A Review on High-Capacity and High-Voltage Cathodes for Next-Generation Lithium-ion Batteries

Shuvajit Ghosh ^† , Udita Bhattacharjee ^† , Subhajit Bhowmik ^† , Surendra K. Martha ^*

Department of Chemistry, Indian Institute of Technology Hyderabad, Kandi, Sangareddy, 502285 Telangana, India

† These authors contributed equally to this work.

* Correspondence: Surendra K. Martha

Academic Editor: Ahamed Irshad

Special Issue: Batteries: Past, Present and Future

Received: June 30, 2021 | Accepted: November 22, 2021 | Published: January 10, 2022

Journal of Energy and Power Technology 2022, Volume 4, Issue 1, doi:10.21926/jept.2201002

Recommended citation: Ghosh S, Bhattacharjee U, Bhowmik S, Martha SK. A Review on High-Capacity and High-Voltage Cathodes for Next-Generation Lithium-ion Batteries. Journal of Energy and Power Technology 2022; 1(1): 002; doi:10.21926/jept.2201002.

© 2022 by the authors. This is an open access article distributed under the conditions of the Creative Commons by Attribution License, which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is correctly cited.

1. Introduction

Research on energy conversion and storage has been rising for the last few decades, and lithium-ion batteries (LIBs) have dominated this space ever since their invention. First marketed by SONY in 1991, LIBs have established a monopoly in the market of portable electronics devices [1]. LIBs come with a unique blend of characteristics such as 150-200 mAh g^-1 specific capacity, 3.3-4 V average voltage, 200-250 Wh kg^-1 gravimetric energy density, 350-600 Wh L^-1 volumetric energy density, at least 3 years of life, >99% energy efficiency, 0.3-2.5% self-discharge rate per month, and almost no maintenance [2,3,4]. However, recent concerns over global warming and other environmental issues incentivized the demands of electric mobility (e-mobility) to shrug off the dependence on pollution-causing fuels. Superior electrochemical properties make LIBs the automatic choice for e-mobility. But the energy density and safety aspects of LIBs need to be reconsidered for applications like long-haul electric trucks, electric trains, electric aviation, etc. At the same time, the cost is required to be lowered for stationary grid storage [5,6,7].

Conventional LIBs comprise a Li-ion insertion anode (e.g., graphite, Li₄Ti₅O₁₂, metal oxides, alloys, etc.) and a cathode (e.g., layered metal oxide like LiCoO₂, spinel oxides like LiMn₂O₄, or olivine LiFePO₄, etc.). The electrolyte used for the traditional system is a Li-salt (e.g., LiPF₆, LiBF₄) in organic solvents like a mixture of alkyl carbonates (e.g., ethylene carbonate, dimethyl carbonate, etc.). A full Li-ion cell consists of 10-15% of lithium [8,9]. Reversible to and fro movement of Li⁺ and e^- via an inner and outer circuit, respectively, results in the interconversion between electrical and chemical energy.

The cathode materials for LIBs can be classified into insertion type and conversion type based on mechanisms [4,10,11]. Conventionally, the cathode materials used for LIBs are insertion-type metal oxides such as layered LiMO₂ (M = Co, Ni, Mn), LiNi_1-y-zMn_yCo_zO₂ (Li-NMC), Li_xMn₂O₄ type spinel, polyanionic based transition metal phosphates (e.g., LiFePO₄), etc. The electrode reaction for insertion-type materials does not involve any structural change in the host matrix and hence performs with stability during prolonged cycling. But the redox reaction for this type of material is Li-site limited and involves the transfer of a maximum of one electron per formula unit. The LiMO₂ type cathode materials deliver practical capacities in the range of 140-220 mAh g^-1. The upper cut-off voltages for LiCoO₂ and LiNiO₂ cannot be increased beyond 4.2 and 4.3 V, respectively, as oxygen gas is released due to their thermal instability. LiCoO₂ further undergoes structural distortion to tetragonal phase from the electrochemically active hexagonal phase beyond 4.2 V. LiNiO₂ also suffers degradation through Li/Ni cation mixing (antisite disorder) [12,13,14,15,16]. LiMnO₂ is chemically stable at higher voltages (till ~ 4.8V) as no exothermic gas evolution occurs. But during cycling, performance degradation is observed due to the dissolution of active form to inactive components [4]. Thus, a mixed oxide in different compositions of LiNi_1-y-zMn_yCo_zO₂ (Li-NMC) was reported to cycle with stabile capacity (140-180 mAh g^-1 as compared to LMO due to the synergistic effects of the conductivity of Co, high voltage of Ni, and thermal stability of Mn. Ni and Co are electrochemically active, and Mn acts as a stabilizer). Li and Mn-rich NMCs are stable up to 4.8 V and delivered a capacity of ~250 mAh g^-1 during initial cycles. Still, the performance degrades due to the formation of various by-products through secondary reactions such as activation of Li₂MnO₃ beyond 4.4 V [17,18,19]. Li_xMn₂O₄ shows a theoretical capacity of 148 mAh g^-1 with the extraction of 1 lithium at an average potential of 4 V, indicating higher thermal stability than LCO. The main reason behind the degradation of spinel Li_xMn₂O₄ (03+ disproportionation reactions below 3V [20,21,22]. The disproportionation reaction can be controlled by maintaining the oxidation state of Mn above 3.5 through doping other metals like Ni, Ti, Mg, etc., and surface coatings with highly conductive carbons like carbon nanotubes (CNTs), graphene, etc. [10]. Olivine structures, such as those of LiMPO₄ (M = Fe, Mn, Ni, Co), are characterized by high structural stability and Li-ion storage capacities around 170 mAh g^-1 [23]. LiFePO₄ has an average voltage of 3.5 V vs. Li/Li⁺, and LiMnPO₄, LiCoPO₄, and LiNiPO₄ have a Li insertion potential of 4.1, 4.8, and 5.1 V, respectively [10,23,24]. Phospo-olivines composed of stable mixed metallic compositions are being explored as high-energy cathode materials due to their higher lithium insertion potential.

The overall performance of LIBs is cathode-limited in terms of ion storage capacity and voltage output. The reversible capacity of commercial graphite and silicon anode is much higher than that of conventional cathodes. Thus, to balance the capacity ratio of cathode and anode in a lithium-ion cell, an excess amount of active cathode material must be used to increase the total weight of the cell. The extra cathode portion comes with a cost as it contains expensive transition metals, also increasing the total cost of the cell. High-capacity cathode materials may cut down that extra cost and weight per cell of a LIB module. High voltage cathode materials are advantageous from the perspective of LIB modules. If the voltage of each cell is relatively higher, a lesser number of cells are required to reach the targeted module voltage decreasing the complexity of cell packing. Moreover, the requirement of a lower current to deliver a particular power value reduces IR (internal resistance) loss. The criteria to balance all the properties like higher energy density, improved safety, low cost, etc., in a single material is challenging, and the research window on novel cathode materials is always open [25,26]. The energy density of the material can be increased by either pushing the capacity or lifting the voltage. Techniques for capacity increment are exclusively studied, but the paths for voltage upliftment are less explored. Elaborate discussions on commercial cathode materials are available in the literature [27,28,29,30,31]. Scope of this article is stressed on the materials that show either capacity >250 mAh g^-1 (lithium excess layered oxides, lithium excess cation disordered rock salt oxides, conversion cathodes) or voltage >4 V (polyanionic oxides, spinel oxides) and yet to be commercialized. Figure 1 depicts the material that follows the selection criteria. This review focuses on understanding the crystal structure, lithium storing mechanism, potential shortcomings, and probable way-outs of cathode materials. Also, a historical timeline of the development of cathode materials for LIBs is given briefly, and the remaining challenges for commercialization are sketched.

Figure 1 Approximate range of average discharge voltage and specific capacity of LIB cathode materials. Materials that lie beyond the 4 V-250 mAh g^-1 range (dotted line square) are chosen to discuss in this review.

2. Historical Development of LIBs

During the decade of 1960-1970s, research on energy storage and conversion became concentrated towards the electrochemistry of H⁺/alkali metal ion storing at solid state [32,33]. The electrochemical concept of the modern-day LIBs came into the limelight after the unveiling of guest-host intercalation into the metal oxide matrix by Armand et al. during the 1970s [34]. This work demonstrated the intercalation of ions into Prussian blue-type materials such as iron cyanide bronzes in 1972 [35]. During the same year, Brian Steele at the NATO conference in Italy and other group researchers, including Gamble et al., and Dines et al. (Exxon USA) indicated the applicability of transition metal dichalcogenides as intercalation electrode materials [36,37,38,39] In the same conference, Armand revealed demonstrated CrO₃ embedded within graphite planes as the electrodes for the first reported Li/Na solid-state batteries with β-alumina electrolyte [39]. The metal chalcogenides showed faster intercalation kinetics according to the reports by Rao et al. (IBM, USA) and Rouxel group in France during 1974 [40,41]. During that same period, M.S. Whittingham used layered metal dichalcogenides to store Li-ions. He named the Li-ion storing mechanism as ‘the intercalation mechanism’, and showed reversible at room temperature (298 K)[42]. In 1976, Exxon, USA, developed Li||TiS₂ batteries with energy densities of ~130 Wh kg^-1 (280 Wh L^-1). TiS₂ is a hexagonal packed crystal structure having lithium intercalation sites with a theoretical capacity of 240 mAh g^-1. But the nominal voltage for these cells is ~ 2 V, which limited their deliverable energy density. Also, difficulty in material production and spontaneous secondary reactions in the cell resulting in the release of H₂S gas became an issue for practical use [10].

In the later part of 1979, Godshall et al. demonstrated the intercalation property of LiCoO₂ (popularly abbreviated as LCO) at a high potential of ~ 3.7 V at elevated temperature conditions (400-450 °C) [43]. Around the same time (in the early 1980s), J.B. Goodenough and coworkers reported the Li-ion intercalation property of LiCoO₂ at room temperature and thus, paved the way for feasible high voltage cathode materials for LIBs [44]. In 1983, the lithium intercalation property of the spinel-type oxide, LiMn₂O₄ cathodes with higher thermal stability compared to LCO was studied by Thackeray et al. [45]. Meanwhile, reversible Li-ion intercalation into graphitic carbon was being studied by Samar Basu (1981) [46] and Yazami in 1983 [47], which led to the development of the Li-ion battery by Akira Yoshino using the LCO cathode and intercalation-type carbon coke anode in the year 1985 [48]. Around the same time (during 1987), Moli energy Ltd commercialized the Li-MoS₂ battery (MOLICEL) with an energy density of 60-65 Wh kg^-1 and potential in the range 2.3 - 1.3 V, but all the batteries were called off due to fire incidents. This unfortunate incident surceased the development of lithium metal batteries (LMBs) during that era [49].

The dominance of LIBs in the energy industry began with the first commercialization of the 18650 cells by Sony Ltd. in 1991 [50]. Following this path-breaking discovery, several attempts were made to improve the energy density and cycling performance using different cathode materials. J.B. Goodenough introduced polyanionic oxides and phosphates for Li-ion insertion cathode, and in 1996 Padhi et al. demonstrated the performance of 1D olivine LiFePO₄ as LIB cathodes [23]. Other polyanionic compounds such as silicates, sulfates, phosphates, vanadium phosphates, fluorophosphates, and fluorosulphates were studied as high voltage cathodes for LIBs thereafter. In 1998, layered lithium manganese oxide cathodes derived from rock salt precursors were introduced [51].

VL18650 based on lithium mixed oxide LiNi_0.8Co_0.15Al_0.05O₂ termed as NCA cathodes were commercialized by SAFT for longer temperature range (-20 to 60 °C) applications. In the early 2000s, layered mixed oxides LiNi_xMn_yCo_zO₂, where x + y + z = 1 (NMC) were demonstrated as LIB cathodes with higher stability, and since then, various compositions of these mixed oxides are studied. Thackeray and the group introduced Lithium-rich NMC with higher capacity in 2005 [52], and the conductivity was enhanced using carbon coating to increase stability. Thereafter, various strategies, including doping with metals such as Ni, Al, Mn, compositional optimization, nano-structuring, and composite formation with the carbon of the layered, spinel, and polyanionic cathodes, have been explored and optimized for commercial applications. Other than these types of insertion cathodes, conversion type cathodes with higher capacity have also been studied for LIBs. In 2005, M.S. Whittingham demonstrated two lithium transfer reactions from Li₂VOPO₄ [10,53,54]. In the late 2000s, the concept of anionic redox was introduced, and for the first time, reversible cathode capacity beyond 250 mAh g^-1 was achieved. Measuring the huge success of rechargeable LIBs in all portable electronic sectors, electric mobility, and other potential fields, Nobel Prize was awarded to J.B. Goodenough, Akira Yoshino, and M.S. Whittingham in 2019. Over the last three decades, LIBs have been manufactured by several companies such as Sanyo, Matsushita, Hitachi, Yuasa, Moli, A&T, SAFT, TESLA, etc., and the research and development in this field have increased exponentially with an increase in the demand for energy with growing civilization. Figure 2 highlights the essential discoveries made in the history of LIB cathodes.

Figure 2 Milestones in the history of LIB cathode discovery.

3. Description of Two Basis Performance Measuring Unit of Battery: Capacity and Voltage

3.1 Capacity of a Battery

The capacity of a battery reflects its capability to store energy. From the material point of view, it is the ability to store lithium ions reversibly. Lithium ions can be stored at a particular crystal site in intercalation-type materials, or they can react with the host material in conversion-type materials. The storage capacity is determined by the amount of lithium that can be reversibly inserted into the host via a reversible first-order phase transition. Insertion of charged species, i.e., Li⁺, into a host changes its valence state, which is compensated by oxidation or reduction of TMs and ejection or incorporation of electrons in the metal d-band. The factors for determining energy storage capacity: - a) reversibility of TM redox properties, b) the available space for accommodating lithium ions, and c) reversibility of lithium insertion reactions. Pathways to increase capacity include modifying any such property mentioned above. Generally, conversion-type cathodes demonstrate higher capacity than insertion-type materials. Conversion reaction between Li and host generates high capacities, whereas insertion sites inside a crystal are limited by the orientation of the crystal-forming elements. This article reviews only those materials which show a capacity of >250 mAh g^-1. Lithium-rich ordered, lithium-rich cation disordered rock salt, and conversion type materials fall under this category.

3.2 Potential of a Battery

The difference in chemical potential between the anode (µ_a) and cathode (µ_c) divided by the magnitude of electronic charge (|e| is taken for calculation, neglecting the negative sign) is termed as open-circuit voltage (OCV) or working voltage.

\[ V=\frac{\mu_{a}-\mu_{c}}{|e|} \]

The voltage is limited by the electrochemical potential window of the electrolyte. The energies of µ_a and µ_c should reside in between the energies of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the electrolytes (Figure 3a). Otherwise, electron transfer occurs from HOMO to µ_c and µ_a to LUMO, resulting in the oxidation of electrolyte at the cathode and reduction of electrolyte at the anode, respectively [55]. Decomposed product accumulates on the electrode surface to form a surface film called ‘electrode-electrolyte interphase (EEI)’ or solid-electrolyte interphase (SEI) more commonly at the anode [56]. Porosity, Li-ion conductivity, thickness, chemical stability, and compatibility of the interphase with other cell components are critical for a battery's function. The growth of SEI increases the internal resistance, which causes the voltage and capacity fade. In terms of molecular orbital theory, the relative Fermi level energy in the band structure of LIB cathodes is defined as OCV. Fermi level (E_F) is related to the electrochemical potential. Holes above E_F and electrons below E_F form a redox couple. If the significant TM redox couple in the structure overlaps with the O-2p band, then oxygen evolution concomitantly occurs with TM redox. Therefore, the material should be designed in such a way that overlap can be avoided.

Figure 3 (a) Electrochemical potential window of the electrolyte. Relative energy levels of cathode, anode, and electrolyte at open circuit condition and (b) Relative position of the various transition metal redox couples in between metallic lithium and oxygen energy levels. Reprinted with permission from reference [57]. Copyright 2017 American Chemical Society.

3.2.1 Origin of Potential Hysteresis

Asymmetry between the charge and discharge profile is expressed as potential hysteresis. The origin of hysteresis lies in the difference in the energy transfer between the charge and discharge processes. The charge is associated with lithium extraction from the host lattice and ejection of the electron from the d-orbital of TM . During the discharge process, lithium is reinserted into the designated site, and the electron gets back to the d-orbital. The reinsertion of lithium and electron stabilizes the host, energetically excited after charge, while deinsertion of lithium and electron during charge destabilizes the system. Therefore, prior to the process of deinsertion, lithium-ion and electron must be promoted to the excited state. The extra promotion energy accounts for the relatively higher energy gain in charge than delivered during discharge [58]. This energy difference is attributed to the potential gap between the charge and discharge processes.

3.2.2 Raising the Voltage of a Battery

The electrochemical potential (E) of the electrode material is related to a change in the internal energy due to ion and electron transfer, i.e., Gibbs free energy (ΔG). Ions exert a dominant effect over electrons. The relation can be expressed as follows:

\[ E=\frac{-\Delta \mathrm{G}}{\mathrm{nF}}, \]

where n is the number of electrons transferred during the charge and discharge processes, and F is the Faraday constant. Therefore, the equation predicts that the more energy released during ions or electrons are inserted into the host lattice, or more energy consumed during ions or electrons are released, the higher is the electrochemical potential. The change in the Gibbs free energy depends on the ionic radius, valence state, electronegativity, and local environment of the metal ions in the host.

Insertion Oxides: Going from left to right in a particular period in the transition metal series of the periodic table, the number of d-electrons increases. If the ions are under similar coordination and valence, ionic radius gradually decreases, and the attraction on outermost electrons by atomic nuclei increases. Therefore, greater energy is needed for electron transfer resulting in higher electrochemical potential in going from left to right in a particular period. Extending the similar line of logic, electrochemical potential decreases going down a group in the periodic table as the ionic radius of the ions having the same valence and coordination states decreases down the group. Figure 3b provides the relative position of the transition metal redox couples between Li⁺/Li and O-2p energy levels [57].

Polyanionic Groups: Polyanionic groups are different from layered oxides. Oxygen is coordinated with X atom besides TM (M), unlike oxides, creating M-O-X linkages instead of M-O linkages. For easy understanding, let us focus on Fe. For instance, in Fe₂(XO₄)₃, (X = Mo, Si, P, S, W, B, etc.), the FeO₆ octahedra shares corners with the XO₄ tetrahedra, and extended version on M-O-X linkage can be rewritten as -O-Fe-O-X-O-Fe-O-. Therefore, the presence of a less electronegative X in place of greater electronegative oxygen increases the covalency between the O-X linkage rendering the M-O bond more ionic than earlier. Thus, the inductive effect of X increases the operating voltage of the system [59]. The trends in the operating potential observed in some iron-containing (Fe^+3/+2 couple) polyhedral compounds are depicted in Figure 4. Generally, the operating potential of several polyanionic groups runs as follows -P₂O₇>PO₄>BO₃>SiO₄. A high degree of electron delocalization in the P₂O₇ moiety raises voltage slightly higher than PO₄.

Figure 4 Relative voltage of the Fe^+3/+2 couple in various polyhedral structures.

4. High-capacity Cathodes (Only >250 mAh g^-1)

Materials that show a capacity of >250 mAh g^-1 are only included in this review.

4.1 Lithium Excess Layered Oxides

Lithium and Manganese-rich layered oxides (LMR-LOs) is another class of cathode materials that demonstrate a high capacity of >250 mAh g^-1 along with high voltage to push the specific energy density beyond 1000 Wh kg^-1. Table 1 shows the comparison of LMR-LOs first with other commercial LIB cathode materials in terms of electrochemical characteristics. The crystal structure, lithiation-delithiation mechanism, voltage fade phenomena, poor cycle life, etc., are quite distinct in LMR-LOs than the conventional class of materials. This section sheds light on these aspects.

Table 1 Relative electrochemical properties of standard LIB cathode materials.

4.1.1 Chemical Composition and Structure

The concept of LMR-LO cathodes was first proposed in the 1990s [60]. The chemical formula is denoted as xLi₂MnO₃·(1-x)LiMO₂, where M is in the form of different metals such as Co, Ni, Mn, Cr, Fe, Mg, Al, etc., or a mixture of them give rise to various compositions. Classically, the crystal structure (Figure 5) comprises of two components: a rhombohedral LiMO₂ phase with a space group $R \overline{3} m$ and a monoclinic Li₂MnO₃ phase with a space group C2/m. The formula Li₂MnO₃ can be alternatively expressed as Li(Li_1/3Mn_2/3)O₂, i.e., lithium ions replacing one-third of the Mn ions in the LiMO₂ phase, and the structure can be visualized to be formed by stacking alternating (Li_1/3Mn_2/3) and Li layers [52]. A similar interlayer spacing of ~ 0.47 nm among the (001) monoclinic planes and the (003) rhombohedral planes show the atomic level compatibility between the two phases in a closed-packed layered structure [61]. Therefore, all the peaks in the powder X-ray diffraction (XRD) pattern can be indexed according to the α-NaFeO₂ type rock-salt structure. However, the microstructural complexity has given rise to contradictory theories on whether to categorize it as a solid solution or nanocomposite [52,62]. But the several studies have revealed that the two phases could intergrow, structurally integrable, and coexist in the domain [52,63]. The high angle annular dark field scanning transmission electron microscopy (STEM-HAADF) image in Figure 4a and Figure 4d shows that both phases can coexist even in a single nanoparticle [59,64]. Yu et al. demonstrated the symbiosis between the two phases and the heterointerface along the [001]_rh zone axis direction through annular bright field scanning transmission electron microscopy (ABF-STEM) study [65].

Figure 5 (a) Refined powder neutron diffraction data considering composites of monoclinic Li₂MnO₃ and trigonal LiMO₂ (M = Ni, Co, and Mn) unit cells. The different color of the atoms in the TM layer represents different TM ions present at the same site. The solid blue arrow shows the cation-ordering peaks exclusively from monoclinic Li₂MnO₃ unit cell, and the dotted arrow represents the (101) reflection in the trigonal phase and the (130)/($20 \overline{1}$) reflections in the monoclinic phase. Reprinted with permission from reference [59]. Copyright 2013 American Chemical Society. Atomic model of (b) Li₂MnO₃, (c) Ni-containing Li₂MO₃, experimental Z-contrast image of the C2/m phase; (d) [100], [110], [$1 \overline{1} 0$] zone projection of the C2/m phase. The zone axes projection regions are labeled with different colored lines in the image d and labeled in squared boxes corresponding to the atomic model in panel e. The fast Li-ion diffusion channel is marked in a red arrow in panel d. Reprinted with permission from reference [66]. Copyright 2013 American Chemical Society.

4.1.2 Lithiation-delithiation Mechanism

The lithiation-delithiation mechanism of LMR-LO cathodes are complex than conventional materials due to the coexistence of the two phases. The initial charge-discharge behavior is different - from the subsequent cycles due to several simultaneous changes in the crystal structure. The transition metal redox process is accompanied by the oxygen redox process under conditions of high voltage. The process of oxygen redox can be further divided into two steps: - a) reversible oxygen redox from the bulk and b) irreversible oxygen gas evolution from the surface. The Li-excess Li₂MnO₃ phase supplies extra lithium and triggers the oxygen redox phenomena. This section explores the details of the anomalous charge-discharge mechanism of LMR-LO cathodes.

First Charge-discharge Cycle: The upper charge cut-off voltage for galvanostatic cycling is generally extended up to 4.8 V starting from 2.0 V, for the LMR-LO cathodes to include the capacity contribution at ≥4.5 V from the Li₂MnO₃ phase. The first charge and discharge proceed with two plateaus [67]. Charge up to 4.4 V range is dominated by transition metal redox of LiMO₂ phase. The Ni^+2/+4 and Co^+3/+4 redox allow lithium deintercalation, whereas Mn⁺⁴ does not contribute like in NMC cathodes [68]. The Li₂MnO₃ phase remains invariant at this stage. At around 4.5 V, the Li₂MnO₃ superlattice gets activated and results in surface oxygen evolution along with Li⁺ extraction from the transition metal layers [69]. The charge in the higher potential region is driven by oxygen redox at bulk as seen in Figure 6a. As a result, the transition metals migrate to the interstitial oxygen defects and vacated octahedral lithium sites to maintain structural integrity after deep delithiation. It initiates the formation of the spinel phase at the surface region [70,71]. Irreversibly blocked TMs at defect sites resist further lithium reintercalation during discharge and are responsible for large voltage hysteresis between the charge and discharge curves [72]. During discharge, high voltage lithium reintercalation into the TM layer is supported by TM (Ni⁺⁴, Co⁺⁴, and Mn⁺⁴) reduction and oxygen redox. In contrast, lithium is reinserted into the lithium layer at low voltage by further Mn⁺⁴ reduction (Figure 6b) [73,74]. The irreversible capacity loss that accounts for a value of approximately 50-100 mAh g^-1 is attributed to the loss of lithium in the form of Li₂O from the material surface and the formation of MnO₂, which takes part in lithiation-delithiation in subsequent cycles [75,76,77].

Figure 6 (a) First charge cycle and (b) first cycle dQ/dV profile recorded for Li and Mn-rich 0.35LiMn₂O₃·0.65Li[Mn_0.45Ni_0.35Co_0.20]O₂ electrode in Li-cell at C/20 rate, 30 °C. Reproduced from reference [67].

Subsequent Charge-discharge Cycles: Following the first cycle, the charge-discharge curves evolve differently. Overall, charging and discharging occur with oxygen release, TM migration, accumulation of spinel phase at surface, and other parasitic changes in the crystal structures that drain the capacity faster. Several degradation phenomena are intertwined to complicate the mechanistic interpretation. Advanced operando techniques such as in-situ XRD, SXAS (Synchrotron X-ray absorption spectroscopy), XANES (X-ray absorption near-edge spectroscopy), X-ray tomography, in-situ Raman microscopy, mass spectrometry, HAADF-STEM (High angle annular dark-field scanning transmission electron microscopy) equipped with EELS (Electron energy loss spectroscopy), DEMS (Differential electrochemical mass spectrometry), HAXPES (Hard X-ray photoelectron spectroscopy), and multi-length scale X-ray spectroscopic imaging, etc. are employed to elucidate the underlying processes [65,68,74,78,79]. This section is split into a few sub-sections for a clear understanding from the readers’ perspective.

(i) Anionic Redox: Abnormal high capacity of LMR-LOs cannot be provided by TM redox alone. Ni^+2/+4, Co^+3/+4, and Mn^+3/+4 contribute a maximum of 170-180 mAh g^-1 capacity, whereas the bulk oxygen redox accounts for the rest of the capacity. Several theories on oxygen redox are reported in the literature [80,81]. Whatever way it takes part in the charge-discharge mechanism, it also brings in several adverse side effects when oxygen gas is released from the surface. Initially, the highly reactive surface oxygen attacks the organic solvent in the electrolyte to produce O₂ and CO₂ gases. Surface oxygen can also form electrochemically inactive Li₂O, Li₂CO₃, etc., species that accumulate on the surface to impede the lithium diffusion process [52,64]. The resultant oxygen vacancies at the surface lead to TM migration, dislocation, and structural changes. The release of oxygen renders the surface porous, and increased porosity obstructs the electron percolation pathway in the active material [82]. Moreover, the exothermic nature of the process of gas release inside the cell supplies the activation energy towards TM migrations and triggers phase transition [83].

(ii) TM migration and Phase Transformation: Although the processes of delithiation and surface oxygen release cause TM reorganization, there is a threshold for such migration, i.e., tri-vacancy at adjacent tetrahedral sites [84]. TM migration does not create any concentration gradient during the first cycle. During cycling, Mn migrates from the surface to the interior via lithium vacancies and occupies vacant lithium sites in the TM layer. On the other hand, the surface gets densified with nickel cations, while the bulk of the material faces a nickel deficit. Thus, a concentration gradient is established, which slows down the lithium transportation rate [85]. X-ray tomography images in Figure 7 with 3D transition metal maps and percentage concentration of the TMs in the pie charts below show the relative distribution of the TMs in the cycled particles with respect to the pristine ones. In the pristine Li_1.2Mn_0.525Ni_0.175Co_0.1O₂ particles, a 76% Mn-Co-Ni contribution suggests a major phase to be layered NMC. Following the first cycle, surprisingly, the content of the pure Mn phase was found to be only 7.5%, and the content of the Mn-Co-Ni phase decreased to 44%. This suggests that Mn segregation begins after the first cycle and the Li₂MnO₃ phase gradually forms throughout subsequent cycles. After 200 cycles, the Mn-Co-Ni content drops down to 37.3%, while the pure Mn and Ni phase content was 10%. The stark difference in the elemental distribution with the progression of cycling is the direct evidence of TM migration [78].

Figure 7 Distribution of transition metals in the Li_1.2Mn_0.525Ni_0.175Co_0.1O₂ cathode particles after electrochemical cycling. Panel a, b, and c represent the selected particles at the pristine state, after 1^st cycle, and 200 cycles, respectively. Images are obtained from energy-tunable synchrotron 3D X-ray tomography. Elemental distribution data resolved from tomography above and below the K-edges of Mn, Co, and Ni. The color legends represent the relative concentration gradient of the TM elements. Scale bar in the panel a is 5 µm. The corresponding pie charts were calculated using the absorption correlation tomography. Reprinted with permission from reference [78]. Copyright 2014 American Chemical Society.

Figure 8a represents a schematic illustration of the phase evolution pathways. Transformation of the pristine rhombohedral LiMO₂ phase ($R \overline{3} m$) to the spinel (Fdm) phase is initiated at the surface during the first cycle and propagates to the bulk upon further cycling. Migration of the TMs to the Li sites and the migration of Li⁺ to the tetrahedral sites leads to forming spinel grains on the surface at first and finally consuming the entire layered structure. On the other hand, the monoclinic Li₂MO₃ (C2/m) phase also shows a propensity towards the spinel (Fdm) phase transition. The significant strain generated after Li and oxygen removal leads to lattice breakdown [66]. After few charge-discharge cycles, nanoscale spinel islands are formed and float randomly on the bulk particle. The nucleation and growth of the spinel domains eventually amorphized the material, create nanopores and thread-like cracks that widen and deepen with cycling [86].

Figure 8 (a) Schematic illustration of phase evolution in LMR-LO materials during cycling finally leading to spinel structure. Reprinted with permission from reference [66]. Copyright 2013 American Chemical Society. (b) Schematic of the degradation loop resulting from JT distortion and TM dissolution.

(iii) Modulation in Redox Contribution: Synergistic contribution from cationic and anionic redox processes helps the LMR-LO cathodes for delivering abnormally high capacity. However, the interplay between these two redox processes is not as straight forward as it might seem. Their percentage contribution to overall capacity alters upon cycling. Among all redox couples, oxygen $\mathrm{O}^{2-} / \mathrm{O}_{2}^{2-}$ and Ni^+2/+4 are the most significant contributors to capacity, whereas Mn^+3/+4 and $\mathrm{Co}^{+3 /+4}$ lie latent or contributes negligibly during the initial cycles. The loss of oxygen with cycling forces reduction of TM. This implies the formation of a surface reconstruction layer as reported in the literature. This layer having a Ni concentration gradient reduces the contribution of the Ni redox process. Moreover, a decrease in the TM valency weakens the TM-O bond covalency, and consequently, capacitive involvement of oxygen redox also decreases. However, the reduction of Mn and Co activates these two redox processes [82]. Thus, the overall discharge capacity remains almost the same as the sacrifice in Ni and oxygen redox is compensated by Mn and Co redox. Such a modulation in the redox contribution leads to gradual voltage decay [87]. Theoretical studies show that the Fermi level in the LMR-LO electronic structure initially lies above the Ni^+2/+4 redox. A gradual evolution of the Mn and Co redox processes and the decay in Ni and oxygen redox upshifts the Fermi level around Mn. Eventually, the differences in the relative Fermi level energies between the Li metal and TM decreases. Thus, voltage decay originates.

(iv) Voltage Hysteresis and Voltage Decay: These two terms are used interchangeably in many literatures, but they connote different senses. Voltage hysteresis refers to the voltage gap between the charge and discharge curve in a particular cycle, whereas voltage decay implies a gradual drop in the average voltage upon cycling. Their microscopic origin is different and mostly related to structural reorganization due to TM migration. However, these two phenomena are strongly correlated [88]. In the LiMO₂ structure, octahedral sites are connected by interstitial tetrahedral sites. Octahedral sites are more energetically favorable positions for cations in cubic close-packed LiMO₂ structures. However, tetrahedral sites may become accessible if excess energy in the form of heat or pressure is supplied externally or evolved in-situ. In stoichiometric oxides, octahedral Li and TM sites exclusively accommodate all the Li and TM cations, respectively. However, cation excess, i.e., primarily lithium excess, forces some amount of lithium to occupy octahedral TM sites. When the cell is charged beyond 4.3 or 4.4 V, a significant amount of Li is removed from the Li layer, lowering the barrier for transition metal migration. TMs migrate from the octahedral TM sites to the tetrahedral Li sites under conditions of high electrode potentials. During discharge below 3.5 V, a fraction of the TMs is forced back to the original sites. This migration is partially reversible during the initial cycles and causes voltage hysteresis during the charge and discharge processes [89]. If the TMs are irreversibly trapped in tetrahedral lithium sites in the lithium layer, they block three adjacent octahedral lithium sites due to coulombic repulsion and hinder lithium reintercalation. This results in voltage decay. On the flip side, TM migration buffers the structural collapse during excessive delithiation. Therefore, a certain extent of voltage decay is unavoidable. Voltage decay also arises from the gradual layered to spinel transition upon cycling, decreases the average voltage continuously as spinel has a redox activity around ~ 3 V vs. Li/Li⁺. Further, the modulation in redox contribution is also described in Section 5.1.2.b.iii [82].

Bettge et al. demonstrated that Li_1.2Ni_0.15Mn_0.55Co_0.1O₂ under standard cycling conditions (2.0-4.7 V vs. Li/Li⁺, 30° C) shows a drop in resistance corrected average voltage by ~ 9 mV per cycle during charge and ~ 6 mV per cycle during discharge, which accounts for 3-5% overall drop approximately [90]. Similar voltage fades are also common in commercial LIB cathodes (NMC, NCA, etc.), but the drop per cycle realized is lower than that realized in Li-rich chemistries. Voltage decay depends on several factors: - a) It is composition-dependent for particular chemistry. Under similar testing conditions, Li_1.2Ni_0.4Mn_0.4O₂ and Li_1.2Ni_0.2Mn_0.6O₂ show 3.8 and 7.2 mV of voltage fade per cycle, b) voltage decay is faster at higher upper cut-off voltages due to induced side reactions, c) at elevated temperatures, the gap between the average charge and discharge voltage becomes more significant, revealing the fact that side reactions during charge override kinetics factor, and d) decay is faster during the initial cycles, but continues at a measurable rate up to several hundreds of cycle until the layered phase is completely transformed to a stable spinel structure [90,91].

Voltage hysteresis reveals asymmetry between the charge and discharge profile and is caused by the metastable path of ion migration during charging-discharging. From a thermodynamics point of view, electrical work consumed during voltage hysteresis is dissipated as heat. Therefore, considerable voltage hysteresis is translated into lower energy efficiency in the case of full cells [68,92]. Voltage decay during long-term cycling is proved to be one of the significant contributors to the energy fade mechanism of LMR-LO cathodes besides capacity loss and resistance build-up.

(v) TM Dissolution: TMs change oxidation states with lithiation-delithiation. The physio-chemical properties of a metal center differ from one oxidation state to another. Thus, with the gradual progress of charge-discharge, several asymmetries arise in the system originating from TM cations. Metal ions are symmetrically surrounded by six oxygen atoms to form the MO₆ octahedra in a pristine layered oxide structure. According to crystal field theory, the electrostatic field of an octahedral ligand arrangement splits five degenerate d-orbitals into two sets: the triply degenerate low-energy t_2g set (d_xy, d_yz, and d_xz) and the doubly degenerate high-energy e_g set (d_x²_-y² and d_z²). Metal ions in cubical electron distribution, i.e., completely filled, half-filled, or empty orbitals ($\mathrm{t}_\mathrm{2g}^\mathrm{n}$, where n = 3, 6 and $\mathrm{e}_\mathrm{g}^\mathrm{m}$, where m = 0, 2, 4) prefer such an arrangement. But if the metal ion is in a non-cubical electronic distribution, i.e., unsymmetrically filled orbitals ($\mathrm{t}_\mathrm{2g}^\mathrm{n}$, where n = 1, 2, 4, 5 and $\mathrm{e}_\mathrm{g}^\mathrm{m}$, where m = 1, 3) experience Jahn-Teller (JT) distortion. According to the JT theorem, unsymmetrically filled orbitals further split to remove the degeneracy and yield an additional stabilization to the system. JT splitting distorts the bond lengths of the axial and equatorial M-O bonds so that the overall energy of the system reaches the minimum value [93]. Therefore, the surface accumulation of JT active high-spin Mn⁺³ in the LMR-LO cathodes due to charge-discharge affects the structural stability. Local JT distortion and lattice instability increase the crystal's reactivity towards electrolytes and causes consequent TM dissolution.

The JT effect relaxes the degeneracy in the e_g orbitals by placing an electron in the d_z². The axial M-O bond elongates, weakening the overlap between the TM and oxygen units. It induces more ionic character of the axial M-O bonds than equatorial and densifies negative charge on the axial oxygens. The proton of strong acid HF originating from the hydrolysis of the LiPF₆ salt readily attacks the axial oxygens at the cathode electrolyte interphase (CEI). Further, the protonation of the axial oxygen is accompanied by the metal (TM d_z²) to ligand (O2p) electron transfer to result in the formation of H₂O. Electron transfer oxidizes the metal center, i.e., Mn⁺³ to Mn⁺⁴ in LMR-LOs. Mn in the highest oxidation state (+4) is highly oxidizing in nature. It decomposes electrolyte solvent molecules and converts itself to a JT-free Mn⁺² state via a two-electron pathway. This stable low valent Mn⁺² gets dissolved by electrolyte, floats towards the anode via the electrolyte, and accumulates on the anode surface to thicken the SEI [94]. Therefore, the cyclic loop of the degradation pathway, once get activated, damages all the interfaces, increases cell impedance, and deteriorates the electrochemical performance.

Way outs from the degradation loop, as shown in Figure 8b include either surface coating at the expense of increased cell impedance or tuning uneven electron distribution of JT active centers by doping metal ions. Substituting JT active metal ions with inactive Al⁺³ or Mg⁺² perturbs the long-range metal-metal order. The newly introduced localization effect reduces oxygen and electron mobility over the M-O bonds, thereby suppressing TM dissolution. Other cell-level indirect measures can also be taken care of to avoid such reactions. The electrolyte salt LiPF₆ can be replaced with imide anions such as bis(trifluoromethane)sulfonimide (TFSI), bis(fluorosulfonyl)imide (FSI), bis(pentafluoroethylsulfonyl)imide (BETI), etc. More covalent C-F bonds in the linear or cyclic imide ions restricts HF generation than polar P-F bonds in LiPF₆. Moreover, limiting the upper charge cut-off can decrease the reactivity of oxygen at the cathode electrolyte interphase. Electrolyte additives are also used to scavenge acidic byproducts.

The extent of Mn-dissolution depends on the crystal structure of the Mn host. Krishna Kumar et al. showed that, when pristine materials are aged inside a glove box under pure argon atmosphere at stirring condition (50 °C for two weeks) dissolved in 1M LiPF₆ in 1:2 EC/DMC, Mn dissolution decreases from ~ 1.1 wt.% in Li_1.2Mn_0.55Ni_0.15Co_0.1O₂ to ~ 0.5 wt.% when blended with carbon-coated LiMnPO₄. The presence of the robust polyanionic phosphate framework hinders the reaction between the acidic proton and the basic oxygen units.

4.1.3 Modification

All the problems mentioned above, along with electrolyte degradation, electrolyte instability at high voltages, JT distortion, Mn dissolution, etc., are also severe shortcomings that result in various problems such as initial irreversible capacity, voltage hysteresis, and voltage decay, poor cycle life, inferior high-rate performance. Modifications can be applied to surface, bulk, or both at a time.

Composition Optimization: The chemical composition of xLi₂MnO₃·(1-x)LiMO₂ varies with the value of x, and so the electrochemical performance. Optimizing the value x is a trade-off between higher capacity and longer cycle life. When the compositional percentage of Li₂MnO₃ dominates over LiMO₂, the specific capacity (≥ 4.5 V) increases along with oxygen evolution, but cycle life and coulombic efficiency deteriorates. Increasing the concentration of LiMO₂ in the composition enables to draw more capacity below 4.5 V, especially for Ni-rich LiMO₂. Manthiram and coworkers compared the charge-discharge performances of (1-z)Li[Li_1/3Mn_2/3]O₂·(z)Li[Mn_0.5-yNi_0.5-yCo_2y]O₂, where y = 1/3 and 0.25≤z≤0.75 when cycled at 12.5 mA g^-1 (~C/20 rate) in the current density range of 2.0-4.8 V (Figure 9a). Except for z = 0.25, the first discharge capacity decreases with an increase in z or a decrease in the lithium content but the amount of oxygen loss increases. The highest first discharge capacity is obtained for z = 0.4, i.e., 254 mAh g^-1 with an irreversible loss of 80 mAh g^-1 [95]. Moreover, the Li₂MnO₃ component can be generalized as Li₂MO₃, where M is the form of Ru, Sn, Mo, Ir, Ti, etc. are also active and show more stable oxygen redox as non-transition and heavy transition metals (4d and 5d) form less directional and more flexible TM-O bonds [96,97,98]. However, heavy atoms lower the specific capacity value and being non-abundant increases the material production cost. This design concept of the solid solution has been used to synthesize different types of combinations. Ramesha et al. showed that the solid solution xLiCoO₂·(1-x)Li₂RuO₃ delivers an initial discharge capacity of ~ 232 mAh g^-1 and retained 85% capacity at 0.2 C rate after 100^th cycles when x = 0.3 [99]. Wang et al. demonstrated that the ternary mixture of Li₂RuO₃-Li₂SnO₃-Li₂MnO₃ dramatically increases the discharge capacity from 109 to 223 mAh g^-1 once the system goes from the composite phase to the solid-solution phase [100].

Figure 9 (a) Comparison of the first charge-discharge profiles (recorded at 12.5 mA g^-1 between 2.0 and 4.8 V) recorded for the (1-z)Li[Li_1/3Mn_2/3]O₂·(z)Li[Mn_0.5-yNi_0.5-yCo_2y]O₂ (y = 1/3 and 0.25≤z≤0.75) samples. The dashed vertical lines separate the initial sloping region A from the plateau region B. Reprinted with permission from reference [95]. Copyright 2007 American Chemical Society, (b) Cycling of 5 wt.% AlPO₄ coated Li[Li_0.2Fe_0.1Ni_0.15Mn_0.55]O₂ sample at different current densities, (c) Charge-discharge profile recorded for pristine, 3 wt.% (AC-3), 5 wt.% (AC-5), 7 wt.% (AC-7) AlPO₄ coated Li[Li_0.2Fe_0.1Ni_0.15Mn_0.55]O₂ at 0.2C rate. Reprinted with permission from reference [101]. Copyright 2015 American Chemical Society, and (d) C/10 rate cycling followed by C/30 and C/20 rate cycling of pristine Li_1.2Ni_0.15Mn_0.55Co_0.1O₂ and Mg⁺²-F^- codoped samples. Reprinted by permission from Springer (reference [102]): Ionics 2017 (DOI 10.1007/s11581-017-2018-9)

Surface Coating: Surface coating can suppress oxygen release, electrolyte corrosion, and phase transition. Coating materials require high electrochemical stability, high voltage stability, inert towards electrolyte and other cell components, surface wettability, etc. Surface coating reaches its maximum efficiency when the lattice constant of the coating material matches the host crystal; the coating is smooth, well-dispersed, covering the entire surface with optimal thickness. Mismatch of lattice constants exerts asymmetry effect to obstruct Li conduction through it and increases contact resistance. The coating material should be chosen in such a way so that it does not lithiate itself in excess as then the volume change may introduce a peel-off effect. Surface coating can minimize surface defects, and its healing function confined only to the surface. The inaccessibility of the crystal interior fails to solve bulk defects. Type of coating materials have been reported in the literature for LMR-LO cathodes includes electrochemically non-active substances such as oxides (Al₂O₃, Ti₂O₃, Er₂O₃, Pr₆O₁₁, etc.), phosphates (AlPO₄, CoPO₄, etc.), fluorides (AlF₃, LiF), and solid-state fast ion conductors (LiPON, LiV₃O₈, Li_xLa_yTiO₃, etc.) [19,103,104,105,106,107,108,109]. Wu et al. demonstrated that pristine Li[Li_0.2Mn_0.54Ni_0.13Co_0.13]O₂, its modified versions such as 3 wt.% Al₂O₃, CeO₂, ZrO₂, ZnO, AlPO₄ treated, and 0.05 F^- atom per formula unit provides first discharge capacity of 253, 285, 253, 252, 252, 261, and 270 mAh g^-1, respectively and the first irreversible capacity of 75, 41, 36, 48, 58, 22, and 58 mAh g^-1, respectively. The retention of oxide ion vacancies in the surface treated samples reduces the first cycle irreversible capacity loss and significantly modifies the first discharge capacity [110]. Optimization of coating thickness is very critical for electrochemical performance as thicker coating leads to enhanced interfacial impedance. Therefore, the weight percentage of the coating agent is a contributing factor for wet-chemical methods of coating. A report suggests that 5 wt.% AlPO₄ coatings (Figure 9b and Figure 9c) on Li[Li_0.2Fe_0.1Ni_0.15Mn_0.55]O₂ to be the optimum, showing the reversible capacity of 220.4 mAh g⁻¹ after 50 cycles at 0.1C, accounting for 74.4% retention and enhanced rate capability (175.3 mAh g⁻¹ at 1C, and 120.2 mAh g⁻¹ at 10C after 100 cycles) [101]. Zhang et al. coated Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ using different weight percentages of the Li-La-Ti-O (LLTO composite) solid-state fast ion conductor. The study shows that 5 wt.% is the optimum coating that gives the best electrochemical performance, i.e., decay of only ~ 35 mAh g^-1 capacity compared to 111 mAh g^-1 capacity loss of untreated sample over 100 cycles at 30 mA g^-1 current density between 2.0-4.8 V range. However, all coated samples show better C-rate performance and higher capacity retention than a pristine one. Highly conductive (10^-3 S cm^-1) amorphous LLTO surface layer lowers the charge transfer resistance, and the fast ionic conduction process through the thin layer results in better C-rate capability. The differential scanning calorimetry (DSC) study implies that better thermal stability provided by the coating layer helps to maintain the perfect crystalline structure even after 100 cycles [111].

Ion Doping: Ions can be doped into the crystal lattice. The electronic structure is tuned in this way and may solve the fundamental voltage decay issues. Alkali metal cations (Cs⁺, Na⁺, Al⁺³, Mg⁺², etc.) normally occupy tetrahedral sites in the Li layer. The doping with larger ion than lithium increases the barrier for TM migration and exerts ‘pillar effect’ to maintain structural stability [112,113,114,115]. Divalent and trivalent cations help increase the average positive valence to strengthen the TM-O bond. It delays the activation of the Li₂MnO₃ component and stabilizes oxygen redox. However, incorporating alkali metal ions cannot drive away from the layered to spinel transformation completely as the spinel phase is more stable than the layered in the delithiated state. Therefore, fluorine doping along with alkali metal ions is demonstrated to be an effective strategy. Na and F-doped in Li_1.2Mn_0.54Ni_0.13Co_0.13O₂ to form Li_1.15Na_0.05Mn_0.54Ni_0.13Co_0.13F_0.01O_1.99 shows 97% capacity retention and 91% voltage retention over 100 cycles with an initial discharge capacity of 260 mAh g^-1 at C/10 current rate. The partial substitution of Li and O with Na and F, respectively, provides structural stability and widens the Li slab to enable rapid Li-ion diffusion [116]. A Stronger M-F bond suppresses oxygen evolution. Similar effects were also exerted when Mg and F were co-doped, where 10-15% excess capacity and 14-15% improvement in capacity retention was observed over 50 cycles (Figure 9d) [102]. Transition metal ions (Cr⁺³, Ti⁺⁴, Nb⁺⁵, etc.) occupy the space in transition metal layers [117,118]. Large TM ions expand the lattice parameters and improve lithium diffusion. Nb⁺⁵ doped Li_1.2Mn_0.54Ni_0.13Co_0.13O₂ demonstrates 7% improvement in capacity retention over 100 cycles than pristine sample due to increased ordering in cation distribution and delivered ~ 20 mAh g^-1 higher capacity than the pristine one at 5C rate due to lesser amplification in charge transfer resistance upon cycling [119]. Similarly, anion doping like fluorine can also stabilize cation-anion interaction, reduces oxygen loss, also increases average voltage as F^- is more electronegative than O^2- [120]. Polyanionic groups have robust moieties that are able to hold TMs at even extreme voltages. Strong interactions between MO_x and TMs can be explored by polyanionic group substitution. These groups (XO₄)^n- or (X_mO_3m+1)^n-1 (X = B, P, As, Mo, Si, etc.) also reduce the number of oxygen ions in the structure. The phosphate polyanion doped material Li_1.17Ni_0.2Mn_0.58Co_0.05O_2-x(PO₄)_x, where x = 0.01, 0.03, and 0.05 minimizes the local structure change during cycling and also minimizes the growth in diffusion impedance during cycling. The sample with x = 0.01 provides a very stable energy density and retains ~ 70% of the initial value after 300 cycles at a C/10 rate [121]. Table 2 summarizes the modification results for the LMR-LO cathode materials.

Table 2 Summary of surface and structural modifications carried on LMR-LO cathode materials on the electrochemical performance against lithium metal anode in coin-type half cells.

4.2 Lithium Excess Cation Disordered Rock Salt Oxides

Commercialized cathode materials possess high theoretical capacities i.e., LiCoO₂ (274 mAh g^-1), LiNi_xMn_yCo_zO₂, where x + y + z = 1 (270-280 mAh g^-1), and LiNi_0.8Co_0.15Al_0.05O₂ (278 mAh g^-1), but the practical useful capacity is limited to ≤ 180 mAh g^-1. In all these materials, the Li: TM ratio is 1:1. Lithium excess cation disordered rock salt oxides have a Li: TM ratio of >1 can deliver >250 mAh g^-1 capacity. The crystallographic orientation, Li-hopping mechanism, and redox behavior of these samples differ from those of the conventional ordered layered oxides. A deep understanding of these phenomena is essential to elucidate higher obtained capacity from lithium excess cathodes.

4.2.1 Crystal Structure and Composition

Ordered rock salts are one of the most geologically abundant crystal structures found in nature. They crystallize in $F m \overline{3} m$ space group and generally referred to as NaCl-like structure (Figure 10a). Common inorganic salts, such as NaCl, KBr, NiO, FeO, LiF, AgCl, etc., fall under this category. Anions occupy the 4b site forming a face-centered cubic lattice in the crystal lattice, where cations fill half of the octahedral vacancies (4a site). The densely packed unit cell of the pure rock salt phase restricts the entry of the guest cations (Li⁺, Na⁺, K⁺, etc.) into its lattice. Therefore, any rock salt phases are strictly avoided during battery cycling due to their electrochemical non-responsive nature towards lithiation-delithiation. The presence of NiO during the synthesis of the Ni-rich cathodes and its formation and surface accumulation during cycling are considered detrimental to battery performance. However, in 2014, Ceder group first showed that supplementing an extra amount of lithium beyond stoichiometry. This results in cation disordering between Li and transition metals, opening up the pathways for long-range Li-ion diffusion throughout the crystal structure (Figures 10c and d). Li and transition metal ions were randomly distributed at the 4a site in the cationic sublattice (Figure 10b) [127,128].

Figure 10 Crystal structure of (a) classic rock salt, (b) cation disordered rock salt. Reproduced with permission from reference [128]. Copyright 2021 American Chemical Society, (c) Galvanostatic cycling of LiNi_0.67Sb_0.33O₂ (black curves) and Li_1.15Ni_0.47Sb_0.38O₂ (red curves) at 1C rate between 2.5-4.6 V range, (d) Schematic illustration demonstrating the coexistence of the two types of ordering and 0-TM diffusion channels at the domain interfaces. Reproduced with permission from reference [129]. Copyright 2014 American Chemical Society.

Unlike conventional layered oxide cathodes, cation disordered rock salts have a wide compositional range. It is mainly synthesized in the oxide or oxyfluoride form. The general chemical formula is expressed as Li_1+x(MM)_1-xO₂ or Li_1+x(MM)_1-xO_2-yF_y (x, y>0). M is a redox reservoir and M is a redox inactive charge neutralizer. Both M and M can be single or multiple transition metals. Typical redox-active M centers include Mn^+2/+4, Mn^+3/+4, Ni^+2/+4, V^+3/+5, V^+4/+5, Fe^+2/+3, Cr^+3/+6, Mo^+3/+6, etc., whereas it is only confined to Ni and Mn centers in the case of conventional layered oxides. The Ni^+2/+4 center provides high discharge voltage, but its overlap with oxygen redox brings instability issues due to irreversible oxygen redox. High valence V, Cr, and Mo center suffer from low average voltage. Fe-based systems undergo significant hysteresis during cycling. Hence, earth-abundant Mn-based systems are most explored, especially Mn^+2/+4 delivers high capacity. To stabilize the Mn center at the +2 state in the as-synthesized material, high valent charge neutralizers, such as Mo⁺⁶, V⁺⁵, Sn⁺⁴, Ti⁺⁴, etc., are effective [130].

4.2.2 Lithium Percolation Theory

The capacity of the cathode material is directly proportional to the amount of lithium that can undergo reversible intercalation-deintercalation. To support facile lithium-ion diffusion, the material should contain properly distributed low barrier channels throughout its structure. The low barrier channels allow microscopic lithium diffusion, whereas a well-distributed network of such channels leads to macroscopic percolation. In rock salt type oxides, lithium migration occurs via the octahedra-tetrahedra-octahedra (O-T-O) route, as the direct movement of lithium between the two edge-shared octahedra is associated with a high energy barrier. Lithium in the intermediate tetrahedral site represents the activated state of lithium migration. The lower the energy of the activated state, the more facile is the migration. The energy of the activated state depends on its interaction with the adjacent atoms. Each tetrahedron has four faces, and all faces are shared by octahedrons. The octahedral sites can accommodate lithium (or transition metals) or remain vacant. Based on the number of transition metals in the surrounding environment, the activated tetrahedral sites can be theoretically classified into the following categories: 4-TM, 3-TM, 2-TM, 1-TM, and 0-TM sites. Figure 11a demonstrates 0, 1, and 2-TM channels [131]. The energy of the activated lithium site is determined by two factors: - a) electrostatic interaction between the activated lithium and transition metal, i.e., the valency of the transition metals, and b) tetrahedral height or slab distance, i.e., distance between lithium and the transition metals. Among the four face sharing sites, two sites must be occupied by lithium, e.g., one site from which lithium is coming (a-site) and another site from which lithium will migrate (b-site). Therefore, the 4-TM and 3-TM sites are inactive or dead or blocked channels through which lithium cannot diffuse. Considering that the two sites are occupied by lithium, the other two face sharing sites (also called ‘gate sites’) can be accommodated by the two TMs (2-TM site), one TM, one Li (1-TM site), or two lithium (0-TM site). Electrostatic repulsion exerted by the two high-valent (generally valency of +3) TM cations on activated Li⁺ is too high for facile migration. Therefore, the 2-TM sites having significant energy barriers are usually inactive. Stoichiometric γ-LiFeO₂ type structures (e.g., LiTiO₂) contain 2-TM channels and are poor lithium-ion conductors. Stoichiometric layered α-NaFeO₂ type oxides (examples - NMC, LCO, NCA, etc.) conduct lithium-ion via the 1-TM channel, where one of the gate sites is occupied by lithium and thus, the energy barrier for lithium migration is relatively lower. The energy barrier for 1-TM diffusion is sensitive towards the state of charge. During charge, delithiation occurs along with oxidation of TMs. The higher valence of TMs in the charged state increases electrostatic repulsion. Furthermore, the migration of TMs into the lithium layer (antisite disorder) during charge and collapse of slabs at a higher depth of delithiation hinders lithium migration. The 0-TM channels have the minimum energy barrier. Lithium migration through it is irrespective of the size of the diffusion channel and valence of transition metals. Low-temperature (LT) spinel-like LiCoO₂ structure possesses percolating networks of 0-TM channels, and macroscopic lithium percolation occurs through its bulk. On the other hand, cation disordered rock salt structures (α-LiFeO₂) contain 0-TM, 1-TM, and 2-TM channels. As described earlier, the 2-TM channels are inactive [132]. Due to the disordered arrangements of the cations, the average height of the tetrahedron sites is lower than that of the ordered rock salt structure. It results in a higher energy barrier for lithium migration via a 1-TM channel. Therefore, only 0-TM channels sustain lithium percolation in the case of disordered rock salt structures. 0-TM channels, i.e., activated tetrahedral site surrounded by four lithium atoms, demand a lithium-rich environment. Their concentration increases with increasing lithium content in the composition. There is a threshold value of lithium content for different structures beyond which the 0-TM channels form a percolating network to enable facile macroscopic diffusion. Monte-Carlo simulations suggest that approximately 10% excess of lithium is required for layered and disordered rock salt structure, >30% for γ-LiFeO₂ type structure, while ~ 23% deficiency is enough for low-temperature LiCoO₂. Therefore, spinel-like lithium oxide is the ideal situation for 0-TM percolation. It is not only capable of acting as a cathode but also as a lithium insertion anode in partially delithiated form.

Figure 11 (a) Possible environments within Li-rich cation disordered layered oxide cathodes. O-T-O diffusion route and activated states of Li migration. Reproduced with permission from reference [131]. Copyright 2014 AAAS, and (b) Schematic band structure of stoichiometric lithium layered oxide and lithium-rich layered oxides.

4.2.3 Contribution of Anionic Redox

Li⁺ and charge-compensating electrons are simultaneously generated from layered oxide cathodes during charge. Both move towards the anode, the former via the inner circuit, while the latter via the outer circuit and return during discharge to complete a cycle. In stoichiometric layered oxides, the amount of extractable lithium is balanced with available transition metal redox, which supplies the electrons. Transition metal redox is the only source of capacity here as each lithium-ion extraction must be balanced by one electron ejection to maintain charge neutrality. In lithium excess oxides, the extra capacity is attributed to the fact that oxide ions can also take part in a reversible redox reaction. The assistance of oxygen redox pushes the capacity beyond the transition metal redox. This section discusses the aspects of oxygen redox in lithium excess materials.

If we consider the case of an oxygen atom, it is surrounded by three metal atoms (M) and three lithium (3M-O-3Li) in the stoichiometric layered LiMO₂ [133]. Oxygen is in a symmetric Li-O-M local environment. However, in lithium excess oxides, the presence of extra lithium can distort the local oxygen environment, where oxygen is coordinated to four lithium atoms and two metal atoms (4Li-O-2M) and gives rise to liner Li-O-Li configuration. Advanced density function theory (DFT) calculation demonstrates increased charge density around oxygen in such configuration. It indicates labile electrons (also denoted as oxygen holes), the extraction of which is the origin of oxygen redox. Reversibility of oxygen redox ($\mathrm{O}^{2-} / \mathrm{O}_{2}^{\mathrm{n}-}$) is critical as it competes with irreversible oxygen (O₂) gas evolution. Formation of peroxo ($\mathrm{O}_{2}^{\mathrm{n}-}$, n = 2) species require O-O σ- interaction [134]. It is only possible when two neighboring oxidized oxygen in the Li-O-Li configuration rotates and hybridizes without sacrificing much M-O bonding. Rotation and subsequent interaction of the metastable oxygen holes enable the formation of a peroxide network avoiding the oxygen gas evolution pathway. Therefore, the rotation of the Li-O-Li configuration is the key step in the mechanism, which depends on adjacent Li-O-M configuration or, more precisely, on the nature of M-O bonding. The material cannot afford the loss in M-O bonding to a large extent as it may affect the structural integrity and probable collapse of the (MO₂)_n slabs. Therefore, the flexibility of the M-O bond governs the energy barrier of Li-O-Li rotation. Transition metals with partially filled d-orbitals form directional M-O bonds, which poses a difficulty in rotation, whereas metals in their d⁰ and d¹⁰ configuration form much weaker bonds and facilitates the rotation. Moreover, in the transition metal series, 4d and 5d metals form more flexible bonds than the 3d metals and lower the energy barrier for rotation [135]. Another important factor worth mentioning is the competition between transition metals and anionic redox. Maximum capacity can be obtained when this two redox does not overlap. O2p in Li-O-Li configuration cannot hybridize with Li2s due to the large gap in energy. This unhybridized O2p state has higher energy than the hybridized O2p bonding states. That is why oxygen oxidation occurs preferentially from the Li-O-Li configuration. On the other hand, the unhybridized O2p state lies in a lower position than the metal antibonding state in the LiMO₂ molecular orbital diagram (Figure 11b) [81]. Overlap between these two indicates initiation of oxygen redox at partial completion of TM redox leading to lower extractable capacity. Early transition metals (V, Cr, Mn, etc.), having high energy 3d orbitals raise the energy of the metal antibonding states preventing the overlap with unhybridized O2p states. Lithium excess oxides containing late 3d transition metals (Ni, Fe, Co, etc.) do not allow the full utilization of TM redox in some cases.

Although anionic redox is a significant step in the capacity generation mechanism of lithium excess materials, Li-O-Li configuration also occurs in stoichiometric layered oxides following cationic disorder at a high depth of delithiation ≥ 4.5 V, especially observed in the case of LiNiO₂. However, oxygen oxidation is substantial at ~ 4.3 V for lithium excess materials [136].

4.2.4 High Capacity vs. Better Reversibility

Lithium excess cathode materials are able to show ≥250 mAh g^-1 capacity, but stability is confined to only a hundred galvanostatic charge-discharge cycles for most of the materials. They particularly suffered from capacity fade and increased polarization under conditions of charge-discharge upon cycling. The poor electrochemical reversibility can be ascribed to a high degree of oxygen redox utilization leading to oxygen gas evolution, metal dissolution, and consequent structural deterioration. Oxygen gas release leaves behind cations on the surface, and the cation-densified surface blocks the passage of lithium migration into the bulk, increasing the impedance of the cell [137]. The evolved gas can trigger side reactions with the electrolyte. Protective surface coating and doping improve the cycling performance [138,139]. One of the most followed modification approaches is fluorine doping that provides multifaceted benefits. Substitution of O^2- with F^- lowers the average anion valence so that the pristine material can accommodate a more significant number of metals at their lower oxidation states (Mn and Ni in their +2 states) [140]. Incorporating low valent metals widens the capacity obtained from transition metal redox. The decreased dependency on oxygen redox to draw more capacity reduces the probability of irreversible oxygen deposition. F-substitution also hinders asymmetric structural distortion during cycling for Jahn-Teller active redox sites [141].

4.2.5 Enhanced C-rate Performance

C-rate performance is the measure of how fast lithium can migrate during galvanostatic cycles. Although the movement of Li⁺ is fast enough through liquid electrolytes, it is impeded by a solid-state diffusion barrier. If delithiation is accompanied by a shift of other atoms in the host, the lithium transport pathway becomes crowded, resulting in kinetic complexities. Polyanionic frameworks (LiFePO₄, LiMn₂O₄) are the ideal examples where atomic rearrangements during delithiation are minimal and show superior rate capability. In contrast, the high current rate performance of LiCrO₂ is severely hampered by Cr migration [142]. Therefore, the traditional view supports the fact that these non-topotactic movements are detrimental to the fast lithium-ion migration. But a recent study suggests an opposite trend for Li-rich cation disordered rock salt oxides. Cr-doped Li_1.2Mn_0.2Ti_0.4Cr_0.2O₂ shows a 40% capacity reduction when current density increases from 20 mA g^-1 to 1 A g^-1 with respect to a 57% decrease in Li_1.2Mn_0.4Ti_0.4O₂. Theoretical studies identified the movement of Cr (from octahedral to the tetrahedral site) as the reason for high capacity retention. Non-topotactic migration of Cr generates additional 0-TM sites, and the resulting expansion in the lithium percolating network improves the kinetics. One necessary condition for such octahedral to tetrahedral site migration is that the metal ion must prefer tetrahedral coordination under conditions of a high oxidation state. According to the theory of coordination complexes, metal ions in d⁰, d¹⁰, and d⁵ (V⁺⁵, Mo⁺⁶, Fe⁺³, Mn⁺², etc.) configuration gather no gain in octahedral site stabilization energy. They may undergo similar movement towards adjacent tetrahedral sites if their size matches the tetrahedron height [143].

Figure 12 depicts the contribution from each atom in lithium excess cation-disordered rock salt oxide materials. The choice of elements is critical, and many inter-related factors must be taken into consideration to reach the optimized composition. Table 3 is a collection of some literature-reported compounds and their electrochemical performances.

Figure 12 Summary of design principles for lithium excess cation-disordered rock salt oxide materials.

Table 3 Electrochemical performances of selected lithium-rich cathode materials.

4.3 Conversion Cathodes

Conversion type cathodes involve chemical reactions with Li-ion through bond breaking and bond formation. Thus, more energy can be stored and transformed using these types of reactions. The Li-ion storage capacity in these cathodes is not site-limited. Therefore, more capacity can be achieved and hence, higher energy can be recorded. Conversion electrode reactions involve two to six electron transfer processes. The conversion reactions result in considerable structural changes of the host matrix, unlike insertion-type ones. The disadvantage with these types of cathodes is the inherent low conductivity leading to low active mass utilization, low reversibility of the conversion phases, and high hysteresis. These also involve high volume expansion during cycling. As a consequence of these shortcomings, the reversibility of the redox reaction decreases rapidly under conditions of repeated cycling, and hence, the energy conversion efficiency also becomes inferior. The redox reactions occurring in the conversion cathodes during the lithiation-delithiation process is given below [10,153,154]

Type 1: MX_Z + yLi ↔ M + zLi_y/zX, M = transition metal (Fe, Ni, Mn, Cu); X = F, Cl, Br, I

Type 2: yLi + X $\leftrightarrow$ LiyX, X = O, S, Se, Te

4.3.1 Type 1-Category Conversion Cathodes

This category of conversion cathodes includes MX_z type of compounds, where M represents transition metals such as Fe, Ni, Mn, Cu, and X represent F, Cl, Br, I. For instance, Iron trifluoride (FeF₃), which has a high theoretical capacity of 712 mAh g^-1 (involving 3 electrons) and a working voltage of 2.7 V, has been studied as a LIB cathode [155]. FeF₃ reacts with three Li-ions during lithiation, i.e., F^- ions diffuse to form LiF and Fe. The structure appears to be Fe metals scattered in a pool of LiF [17]. The reaction for the above process is denoted as below.

\[ \mathrm{FeF}_{3}+3 \mathrm{Li}^{+}+3 \mathrm{e}^{-} \leftrightarrow 3 \mathrm{LiF}+\mathrm{Fe} \]

Similarly, metal fluorides, including nickel (II) fluoride (NiF₂), cobalt (II) fluoride (CoF₂), FeF_2, etc., have also been explored as LIB cathodes. Some common Type 1-category conversion cathodes are summarized in Table 4 [10,153].

Table 4 Properties of some typical Type-1 conversion-type cathodes [17].

The feasibility of reversible electrode reaction in these electrodes is questionable due to the added limitations like poor ion transport kinetics, lesser structural stability, and metal cluster formations at a fully delithiated state, which catalyzes the electrolyte decomposition. Electrode engineering, nano-structuring, conductive coating, and composite formation using conductive additives like carbon and conducting polymers can significantly improve their electrochemical performance [10,153]. A 3-dimensional (3D) carbon fiber electrode architecture with nano-size active iron fluorides and multilayer graphene (MLG) as connectors can deliver capacities of 595 mAh g^-1 and 445 mAh g^-1 in the operating voltage range of 4.5-1.0 V and 4.5-1.5 V, respectively. Besides, the hysteresis voltage is reduced from 2 V for pristine FeF₃ to about 0.9 V for the 3D carbon fiber architecture electrodes, improving the reversibility. A capacity of around 600 mAh g^-1 was maintained till 25 cycles even when the electrodes were cycled at an elevated temperature of 60 °C due to the enabling of reversible 3 electron kinetics [156]. But the cycle life and C rate performances reported were still lower compared to conventional insertion cathodes (Figure 13).

Figure 13 Slurry coated on Al-foil by conventional method (a) Voltage profiles of Li/FeF₃ cell at 25 °C and (b) Capacity plotted as a function of cycle number for the Li/FeF₃ cell. FeF₃ on carbon fiber 3D electrode. (c) Voltage profile (C/50 rate) between 1.5-4.5 V at 25 °C and (d) Capacity plotted as a function of the cycle number. The inset of figure d represents a scanning electron microscopy image of FeF₃/carbon fiber electrode. Adapted from reference [156] with permission from the Royal Society of Chemistry.

4.3.2 Type 2-Category Conversion Cathodes

This type of conversion electrodes includes X = O, S, Se, Te, etc. The most extensively studied conversion cathodes of this category include the oxygen cathode in the lithium-air battery and the sulfur cathode in the lithium-sulfur battery.

Lithium-sulfur battery utilizes sulfur as a cathode and lithium metal as the anode material. The common electrolyte used is 1M LiTFSI in a 1:1 volume ratio of DOL: DME with 1% w/w LiNO₃ as an additive for lithium surface passivation [157]. Sulfur has a high theoretical capacity of ~1600 mAh g^-1 and high abundance in the earth’s crust, but it has poor electronic conductivity (10^-30 S cm^-1). High volume expansion (~80%), dissolution of intermediate polysulfides in the electrolyte with effective loss of active S, Li₂S deposition, and dendrite formation on Li anode during cycling results in capacity fading and cycling instability, which limit the practical application [158,159,160]. The electrochemical cell reaction involved during the charge-discharge process is as follows [10]:

Cathode reaction: S + 2Li₊ + 2e₋ ↔ Li₂S

Anode reaction: 2Li ↔ 2Li₊ + 2e₋

Overall reaction: 2Li + S ↔ Li₂S E⁰ = 2V

During charge-discharge, irreversible by-products such as methanol, ethanol, thiocarbonates, and ethylene glycol are formed as a result of intermediate polysulfide formation. The formation of polysulfides through the intermediates can be represented as follows:

\[ \mathrm{Li}_{2}\mathrm{S} \rightarrow \mathrm{Li}_{2}\mathrm{S}_{2} \rightarrow \mathrm{Li}_{2}\mathrm{S}_{4} \rightarrow \mathrm{Li}_{2}\mathrm{S}_{6} \rightarrow \mathrm{Li}_{2}\mathrm{S}_{8} \rightarrow \mathrm{S}_{8} \]

Due to polysulfide formation and resulting low cycle life, instead of taking bare S, the Li₂S cathodes are also reported. Li₂S nanoparticles embedded with conductive carbon show little structural change after 400 cycles. Self-standing 3D printed sulfur/carbon composite cathodes with high sulfur loading were found to deliver a stable capacity of 564 mAh g⁻¹ within 200 cycles at a 3 C current rate by facilitating Li⁺/e^- transport [161,162,163,164].

Se and Ti-based conversion cathodes have high electronic conductivity compared to S and show higher rate capability. But they also suffer from issues similar to S-based conversion reactions.

The Li-air battery is one of the most promising Li-based batteries with regard to the ultrahigh specific energy density of 11,650 Wh kg^-1 that is much higher than conventional LIBs (150-200 Wh kg^-1) [165]. The electrochemical cell reaction of Li/O₂ (air) in an organic electrolyte (Li salts such as LiPF₆ in ether-based solvents such as dimethyl ether, DMSO, acetonitrile, pyrrolidone, ionic liquids, etc) [166] is as follows [167]:

\[ \text { Cathode reaction: } \mathrm{O}_{2}+2 \mathrm{Li}^{+}+2 \mathrm{e}^{-} \leftrightarrow \mathrm{Li}_{2} \mathrm{O}_{2} \text { (solid) } \]

\[ \text { Anode reaction: } 2 \mathrm{Li} \leftrightarrow 2 \mathrm{Li}^{+}+2 \mathrm{e}^{-} \]

\[ \text { Overall reaction: } 2 \mathrm{Li}+\mathrm{O}_{2} \leftrightarrow \mathrm{Li}_{2} \mathrm{O}_{2} \text { (solid) } \mathrm{E}^{0}=2.96 \mathrm{~V} \]

The major drawback of a Li-O₂ battery is air cathode and electrolyte decomposition on reaction with O₂. The deposition of solid Li₂O₂ at the electrolyte- air interface that further converts into Li₂O complicates the oxygen reduction reaction (ORR) and results in a steep decrease in the cell capacity on cycling. Research is dedicated to the ORR mechanism in an organic electrolytic medium and ORR catalysts for the cathode. The major types of catalysts include conventional metal catalysts (e.g., Pt, Pd, Au, Co), metal oxides (various crystallographic phases of MnO_x, CoO_x), composites of metal and metal oxides (e.g., the composites of Ag, Ni, and MnO_x, CoO_x), metal carbide, nitrides, sulfides, or organo-C/N/S composites with metal (e.g., Fe/N/C, TiN/C) [167]. Also, nano-structuring of the metal oxides and metal components is reported to enhance the performance. Moreover, nanostructured, high surface area, porous carbon has also been reported as catalysts and supporters for ORR in Li-O₂ batteries. But the carbon surface suffers from instability due to the oxidative environment and the reaction with the discharge product (Li₂O₂), resulting in the formation of Li₂CO₃. Thus, passivated carbon nanostructures using metal oxides such as Al₂O₃, Ru₂O are utilized to avoid direct contact with the carbon surface [165,166,168,169].

Conversion cathodes are the next-generation cathodes for lithium-ion-based batteries. Still, the limitations of the lower cycling stability and reversibility are being resolved by researchers using various above-mentioned methodologies.

5. High Voltage Materials (Only >4 V)

Materials that show either an average voltage greater than 4 V or demonstrate a significant plateau capacity beyond 4 V are only included in the review. LiFePO₄ is the only material in the polyanionic phosphate group that is still reigning in the commercial market since its first introduction in 1995. Polyanionic frameworks provide robust lattice which undergoes minimal atomic arrangements upon lithiation-delithiation. It shows 150-160mAh g^-1 capacity at more comprehensive current ranges with an average voltage of approximately 3.45 V and easily cyclable up to 5000 cycles with good capacity retention [23]. It is also regarded as one of the stable cathodes (air, thermal and cyclic stability) and, coupled with the LTO anode, results in an ultrasafe LFP-LTO battery pack, that can power electric buses, trains, etc. However, due to low voltage and capacity, the overall energy density is lower than that of oxide cathodes. Hence, LFP analogs having higher voltages can mitigate the energy density issue and worth consideration for commercial applications. Several other polyanionic frameworks such as sulfate, silicate, fluorophosphate, pyrophosphate, etc., that crystallize in a variety of polymorphs have also been studied as high voltage LIB cathodes. Table 5 compares the electrochemical properties of the various classes of polyanionic materials.

Table 5 Electrochemical properties pristine polyanionic materials without modification.

5.1 Polyanionic Oxides

5.1.1 Phosphate Group

The structure of polyanionic cathodes is described in the previous section (4.2.2.b.). The inductive effect of the polyanionic framework alters the iono-covalency of the M-O bond and results in good structural stability. The stronger X-O bond decreases orbital overlap between M-O bonds, increasing the ionic character of the M-O bond and the electrochemical potential. The stronger M-O-X linkage holds the oxygen very tightly and does not release oxygen during abusing conditions (under overcharging, overheating, short circuit, etc. conditions). Therefore, the system is intrinsically safer than layered oxides. The more ionic or less covalent M-O bond stabilizes the M^+n/+(n+1) redox. The stabilization in redox energy allows faster ion migration and generates higher C-rate benefits over layered oxides. LiCoO₂ undergoes non-linear expansion after 50% delithiation during the charge [182]. On the contrary, polyanionic frameworks maintain structural integrity and can be cycled for more than 5000 cycles. Overall, polyanionic frameworks have advantages like safety, fast charge-discharge, long-term cycling, and high voltage, but the capacity is limited (170 mAh g^-1 for phosphates, ~325 mAh g^-1 for silicates).

High Voltage Phospho-olivines: The high voltage phosphate olivines are described as LiMPO₄ (M = Mn, Co, Ni). In the lattice, oxygen atoms are distributed in a hexagonal closed packed (hcp) array, where Li⁺ and M⁺² occupy half of the octahedral sites and phosphorous takes one-eighth of the tetrahedral sites. The structure is crystallized in the orthorhombic phase, either in Pmnb or in the Pnma space group (Figure 14a) [183]. The electrochemical profiles are flat over a large Li-composition range revealing framework stability upon lithiation/delithiation as shown in Figure 14c [184]. Influence of the $\mathrm{PO}_{4}^{3-}$ anion tunes the redox energy of the M^+3/+2 couple and increases the operating potential of LiMPO_4, such as ~ 4.1 V for Mn^+3/+2, ~ 4.8 V for Co^+3/+2, and ~ 5.1 V for Ni^+3/+2. Still, surprisingly, Fe^+2/+3 stands at ~ 3.4 V. Higher electronegativity of Fe ( compared to Mn) should result in the higher operating potential of LiFePO₄ (compared to LiMnPO₄), but the trend is opposite. Fe⁺² (3d⁶) requires extra energy to pair the sixth electron in the t_{2 g} orbital, while Mn⁺² (3d⁵) does not. Thus, pairing energy increases the redox energy of the Fe^+2/+3 couple. Hence, more energy could be released or consumed during the redox process of Mn^+2/+3, and LiMnPO₄ shows higher operating potential [185]. However, the associated problems are plenty. LiMPO₄ has intrinsic lower electrical and ionic conductivity issues. Electrical conductivities of LiFePO₄, LiMnPO₄ and LiCoPO₄ are ~ 10^-9, ~ 10^-11, and ~ 10^-9 S cm^-1, respectively, which are much lower than commercial LCO cathode (~ 10^-3 S cm^-1). LiMnPO₄ shows Jahn-Teller distortion of Mn⁺³. On the other hand, LiNiPO₄ and LiCoPO₄ suffer from major electrolyte degradation during high voltage cycling [186]. In addition, the pure phase synthesis of LiNiPO₄ is difficult due to the segregation of the Li₄P₂O₇ and Ni₃P impurity phases [187]. Several ways of coating and doping are reported in the literature for the performance enhancement of LiMPO₄ [188].

Figure 14 (a) Crystal structure of LiMPO₄ (M = Mn, Fe, Co, and Ni). Reproduced from reference [183]. Copyright 2021 IOP Publishing Ltd, (b) Typical voltage profiles of Carbon coated LiMn_0.8Fe_0.2PO₄ composite electrodes measured galvanostatically at various discharge rates at 30 °C in the standard electrolyte solution considering theoretical capacity of 165 mAh g^-1. Reproduced with permission from reference [189]. Copyright 2021 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim and (c) Discharge profile of LiMPO₄ (M = Mn, Fe, Co, and Ni) olivines and LiFe_0.5Mn_0.5PO₄. Reproduced from reference [184].

Carbon-coated LiMnPO₄ shows a flat voltage plateau around 4.1 V vs. Li⁺/Li and shows a reversible capacity of 140 mAh g^-1 at a C/10 rate in the range of 2.7-4.4 V . The material can sustain hundreds of charge-discharge cycles at elevated temperatures without significant capacity fade. Carbon coating increases the electronic conductivity of the pristine sample and stabilizes the cathode-electrolyte interphase. However, LiMnPO₄ suffers from a shortage of specific capacity and rate capability [190]. 10% carbon-coated LiMn_0.8Fe_0.2PO₄, where 20% Mn sites are substituted with Fe, can mitigate the issues mentioned above. Fe doping pushes the specific capacity beyond 150 mAh g^-1. The composite electrode delivers ~ 100 mAh g^-1 capacity at a 10 C rate, whereas the undoped composite electrode delivers only 60 mAh g^-1 (Figure 14b). The performance improvement is attributed to fewer defects in transition metal sites after Fe doping, which maintains the open space for Li-ion conduction even after long-term cycling [189].

On the other hand, due to low electrical conductivity, high voltage cycling, structural instability, etc., LiCoPO₄ losses almost 50% capacity below 100 cycles at C/10 rate [191]. Modification strategies include nanoparticle synthesis, carbon coating, metal-ion doping, use of electrolyte additives, etc. Xing et al. showed that carbon-coated LiCoPO₄ nanograins synthesized via the sol-gel method deliver an initial discharge capacity of 137 mAh g^-1 at C/10 rate and retains 68% after 30 cycles when cycled between 3.0-5.1 V using 1M LiPF₆ in 1:1:1 ratio of EC-DMC-EMC along with 0.1 wt.% thiophene additive [192]. Literature reports on LiNiPO₄ are rare due to difficulty in pure phase synthesis. LiNiPO₄ soaked in Graphitic carbon foam delivers a stable discharge capacity of ~ 80 mAh g^-1 with 8% fade after the 10^th cycle when cycled at a C/10 rate between 3.5-5.4 V. One vital point to note here is that the discharge voltage profile is unexpectedly sloping, unlike the flat voltage region in other phospho-olivines. Other drawbacks are electrolyte decomposition and inferior high-rate performance (almost no capacity at C/2 rate) [193].

Rhombohedral or Monoclinic Li₃M₂(PO₄)₃: The vanadium analog, i.e., Li₃V₂(PO₄)_3, shows two room temperature polymorphs: α-monoclinic and β-rhombohedral (NASICON - Sodium superionic conductor). The monoclinic phase is isotypic with the α-form of the monoclinic Li₃Fe₂(PO₄)_3, whereas the rhombohedral form is isotypic with the NASICON framework. The α-form is thermodynamically more stable and can be directly synthesized via the chemical reaction route, while β-form is synthesized from Na₃V₂(PO₄)₃ through the ion-exchange method [194]. Conversion of the rhombohedral Li₃V₂(PO₄)₃ from the Na₃V₂(PO₄)₃ crystals via the ion-exchange method can be seen in Figure 15a [195]. The main structural difference between the two polymorphs is that the ‘lantern unit [V₂(PO₄)₃]’ is interconnected in the β-form, while in the α-form, they are alternatively oriented in a perpendicular fashion leading to more closed packed structure than the β-form [196]. All the three lithium can be electrochemically extracted from the monoclinic phase, while deinsertion of only two lithium is possible from the rhombohedral phase. Rhombohedral Li₃V₂(PO₄)₃ under potentiodynamic testing conditions (10 mV/1.5 h followed by equilibration) between 3.0-4.5 V range shows two lithium extraction accompanied by the oxidation of V⁺³ to V⁺⁴ and occurs via a biphasic process with an average voltage of 3.77 V vs. Li⁺/Li (Figure 15b). On the contrary, the monoclinic phase undergoes three lithium extraction in the voltage range of 3.0-4.8 V due to V^+4/+5 redox at ~ 4.5 V. It occurs via a multiphasic process with an average voltage of 3.85 V vs. Li⁺/Li [197]. Figure 15c shows the Li-ion migration pathway in the monoclinic structure [198]. The obtained capacity of ~ 200 mAh g^-1 and high voltage pushes the specific energy density to 2330 mWh cm^-3, which is comparable to that of other commercial materials such as LiCoO₂ (2750 mWh cm^-3) and LiFePO₄ (2065 mWh cm^-3) [196]. However, both the phases suffer from similar poor electronic and ionic conductivity issues like other phospho-olivines (electronic conductivity is lying in the range of 10^-8-10^-9 S cm^-1). Carbon coating can enhance electronic conductivity. Ding et al. performed dual carbon coating as an effective strategy and achieved high-rate performance in aqueous electrolytes [199]. Metal-ion doping can address the challenge of bulk ionic conductivity. Doped Li₃V_2-xM_x(PO₄)₃/C (M = Cr⁺³, Co⁺², Na⁺, Nb⁺⁵, etc.) have been studied where electrochemical performances have been upgraded as a result of doping [200,201,202,203]. One special mention is high valent Mo⁺⁶ doping. The Li₃V_1.97Mo_0.03(PO₄)₃/C composite retains 98.3% of the initial capacity (122.8 mAh g^-1) after 150 cycles at 20 C current rate compared to < 70% retention for the pristine material. Substitution of low valent V⁺³ with Mo⁺⁶ strengthens the Mo-O bond to stabilize the structure. It also weakens the Li-O interaction to increase Li-ion diffusivity. Moreover, Mo-doping leads to Li-deficient hole-conducting material, and the donor effect increases ionic conductivity [204]. SiO₂ modification on monoclinic Li₃V₂(PO₄)₃/C improves the cycling performance, although multiple plateaus were still visible (Figure 15d). The transmission electron microscopy (TEM) image in Figure 15e confirms that the Li₃V₂(PO₄)₃ particles were wrapped in amorphous silica and carbon layers having a 3-5 nm thickness [205].

Figure 15 (a) Conversion of the rhombohedral Na₃V₂(PO₄)₃ crystal structure to the rhombohedral Li₃V₂(PO₄)₃ crystal structure via the ion-exchange method, (b) Charge-discharge curves of the rhombohedral Li₃V₂(PO₄)₃/carbon nanofiber (CNF) composite at various current rates ranging from 0.5C to 100C. Reprinted with permission from reference [195]. Copyright 2020 American Chemical Society, (c) Schematic illustration of the possible Li-ion migration pathways in monoclinic Li₃V₂(PO₄)₃ crystal structure. Reprinted with permission from reference [198]. Copyright 2012 American Chemical Society, (d) Charge-discharge voltage profiles of monoclinic Li₃V₂(PO₄)₃/C and 1, 2, 3 wt.% SiO₂ coated samples at C/10 current rate in between 3.0-4.8 V range and (e) Transmission electron microscopy (TEM) image of 2 wt.% SiO₂ coated Li₃V₂(PO₄)₃/C. The lattice fringes of pristine material and coating region are clearly visible. Reprinted with permission from reference [205]. Copyright 2012 American Chemical Society.

5.1.2 Tavorite Compounds and Derivatives

The tavorite phase generally defines a large class of phosphates and sulfates of the general formula Li_yMXO₄Z, where y = 0, 1, 2; M = Co, Ni, Mn, V, Fe; X = P, S; Z = F, O, OH. The presence of ‘Z’ helps the two lithium and two electrons exchange per formula unit theoretically. However, practically only one electron exchange is observed in most cases [206].

Vanadium-based fluorophosphate (LiVPO₄F) and oxyphosphate (LiVOPO₄) are examples that can exchange two electrons per formula unit. However, besides their high voltage redox, e.g., at 4.2 V for V^+3/+4 in LiVPO₄F and 4.0 V for V^+4/+5 in LiVOPO₄, the low voltage redox at 2.2 V for V^+3/+4 in LiVOPO₄, and 1.8 V for V^+2/+3 in LiVPO₄F lowers the average voltage down to ~ 2.4 V for LiVPO₄F and ~ 1.8 V for LiVOPO₄ [207,208,209]. Due to this fact, proper attention was not provided on these materials like insertion and other polyanionic cathodes, although having high theoretical capacity. LiVPO₄F can show reversible capacity around 120-150 mAh g^-1 when cycled within C/2 to 10 C current rate. Further optimization is needed to scale up vanadium phosphates to reach the level of LiFePO₄.

Fluorophosphates (Li₂MPO₄F) is a fluorine substituted phosphate analog and comes under the category of the tavorite phase. The edge shared MO₄F₂ infinite chains interconnected by PO₄ tetrahedra opens two sites for Li, out of which only one is electrochemically mobile [176]. The presence of fluorine upshifts the average voltage to ~ 5 V for Co^+3/+2 and ~ 5.3 V for Ni^+3/+2, i.e., approximately 200 mV is gained due to fluorination. Li₂CoPO₄F is reported to show a capacity of ~ 110 mAh g^-1 in a sulfone-based electrolyte with an upper cut-off voltage of ~ 5.5 V [210]. Li₂NiPO₄F is hard to synthesize in the pure phase, and related reports are scarce in the literature [177]. The study on these materials is mainly limited due to the electrode-electrolyte instability at high voltages.

5.1.3 Pyrophosphates

Pyrophosphate (Li₂MP₂O₇) combines a high theoretical capacity of 220 mAh g^-1 due to the two-electron transfer per metal cation along with high average voltage (>4 V). As mentioned earlier, pyrophosphate moiety has more resonance structures arising from its extended P-O backbone, and the resulting covalence is the reason behind higher average voltage than phosphates. The crystal structure can be described as 3D frameworks with MO₅ square pyramids sharing an edge with the MO₆ octahedra, which further interlinked via P₂O₇ groups (Figure 16a and Figure 16b). The electrochemical performance of Li₂FeP₂O₇ is comparatively inferior to that of LiFePO₄. The average voltage is almost similar, i.e., ~ 3.5 V vs. Li⁺/Li. But the higher atomic weight of the P₂O₇ moiety lowers the capacity down to ~ 110 mAh g^-1 (Figure 14d) [211]. Similar specific capacity lowering effect due to higher molar mass per formula unit is also obvious for Li₂MP₂O₇ other than Fe. Other analogs such as Li₂VOP₂O₇, Li₂CoP₂O₇, and Li₂MnP₂O₇ draw attention due to the high average discharge potentials of 4.1 V, 5.0 V, and 4.4 V, respectively, but the reversible capacity lies around 100 mAh g^-1 (Figure 16c and Figure 16d) [212,213,214]. Ng et al. showed by first-principles calculations that lithium diffusion and the charge capacity of lithium cobalt phosphate depends on the intrinsic defects. Doping Si in place of P improves charge carrier mobility without altering other electronic properties of the pristine material [215]. However, cyclic stability is also inferior resulting from poor stability of the electrolytes at ~ 5 V . Moreover, difficulty in extracting lithium due to sluggish kinetics is the other reason behind these materials' poor cycling performance. Whittingham and coworkers synthesized molybdenum (oxy) pyrophosphate [δ-(MoO₂)₂P₂O₇] that can accommodate four lithium ions upon discharge to 2 V based on Mo^+6/+4 redox accounting for a capacity of ~ 250 mAh g^-1. Intercalation of more than two lithium ions is not electrochemically reversible and leads to a loss in crystallinity. The reversible two lithium intercalation capacity of ~ 110 mAh g^-1 between 4.0-2.3 V have been achieved for this material [216].

Figure 16 (a) Crystal structure of Li_2-xCoP₂O₇. Reprinted with permission from reference [215]. Copyright 2019 American Chemical Society, (b) crystal structure of Li₂MnP₂O₇. Reprinted with permission from reference [211]. Copyright 2010 American Chemical Society, (c) Initial voltage profiles of Li₂CoSiO₄ (red curve) and Li₂CoSiO₄/C (black curve), (d) Charge-discharge capacity as a function of cycle number for Li₂CoSiO₄/C. Reprinted with permission from reference [217]. Copyright 2013 American Chemical Society.

5.1.4 Silicates and Fluorosilicates

Silicates or orthosilicates Li₂MSiO₄ (where M = Mn^2+, Fe^2+, and Co²⁺⁾ is a class of poly-oxyanion cathodes that have attracted much attention after the first proposal in 2005 by Nyten et al. [218]. It was reported as a phase pure Li₂FeSiO₄, with an orthorhombic structure that showed an initial discharge capacity of 130 mAh g^-1. Li₂MnSiO₄ is another extensively studied silicate cathode that shows a high voltage of ~ 4.1 V and a high theoretical capacity of 333 mAh g^-1 as the extraction of two lithium ions is possible, which is accompanied by the Mn⁺²/Mn⁺³ and Mn⁺³/Mn⁺⁴ redox couple. But practically it shows only 120 mAh g^-1 capacity [219]. Fe and Si are among the most abundant and low-cost elements in the earth's crust. This is also one of the lying reasons behind quick popularity gain after the first report [218]. Other features of the silicate materials are as follows: a) unlike the olivine phosphate materials (where at most only one lithium can be extracted), silicates can undergo two lithium ions per formula unit extraction accompanied by a two-electron redox process (i.e., resulting from both M^2+/M³⁺ and M^3+/M⁴⁺ redox couples) leading to the higher theoretical specific capacity of this type of materials (300-330 mAh g^-1), b) due to strong Si-O bond this material shows high thermal stability and exhibits good safety under extreme testing conditions. However, low electronic conductivity (10^-14 S cm^-1), thereby slow lithium-ion kinetics; amorphization of the active material, and structural reversibility are the major drawbacks of these types of materials.

Among silicates, Li₂CoSiO₄ has a high voltage of ~ 4.5 V and a capacity of about ~ 325 mAh g^-1, which leads to a high energy density of ~ 1463 Wh kg^-1. The 2^nd Li-ion deintercalation voltage plateau at around 5 V is difficult to achieve [220]. Co is expensive and environmentally toxic. Therefore, the material lost commercial interest as the battery industry is seeking to reach a Co-free zone. However, there are some excellent reports on this material. Gong et al. have reported a charge capacity of 234 mAh g^-1 for the Li₂CoSiO₄/C composite, which corresponds to 0.46 lithium per formula unit. They have used the hydrothermal route for synthesis and found the reversible Li extraction and insertion at 4.1 V vs. Li. [220]. Lyness et al. synthesized all three polymorphs of Li₂CoSiO₄ and showed that the βI polymorph exhibits electrochemical activity. They reported a charge capacity of 170 mAh g^-1, which is equivalent to 1.1 Li per formula unit for the βI polymorph of Li₂CoSiO₄ coated with carbon [221]. Linda Nazar and coworkers synthesized the Li₂CoSiO₄/C nanocomposite via hydrothermal method using highly ordered mesoporous carbon/silica (MCS) framework as both carbon and silica precursors. However, the electrochemical performance is very poor, as shown in Figure 15c and Figure 15d [217].

5.2 Spinel Oxides

Spinel oxide as cathode for lithium-ion battery is primarily limited to LiMn₂O₄ (LMO). This compound has a spinel-like structure with anionic lattice having an α-NaFeO₂ type structure in which Li and Mn reside in the tetrahedral and octahedral sites, respectively [222]. In LMO, lithium-ion shuttles between the tetrahedral sites by hopping via the intermediate octahedral site. Due to this spinel structure and diffusion behavior, the lithium diffusion pathway is three-dimensional. This type of stable 3D geometry helps in the Li-ion diffusivity and offers high-rate capability. The material shows a theoretical capacity of 148 mAh g^-1 and a discharge potential of 4 V [223]. But LiMn₂O₄ cathode has a significant issue of Mn dissolution [224]. The insertion of Ni into the spinel structure of LiMn₂O₄ forming LiNi_0.5Mn_1.5O₄ (LNMO) leads to an increased operating voltage of 4.7 V vs. Li⁺/Li. LNMO has gained much more attention than LMO because of its high operating potential, low cost of synthesis, and Co-free nature. It shows a reversible specific capacity of 148 mAh g^-1 [225]. The compound crystallizes in an ordered and disordered configuration depending on the synthetic condition (Figure 17a) [226]. The disordered configuration exhibits a more stable cycling performance due to its better ionic and electronic conductivities. But the practical application of LNMO is hampered due to electrolyte instability at such a high working potential of ~ 4.7 V vs. Li⁺/Li.

Figure 17 Crystal structure of LiNi_1.5Mn_0.5O₄. Reprinted with permission from reference [226]. Copyright 2020 American Chemical Society, (b) Discharge capacities of uncoated and LiPON coated LNMO at C/5 and C/3 current rates. Reprinted with permission from reference [227]. Copyright 2013 The Electrochemical Society, (c) Charge-discharge profiles of undoped, Cr-doped, and Nb-doped LNMO at C/5 rate, and (d) Cycling performance of undoped, Cr-doped, and Nb-doped LNMO up to 500 cycles. Reprinted with permission from reference [228]. Copyright 2016 American Chemical Society.

5.2.1 Metal Cation Doping

One of the most effective methods is the partial replacement of Ni or O in LNMO with a metal cation or anion. Mg doping into LNMO confers structural stability and improves the voltage profile (4.70-4.75 V vs. Li⁺/Li) [229]. The nanosized LiMg_0.05Ni_0.45Mn_1.5O₄ material is synthesized by a modified solid-state synthesis with an ordered cubic spinel structure that shows good capacity retention upon cycling (131 mAh g⁻¹ at C/10 and 117 mAh g^-1 at 1 C) [230]. There are also some essential outcomes being obtained by Cr-doping. The strong oxygen affinity of the Cr atom helps to form strong Cr-O bonds favoring good stability of the spinel structure during cycling. A single-phase LiMn_1.4Cr_0.2Ni_0.4O₄ spinel synthesized by simple sucrose assisted combustion method exhibits the maximum rate capability with capacity retention of 92% at 60 C (1C = 147.5 mA g⁻¹ or 0.260 mA cm⁻²) discharge rate among LiNi_0.5Mn_1.5O₄- type cathodes [231]. Mao et al. synthesized Cr-and Nb-doped LNMO via a polyvinylpyrrolidone (PVP) combustion method. The Cr doped sample LiCr_0.1Ni_0.45Mn_1.45O₄ retains 94.1% capacity after 500 cycles at 1C, the coulombic and energy efficiencies remain at over 99.7%, and 97.5%, respectively. The Nb-doped sample LiNb_0.02Ni_0.49Mn_1.49O₄ shows slightly higher capacity but lesser retention (Figure 17c and Figure 17d) than other materials. The performance improvement is ascribed to the fact that Cr doping sharpens the edges and decreases particle size, while Nb doping results in smoother surface and more rounded large particles [228].

5.2.2 Surface Coating

The presence of ppm level HF produced due to reaction of salt LiPF₆ and trace level water in the organic electrolyte (e.g., 1M LiPF₆ in 1:1 ratio of EC: DEC) causes unavoidable side reactions. The steps of the reaction mechanism are discussed below:

1) LiPF₆ decomposes at >4.4 V and forms PF_5, which reacts with water leading to the formation of HF according to the reaction:

LiPF₆ → LiF + PF₅; PF₅ + H₂O → 2HF + PF₃O

2) The resulting HF causes high voltage LNMO to dissolve in the electrolyte solvent through the following reaction:

4HF + 2LiNi_0.5Mn_1.5O₄→3Ni_0.25Mn_0.75O₂ + 0.25NiF₂ + 0.75MnF₂ + 2LiF + 2H₂O [232]

Therefore, capacity fade is observed over cycling due to the loss of the active materials following the reactions mentioned above. The surface coatings have been proven to be the effective materials to mitigate this HF corrosion. In this regard, some oxide and fluoride materials draw special attention as surface coating materials for LNMO [233]. SiO₂ is known as an HF scavenger. Therefore, SiO₂ helps to protect LNMO electrode materials by neutralizing the HF in the electrolyte. Wei Pang et al. have studied the effect of SiO₂ coatings on LNMO. SiO₂ coating layer lowers the capacity decay rate by 45% and 65%, at 25 C and 55 C, respectively, and enhances the Li diffusivity by 15% [234]. Tao et al. reported TiO₂ coated LNMO synthesized by simple wet chemical method. The coated material shows better cycling stability, about 88.5% of the capacity was retained at a 2 C rate at the end of 500 cycles [235]. Wu et al. investigated AlF₃ coated LNMO using 1 wt.% AlF₃, which shows good reversibility, with 93.6% retention after 50 cycles [236]. Sub-nanometer LiPON coating also improves the high voltage (~ 4.9 V) cycling stability of the LNMO cathodes in conventional carbonate electrolytes, as shown in Figure 17b [227].

5.2.3 Rare Earth Metal Doping

Partial substitution of Mn by rare-earth elements, such as La, Nd, and Ce, have also been investigated to improve the electrochemical performance of LNMO. Ce substituted LiNi_0.495Ce_0.005Mn_1.5O₄ delivers better C rate performance than LiNi_0.5Mn_1.5O₄ [237]. This is due to the synergistic effects of the decrease in the extent of mixing between lithium ions and transition metal ions, an increase in the lattice parameter, and Ni/Mn disordering degree after an appropriate amount of Ce doping. The doped materials exhibit a specific capacity of 123 mAh g^-1 at 1C rate, capacity retention of 94.51% after 100 cycles, and 115.4 mAh g^-1 capacity at 10C rate, indicating good rate capability. Besides, it has been observed that Ce substitution decreases the formation of Mn³⁺ and stabilizes the structure, and improves the electrical conductivity. The LiNi_0.495Mn_1.495Ce_0.02O₄/graphite full-cell can deliver a capacity of 113 mAh g^-1 and capacity retention of 98% after 100 cycles at 1 C rate and 25 °C [238]. Nb doping also improves performance, as described in Figure 16c and Figure 16d [228].

5.2.4 Anion Doping

Similar to cation doping, few reports of anion doping, e.g., F-doping, replace some of the oxygen with fluorine and thereby helps suppress the formation of NiO impurity, improves the tap density of the material, and simultaneously reduces the voltage polarization. Moreover, the dissolution of the particles into the electrolyte can be mitigated by fluorine coating (Figure 6). Xu et al. reported the sol-gel synthesis of LiNi_0.5Mn_1.5O_3.975F_0.05. Electrochemical performance shows that fluorine doping enhances the initial capacity from 130 mAh g⁻¹ to over 140 mAh g⁻¹ in the voltage range of 3.5- 5.2 V compared with pristine LiNi_0.5Mn_1.5O₄ [239].

Table 6 Electrochemical performance of LNMO modified by surface coating and dopings.

Spinel LiNi_0.5Mn_1.5O₄ (LNMO) is a promising cathode material due to its high operating voltage of ~ 4.7 V vs. Li⁺/Li, leading to a specific energy of 650 Wh kg^-1, which is 162.5% and 131.3% higher than that of LiMn₂O₄ and LiFePO4, respectively. Moreover, these materials exhibit good cycling stability, excellent rate performance, superior thermal stability, and high ionic conductivity. However, the high operating voltage, which is beyond the oxidation potential of the carbonate solvents, limits its practical applications. Therefore, large irreversible capacity and low coulombic efficiency are observed for these materials with conventional LIB electrolytes. In addition, electrolytes with high voltage additives can enhance the long-term electrochemical performances of this material.

6. Conclusion and Outlook

This review focuses on LIB cathode materials which show either capacity ≥250 mAh g^-1 or voltage ≥4 V. The crystal structure, performance comparison among various compositions, Li-ion storing mechanism, shortcomings of the materials, capacity, voltage fading mechanisms, and the reported methods to overcome these problems have been systematically described.

• The concept of lithium-rich layered oxide cathodes is based on the solid solution of Li₂MO₃ (M = Mn, Ru, Ti, Sn, etc.) and LiMO₂(M = Ni, Co, Mn, etc.). Incorporating oxygen redox beyond the transition metal redox opened up the gateway to achieve high capacity that was beyond conventional layered oxides. But irreversible oxygen redox causes transition metal migration, voltage hysteresis, voltage decay, severe capacity fade, and poor rate performance during charge-discharge cycles. Based on the reported theories, coating, doping, and composition optimization effectively hinder oxygen gas release, thereby balancing high capacity and long-term cyclability. 
• Lithium-rich disordered rock salts are another class of oxide materials that show capacity beyond 250 mAh g^-1 utilizing the benefit from oxygen redox. It has a wide compositional range that provides the opportunity to use metals (Ru, Mo, Nb, V, Cr, etc.) that are not commonly used to synthesize conventional layered oxides. The ease of synthesis via simple solid-state and mechanical route and earth-abundant Mn-rich compositions is desirable for commercial applications. However, it also suffers from similar problems that appear in lithium-rich layered oxides. Cycling stability in most of the literature reports was confined to 100 cycles, and the performance of those 100 cycles significantly relied on the amount of carbon in the coating slurry. Therefore, further optimization of electrode fabrication, slurry composition, and testing protocols are required.
• On the other hand, conversion cathodes can yield a capacity of >300 mAh g^-1 through non-site limited chemical reactions with lithium. However, the performance is limited by the vast volume expansion and low electrical conductivity. The commercialization of silicon-based anodes was made possible due to extensive research to synthesize unique morphology, robust binders, and optimization of carbon contents. Similar studies are also needed for the further development of conversion cathodes. 
• Phosphate frameworks are extensively studied in this category after the commercial success of LiFePO_4. Fe-based systems primarily offer < 4 V of average voltage. On the other hand, V-based materials such as Li₃V₂(PO₄)₃, LiVOPO₄, LiVPO₄F demonstrate ≥4 V plateau and stable capacity. Their electrochemical performance mainly depends on the uniform and well-distributed carbon coating as the hindrance of electron conduction is the major shortcoming. Remaining challenges include further optimization of carbon coating, size and shape of the synthesized nanoparticles, and the synthetic conditions. Ni-and Co-based phosphates, pyrophosphates, and silicate moieties can provide 4.8-5.0 V of discharge voltage. However, most reports are only computationally realized because of the difficulty in synthesizing the pure phase material and the instability of the conventional carbonate electrolytes. The development of the solid-state electrolyte or high voltage ionic-liquid electrolytes may broaden the scope of investigating these materials in the future.
• High voltage spinel oxides, especially LNMO, also deliver 4.7 V vs. Li/Li⁺. But improving electrolyte stability by using additives, formation of artificial interphase, thin and uniform carbon coating is essential for its electrochemical performance. Areas, where LNMO lacks are long-term cycling stability and coulombic efficiency. Therefore, research in high-voltage electrolyte and cathode-electrolyte interphase are the keys to future progress. 

In summary, high-capacity materials mainly suffer from the inaccessibility of reversible lithium sites after a few cycles, whereas high voltage materials experience cathode-electrolyte incompatibility at elevated voltages. Tuning the electronic structure of the host crystals through doping, protective coating on the surface of active materials to hinder the attack of reactive electrolyte, synthesis with suitable morphology for facile ion migration, manufacturing electrode architecture to buffer the volume change, and the use of high voltage electrolytes or electrolyte additive salts are the exploited tactics to improve the electrochemical performance. An important point to note here is that high capacity and high voltage are not the only criteria to judge the commercial suitability of material.

Batteries, being energy storage devices, must meet all the long-term storage requirements. Apart from the energy density, other criteria like power density, cost, safety, recyclability, ease of operation, life span, etc., are also equally vital for a LIB performance matrix. Based on the commercial demand, the performance comparison of the materials is discussed here by considering the parameters mentioned above. It is also shown that how does one material surpasses the other in a particular electrochemical property.

• From the point of view of energy density, LMR-LO cathodes can reach beyond 1000 Wh kg^-1, whereas the current LIB technology stands at ~ 300 Wh kg^-1. To shrug off the complete dependence on fossil fuel, LIBs at some point should be used to substitute gasoline. Still, the inadequate energy storing capability of the currently used LIBs is a major obstacle behind electric aviation. The energy density of gasoline is ~ 13 kWh kg^-1, more than 40 times higher than that of LIBs. Considering ~ 12.6% of the average gasoline tank-to-wheel efficiency of the US fleet, the practical energy density becomes ~ 1700 Wh kg^-1, which is also approximately ten times higher than that of the stated system. A lithium-air battery, where oxygen is used as a cathode, is the only system that can theoretically satisfy the energy density demand (11,640 Wh kg^-1 for a nominal voltage of 3 V). 
• The power density is another important factor that must be kept in balance with the energy density. Although supercapacitors, hybrid ultracapacitors, etc., are specially designed devices to achieve high power output, a battery should possess a threshold value of power at the time of quick energy supply. The robust structural moiety of the polyanionic materials supports fast lithium-ion migration and is thereby used in high-power commercial tools. Next-generation high voltage polyanionic frameworks coupled with the Li₄Ti₅O₁₂ (LTO) anode is the best choice in terms of power output. Cationic reorganization in high-capacity layered cathodes and irreversible phase formation in conversion materials are problematic towards power delivery. 
• Cost is an essential criterion for any material to gain wide acceptability in the mass market. The lower the cost or lesser the $/kWh for an electric vehicle, the more attractive it is to the customers. Due to uneven geological distribution, geopolitical tensions, and unethical exploitation of child labor, Co and Li have been considered under the critical element category for the last few years. Mainly, the fluctuation in the cobalt price results in a sudden upsurge or rapid downfall in the stock market price. On the other hand, too much reliance on Ni for NMC-811 in EVs may cause a vast Ni deficit in the no-distant future. Mn-rich disordered rock salt materials and Ni-Co free polyanionic frameworks can be a viable option. However, if lithium-oxygen comes into the picture, the worries about cathode cost can be relived as oxygen is ubiquitous. Another way to regain the material value is to reuse or recycle spent LIBs. But the consideration of recycling comes from a commercialized material. Materials discussed here are still being studied at a laboratory scale, and literature reports on recycling are rare. On a general note, polyanionic frameworks are resistant to chemical attacks and are difficult to recycle due to the covalently bonded network. Moreover, it will not be attention-worthy from an industrial perspective if it does not contain any critical metals. Therefore, the recycling of oxide materials is only industry-relevant. 
• Safety must be given the utmost priority for designing a device. Customers do not embrace any unsafe gadget at all for any type of use. Heat generation is a common phenomenon in the case of LIB-run devices. Ni-rich content of current commercial cathodes may cause massive heat evolution, and Li-plating at graphite during fast charging, at low state-of-charge (< 20%), and low-temperature may pose a potential threat. From a chemistry point of view, a polyanionic cathode coupled with an LTO anode is the most reliable system. However, the future commercialization of solid-state batteries may solve the problem. 
• Finally, the durability is an essential factor for commercial success. For example, LIBs in the mobile phone can sustain 3-5 years, and the threshold is a minimum of 10 years for battery-run electric vehicles. Polyanionic materials can perform for a long time but at the expense of low energy density, which means adding more weight or putting excess volume in the battery of equivalent energy-power rating. On the contrary, irreversible lithium loss from lithium cathodes, parasitic side reactions in conversion cathodes, polysulfide shuttling in the Li-S system, and overvoltage and polarization during ORR, OER, etc., in the case of Li-O₂ raise a question mark in terms of reversibility. Overall, the high energy materials need to cover the far distance before it becomes a commercial reality. But the scientific concept, electrochemical mechanism, failure processes corresponding to these materials are worth consideration. 

Author contribution

SG, UB, and SB has written the various parts of the review. Dr. Surendra K. Martha has conceptualized, planned, and coordinated the work.

Funding

SG acknowledges CSIR, Govt. of India (File No: 09/1001(0067)/2019-EMR-I) for fellowship. UB and SKM acknowledge DST-SERB Sanction Order: CRG/2018/003543 for the fellowship and financial support to this work. SB acknowledges DST-IISc Energy Storage Platform on Supercapacitors and Power Dense Devices through the MECSP-2K17 program under grant no. DST/TMD/MECSP/2K17/20, Government of India for fellowship.

Competing Interest

The authors declare no competing interest.