Recent advances of cobalt-free and nickel-rich cathode materials for lithium-ion batteries

Crystal structure instability

A stable crystal structure is essential for Ni-rich layered oxide cathode materials including CFNR. Ideally, these materials should possess a well-defined crystal structure, where all elements in the lattice only play their unique roles without affecting each other. However, this is challenged by the issues mainly including Li/Ni mixing and oxygen loss of these materials, leading to their crystal structure instability.

Li/Ni mixing inherently exists in and has been attributed to similar ionic radii, superexchange interaction, magnetic frustration, and kinetic advantage of Ni-rich layered oxides^[3,20,42].

Specifically, same as their prototype of LNO, Ni-rich layered oxides have an α-NaFeO₂-type hexagonal layered crystal structure (R$$\overline{3m}$$ space group), where the Li ions and TM ions are coordinated by oxygen to construct an alternately-arranged layered structure with Li and TM layers [Figure 3A]^[2]. The different sizes of Li⁺ (0.76 Å) in the Li layers and Ni³⁺ (0.56 Å) in the TM layers are favorable for stabilizing the structures of these materials. However, the low valance state of Ni, i.e., Ni²⁺, is always found in these materials, which has been largely attributed to the inadequate synthesis conditions during the precursor calcination of these materials (e.g., inappropriate temperature and deficient oxygen atmosphere)^[13,43,44]. Also, the Ni³⁺/Ni⁴⁺ generated on the surfaces of these materials during their charge/discharge can react with the electrolyte to produce Ni^2+[45,46]. As a result, due to their similar sizes in radii, Li⁺ (0.76 Å) can enter into the TM slab and Ni²⁺ (0.69Å) into the Li slab, resulting in the so-called Li/Ni mixing^[47,48].

Figure 3. (A) Model of the layered structure with R$$\overline{3m}$$ space group. This figure is quoted with permission from Yu et al.^[2]. (B) Schematic for 180° superexchange interaction. This figure is quoted with permission from Zheng et al.^[26]. (C) Schematic diagram of magnetic frustration effect. This figure is quoted with permission from Li et al.^[25]. (D) Migration routes of Ni ions to Li sites in the layered structure for the Oh → Td → Oh route (i) and the Oh → V_O site → Oh route (ii), (E) Energy barriers calculated for migration of a Ni ion for the Oh → Td → Oh route (i) and the Oh → V_O site → Oh route (ii), the x-axis denotes the positions of Ni ions on the migration paths in (D). This figure is quoted with permission from Kim et al.^[52].

Superexchange interaction refers to the enhanced exchange between two TM cations, one from the TM layer and the other from the Li layer, through a bridged anion (i.e., O^2-) of a layered oxide^[26,49]. Through its unpaired spin electrons, the p orbitals of this intervening oxygen can bind with the d orbitals of the TM cations to form σ bonds. Using a Ni-rich layered oxide as the example [Figure 3B]^[26], the antisite Ni²⁺ in the Li layer tends to form a 180°-exchange interaction stronger with the Ni²⁺ than those with other metal ions in the TM layer. Consequently, the preference of this interaction (due to its low energy) can enhance the Li/Ni mixing, which, in turn, makes the Ni-involved exchange interaction “super”, resulting in the so-called superexchange interaction.

Magnetic frustration originates from the strong magnetic interaction due to the conflicting electron spins of TM ions in the hexagonal lattice of a TM layer. Using three triangularly placed Ni³⁺ (having their d shells half-filled) as the example [Figure 3C]^[25], the Ni³⁺ in the center of the hexagonal lattice has a spin state parallel (in the same direction) to those of its three surrounding Ni³⁺ at the apexes of the lattice. This arrangement of conflicting electron spins is a higher-energy and hence unstable state, leading to the so-called magnetic frustration that strengthens the Li/Ni mixing^[50,51].

Kinetic advantage represents a situation in which the migration of Ni²⁺ from the TM layer to the Li layer has a low barrier in energy. Particularly, with LiNi_1-xCo_xO₂ as the example^[52], the migration of Ni²⁺ from the TM layer to the Li layer is favorable to follow a route of O_h (octahedral Ni site)-T_d (tetrahedral site)-O_h (octahedral Li site) [Figure 3D(i)] owing to its low energy barrier of between 0.303 and 0.598 eV for the Li 0% (fully delithiated) and Li 100% (fully lithiated) states, respectively [Figure 3E(i)]. Even upon the oxygen loss of LiNi_1-xCo_xO₂, Ni²⁺ can also find a favorable path of Oh (Ni site)-Vo (oxygen vacancy)-Oh (Li site) for migration [Figure 3D(ii)] because of its low energy barrier of between 0.567 and 0.617 eV for the Li 100% (fully lithiated) and Li 0% (fully delithiated) states, respectively [Figure 3E(ii)]^[52]. In this regard, compared to the migration of Co (e.g., with a high energy barrier in the route of O_h-T_d-O_h of 1.6 eV for Li_0.5CoO₂^[53]), that of Ni²⁺ has a remarkably lower energy barrier and thus is considered kinetically advantageous for Li/Ni mixing^[52].

Due to the strengthened Li/Ni mixing by the factors above, the crystal structures become unstable (with a decreased c parameter)^[54], the Li⁺ diffusion pathways are blocked (by the antisite Ni²⁺ in the Li slab)^[1], the surface metals are lost (due to the migration of antisite Ni²⁺ to the particle surface)^[26], and the layered structure is damaged (to disordered spinel and eventually rock salt)^[55,56] for Ni-rich layered oxides.

Apart from Li/Ni mixing, oxygen loss is another important cause influencing the crystal structure instability of Ni-rich layered oxides^[55]. Using the prototype of these materials (i.e., LNO) as the example, upon delithiation, its oxygen arrangement is transformed from the original face-centered cubic (O3, ABCABC) structure to the final hexagonal close-packed structure (O1, ABAB) [Figure 4A]^[55,57]. This transformation is accompanied by oxygen loss, which can be quantitatively interpreted by the formation energy of oxygen vacancy with thermodynamic calculations^[58]. As can be seen in Figure 4B, the formation energy of oxygen vacancy of LNO decreases rapidly from 1.8 eV for the fully lithiated state to 0.35 eV for the 75% delithiated state, providing a strong driving force for oxygen loss.

Figure 4. (A) Atomic models of O3 phase (space group R$$\overline{3m}$$), O3 phase with partially transited O1 phase, and pure O1 phase. Li, Ni, and O are denoted by green, gray, and red symbols, respectively. This figure is quoted with permission from Wang et al.^[55]. (B) Oxygen vacancy formation energy during delithiation of LiNiO₂, (C) Configuration for Ni migration in Li_0.5NiO₂, Vo1, Vo2, and Vo3 represents oxygen vacancies neighboring with diffusive Ni ion, (D) The migration barriers for Ni ions to diffuse from transition metal layer to Li layer, under different situations (Pristine: no neighboring oxygen vacancy; Vo1: with neighboring vacancy Vo1; Vo1 + Vo2: with neighboring vacancies Vo1 and Vo2; Vo3: with neighboring vacancy Vo3). This figure is quoted with permission from Kong et al.^[58].

Furthermore, oxygen loss promotes Li/Ni mixing, as depicted in Figure 4C, in which the oxygen loss-deduced vacancies (denoted as Vo) can reduce the energy barrier for the migration of their adjacent Ni ion from the Ni layer to the Li layer. Following the path of Oh-Td-Oh [Figure 4C], the migration energy barriers for a half-delithiated LNO with and without oxygen vacancies are calculated and plotted in Figure 4D^[58], showing the strong effect of oxygen vacancies in reducing the energy barrier for Ni migration and thus enhancing the Li/Ni mixing.

As a result, along with the promoted Li/Ni mixing, oxygen loss destroys the crystal structures and produces unwanted gases, jointly leading to inferior electrochemical performance, poor thermal stability, and even safety hazards for Ni-rich layered oxides^[52,55-60].

Phase transition

Along with their delithiation upon charge, Ni-rich layered oxide cathode materials, including CFNR, undergo a phase-transition sequence of Hexagonal 1 (H1) - Monoclinic (M) - Hexagonal 2 (H2) - Hexagonal 3 (H3) - Hexagonal 4 (H4) [Figure 5A]^[4,61]. The oxygen arrangement of the H1, M, H2, and H3 phases is in an O3 type but that of the H4 phase is in an O1 type. The presence of these phases during the charge/discharge of LNO, as an example, has been verified by the in-situ synchrotron diffraction, as shown in Figure 5B. Although a complete reversibility is desired for the transition between these phases, the real situation is complicated. Specifically, at a high delithiation state (x > 0.95), the layered slip of the H3 phase results in the formation of the H4 phase (NiO₂, CdI₂-type structure), unfortunately creating lattice mismatch and volume changes that can impede Li⁺ transport and degrade the cathode materials^[23]. In practice, therefore, the cutoff voltage during charge (i.e., the depth of delithiation) is usually limited to avoid the appearance of the H4 phase for these materials^[7,8,11].

Figure 5. (A) Schematic diagram of the structural evolution (H1 → M → H2 → H3 → H4) during the charging of layered LiNiO₂ materials. This figure is quoted with permission from Li et al.^[4]. (B) Contour maps of in situ synchrotron diffraction of LNO collected during the first cycle. Cell voltage as a function of state of charge or discharge of the first cycle, which is aligned with the diffraction patterns (i). XRD patterns of selected scattering angle regions: 4.68°-5.48° (ii), 9.59°-10.45° (iii), and 11.36°-12.26° (iv). Based on the refinement results, the phases in each region are labeled in (i). The peak splitting for phase M is also indicated in (iii) and (iv). This figure is quoted with permission from Li et al.^[61]. (C) Illustration of the reactions of the surface residual lithium of a LiNi_0.96Mg_0.02Ti_0.02O₂ (NMT) cathode with H₂O and CO₂ in the air, forming LiOH and Li₂CO₃, respectively. This figure is quoted with permission from Kim et al.^[15]. (D) Cross-sectional SEM images of a LiNi_0.9Mn_0.1O₂ (NM90) cathode after 1,000 cycles in full cells featuring EF91 electrolyte, cycled in the voltage range of 3.0-4.3 V. This figure is quoted with permission from Park et al.^[9].

During lithiation/delithiation, largely owing to their smaller change in lattice structure (using the c-axis parameter change as the example as shown in Figure 5A), the transition among the H1 → M → H2 phases is more reversible than that between the H2 → H3 phases^[11,62]. Indeed, upon the increased depth of delithiation, the H2 → H3 transition can cause a large anisotropic contraction (as schematically presented by the dramatic decrease in the c-axis parameter in Figure 5A) and hence the collapse of the layered structure. This structural strain, in turn, induces inter-/intra-granular cracking, loss of interparticle contact, and loss of active materials through defect and rock-salt formation, thereby leading to inferior and irreversible electrochemical performance for the cathode materials^[1,63]. Therefore, owing to its great impact on the structure stability and electrochemical performance, the H2 → H3 phase transition for Ni-rich layered oxide cathode materials, including CFNR, has been extensively studied and should be paid more attention in the future research^[9,11].

Surface side reactions

An ideal surface able to ensure the stable mechanical, thermal, and electrochemical properties is required for Ni-rich layered oxide cathode materials including CFNR. In practice, however, this remains challenging because of the possible surface side reactions of these materials attributable to their high reactivities with the electrolyte and the environment.

On the one hand, the high reactivity of these materials refers to Ni^3+/4+ generated upon their delithiation to a highly-charged state, which can react with the electrolyte to produce the undesirable species such as Ni²⁺, O₂, and CO₂^[15,42]. While the generated Ni²⁺ can enhance the Li/Ni mixing to deteriorate the crystal structure, the generated O₂ means oxygen loss and induced crystal structure instability for these materials; both have been discussed under Section "Crystal structure instability". Moreover, all gases produced imply the gassing inside the battery that may cause serious safety hazards^[46]. The surface of the cathode materials is prone to erosion/dissolution by HF derived from the LiPF₆ decomposition and hydrolysis of the electrolyte, leading to the loss of active TM and, hence, the capacity degradation for these materials^[64]. Also, the dissolved TM can be reduced into metals and participate in forming a solid electrolyte interface (SEI) on the anode. These will hinder the Li⁺ intercalation into the anode and disrupt the original cathode/anode capacity ratio, thereby raising the impedance and deteriorating the battery performance^[64].

On the other hand, the high reactivity of these materials is expressed by their environmental instability. Specifically, the residual lithium on the surface of these materials, which is left during the material synthesis^[46,65], can initiate some unwanted side reactions, such as those with H₂O and CO₂ in the air to form LiOH and Li₂CO₃, respectively [Figure 5C]^[15]. In the subsequent slurry process for electrode preparation, the LiOH and Li₂CO₃ thus formed can react with polyvinylidene fluoride (PVDF), the commonly used binder for cathodes, to form gels that will increase the difficulty and even disable the production of the electrodes^[65,66]. Therefore, cathode materials without any side reactions with the electrolyte and the environment are highly desired.

Microcracking

Mechanical stability/integrity of Ni-rich layered oxide cathode materials including CFNR play an important role in ensuring their stable electrochemical performance. This is, however, affected by a mechanical degradation mechanism (called microcracking) that occurs during the charge/discharge of these materials^[67,68]. Typically, the secondary particles of layered oxides comprise multiple primary particles in the size of sub-micrometers (even nanometers)^[7,67]. Upon charge/discharge, the Li⁺ intercalation/deintercalation into/out of these materials may cause expansion and contraction in volume for their primary particles, resulting in intragranular fracturing within and intergranular fracturing between the primary particles (i.e., microcracking) as marked in Figure 5D by red and blue arrows, respectively^[9]. The continuous expansion and contraction of the primary particles exacerbate the growth of the intergranular fractures (microcracks) to form isolated particles. Without contact with the electrolyte, these isolated particles cannot undergo Li⁺ intercalation/deintercalation and hence become electrochemically inactive. As a result, microcracking will deteriorate the mechanical stability/integrity and degrade the electrochemical performance of the cathode materials^[67,68].

Moreover, upon charge/discharge cycling, the proliferation of microcracks allows the electrolyte to infiltrate/penetrate into the newly formed surfaces, making the surface side reactions, as discussed under Section "Surface side reactions", occur inside the particles of the cathode materials, further degrading their electrochemical performance and even causing safety issues for batteries^[55]. Therefore, mechanically stable/complete cathode materials able to suppress such kind of microcracking are highly desired.

The degradation mechanisms of Ni-rich layered oxide cathode materials including CFNR discussed above can occur individually and interact with each other. The appearance of one mechanism may trigger others to occur, jointly deteriorating the structural, mechanical, thermal, and electrochemical properties of these materials and even deducing safety hazards for their batteries. Therefore, when researching a new strategy for improving the properties for CFNR cathode materials, these degradation mechanisms should be thoroughly considered/addressed.

Bulk-phase doping

Bulk-phase doping refers to incorporating small amounts (typically less than 5%) of additional elements as dopants into the host lattice of a cathode material without introducing impurity phases. With this strategy, substantial efforts have been made to dope non-Co elements into Ni-rich layered oxides to fabricate CFNR cathode materials^{[2,3,18,19,44,74-81]}. This has been demonstrated to be effective in stabilizing the crystal structure and suppressing the phase transition to enhance the properties for CFNR cathode materials^[7,9]. In this regard, as schematically presented in Figure 6A, bulk-phase doping has been realized for CFNR cathode materials on a range of locations including Ni-site (TM layer), Li-site (Li layer), Ni/Li dual-sites, O-site, and interstitial sites^{[16,27,71,72,77,81-85]}.

Figure 6. (A) Schematic diagram of possible doping sites on a CFNR cathode material. (B) Capacity retention of the Ni-rich Li(Ni_1-xMn_x)O₂ (0.1 ≤ x ≤ 0.5) electrodes in a voltage window of 2.7-4.3 V at 30 °C for 120 cycles. 1-20 cycles: 0.2 C; and 21-120 cycles: 0.5 C. This figure is quoted with permission from Sun et al.^[7]. (C). Summary of anion doping effects on the battery performance of LNO. This figure is quoted with permission from Kong et al.^[84].

Ni-site doping has been considerably studied for CFNR cathode materials owing to its effectiveness in strengthening the metal-oxygen binding thus suppressing the Li/Ni mixing, oxygen loss, and phase transition^{[16,19,36,77,84,86]}. It can be carried out either during the synthesis of precursors or in the calcination of the pre-formed precursors. In the former, the dopant ions are introduced into the bulk metal salt solution to prepare precursors, followed by calcinating the precursors thus obtained with a lithium source to produce the cathode materials^[8,41,87]. This allows the uniform distribution of dopants in the cathode materials. The latter physically mixes the dopants with the pre-formed precursors and a lithium source, followed by calcination to produce the cathode materials^[9,10,16,30]. A large variety of metal ions having ionic radii similar to that of Ni³⁺ (0.56 Å) have been employed for Ni-site doping, which, depending on their valence states, can be classified into low- and moderate-valence dopants, respectively.

Fe³⁺ and Al³⁺ are frequently used low-valence dopants for Ni-site doping for CFNR cathode materials^{[18,38,77,88,89]}. For example, Xi et al. used Fe³⁺ as the dopant to replace Co³⁺ of NCM811 to prepare a new CFNR cathode material of LiNi_0.8Fe_0.1Mn_0.1O₂^[89]. With a traditional co-precipitation method, Fe was first incorporated into the hydroxide precursor with Ni and Mn, followed by calcinating the resultant precursor with LiOH to produce the target LiNi_0.8Fe_0.1Mn_0.1O₂. Even without Co, this material realized a low energy barrier of Li⁺ diffusion (0.969 eV), similar to that of NCM811 (0.979 eV), thus exhibiting a high capacity of 202.6 mA h g^-1 at 0.1 C in 3.0-4.5 V and a capacity retention up to 81.1% after 10 C cycling for 150 cycles in 3.0-4.3 V.

Moderate-valence dopants, such as Zr⁴⁺, Ti⁴⁺, and Mn⁴⁺, are also employed for Ni-site doping for CFNR cathode materials, which is largely encouraged by their capability to form strong TM-O bonds^{[7,18,78,90,91]}. Among them, Mn⁴⁺ has been paid more attention^[7,31,78]. For instance, Sun et al. studied the effect of the Mn⁴⁺ dopant using the conventional co-precipitation process to prepare a range of Ni_1-xMn_x(OH)₂ (0.1 ≤ x ≤ 0.5) precursors, followed by calcinating these precursors with LiNO₃/LiOH to obtain the LiNi_1-xMn_xO₂(0.1 ≤ x ≤ 0.5) cathode materials^[7]. These materials comprised dense and spherical secondary particles with high crystallinity and thus low cation mixing. Therefore, having the lowest cation mixing (2.9%), the optimized composition, i.e., LiNi_0.9Mn_0.1O₂, delivered a capacity of up to 200 mAh g^-1 at 0.5 C (2.7-4.3 V, 30 °C) and retained this capacity at ~86% after cycling for 100 cycles [Figure 6B]. Nevertheless, it should be noted that doping of the moderate-valence elements (along with their induced charge imbalance) may result in the formation of Ni²⁺ and hence the promoted Li/Ni mixing^[31,90]. This indicates a need for the careful content optimization of moderate-valence elements when doping them into CFNR cathode materials.

Similarly, with its ionic radius close to that of Li⁺ (0.76 Å)^[39], Mg²⁺ (0.72 Å) was extensively studied for doping the Li-site of CFNR cathode materials^[85,92-95]. Indeed, Mg²⁺ has been demonstrated to be effective as a pillar ion in mitigating the internal stress of the crystal lattice during the anisotropic lattice contraction at a high charging state, thereby enhancing the structural stability and electrochemical performance for CFNR cathode materials^[92,96]. For instance, Laine et al. used co-precipitation to incorporate varied amounts of Mg²⁺ into Ni(OH)₂, the hydroxide precursor of LNO, followed by calcinating the precursors thus prepared with LiOH to produce Mg-doped LNO cathode materials^[95]. Although having a lowered initial capacity at 0.1 C in 3.0-4.3 V, these Mg-modified materials displayed a significantly enhanced cycling stability, for example, with the boosted capacity (198.3 vs. 193.6 mAh g^-1) and capacity retention (98.9% vs. 87.6%) for a 2 mol % Mg-doped LNO cathode over its pristine counterpart in the subsequent cycling for 62 cycles.

Logically, combing their benefits as elucidated above, Ni-site doping and Li-site doping have been conducted simultaneously, resulting in a range of Ni/Li dual-site doping combinations, such as Al/Mg^[41,97], Ti/Mg^[55,81,96], and Mn/Mg^[72], to further improve the properties for CFNR cathode materials. For example, Mu et al. prepared a Ni/Mg/Mn-incorporating hydroxide precursor via co-precipitation first and then calcinated the as-prepared precursor with LiOH to synthesize a Mg/Mn-LNO cathode material^[72]. Owing to the occupations of Mg²⁺ on the Li site and Mn⁴⁺ on the Ni site, respectively, for co-doping, the Mg/Mn-LNO cathode delivered a smooth voltage profile, enhanced structural stability, elevated self-discharge resistance, and inhibited nickel dissolution. These enabled the Mg/Mn-LNO cathode with improved cycling stability, presenting high capacity retentions of 80% after 350 cycles at C/3 and 67% after 500 cycles at 2 C in 2.5-4.4 V.

Apart from cation doping on the metal sites, as discussed above, anion doping on the O-site has also attracted considerable attention for CFNR cathode materials, with the frequently studied anions being F, Cl, and S^[84,98]. Although anion doping can be expected to enhance the properties for cathode materials, the selection of anions should be carefully considered because of their possible conflicting effects. In this regard, Kong et al. used density functional theory (DFT) to systematically investigate the effects of F, Cl, and S on the properties (including redox potential, ionic conductivity, Li/Ni exchange, lattice distortion, and Ni migration upon delithiation) of LNO as a model cathode material^[84]. As schematically summarized in Figure 6C, while F can improve almost all the properties considered, it, unfortunately, facilitates the Li/Ni mixing due to its reduced energy in forming the Li/Ni defects. With their increased Li/Ni defect formation energies, in contrast, Cl and S can suppress the Li/Ni mixing and lattice distortion (especially Cl). However, concerning F, Cl- and S-dopings result in lower redox potentials for their doped LNO, which is attributed to their relatively weaker bonding strength with Li. Moreover, Cl and S have contradictory effects on the structural stability and rate capability for LNO; that is, a high doping concentration can benefit the former whereas a low one is preferred for the latter. To address the conflicting effects from anion doping, a codoping strategy by coming anion doping with cation doping has been suggested for Ni-rich layered oxide cathode materials including CFNR^[84].

Unlike the doping on a specific site, such as the Ni-site, Li-site, or O-site, as discussed so far, interstitial doping refers to that within the space between varied atoms of a cathode material [Figure 6A]^[71,99], which is usually achieved by boron (B)^[16,71]. B-doping has been considered effective in microstructurally engineering the primary particles of a cathode material. Specifically, it can dissolve (insert) into the host structure, leading to a surface-confined distribution to inhibit the growth and hence increase the aspect ratio (length/diameter) for the primary particles of the cathode material. This significantly enhances the microstructural and hence cycling stabilities for CFNR cathode materials. For example, Kim et al. doped B into LNO by firstly synthesizing a Ni(OH)₂ precursor via co-precipitation and then calcinating the pre-synthesized precursor with B₂O₃ and LiOH to fabricate a B-doped LNO (i.e., B-LNO) cathode^[71]. In a full-cell configuration (with a graphite anode), this B-LNO cathode showed an initial capacity of 200 mAh g^-1 and retained this capacity up to 81% even after cycling for 300 cycles (0.5 C, 2.7-4.2 V), significantly outperforming its pristine counterpart (having a similar initial capacity but a lower capacity retention of about 71%).

Surface coating

Surface coating is a strategy that applies a protective layer on the surface of CFNR cathode materials to protect them from the direct contact with the electrolyte and the environment to suppress the side reactions for enhancing their properties^{[15,69,73,100]}. Generally, this method can be carried out on either the secondary or primary particle surface of a target CFNR cathode material.

Secondary particle coating refers to applying a modifying material directly onto the secondary particle surface of a CFNR cathode material. This has been practiced for years with various materials, such as metals, metal oxides, polymers, and carbon/polymer composites, as illustrated below.

For coating metals on CFNR cathode materials, Chu et al. developed a method by combining solid-phase mixing with low-temperature sintering to coat a 35 nm-thick selenium (Se) metallic coating on LNO to fabricate a Se-LNO cathode [Figure 7A]^[73]. Owing to its stabilized interface and enhanced kinetic behavior and the suppressed phase transition, particle pulverization, and HF corrosion enabled by the Se coating, this Se-LNO cathode showed an improved rate capability (capacity at 5 C: 149.6 vs. 135.7 mAh/g) and cycling stability (capacity retention after 300 cycles at 1 C: 82.09% vs. 31.31%) over its pristine counterpart.

Figure 7. (A) Schematic diagram of the synthesis process for Se-LNO. Long-term testing of LNO and Se-LNO at 1 C and cutoff voltages of 2.8-4.3 V over 300 cycles. This figure is quoted with permission from Chu et al.^[73]. (B) Schematic illustration of the advantages and cycling performance of the dual-protection PP-coating layer. This figure is quoted with permission from Kim et al.^[15]. (C) Differences in mechanism and effects of low-valence (Al) and high-valence (Nb, Ta, and Mo) dopants, including the incorporation into bulk and the grain boundary coating, (D and E) Cycling performance (0.5 C, 2.7-4.3V) of half-cells featuring LNO, Al-LNO, Nb-LNO, Ta-LNO, and Mo-LNO cathodes calcined at 650 °C (D) and their respective optimized calcination temperatures (E). This figure is quoted with permission from Park et al.^[10].

Inspired by the previous research about the capability of a metal oxide coating of Al₂O₃ to enhance the properties of layered oxides^[101-105], Kang et al. used first-principles calculations to investigate how Al₂O₃ can improve the thermal stability of CFNR cathode materials^[100]. With the archetype cathode of LNO, their study indicated that the Al₂O₃ deposits undergo a phase transition from the corundum-type crystalline (c-Al₂O₃) to amorphous (a-Al₂O₃) structures as the number of coating layers reaches three. The resultant a-Al₂O₃, in turn, can significantly suppress the oxygen evolution of LNO at 400 K, which is apparently different from the situation of the pristine LNO that immediately releases oxygen upon its exposure within a very short period time of 1 ps at the same temperature. The authors verified the underlying mechanism of this significance to be the strong contacting force at the interface between LNO(012) and Al₂O₃ deposits that originates from the highly ionic chemical bonding of Al and O at the interface.

To pursue the polymer coating strategy for CFNR cathode materials, Kim et al. mechanically mixed LiNi_0.96Mg_0.02Ti_0.02O₂ (NMT) powder in a polyimide/polyvinylpyrrolidone (PI/PVP, denoted as PP)-containing N-methyl-2-pyrrolidone (NMP) solution, followed by a stepwise imidization process to remove the remained NMP solvent to fabricate a PP-coated NMT (PP@NMT) cathode^[15]. Due to the high thermal stability of PI and the high Lewis basicity (to scavenge HF of the electrolyte) of PVP, the resultant dual-protection PP layer enabled the PP@NMT cathode with the superior resistances to the surface contamination in moist air and the interfacial side reactions upon cycling. As a result, the PP@NMT cathode preserved a relatively clean surface with the bare generation of lithium residues, structural degradation, and gas evolution even after exposure to air with ~30% humidity for two weeks compared to the bare NMT, and exhibited a high capacity retention of 86.7% (vs. 63.2% for the bare NMT) after 500 cycles at C/3 using a localized high-concentration electrolyte [Figure 7B].

To study carbon/polymer composite coatings for CFNR cathode materials, Luu et al. prepared a graphene-ethyl cellulose (GrEC)-coated LNO cathode by firstly dispersing LNO powder in a GrEC-containing NMP solution and then casting the resultant dispersion directly onto an Al foil to fabricate the electrode^[69]. The hydrophobic property of the GrEC composite was demonstrated to be effective in protecting the LNO surface from the contact with atmospheric moisture, thereby minimizing the lithium impurity generation and hence enhancing the electrochemical performance for the LNO electrode thus modified^[69]. Even upon exposure to humidified CO₂ for 24 h, the GrEC-coated LNO electrode delivered a capacity of 230 mAh g^-1 (0.1 C, 2.8-4.6 V), slightly lower than that of the electrode without CO₂ exposure (240 mAh g^-1). When compared to the significant capacity degradation of the pristine LNO electrode from 240 to 180 mAh g^-1 upon the CO₂ exposure, this clearly highlights the effectiveness of the carbon/polymer composite coatings for CFNR cathode materials.

With respect to secondary particle coating, primary particle coating is an emerging method, in which the coating materials are directly cladded on the surface of the primary particles of a CFNR cathode material. Direct coating at the primary particle level has been demonstrated to possess some unique advantages as elucidated below.

One approach to realizing primary particle coating is conducted by crushing the secondary particles of a cathode material into primary particles first, followed by coating the primary particles thus formed with a modifying material. For instance, Brow et al. first milled the secondary particles of LiNi_0.9Mn_0.05Al_0.05O₂ into primary particles and then used H₃PO₄ to react with the surface residual lithium of the resultant primary particles to form a Li₃PO₄ layer on the primary particles^[106]. Benefiting from the enhanced air stability and suppressed electrolyte-side reactions by the Li₃PO₄ layer, the LiNi_0.9Mn_0.05Al_0.05O₂ thus modified showed a remarkably enhanced cycling stability (especially at a high voltage) with a capacity of 180 mAh g^-1(vs. 100 mAh g^-1) after 100 cycles at C/3 in 2.8-4.6 V over its pristine counterpart. However, as can be seen, this method needs to crush secondary particles before the coating on primary particles, which inevitably damages the morphology of the secondary particles and, likely, the structure of the primary particles. It also requires extra time for processing and thus raises the practical application cost.

Without crushing the secondary particles, another approach for primary particle coating is achieved by forming the coating in situ during the calcination process of a cathode material. Specifically, this can be done by mixing a Ni(OH)₂ precursor with LiOH and the oxide (or hydroxide) of a high-valence element (such as Nb⁵⁺, Ta⁵⁺, W⁶⁺, or Mo⁶⁺), followed by calcination under oxygen at a high temperature^{[9,10,14,17,70,74,76,80,107-109]}. As schematically illustrated in Figure 7C, unlike the doping of low-valence elements into the bulk structure, these high-valence elements usually accumulate (due to their insolubility into the crystal structure^[10]) along the interparticle boundaries and form Li-X-O compounds (due to their high electronic hybridization with oxygen^[108]) there to accomplish the coating on primary particles. Since the temperature at which the content of Li-X-O compounds declines is higher for high-oxidation-state dopants, the segregation at the grain boundary can be reinforced, the size of primary particles can be refined, and the calcination temperature range can be widened for high-valence-element-doped CFNR cathode materials, thereby maintaining their highly aligned microstructure and high crystallinity over a wide calcination temperature range^[10]. Furthermore, the grain size refinement achieved by high-valence-element doping dissipates the deleterious strain from abrupt lattice contraction through fracture toughening and the removal of local compositional inhomogeneities, and the enhanced cation ordering induced by the presence of high-valence elements stabilizes the delithiated structure through a pillar effect^[9]. Together, these boost the cycling stability for high-valence-element-doped CFNR cathode materials. Using this strategy, Park et al. doped LNO with Nb⁵⁺, Ta⁵⁺, and Mo⁶⁺, respectively, by calcinating their oxides with a Ni(OH)₂ precursor (pre-prepared via co-precipitation) and LiOH^[10]. They clearly demonstrated the superior cycling stability of the resultant Nb-LNO, Ta-LNO, and Mo-LNO cathodes over the pristine and a low-valence-element-doped (i.e., Al-LNO) counterparts regardless of the calcination temperatures (either at a same temperature of 650 °C [Figure 7D] or their respective optimized calcination temperatures [Figure 7E]. Importantly, these works spotlight the excellence of primary particle coating (through high-valence-element doping) in facilitating the practical applications of CFNR cathode materials.

Single crystallization

Single crystallization is a strategy that synthesizes the CFNR cathode materials in a single crystal structure to stabilize the mechanical structure/integrity for improving the properties of these materials^[110]. As discussed earlier under Section "Microcracking" and now schematically illustrated in Figure 8A, upon the microcracking during the charge/discharge of a conventional polycrystalline cathode material, its resultant intergranular fractures (microcracks) allow the electrolyte to infiltrate/penetrate into the newly formed surfaces of the material. This, in turn, promotes a series of issues including surface side reaction, surface pulverization, structure collapse, disordered phase transition, TM dissolution, and cathode/electrolyte interface (CEI) thickening, and eventually degrades the electrochemical performance and may cause safety concerns for this cathode material.

Figure 8. (A) Schematic illustration for the mechanism of single crystal structure mitigating the structure degradation and performance deterioration of LiNi_0.9Mn_0.1O₂ (NM91), (B and C) SEM images of PC-NM91 (B) and SC-NM91 (C), (D) Half-cell cycling tests of PC-NM91 and SC-NM91using Li as the anodes. This figure is quoted with permission from Dai et al.^[67].

Although the strategies of bulk-phase doping^[111,112] and surface coating^[113-115] can alleviate the microcracking to some extent, single crystallization has been demonstrated to be the most effective in doing so^[116,117]. Particularly, since single crystal structures can eliminate the uneven stresses deriving from the intergranular boundaries of polycrystalline structures, single crystallization has been extensively studied to suppress the microcracking-induced issues of polycrystalline materials [Figure 8A] and, more excitingly, to synthesize single crystal CFNR cathode materials in recent years^{[67,86,87,93,118-120]}.

Single crystal CFNR cathode materials can be synthesized using a solid-phase calcination process similar to polycrystalline materials. However, since the calcination temperature for a single crystal cathode material is always higher than that for its polycrystalline counterpart, which would result in lithium evaporation and serious Li/Ni mixing^[68], a low-melting-point lithium salt (in addition to the lithium source used in the conventional solid-phase calcination process) is usually introduced to form a molten salt mixture to facilitate the single crystal growth through a dissolution-recrystallization mechanism^[68,118,119]. For example, Dai et al. developed a LiOH-Li₂SO₄ molten salt method to synthesize a single crystal LiNi_0.9Mn_0.1O₂(SC-NM91) cathode material^[67]. Specifically, SC-NM91 was obtained by calcinating its Ni_0.9Mn_0.1(OH)₂ precursor (pre-prepared via co-precipitation) in a LiOH-Li₂SO₄ molten salt mixture. For comparison, the polycrystalline LiNi_0.9Mn_0.1O₂ (PC-NM91) was synthesized with the traditional solid-phase calcination process without Li₂SO₄. Upon the single crystal formation, the characteristic microstructure of the spherical secondary particles of PC-NM91 [Figure 8B] vanished and a single granular morphology with complete/uniform structure and good dispersion appeared for SC-NM91 [Figure 8C]. As a result, owing to its strengthened structure able to overcome the microcracking-induced issues of polycrystalline cathode materials as illustrated in Figure 8A, SC-NM91 showed excellent cycling stability (capacity retention after 300 cycles at 1 C in 3.0-4.3 V: 87.72% vs. 65.53%), significantly outperforming its PC counterpart [Figure 8D]. More importantly, this emphasizes the benefit of single crystallization in enhancing the mechanical stability/integrity and hence electrochemical performance for CFNR cathode materials^{[67,86,87,93,118-120]}.

As seen above, various modification strategies have been considerably studied to improve the properties of CFNR cathode materials. While bulk-phase doping into various sites can stabilize the crystal structure and restrain the phase transition, surface coating on secondary and primary particles can suppress the side reactions, and single crystallization can strengthen the mechanical stability/integrity for these materials. Moreover, regarding the varied degradation mechanisms of CFNR cathode materials that may trigger each other to take place simultaneously, different modification strategies can/should be incorporated together to jointly overcome these problems to efficiently enhance the overall structural, mechanical, thermal, and electrochemical properties of these materials, towards their practical applications.