-phase reaction to partially replace Li⁺ in bulk. Tests showed that Na⁺ doping significantly improved the electrochemical performance of the material. The main reason is that Na⁺ doping into the layered structure expanded the layer spacing, which promoted the rapid diffusion of Li⁺ and reduced the release of oxygen in the crystal lattice. Meanwhile, Mn⁴⁺ in Li₂MnO₃ was further activated after Na⁺ doping and provided reversible capacity in subsequent electrochemical cycles.

Li et al.^[182] successfully doped boron polyanions into a cathode material to synthesize Li[Li_0.2Ni_0.13Co_0.13Mn_0.54](BO₄)_0.75x(BO₃)_0.25xO_2-3.75x. First-principles calculations show that the doped borate anion can reduce the O 2p orbital energy level, due to the doping of B in the tetrahedral position of the transition metal layer [Figure 13D]. This reduced the covalency of the TM-O bond, inhibited the precipitation of oxygen in the LRMO and prevented structural damage of the material during charge-discharge cycling. Wu et al.^[183] successfully synthesized a Fe-doped Li_1.2Mn_0.585Ni_0.185Fe_0.03O₂ material. The doped Fe instead of Co is located in the TM layer, which increased the lattice distance and enhanced the diffusion of Li⁺ in the host structure. Simultaneously, it inhibited the migration of TMs to the Li layer and hindered the transformation of the crystal structure from the layered to spinel and rock-salt structures. Nayak et al.^[184] successfully synthesized Li_1.2Ni_0.16Mn_0.51Al_0.05Co_0.08O₂ as a cathode material using Al³⁺ instead of Mn⁴⁺. Al doping can suppress the phase change of the material from laminar to spinel during the cycle, thereby stabilizing the crystal structure. The electrochemical impedance of the doped material did not change significantly during cycling, indicating that Al doping exhibits a certain surface effect.

Jiang et al.^[185] used F^- to replace O^2- in Li_1.2Ni_0.2Mn_0.6O₂, which changed the local electronic environment of Mn, reduced the release of oxygen and caused a change in the surface structure. The symmetry of the MnO₆ octahedron is destroyed because of the shorter Mn-F bond than the Mn-O bond, thereby alleviating the Jahn-Teller effect. The expansion of the interlayer spacing after doping caused Li⁺ to be inserted/extracted faster, thereby suppressing the increase in charge transfer resistance during cycling. In addition, the conversion of the layered structure to the spinel phase during cycling was inhibited, which finally improved the electrochemical performance. Li et al.^[186] successfully prepared Li(Li_1/6Mn_1/3Ni_1/3Sb_1/6)O₂ materials with mixed Li@Mn₆ and Sb@Ni₆ superstructure elements [Figure 13E]. The study found that the coordination environment of local oxygen changes, which reduces the energy barrier for Li⁺ diffusion and provides additional capacity. Wang et al.^[187] partially replaced oxygen with anion Cl to change the coordination environment of the TMs, which stabilized the TM-O bond and effectively inhibited the irreversible O₂ release and TM migration [Figure 13F], thereby enabling the material to exhibit excellent capacity and voltage retention.

Other electrode/electrolyte scope

In order to ensure that an electrode exhibits good electrochemical performance, a certain amount of conductive agent is usually added during the production of the electrode slice to reduce the contact resistance and improve the electronic conductivity. The conductive agent should possess good dispersibility and a high specific surface area and electrolyte adsorption capacity to further improve the ionic conductivity of the electrode. The main conductive additive in traditional cathodes is acetylene black; however, acetylene black is difficult to disperse and exhibits strong oil absorption, resulting in an uneven local composition of the slurry. Yu et al.^[188] used carbon nanotubes (CNTs) to prepare composite LRMO cathode electrodes, which displayed the characteristics of a binder, lightweight and self-standing [Figure 14A]. Based on long-period electrochemical tests, the self-supporting structure can reduce polarization, stabilize the electrode to inhibit structural deformation and simultaneously promote the rapid transmission of electrons/ions. Luo et al.^[189] prepared a binder-free LiCoO₂ electrode with excellent flexibility and conductivity by constructing a continuous 3D super-aligned CNT frame with embedded LiCoO₂ particles. This binder-free cathode electrode showed better cycle stability, rate performance and energy density than the typical electrode with a binder.

Figure 14. (A) Schematic synthesis process of self-standing CNT-LRMO (i) with photographic (ii) and SEM images (iii)^[188]. (B) Chemical structure of GG and PVDF (i), photographs of electrolyte (ii) and GG (iii) and the mixture of GG and electrolyte after 0 h (iv) and 24 h (v)^[190]. (C) Schematic of the effect of PAA binder for LRMO electrode^[191]. (D) SEM micrograph of PVDF-HFP separator (i-iii) and modified PVDF-HFP/GO nanosheet (iv-vi) in its original state and stored at 150/170 °C for 1 h^[192]. CNT-LRMO: Carbon nanotube-lithium-rich manganese-based layered oxide; GG: guar gum; PAA: polyacrylic acid; HFP: hexachloropropylene; GO: graphene oxide.

As an inactive ingredient, the binder cannot provide capacity. Currently, the matching binder for LRMOs is mainly based on a fluorocarbon (PVDF). PVDF is sensitive to the water content during the electrode preparation process and will swell in the electrolyte. Furthermore, its exothermic reaction with metal lithium and Li_xC₆ at higher temperatures is unfavorable from a safety perspective. Therefore, designing and synthesizing a binder with excellent conductivity and viscosity is very important for improving the energy density, rate and cycle performance. Zhang et al.^[190] used guar gum (GG) as the binder of a LRMO, which exhibited low solubility and swelling ability in the electrolyte [Figure 14B] and effectively reduced the volume expansion of the electrode during cycling. Polyacrylic acid (PAA) was also used by Yang et al.^[191] as a binder for LRMO electrodes [Figure 14C]. The H⁺/Li⁺ exchange between the LRMO and PAA generates protons on the surface layer of the cathode, which effectively suppress the migration of TMs, thereby forming a stable structure. This surface modification represents a new method for stabilizing LRMOs.

Currently, microporous polyolefin membranes, such as polyethylene and polypropylene, are the most commonly used separators for lithium-rich cathode materials, because of their excellent mechanical strength and chemical stability. However, the low melting point of polyolefin membranes results in high thermal shrinkage at high temperatures, causing an internal short circuit and even explosion. Simultaneously, the non-polar polyolefin separator exhibits poor wettability, which has a negative impact on the rate and cycle performance of the battery. Choi et al.^[192] incorporated graphene oxide (GO) nanosheets into a PVDF-hexachloropropylene (HFP) framework through phase inversion during the pore formation process [Figure 14D]. The addition of GO greatly improves the thermal stability of the composite separator. Shin et al.^[193] reported a new concept of coating nitrogen and sulfur co-doped graphene (NSG) nanosheets on a polyethylene separator. The thin NSG layer acts as a physical barrier to inhibit Li dendrite growth due to its excellent mechanical properties. Due to the local strain caused in the carbon skeleton, the incorporation of N and S can cause structural deformation, thereby improving ion migration while maintaining a uniform ion flux on the surface of the Li metal.

The high working voltage of an LRMO exceeding 4.5 V causes the traditional carbonate-based electrolyte^[194,195] to decompose severely on the surface of the cathode materials^[196], forming an uneven, unstable and unprotected CEI, which hinders the diffusion of Li⁺ and reduces the electrochemical efficiency. Liu et al.^[21] studied the effect of concentrated electrolyte (with a solute concentration of 3 mol L^-1) on the surface structure of a cobalt-free LRMO and found that high-concentration electrolytes can form a CEI rich in LiF and other inorganic species on the surface of the electrode, which significantly inhibited the dissolution of TM and effectively protected the cathode material from corrosion. Cui et al.^[197] added lithium difluoro(oxalato)borate to a perfluorinated electrolyte and the strong fluorinated CEI formed in situ effectively reduced the side reaction between the O2-type LRMO electrode and electrolyte, and further suppressed O2 structural transformation and oxygen release. Lou et al.^[198] first studied methyl diphenyl phosphonate as an electrolyte additive to form a strong artificial mesophase layer in situ on the surface of the cathode electrode. The intermediate phase can effectively inhibit the decomposition of the electrolyte and greatly improve the stability of the interface between Li_1.16Ni_0.2Co_0.1Mn_0.54O₂ and the electrolyte under high voltage and temperature.