Fe-based cathode materials show great potential for the battery applications due to low cost, high abundance in the Earth’s crust and nontoxic feature. α-NaFeO₂ is an representative O3-type layer cathode, it delivers a reversible capacity of 80 mAh g⁻¹ (0.33 mol Na extraction/insertion) along with a flat discharge platform about 3.3 V in the voltage range of 2.5-3.4 V [48], which is similar to that of isostructural LiCoO₂ [34]. The electrode performance shows high relevance with cut-off voltage. When the cut-off voltage is surpassed to 3.5 V, the reversibility gradually decreases and polarization drastically increases. The α-NaFeO₂ become almost completely inactive when about 0.7 mol Na is extracted. Komaba and co-workers [48] believed that the poor reversibility is resulting from irreversible structural transition and iron ion migration, thus hindering the Na⁺ re-insertion. Later, chen et al. [77] directly observed iron ion migration to Na layer at the atomic scale and presented the Fe migration from the O_Fe to T_Na (tetrahedron site in the Na layer) is energetically favorable, which is also verified by DFT calculations (Fig. 2a, b). Recently, Chung and co-workers [78] investigated comprehensively the irreversible phase transition process when more than 0.5 Na is extracted in α-NaFeO₂ (Fig. 2c). It is demonstrated that formation of Fe₃O₄ in the surface blocks Na insertion into the structure accompanied with oxygen release. More recently, Johnsonand co-workers [79] found that more than 20% of Fe⁴⁺ have spontaneous tendency to reduce back to Fe³⁺ in a charged state, accompanied by electrolyte decomposition, which is evidenced by exsitu Mössbauer spectroscopy (Fig. 2d). This self-discharge behavior can trigger rapid performance degradation and increase storage cost. Meanwhile, they also presented that desodiation process of α-NaFeO₂ is accompanied by coherently nucleated O''3-Na_1−xFO and m-Na_1−xFO phases in 2.0-3.6 V, rather than to form P3 phase via layer gliding, which is common for most O3-type cathode, such as NaCrO₂, NaCoO₂ and NaNi_0.5Mn_0.5O₂. The unique phase transition process may ascribe that tetrahedral Fe³⁺ in Na layer resist layer glide and lead to nonuniform extraction of sodium (Fig. 2e).

It is known that solvothermal method can make reactants more uniform in the solution, compared with the common solid state method. Specially, in the synthetic process of NaFeO₂, reactive Na₂O₂ was usually used as sodium source rather than Na₂CO₃ owing to the latter usually easily to form tiny electrochemically inactive β-NaFeO₂ [80]. Meanwhile, Na₂O₂ is highly sensitive to H₂O and CO₂ in ambient air. Thus, it increases difficulty in preparation. Tabuchi [81,82] et al. prepared high-crystallinity α-NaFeO₂ via solvothermal method without Na₂O₂, it delivers improved cyclic performance compared to the solid state method. Xia [50] came up with a carbon-coated single crystal O3-NaFeO₂ nanoflakes by facile solvothermal route with subsequent carbon-coated (Fig. 3a), which exhibits a high capacity of 89.6 mAh g⁻¹ in the voltage range of 2.0-3.4 V and capacity retention of 87.3% after 100 cycles at 0.1C. In addition, carbon coated NaFeO₂//hard carbon full cell also presented good cyclic performance with 81.9% capacity retention after 100 cycles.

Thermal stability and air stability are key component for future commercial applications, which is closely relevant to safety and storage performance. Nishida and co-workers [83] compared thermal properties of α-NaFeO₂ and LiCoO₂ counterparts by DSC measurements. The results indicated that fully-charged α-NaFeO₂ presents better thermal stability in the electrolyte than LiCoO₂ counterparts with higher exothermic onset temperature and less heat generation. Of note is that α-NaFeO₂ have hygroscopic character and Na⁺ have spontaneous tendency to diffuse to the surface [48], leading to formation of electrochemically inactive phase, such as NaOH, Na₂CO₃. It drastically impedes their practical viability (gelation of the slurry) and decrease the electrochemical performance. Thus, it is essential to incorporate proper ion or surface coating to realize improved air stability. Nevertheless, the focus on air stability for NaFeO₂ is still few.

Although α-NaFeO₂ have been widely developed depending on the economic efficiency, high abundance, thermal stability, and high voltage platform, the cyclic performance is still unsatisfactory for potential application, which is closely relevant to Fe migration and irreversible phase transformation. Zhou and co-workers [49] reported an Ru-substitution Na₄FeRuO₆ cathode which exhibited a high reversible capacity of 120 mAh g⁻¹ and superior capacity retention of 80% over 100 cycles based on reversible O3-P3-O1 phase transformation (Fig. 3b). In contrast, NaFeO₂ severely decay to about 10 mAh g⁻¹ after several cycles and presents an irreversible phase of O'3 and O''3. The larger size of Ru⁴⁺ effectively suppress the irreversible migration of Fe ions, which is clearly characterized on an atomic scale by HAADF-STEM (Fig. 3c). Then, the less Fe ion migration maintain the well structure integrality and decrease the local lattice strain, thus particle cracks was remarkably reduced. For all we know that particle crack is one of the main reasons for performance deterioration. Once the mass of cracks generates, the Na⁺ transport channel will be blocked and layer structure may be destroyed during long term cycling, leading to the battery failure eventually [12,84]. Thus, it is vital to reduce the micro-crack by proper strategy. In addition, Mg introduction to α-NaFeO₂ also was demonstrated to greatly reduces migration of Fe ions and increases capacity and cycling stability [85]. Guo et al. prepared a new cathode NaFe_0.45Co_0.5Mg_0.05O₂ based on the synergetic effect of multi-metal ions, in which Fe³⁺ support the high redox potential, Co³⁺ can realize fast kinetics and Mg²⁺ show a pinning effect to stabilize the structure with compatible radii. It presented a discharge capacity of 140 mAh g⁻¹ and superior capacity remain (about 95 mAh g⁻¹) after 500 cycles at 1 C in 2.5-4.0 V. Meanwhile, the high-rate performance also is excellent, which delivers a capacity of 118.2 and 73.9 mAh g⁻¹ at 5 C and 10 C, respectively. It can be explained by a wide capacity range (> 70%) of formed P3 and P'3 phases, in which prismatic paths can greatly enhance the Na⁺ migration compared with octahedral paths resulting from O3 phases [86,87]. The low Na⁺ migration barrier (0.22 eV) resulting from DFT calculation also support this opinion. Some other elements substitution such as Ni [88,89], Ti [90], Co [91,92] have been reported, but the electrode performance is still unsatisfactory in long term lifespan or average operating voltage.

Mn-based cathode have received great attention due to the low cost and high abundance [93,94]. Na_xMnO₂ with 2D structures have two polymorphs, which is classified into α-NaMnO₂ (space group C2/m) and β-NaMnO₂ (space group Pmnm) at low-temperature and high-temperature form, respectively [35,95,96]. α-NaMnO₂ is monoclinic O3 layered structure owing to Jahn-Teller distortion of Mn³⁺ [51]. While β-NaMnO₂ has a zigzag layer structure, which is made up of two edge-sharing stacks of the MnO₆ octahedra and Na ion [97]. Thus, we will focus on the α-NaMnO₂ in the article. O'3-type α-NaMnO₂ as electrode materials, was first reported by Hagenmuller and co-workers [35] in 1985, there are reversible de/interealation at the Na content between 0.77 and 1.0 with large voltage hysteresis. It is due to high resistivity resulting from intrinsic semiconductor characteristic for α-NaMnO₂, which is also evidenced from first principles computations [98].

In 2011, Ceder et al. [51] revisited the monoclinic α-NaMnO₂, it delivers 185 mAh g⁻¹ of reversible capacity at 0.1 C rate in the voltage range of 2.0-3.8 V, corresponding to 0.8 mol Na intercalated back. The charge-discharge curves have a character of muti-plateaus indicating different reaction paths with different intermediate phase. However, it only remains a capacity of 132 mAh g⁻¹ after 20 cycles, mainly resulting from strong Jahn-Teller effect of Mn³⁺. According to crystal field theory, the Jahn-Teller effect is related to the distribution of transition metal d electrons in the energy level. In octahedral structure, the d orbits are divided into low-energy orbits t_2g (d_xy, d_zx and d_yz) and high-energy orbits e_g (d_z² and d_x²-_y²). The electrons occupy the d orbits in the order of low to high energy. Three outermost layer electrons of Mn³⁺ occupy preferentially the orbit t_2g and another electron occupies the d_z² or d_x²-_y² orbits, which generally produce elongated or compressed distortions, leading to distortion of the layered structure. Such distortions are usually detrimental to structural stability and may be accompanied by Na⁺/vacancy ordering [22,99,100]. Thus, it is essential to regulate Jahn-Teller effect to improve the cyclic stability. Li et al. [52] designed a O'3-NaMn_0.6Al_0.4O₂ cathode with queue-ordered superstructure (Fig. 3d). On the one hand, the special superstructure reinforces the layer structure and decrease the harmful effect from Jahn-Teller distortion. On the other hand, the Mn³⁺ migration from transition metal layer to Na layer is restrained. Thus, it delivers a modest capacity of 160 mAh g⁻¹ and enhanced rate and cyclic performance, compared with the O'3-NaMnO₂. In addition, the queue-ordered superstructure is demonstrated that it has the lowest energy and strengthen ionic bonds between Mn and O by DFT calculations, which can increase the structure stability during cycles, which further supports the improved cyclic performance. The anionic redox chemistry has been developed in Li-rich materials due to the additional capacity delivery along with higher energy. However, it usually accompanied with O₂ release in the high voltage regions owing to the weak covalence bond from 3d transition metal. Zhou’s group [53] predicted that Ir doping in Mn-based cathode has the most favorable oxygen absorption and compared with Cd and Ru doping by DFT calculations, and increasement of Mn 3d band as well as large hybridization between Ir and O can effectively decrease O₂ release (Fig. 3e). Based with the deduction, they prepared Na_1.2Mn_0.4Ir_0.4O₂ O3-type Na-rich materials, which delivers a capacity of 179 mAh g⁻¹ in the voltage range of 1.5-4.4 V by employing cationic and anionic redox. The highly reversible O redox was confirmed by in operando Raman and XAS tests. It provides a new strategy to increase the capacity delivery for Mn-based cathode. Cho and co-workers [101] employ first-principles calculations to predict electrochemical performance for Na(Li_1/3Mn_2/3)O₂ cathode. It shows high redox potentials (≈4.2 V vs Na/Na +) with high charge capacity (190 mAh g⁻¹) resulting from the anionic redox reactions.

Moreover, multielement co-doping can also effectively improve the cyclic performance, average operating voltage, energy density and so on depending on synergistic effect. Deng [54] synthesized an air-stable Na[Li_0.05Mn_0.50Ni_0.30Cu_0.10Mg_0.05]O₂ cathode by coprecipitation method and proposed that Li and Mg are contributing to enhancement of electronic conductivity, Cu can improve the surface structure and regulate phase transition, leading an outstanding air stability, Inactive Mn and Li substitution can remain structure stability. It delivers a discharge capacity of 172 mAh g⁻¹ and superior rate capacity up to 50 C. After 1000 cycles at 20 C, 70.4% capacity retention was maintained, which surpass many other reported O3-type cathode. The pouch cell assembled by this cathode/hard carbon delivers discharge capacity of 70 mAh g⁻¹ and good stability. Recently, a O3-type Na_0.5Mn_0.65Ni_0.15Al_0.1Mg_0.05Co_0.05O₂ nanoflower [102] was presented based on multi-metal synergetic mechanism, in which Al³⁺ and Mg²⁺ have a positive effect on structure stability, Co²⁺ can enhance rate capacity, Ni²⁺ and Mn³⁺ is mainly responsible for charge compensations. A capacity loss of only 0.018% in every cycle was realized after 1000 cycles at 1 C.

NaNiO₂ is stable in two polymorphs: a low-temperature layered O3 structure with a monoclinic distortion, and a high-temperature rhombohedral phase. Monoclinic NaNiO₂ (space group: C2/m) has been studied extensively owing to the relatively high voltage platform and redox reaction [103]. The structure is categorized as O'3 due to the Jahn-Teller effect of low spin of Ni³⁺ (t6 2 g e1 g), which is composed of alternately stacked Ni-O layer containing edge sharing elongated NiO₆ octahedra and Na layer [104,105]. Ceder [55] investigated electrochemical deintercalation of sodium from NaNiO₂ in 2012. It delivers a capacity of about 120 mAh g⁻¹ in 1.25-3.75 V with multiple plateaus, corresponding to 0.52 mol Na back. The complex phase transitions O'3-P'3-P''3-O''3-O'''3 is confirmed by in situ XRD [103,106]. When cycle in 2.0-4.5 V, it shows significant capacity decline in several deintercalation/intercalation, which is due to the formation of an inactive phase at high voltage zone. Unfortunately, the phase can't be identified. Recently, to understand in depth the structural evolution along with irreversible capacity loss in high voltage, Wang et al. [107] visually observed the structural evolution process by in operando TXM-XANES. It exhibits a dissymmetric spatial distribution with a “core-shell” reaction mechanism and one irreversible phase (Na_0.17NiO₂) result in discrepant intercalation (Fig. 4a, b). The irreversible capacity loss is resulting from two regions, below 3.0 V and above 4.0 V. The former is caused by increased c spacing after the first charge process whereas it cannot shrink back. The latter are affirmed to owing to unrecoverable lattice distortion and possible electrolyte decomposition. Therefore, it is vital to adopt effective strategy such as cation-doping or optimizing cutoff voltage to obtain higher reversibility and structure stability.

It is worth pointing out that pristine NaNiO₂ still shows inferior cycle stability and reversibility, and it is very hygroscopic against moist atmosphere, which slow down their practical application in sodium ion batteries. Many investigations have confirmed transition metal substitution have the positive effect in electrochemistry performance, such as Na_0.8Ni_0.6Sb_0.4O₂ [56,108], NaNi_1-yCo_yO₂ [109,110,111], Na_0.7Ni_0.35Sn_0.65O₂ [57,112], NaNi_0.5Ti_0.5O₂ [58,113]. Xu et at. prepared a Na_0.8Ni_0.6Sb_0.4O₂ cathode [56], delivering a capacity of 106 mAh g⁻¹ at 0.1 C based on the Ni²⁺/Ni³⁺ redox couple along with a high average working voltage of 3.5 V, which is due to the reduced orbital hybridization between O 2p and M 3d. In addition, the phase evolution process was believed to simplified O3-P3 with small volume change (1.0%), thus shows an excellent cycling life, compared with NaNiO₂ [106]. However, the rate performance is inferior may resulting from poor electronic conductivity. Trivalent cobalt also can suppress the cooperative Jahn-Teller distortion and enhance structure stability, whereas the content of cobalt haves to exceed 0.22 (NaNi_1-yCo_yO₂). It is well-established that the price of cobalt is expensive and the distribution is uneven in the earth's crust. Overmuch cobalt addition will make Na-ion batteries lose competitiveness compared with Li counterpart.

Guo et al. [57] introduce Sn into Ni-based layered oxide, it shows the highest redox potential of 3.7 V among all O3-type cathode. It can be explained by electronic structure and molecular orbitals. Sn⁴⁺ ([Kr] 4d¹⁰) does not hybridize with O 2p orbitals due to no single d electrons. Orbital hybridization only exists in O 2p orbitals and Ni e_g orbitals and charge delocalization is thus prevented in the metallic layer, leading to an increasement of Ni-O bond ionicity and Ni redox reaction. In contrast, other traditional transition metals (M') substitution such as Fe, Mn, O 2p orbital hybridization can simultaneously happen with both Ni e_g and M' e_g, thus the contribution from the Ni e_g orbital became smaller, resulting in a relatively lower voltage. The charging and discharging curve is smooth and shows a considerably flat platform. Unfortunately, the reversible capacity is only 64 mAh g⁻¹ in 2.0-4.0 V, which is relevant to internal Na-deficient (0.7).

It is confirmed that titanium as structure stabilizer can reduce the series of structural changes and increase the average potential, compared to half cells using NaNiO₂ cathodes. The discharge curves of NaNi_0.5Ti_0.5O₂ is composed of a slope and a platform rather than multistep potential curve [113]. Zhou’s research [58] demonstrated its good performance delivery with a reversible specific capacity of 102 mAh g⁻¹ at 0.2 C and capacity retention of 75% at 1 C after 300 cycles in the voltage range of 2.0-4.0 V, which is well ahead of NaNiO₂. The full-cell assembled by NaNi_0.5Ti_0.5O₂/ pre-sodiated hard carbon delivers a reversible capacity of 93 mAh g⁻¹ and good cyclic performance. However, the phase of NaNi_0.5Ti_0.5O₂ is always accompanied by a small quantity of NiO about 7 mol %, which is inevitable regardless of all synthesis conditions. Recently, Zhou’s group [59] design a cation-mixing stabilized cathode (Na_0.8Ni_0.3Co_0.1Ti_0.6O₂) by composition adjustment. It is verified that there is a typical phase transformation of O3-P3 in the first charge along with cation mixing. However, it unexpectedly maintains the P3 structure in the subsequent cycles (Fig. 4c), which is due to the TM ions in Na layers have a grasp effect on TM sheets. P3-type structure with trigonal prismatic sites can enable fast Na⁺ migration compared to the octahedral sites in O-type structure, thus can deliver great rate capability. In addition, the P3-type structure shows an extremely low lattice strain in desodiation and sodiation. Therefore, it delivers an excellent rate capability and superior capacity retention of 98% at 5 C after 1000 cycles. More recently, they presented reversibility of phase transition is highly relevant to oxygen vacancy concentrations and thus have a crucial role on electrochemistry performance [60]. The synthetic Na_0.9Ni_0.3Co_0.15Mn_0.05Ti_0.5O₂ sample in O₂ atmosphere have abundant oxygen vacancies but the oxygen vacancy decreases in Ar atmosphere. The vacancy-rich sample process an irreversible phase transition of O3-P3. However, the vacancy-poor sample is a mixed phase containing O3 and partial P3 after cycling. In other words, it is irreversible process. Moreover, the electronic and ionic conduction are facilitated due to incorporation of oxygen vacancy. As expected, it shows the splendid rate capability (about 50 mAh g⁻¹ at 100 C) and excellent cycling performance (capacity retention of 81.6% over 1000 cycles).

Recently, the new material families of Na₃Ni₂XO₆ (X = Sb, Bi, Ru, etc.) are presented owing to the superior structure stability and long lifespan [61,62]. Unlike disordered Ni-based layer oxides, the Ni²⁺ and X⁵⁺ are preferably formed an ordered distribution within the transition metal layers due to the unique proportion (2:1) and notably different ionic radius, thus usually form honeycomb layered structure with ordered superlattice [114,115], where each XO₆ octahedron is surrounded by six edge-sharing NiO₆ octahedra in the “ab” plane. It is worth pointing out that there are some weak reflections around 2θ = 20° in the XRD pattern based on the C2/m model, which is different from the diffraction peaks resulting from O3-type structure [116]. Honeycomb-layered Na₃Ni₂SbO₆ was first applied for an cathode in 2014 [62], it shows a capacity of about 120 mAh g⁻¹ based on Ni²⁺/³⁺ redox couple along with an average high working voltage at 3.3 V. Meanwhile, the rate and cyclic performance are superior compared to the NaNiO₂. However, the underlying mechanism is still not clear. Recently, Guo’s group [117] revisited the NaNi_2/3Sb_1/3O₂ and presented that ordered Ni₆-rings with super-exchange interaction form symmetric atomic configuration and degenerate electronic orbital in layered oxides (Fig. 4d), thus enhancing the structural stability in terms of air, thermal and phase stability, which is also supported by DFT calculations. In addition, the redox potential was increased and phase-transition process became simply compared with pristine NaNiO₂. Zhou’s group developed the superlattice-stabilized layered oxide cathode (Na₃Ni₂RuO₆) [61], which shows an ultralong voltage platform and a high discharge capacity of 130 mAh g⁻¹ along with reversible O3 to P3 phase transition. Meanwhile, it also presents excellent capacity retention (81%) after 1000 cycles with cation-migration-free. So stable structure can be explained visually by a “puzzle model”. TM ions can generate an interactive force resulting from ordered honeycomb-type arrangement, thus modify the crystal strain and remain the structural integrity during the numerous desodiation and sodiation (Fig. 4e). However, due to the price of raw material (RuO₂) is very expensive and abundance is low (only 0.001 ppm in the earth's crust), so it may be not suitable for large-scale commercial application.

NaNi_0.5Mn_0.5O₂ is a typical O3-type layered cathode, which was first developed by Komaba [32] in 2009. It shows great potential to achieve commercial application relying on high discharge capacity resulting from Ni^2+/3+/4+ redox couple, low cost for inactive Mn⁴⁺ and facile synthesis. It is worth pointing out that there is not almost cation mixing in NaNi_0.5Mn_0.5O₂ due to the comparatively different ion sizes of Na⁺(1.02 Å) and Ni²⁺(0.69 Å), compared to the Li counterpart LiNi_0.5Mn_0.5O₂ which containing Li⁺/Ni²⁺ disordering, thus making sure the better ion transfer [118]. In the voltage ranges of 2.2-3.8 V, it delivers a high reversible capacity of 125 mAh g⁻¹ at the current density of 4.8 mA g⁻¹. When the upper voltage is 4.5 V, a capacity delivery of 185 mAh g⁻¹ is realized, but the capacity deteriorated only after 20 cycles. The charging and discharging curve contained several potential plateaus resulting from sliding of slabs and Na⁺/vacancy ordering [119]. A structure evolution of O3-O'3-P3-P'3-P''3 was determined by ex situ X-ray diffraction [31]. Komaba presented fluorinated ethylene carbonate [120] as an electrolyte additive can effectively improve coulomb efficiency and cycling stability owing to the passivation and suppression of side reactions. The full cell NaNi_0.5Mn_0.5O₂ /hard carbon delivered a high capacity over 200 mAh g⁻¹ (based on negative electrodes) along with operating voltage of about 3 V. Even though the current density is 300 mA g⁻¹, it remains an almost similar capacity. Another full cell employed with Sn-C as anode [121] presented a high energy density of 300 Wh kg⁻¹ with good retention, further confirms commercial potential of NaNi_0.5Mn_0.5O₂.

Although NaNi_0.5Mn_0.5O₂ cathode materials show excellent superiority, some shortages have to be overcame for commercialization progress. Multiple phase transformation not only induce stepwise plateaus leading an unstable power export, but also disturb diffusion of sodium owing to the TM ions in different sites, leading terrible rate and cyclic performance [21,94]. Thus, suppressing phase transition and extending solid solution zone by appropriate elements addition can effectively alleviate the above problems. In addition, the intrinsic hygroscopic is also a crucial factor to hinder their commercial application, which will increase production and storage costs along with performance degradation. The Na-deficient phase is formed due to the spontaneous Na⁺/H⁺ exchange in ambient air [122,123]. Meanwhile, some insulating alkaline species such Na₂CO₃ and NaHCO₃ on the surface of bulk particle can decrease capacity delivery and trigger gelation of the slurry [124,125,126]. Proper elements substitution and surface modification can improve the problem.

Wang et al. [63] apply partial Ti to replace Mn in NaNi_0.5Mn_0.5O₂. Due to the similar ionic radius and same valence, it still maintained O3-type structure along with homogeneous distribution after introduction of Ti. Series of sodium-sufficient O3-NaNi_0.5Mn_0.5-xTi_xO₂ (0 ≤ x ≤ 0.5) materials was investigated and it can be concluded that NaNi_0.5Mn_0.2Ti_0.3O₂ cathode delivers the best electrochemical performance, which presents a high discharge capacity of 134 mA h g⁻¹ at 0.05 C in 2.0-4.0 V along with 85% capacity retention at 1 C after 200 cycles based on Ni²⁺/Ni³⁺ active redox couple. Meanwhile, the charge/discharge profiles become smoother and plateau potential is increased. The desodiation and sodiation process is corresponding to O3 − P3 transition via gliding of the TMO₂ layers (Fig. 5a). In other words, the Ti incorporation in O3-type NaNi_0.5Mn_0.5O₂ suppress the multiple phase transformation and decrease of the energy barrier for the two-phase conversion, leading a highly reversible phase transition along with excellent rate and cyclic performance. Komaba et al. [127] presented Mg and Ti co-substituted cathode Na[Ni_4/9Mn_1/3Mg_1/18Ti_1/6]O₂, which delivers a better capacity of 200 mAh g⁻¹ without any capacity loss due to substitution and enhanced cyclic performance in the voltage range of 2.2-4.5 V. In contrast to the NaNi_0.5Mn_0.5O₂, it demonstrates delayed O3-O'3 transition and decreased lattice parameter mismatch between O3 and P3 phases, resulting in gradual structural changes along with smoother curves. Furthermore, an OP2-like P-O_e-O_f intergrowth phase as an intermediate buffer phase effectively suppress the dramatic shrinkage in interslab spacing, leading a high reversibility of structure (Fig. 5b). After cycling, the surface of Mg and Ti co-substituted cathode is as smooth as pristine, but the surface deterioration is significant for NaNi_0.5Mn_0.5O₂. Hong and co-workers [128] demonstrated that Al-doped effectively facilitate Na⁺ mobility and improve structural stability due to increasement of the interslab distance resulting from stronger Al − O bond. O3-NaAl_0.02Ni_0.49Mn_0.49O₂ presented a high capacity retention of 63.2% after 200 cycles at 1 C along with coulombic efficiency of over 99% in the voltage range of 2.0-4.0 V. Recently, Zn²⁺ doping [64] is verified that can not only enhance electrochemical reversibility in the high voltage range of 4-4.2 V, but also improve air stability. The O3-NaNi_0.47Zn_0.03Mn_0.5O₂ cathode presents a reversible capacity of 113 mAh g⁻¹ at 0.5 C and good capacity retention. It is worth mentioning that a small volume change of approximately 2.00% in the voltage range of 2-4 V was demonstrated. Meanwhile, the average sodium ion diffusion coefficients of Zn²⁺ doping cathode is 2.22 × 10⁻¹⁰ cm² s⁻¹, which far beyond the pristine sample (1.37 × 10⁻¹² cm² s⁻¹). Tarascon [65] replaced Mn⁴⁺ by a non-transition metal ion Sn⁴⁺, preparing a series of NaNi_0.5Mn_0.5−ySn_yO₂ (y = 0-0.5). It is believed that Sn⁴⁺ substitution could reduce orbital overlap, favor charge localization in the metallic layer and increase bond ionicity, resulting in higher redox potential. Moreover, destabilization of P3 phase could suppress the phase transition from O3 to P3 due to increasing the ionic nature of the crystal lattice. Thus, it is a promising strategy to simultaneously increase the redox voltage and delay the phase transition. The fully substituted NaNi_0.5Sn_0.5O₂ phase shows a high redox potential of ≈3.2 V with a single-phase transition from O3 − P3. When y = 0.4, NaNi_0.5Mn_0.1Sn_0.4O₂ cathode delivers a capacity of 91 mAh g⁻¹ and superior capacity retention (85% after 200 cycles). However, H₂O insertion is present after Sn⁴⁺ substitution due to the increased interlayer distance, the application has been limited a certain extent.

Surface coating is considered an effective way to suppress the unfavorable side reactions during cycling and remain structure stability. Hwang [66] prepared MgO coating NaNi_0.5Mn_0.5O₂ with partial Mg doping. The coating layer of 70-100 nm formation on the particle surface effectively prevent the bulk particle HF erosion. Meanwhile, Mg²⁺ incorporation into the transition metal (TM) layers reduce the extent of the irreversible multiphase transformation (Fig. 5c). In the voltage range of 2.0-4.2 V, the MgO coating NaNi_0.5Mn_0.5O₂ shows improved capacity retention of 75% after 100 cycles at 0.5 C, though the discharge capacity slightly decreases due to the inactive MgO. The pouch-type full cells were assembled with MgO coating NaNi_0.5Mn_0.5O₂ and hard carbon, delivers excellent capacity retention of 70% after 200 cycles at 1.0-4.1 V (Fig. 5d). Zhang [129] performed a surface modification process by a rotary evaporator, which make AlPO₄ protective layer uniformly deposit on the cathode material (Fig. 5e). It can protect bulk particle from corrosion of electrolyte, decrease transition metal dissolution, and remain crystal structure. Meanwhile, it can protect the anode from increased resistance by suppressing the dissolution-migration-deposition process. The half-cell and full-cells show excellent cycling lifespan (95% after 400 cycles at 1 C, 78% after 200 cycles at 1 C, respectively) (Fig. 5f). In addition, the protective layer can further improve its environmental stability. Atomic layer deposition (ALD) technology is an effective way to realize surface coating at atomic scale with a precise and uniform thickness control, have been widely applied in many energy storage systems [7,130,131,132]. Yang [133] prepared thin Al₂O₃ coating (about 3 nm) NaNi_0.5Mn_0.5O₂ via ALD, which presents an excellent capacity retention of 68.0% after 300 cycles at 0.5 C, along with improved rate capability. In addition, the air sensitivity test was performed and showed more stable structure compared to the pristine sample. There is obvious impurity and hydrated compounds formation in the uncoated sample by XRD analysis. So, the Al₂O₃ layer effectively avoid the direct contact of ambient atmosphere, leading less degradation (Fig. 6a). Although the cyclic performance of layer cathode is obviously improved, it's essential to point out that most of coating layers are electrochemically insulated and discharge capacity is decreased, leading a lower energy density to a certain extent.

Hygroscopic characteristic is a crucial factor to hinder their commercial application. The spontaneous Na⁺/H⁺ exchange is easily happened in the O3-type NaNi_0.5Mn_0.5O₂ cathode in ambient air [122,123]. As the interlayer spacing become larger due to the increased repulsion between adjacent oxide layers, more Na⁺ are extracted, eventually resulting in a Na-deficient phase and deteriorative electrode performance[21]. Meanwhile, some insulating alkaline species on the surface of bulk particle are formed [124,125]. On the one hand, they decrease capacity delivery and deteriorate rate and cyclic performance. On the other hand, they will trigger defluorination of polyvinylidene fluoride (PVDF) binder, then cause the gelation of the slurry and rough surface, impeding their practical viability [135]. Surface coating strategy can avoid cathode exposing in ambient air containing H₂O and CO₂ to a certain extent. But it is usually accompanied by decreased energy density. Yao [11] designed an air-stable O3-type cathode NaNi_0.45Cu_0.05Mn_0.4Ti_0.1O₂ by Cu/Ti co-substitution, in which Cu has comparable electronegativity and Ti has substantial difference in Fermi level compared to the Ni²⁺/Mn⁴⁺, resulting in contractions of Na layers and increasement of the valence state of Ni. After exposed to air for 2 days, the structure is not almost changed and even retains its structure after being soaked in water (Fig. 6b). In contrast, O’3 and P3-type Na_1−yNi_0.5Mn_0.5O₂ phase arose just after 2 h for NaNi_0.5Mn_0.5O₂ cathode resulting from spontaneous Na⁺ extraction and insertion of H₂O molecules. In addition, NaNi_0.45Cu_0.05Mn_0.4Ti_0.1O₂ still shows high capacity retention after aging experiments. The Cu/Ti co-substitution cathode shows higher reversible capacity along with smoother discharge/charge curves and significantly enhanced cyclic performance (70.2% of capacity retention after 500 cycles). The phase transition was determined by in situ XRD, corresponding to highly reversible O3-P3 evolution, thus Cu/Ti co-substitution also effectively suppressed complex multistage evolution. Xiao [134] incorporate Mg²⁺ into the NaNi_0.45Cu_0.05Mn_0.4Ti_0.1O₂ cathode. The NaNi_0.40Cu_0.05Mg_0.05Mn_0.4Ti_0.1O₂ cathode presents outstanding capacity retention of 76.4% after 1000 cycles at 5 C and a stable midpoint voltage (Fig. 6c). Meanwhile, the O3-type phase could be maintained after 3 days (Fig. 6d). Zheng [136] demonstrated that ethanol washing can remove most of the alkaline species such as Na₂CO₃ and NaHCO₃ and remain original NaNi_0.5Mn_0.5O₂ structure.

Although O3-type Ni-Mn based oxides are superior to other oxides, but their practical application is limited by irreversible multiple phase transformation and hygroscopic characteristic. The frequent irreversible phase transitions are mainly due to the lack of shielding by the electrostatic repulsion of lattice oxygen and the gliding of transition metal layers. The hygroscopic characteristic is mainly related to the excess surface residual alkali and the large layer spacing, resulting in the insertion of water into the layered structure. In addition, the price of nickel is also rising year by year, which may trigger higher raw material costs, and it is necessary to properly reduce the Ni content and use low-cost element substitution. Among many elements, Fe has the advantage of low cost and the ionic radius is similar to Ni and Mn, which may be easily introduced into the NiMn-based material to reduce layers gliding. Yuan [67] prepared a serious of Fe-substituted NaNi_0.5Mn_0.5O₂ cathode materials. It is confirmed that Fe substitution to Ni-Mn layered oxide not only could effectively suppress the suppression of MO₂ gliding and Na/vacancy ordering resulting in reversible O3-P3 transformation, but also could expand the interslab distance which leads to fast ionic diffusion. The O3-type NaFe_0.2Mn_0.4Ni_0.4O₂ showed smoother the charge/discharge profiles compared with NaNi_0.5Mn_0.5O₂ and remarkably improved cycle stability with a capacity retention of 96.4% after 30 cycles in a voltage range of 2.0-4.0 V. A superior reversible capacity of 133, 133, 131, 125, 120, 112, 100, and 86 mAh g⁻¹ at current rates of 0.05, 0.1, 0.2, 0.5, 1, 2, 5 and 10 C (1 C = 240 mA g⁻¹) were also delivered. Furthermore, in the high voltage range (> 4.0 V), the O3-type NaFe_0.2Mn_0.4Ni_0.4O₂ undergo a different OP2 phase transformation with smaller interslab distance (5.13 Å) than the P3'' phase in NaNi_0.5Mn_0.5O₂ (5.72 Å), which effectively restrain co-insertion of the electrolyte solvent, leading improved cyclic performance. Sun et al. [137] proposed that with Fe contents increasement in the Ni-Fe-Mn based materials, the electrode exhibits more stable cyclic performance and perfect thermal stability although the reversible capacity slightly decreases. This further proves that addition of Fe has a positive effect on stabilization on crystal structure. Based on the DFT calculations, Guo et al. [68] proposed Fe³⁺ with slightly larger radius is introduced into NaNi_0.5Mn_0.5O₂ could enlarge transition metal layers and facilitate electronic delocalization. The phase transformation process of the NaFe_0.3Ni_0.35Mn_0.35O₂ is single O3 − P3 without any monoclinic distortion (Fig. 7a). The capacity retention ratio of NaFe_0.3Ni_0.35Mn_0.35O₂ could reach 80% after 100 cycles at 1 C (1C = 240 mA g⁻¹) within 2.0 − 4.0 V voltage range and the rate performance were also obviously enhanced. In addition, the air stability of Ni-Fe-Mn based materials was effectively improved and it remained the O3 crystal structure after exposed to air for 1 day which was verified by directly at atomic-scale by ABF-STEM. Due to the excellent performance and lower price of Ni-Fe-Mn based material, the commercialization process is also advancing, Ma's group [27] realized kilogram-scale production of NaFe_1/3Ni_1/3Mn_1/3O₂ material by co-precipitation method and following calcination, which widely used in the production of layered oxide cathode materials for lithium-ion batteries. Meanwhile, they successfully prepared 1 Ah soft-packed batteries using NaFe_1/3Ni_1/3Mn_1/3O₂ and hard carbon as cathode and anode, respectively. It remains outstanding long term cycling performance with capacity retention over 73% after 500 cycles between 1.5 V and 3.8 V at 1C rate (Fig. 7b). The corresponding XRD diffraction pattern is almost same as the initial sample, further demonstrating its excellent structural stability. Afterwards, they used optimized electrolyte [28] (1 M NaPF₆ in 1:1 PC-EMC with 2wt% FEC, 1wt% PST, and 1wt% DTD) to enhance the long cycling life with 92.2% capacity retention after 1000 cycles at 1C. On the one hand, the PST and DTD additives can produce more organic molecular moieties, resulting in a burly SEI on the anode, thus irreversible reduction (decomposition) of the electrolyte solvents could be restrain. On the other hand, transition metal ions could combine with decomposition products such as RSO₃⁻, ROSO₃⁻ and SO₃²⁻ anions to form insoluble transition metal sulfates or sulfites, resulting in a dense and thick CEI, which will reduce dissolution of transition metal ions (Fig. 7c).

Generally, Ni-Fe-Mn-based materials exhibit good cycling performance, but significant performance deterioration along with structure monoclinic distortion often occur in the high voltage region (> 4.0 V), while the plateau voltage needs to be increased to obtain higher energy density for practical applications. In addition, dissolution of transition metal ions and side reactions between active materials and electrolytes occurs during cycles, leading to loss and fracture of active material. For the above problems, ion doping, surface modification and composite structure construction are effective strategies.

The ion doping can modulate the crystal structure and make the phase transition highly reversible, while the dopant ions can combine with oxygen ions to form stronger chemical bonds, increasing the solidity of the layered structure. The Ti instead of Mn not only increases the lattice spacing, but also can lift the working potentials [69], which delivered improved cyclic performance (capacity retention of 84% after 200 cycles at 0.1C) and higher energy density of about 460 Wh kg⁻¹. Ding and co-workers [138] realized atomic-level mixing of Ti in Ni-Fe-Mn based materials by sol-gel method. The Na[Ni_0.4Mn_0.4Fe_0.2]_0.95Ti_0.05O₂ showed excellent rate capacities of 160.7, 132.1, 125.1, 120.9, 114.7 and 99.6 mAh g⁻¹ at 0.1, 0.2, 0.5, 1.0, 2.0 and 5.0 C at 2.0-4.0 V, respectively (1C = 140 mA g⁻¹) due to the larger parameter c from Ti introduction. The corresponding capacity retention also increased from 69.6% (NaNi_0.4Mn_0.4Fe_0.2O₂) to 83.8% after 200 cycles at 1 C. In addition, it is verified that the irreversible multiphase transformation of the P3 to O3 phase was effectively restrained. Recently, Park [139] confirmed the energy barrier of Na⁺ migration decreased after Ti substitution into O3-type NaFe_0.25Ni_0.25Mn_0.5O₂ based on DFT calculations.

The main factors limiting the performance of O3 Ni-Fe-Mn based layered oxide materials are migration of transition metal ions and irreversible phase change at high voltage. The transition metal ions in TM layer migrate to adjacent Na layer, hindering the transport of sodium ions. In a high cut-off voltage (> 4 V), with most Na⁺ extracting from the unit cell, the layer structure tends to glide and distort, leading irreversible phase transformation, which is accompanied by large variation of the lattice volume and a fast capacity attenuation. Sun [140] proposed Li introduction into the transition metal layer could remain structure stability due to stronger Li − O bond relative to Ni − O and Mn − O bonds and the migration of Fe³⁺ could be restrain. In addition, electronic conductivity, capacity, its retention, and rate capability were improved after Li incorporation. Guo [70] put forward the dual-stabilization effect of Li substitution in high-voltage in detail. One is unfavorable P3-P'3 phase transition at high voltages (> 4.2 V) was suppressed due to the weak Jahn-Teller distortion of Ni³⁺. The other is Li may bond with F⁻ anion, leading a stable electrolyte-cathode interphase, which is useful for less transition metal ions dissolution (Fig. 7d). Thus, the Li doped sample Na_0.85Li_0.1Ni_0.175Mn_0.525Fe_0.2O₂ displayed a high discharge capacity of 160 mAh g⁻¹ at 0.1 C at the voltage of 2.0-4.5 V and excellent capacity of 88% after 100 cycles at 1 C. It is well-established that Mn³⁺ ions are accompanied by strong Jahn-Teller effect, which seriously damage the structural stability and the cyclic performance. By suitable element doping, the content of Mn³⁺ can be reduced and thus the structural stability can be improved. Liu [141] prepared O3-NaMn_0.48Ni_0.2Fe_0.3Mg_0.02O₂ materials, which showed lower Mn³⁺ and minimize Jahn − Teller effect, leading outstanding structure stability. The reversible capacity almost has no attenuation after 100 cycles at 0.1 C in the voltage range of 1.5-4.2 V (the capacity retention of NaMn_0.5Ni_0.2Fe_0.3O₂ is 81%). Huang [71] proposed that F-doped result in a lower Mn³⁺ content, which stabilize the O3-type layered structure. However, the discharge capacity has little decrease. So, it is crucial to have balance for cyclic performance and reversible capacity. In addition to the elements mentioned above, Ca [142], Al [143] and Zn [144] were proved to have remarkable enhancement on rate capability, cycling stability and reversible phase transformation.

Co-doping of multiple elements are an effective strategy to obtain a sample with excellent overall performance, such as high thermal stability and fracture toughness, splendid low temperature performance, high strength and air stability [145,146,147]. In the layered oxide cathode, Ni²⁺, Fe³⁺ and Cu²⁺ could provide charge compensation for high capacity delivery, Mg²⁺, Mn⁴⁺ and Ti⁴⁺ have positive effect on structure stability, Sn⁴⁺ and Sb⁵⁺ could raise the average voltages of the materials, which increase energy density. Co³⁺ has electrochemical activity and may counter the Jahn − Teller effect of Mn³⁺, leading to good structure stability. Hu’s and co-workers [72] employed high-entropy configuration strategy to design O3-type NaNi_0.25Mg_0.05Cu_0.1Fe_0.2Mn_0.2Ti_0.1Sn_0.1O₂ oxide. It was proposed that the multicomponent transition metal ions strengthen the whole skeleton structure of layered oxides by weaker Jahn − Teller distortion, Na⁺/vacancy ordering, and lattice parameter changes (Fig. 7e). Meanwhile, the TMO₂ slabs were expanded to make sure faster ion transport, compared to the baseline sample (NaNi_0.4Fe_0.2Mn_0.2O₂). This material delivered a reversible capacity of about 130.8 mAh g⁻¹ and energy density of 421 Wh Kg⁻¹ in a voltage range of 2.0-4.0 V at 0.1 C (1 C = 140 mA g⁻¹). It remained a capacity retention of 75% over 500 cycles at 1 C, which is much higher than the baseline sample (38%). It also showed outstanding full cell performance when coupled with hard carbon as anode. The discharge capacity and energy density reached 123.5 mAh g⁻¹ (based on cathode mass) and 260 Wh kg⁻¹ (based on the total mass of active materials) in the voltage range of 1.0-4.0 V at 0.1 C. The capacity retention is about 76.5% after 500 cycles at 0.5 C, which is pretty outstanding compared to the other layered oxide cathode. Chen and co-workers [73] designed a five-components O3-type layered oxide cathode (Na_0.94Ni_0.29Cu_0.1Fe_0.16Mn_0.3Ti_0.15O₂) with uniform distribution of elements (Fig. 7f). The charge-discharge curve is consisted of voltage plateau around 3 V and a sloping voltage profile at 3.0-4.0 V, which is corresponding to a highly reversible O3-P3 phase transition (without any monoclinic phase). So, it delivered a reversible capacity of 122 mAh g⁻¹ and high capacity retention of more than 79% after 300 cycles at 0.5 C in the voltage of 2.0-4.0 V. Recently, Lin et al. [148]. also reported a high-entropy O3-type cathode, containing similar chemical composition. The NaCu_0.1Ni_0.3Fe_0.2Mn_0.2Ti_0.2O₂ showed excellent long-term cycling stability (approximately 71% after 500 cycles at 0.5 C), which is attributed to the stabilization effect of entropy [149,150].

In short, ion doping is an effective strategy to suppress irreversible phase transitions, reduce transition metal ion dissolution and improve the structural stability of layered materials. However, the mechanism of ion doping for material property improvement still needs further exploration and in-depth study, especially multi-ion co-doping.

In addition to the ion doping, surface modification is also one of the effective strategies to enhance overall performance of cathode. Uniform coating layer can mitigate the corrosion of cathode from the electrolyte and maintain structural integrity, resulting an improved long cycle performance. In particular, the fast ion conductor such as NaTi₂(PO₄)₃ [151,152], Na₂SiO₃ [74] and etc. as coating layer can accelerate Na⁺ diffusion at the interface, further increasing the rate performance. Xiao and co-workers [74] reported a O3 NaNi_1/3Mn_1/3Fe_1/3O₂ uniformly wrapped by Na-ion conductor Na₂SiO₃ via in situ coating (Fig. 8a). The Na₂SiO₃ coating layer can reduce polarization and suppress voltage drop during charge-discharge process. Meanwhile, Na⁺ diffusion was effectively improved due to the three-dimensional structure of Na₂SiO₃. Thus, 5 mol% Na₂SiO₃ coated NaNi_1/3Mn_1/3Fe_1/3O₂ cathode showed excellent cycling performance. The capacity retention could reach 88.3% after 200 cycles at 0.5 C in the voltage of 1.5-4.2 V. The rate capacity is 58 mAh g⁻¹ at current density of 5 C, while the uncoated sample remain a capacity of only 12 mAh g⁻¹. Yu et al. [36] reported a synergetic strategy to improve electrochemical performance of Ni-Fe-Mn based cathode by integration TiO₂ coating and Ti⁴⁺ doping. On the one hand, the TiO₂ coating on the surface of the particle inhibits the side reaction and reduces the dissolution of transition metal ions. On the other hand, Ti⁴⁺ is successfully doped into transition metal layer of bulk phase structure, while increases the sodium layer spacing and improves the diffusion of sodium ions. In addition, the content of Mn³⁺ is reduced and the Jahn − Teller effect is suppressed, leading entirely reversible phase transformation (Fig. 8b). This material yields fabulous rate capability of 197.2, 160.9, 140.7, 118.3, 103.6, 88.3, 62.6, 41.8, and 30.0 mAh g⁻¹ at 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 8, and 10 C (1 C = 200 mA g⁻¹), respectively. It also delivered good cycling stability with about 56.2% and 71.0% at 0.1 and 1 C after 100 cycles in the voltage of 1.5-4.2 V. As we all know that oxygen ions may participate in redox to provide additional capacity, which is common in the cathode materials for Li-ion batteries. Liu and co-workers [75] adopt similar synergetic strategy to enhance the rate and cyclic performance by simultaneously coating and doping. In particular, Zr⁴⁺ doping adjusts the Na-O-TM structure and the electronic state of O-2p occupies a higher energy level which is supported by DFT calculations, resulting in the higher reversibility and cycle stability of oxygen redox (Fig. 8c). Thus, ZrO₂ coated NaNi_1/3Fe_1/3Mn_1/3O₂ delivered high reversible specific capacity of 152.4 and 157.1 mAh g⁻¹ at 0.05 C in the voltage of 1.5-4.3 and 1.5-4.6 V, respectively. It also remains a capacity retention of 81.9% at 100 cycles at 1 C.

Air stability has a huge impact on electrode assembly, production and transportation. Uniform and dense coating layer can solve this problem very well. Lamb et al. [100]. reported a phosphate-based coating on O3 Na(Ni_0.3Fe_0.4Mn_0.3)O₂. The capacity retention of 1% (NaPO₃)_n coating Na(Ni_0.3Fe_0.4Mn_0.3)O₂ increased 14.0% after 100 cycles at 0.25 C, compared to the pristine Na(Ni_0.3Fe_0.4Mn_0.3)O₂ (Fig. 8d). After exposed to the ambient air for 9 days, it only loss a capacity of 30 mAh g⁻¹ (the capacity loss of pristine sample is 80 mAh g⁻¹). The reasons for showing excellent cyclic performance and air stability could be attributed to the following points: firstly, the coating layer reduced corrosion of electrode materials by electrolyte. Secondly, when a large quantity of phosphate coating is applied, sodium is leached from the bulk to form sodium phosphate, leading to the formation of inert metal oxides on the surface. It is worth pointing out that over 1 wt% coating will cause lower initial capacity. In addition, some other materials such as Al₂O₃ [153] and AlF₃ [45] were also employed as effective coating layer to improve structure stability and integrality. A portion of the coating material will cause a slight decrease in reversible capacity and energy density, so proper addition ratio is critical.