Recent advances in the design of cathode materials for Li-ion batteries

The Li-ion battery (LIB) industry has rapidly developed and dominates the market of electric vehicles and portable electronic devices. Special attention is devoted to achieving higher power and energy densities, along with enhancing safety and reducing cost. Therefore, critical insights should be made on the understanding of the behavior of the components of LIBs under working conditions in order to direct future research and development. The present review discusses the literature on the properties and limitations of different cathode materials for LIBs, including layered transition metal oxides, spinels, and polyanionic positive electrode materials, with critical insights on the structural, thermal, and electrochemical changes that take place during cycling. Besides, the strategies and techniques capable of overcoming current limitations are highlighted.


Introduction
The pressing demand to introduce alternative energy sources to fossil fuels and minimize CO 2 emissions generates considerable recent research interest in the development of renewables along with energy storage systems. Shortly, the Earth is expected to run out of petroleum resources as they are being depleted at a fast rate. In addition, avoiding petroleum resources is necessary because of their negative environmental impacts, such as the water and air pollution that led to global warming problems.
At the global level, renewable sources show a great theoretical ability to cover the whole human need for energy. For instance, Earth receives 1.2 Â 10 5 TW of solar power, which can cover the annual worldwide need for energy in almost one hour. Despite the current limitations in the properties of materials that lead to inefficient collection and conversion of light, the primary current focus is on how to progress rapidly towards highly functional materials to satisfy the ever-increasing energy demand.
Energy storage technologies are divided into stationary and portable systems; both are of great importance, and each type is at a different stage of development. In multiple stationary applications, grid-connected energy storage systems have been Nourhan Mohamed received her M.Sc. degree in Sciences, Technologies, Health from Picardie Jules Verne University. Her master's program was a joint Erasmus program in materials for energy storage and conversion "MESC". She joined the Energy Materials Laboratory at AUC for a year as a Research Assistant. Her research is focused on positive electrode materials for Li-ion batteries.
Nageh Allam received his PhD in Materials Science and Engineering from The Pennsylvania State University. He joined Georgia Institute of Technology as a postdoctoral fellow then Massachusetts Institute of Technology as a Research Scholar. He moved to The American University in Cairo in September 2011 where he is now a professor of Materials Science and Engineering. His research focuses on the fabrication of well-designed nanostructured materials with composition, size, and shape control for use in energy conversion and storage, electrochemical sensors, and biomedical applications, among others. His research comprises both experimental and theoretical activities.
deployed to facilitate the integration of renewable power sources due to their variable nature. Energy storage systems are connected to the grid to provide sustainability by storing energy during the low-energy demand time and supply it during the high-energy demand time. Nowadays, the global energy storage is dominated by one technology, the pumped hydropower, for grid applications, which accounts for 95% of the worldwide energy storage. The rest 5% comprise mainly of thermal, mechanical, and electrochemical storage. In comparison, a project of Li-ion batteries contributes 24 MW that is as low as 0.01% of the overall storage, as shown in Fig. 1.
Advances in renewable energy require modernization of the electricity storage systems, including electrochemical capacitors, lead-acid batteries, high-temperature sodium batteries, lithium-ion batteries, and redox ow batteries. In this review, we focus mainly on Li-ion batteries because of their superior performance in comparison with other conventional battery types. Li-ion batteries dominate the portable devices industry due to their prime properties such as high-energy-density, high charge/discharge rate, low maintenance, and long-lifetime, which make them the most suitable candidates for a plethora of applications. Therefore, they met the market requirements, and their sales increased signicantly between 2008 and 2014 from 0.1 to 22.5 billion dollars in the electric vehicles sector and from 9.1 to 31 billion dollars in the total market. 1, 2 Li-batteries have been widely adopted because Li is the lightest metal (6.94 g mol À1 ) with the smallest radius that can provide high gravimetric and volumetric capacity, leading to a signicant weight and volume reduction of the battery. Furthermore, Li exhibits the lowest reduction potential (À3.04 V versus SHE), allowing the cell to deliver the highest possible potential resulting in a decrease in the number of cells required to operate a device. However, several safety problems caused the withdrawal of Li metal batteries from the market aer some catastrophic accidents because of the high reactivity of the Li metal and the growth of Li dendrite, which results in short circuits. Therefore, Li metal anode was replaced by safer alternatives such as intercalation, alloying, and conversion materials. Consequently, much research has focused on the chemical, thermal, and structural stability of the materials for Li-ion batteries and how to introduce practical solutions to their well-recognized safety problems. These problems are related to the reactions that can take place under extreme conditions such as high over-potential, electrolyte decomposition, SEI decomposition, cathode material decomposition, and generation of ammable gases. The safety stands as an obstacle to the introduction of new high energy cathode materials because they work outside the stability window of the available electrolytes. Improving the performance, along with the cost and safety, are the key factors to expand the Li-ion battery applications signicantly. A signicant interest in Li-ion batteries is given to the cathode materials and how to improve its electrochemical performance along with preserving the mechanical, electrochemical, and chemical stability of the materials upon cycling. Because the performance of the battery, including cell potential, energy density, power density, lifetime, and safety in addition to cost, depends to a great extent on the cathode material. For instance, high energy can be obtained from a battery by increasing the intercalation voltage (cathode material type) or the amount of Li + that can participate in the electrochemical reaction (capacity). In this review paper, we focus on different types of cathode materials and discuss their electrochemical properties considering their structure and morphology as well as thermal and electrochemical performance.

Li-ion batteries (LIBs)
A Li-ion module consists of several cells connected into packs to deliver the required voltage, power, and energy. LIBs convert chemical energy into electricity via a reversible electrochemical reaction. They consist mainly of anode and cathode, which are electronically isolated by an electrolyte. The anode is the negative electrode, which is oxidized during the electrochemical reaction to donating electrons to the external circuit. The cathode is the positive electrode, which is reduced during the electrochemical reaction and accepts electrons from the external circuit. The electrolyte is a solid or a liquid medium that allows the Li ions to move between the anode and cathode and force the electrons to move through the external circuit. Therefore, electrolyte should exhibit high ionic and low electronic conductivity. Moreover, the Fermi level of the anode should be at lower energy than the LUMO of the electrolyte to prevent the electrolyte reduction, while the Fermi level of the cathode should be at higher energy than the HUMO of the electrolyte to prevent the electrolyte oxidation. During the rst charge, the electrolyte undergoes a reduction reaction at the surface of the anode forming an irreversible passive layer, known as solid electrolyte interphase (SEI). It is a complicated layer composed of organic and inorganic components that prevent the further decomposition of the electrolyte upon cycling, but it results in an irreversible capacity loss. The performance of the battery is dependent on the morphology, composition, and thickness of this layer. 4 If this protective layer is damaged, the decomposition of the electrolyte will continue again, leading to fast heating and undesirable chemical reactions that are hazardous. There are several reported mechanisms for the formation and stabilization of the SEI on the surface of both anode and cathode. [5][6][7][8] 3 Challenges and trend in cathode materials for Li-ion batteries The development of the cathode materials for Li-ion batteries remains challenging because the existing materials such as layered transition metals oxides, olivines, or spinel all show upsides and downsides. For example, the layered oxides such as LiCoO 2 suffers from the instability that limits their potential window and capacity. Moreover, the toxicity and the high production cost of the Co-based materials are undesirable. On the other hand, Li-rich layered oxides (Li 1+x M 1Àx O 2 ), where M is a mixture of transition metals (Ni, Mn, and Co), is promising as they can work at high discharge voltages of >4.5 V and deliver high specic capacities. However, they suffer from large voltage decay during cycling and high irreversible capacity loss at the rst cycles, which limit their use. Moreover, olivine materials such as LiFePO 4 are competitive candidates surpassing the stability of layered oxides at elevated temperatures due to their high thermal and structural stability. However, their low electronic and ionic conductivity stand as an obstacle to expand their use for high energy Li-ion batteries. Several steps forward have been made in the development of cathode materials for Liion batteries to overcome current challenges. First, investigating the changes in the electrochemical performance inside the cathode during cycling will lead to understanding the origin of the malfunction of the materials under working conditions. The strategies considered to improve these materials should address the problems related to the electrochemical performance side by side with the safety and cost issues. These strategies mainly include designing nanostructured materials with controlled morphology, introducing articial defects in the nanostructured materials, using nanostructured complexes by doping or surface coating, and surface modication that is indispensable to minimize the undesirable side reaction at the electrolyte-cathode interface. Those are critical factors in jeopardizing the electrode performance. Also, there is a great interest in identifying new cathode materials that can fulll the required balance between the performance and environmental requirements. In this review, we compare the different cathode materials that have been widely used in Li-ion batteries and analyze their performance.

Layered transition metal oxides positive electrode materials
The reason for the continued attention to transition metal oxides LiMO 2 (M ¼ Co, Mn, Ni) as electrode materials is their high Li + mobility in 2-dimensional space, thanks to its layered Fig. 2 The crystallographic framework of the layered LiMO 2 (M is Co or Ni) consisting of MO 6 octahedra (gray) and Li + (orange) "by VESTA". structure, as shown in Fig. 2. The crystallographic framework of Layered LiMO 2 (O3 type) is a-NaFeO 2 -like structure, which consists of a cubic close-packed oxygen array occupying the 6c sites, where Li and M ions occupy the 3a and 3b sites, respectively. M and Li ions occupy alternate layers in the octahedral sites along the (111) plane forming ABC stacking sequence. The cell parameters of LiMO 2 (M ¼ Co and Ni) are shown in Table 1.

LiCoO 2
LiCoO 2 represents a signicant advance in the history of rechargeable Li-ion batteries, as it was the rst commercialized positive electrode material by Sony in 1991. Sony combined the LiCoO 2 cathode and carbon anode to produce the rst successful rechargeable Li-ion battery. From the electrochemistry point of view, Li x CoO 2 is a very attractive material because it exhibits a high theoretical capacity of 274 mA h g À1 and high energy density. However, its performance is strongly affected by structural and thermal instability at high voltage, and high temperature, as well as fast capacity, fade at the high current rate or deep cycling. During charging, several phase transitions take place at different stages of the de-intercalation of Li + . Fig. 3 shows a typical charge/discharge behavior of Li x CoO 2 as a function of the amount of x. The large plateau at 3.93 V vs. Li + / Li 0 corresponds to a rst-order transition (H1 4 H2). The coexistence of two phases between the lithium concentrations x ¼ 0.75 and 0.93 is accompanied by the transition from a semiconductor to the metal, which facilitates the charge transfer. In the x range between 0.75 and 0.25, two order/disorder transitions were observed from hexagonal to monoclinic, then from monoclinic to hexagonal, which correspond to the plateaus b and c. Therefore, almost one-half of the lithium ions can be cycled reversibly in and out of the compound corresponding to the composition range between x ¼ 1 and 0.5. Consequently, a practical specic capacity of 140 mA h g À1 is obtained at a charging voltage of 4.2 V vs. Li/Li + to avoid the structural rearrangement of the material. It was traditionally accepted that the positive charge decrease by Li ions removal from Li x CoO 2 is compensated by the increased charge of transition metal ions. At the same time, O 2À electronic state is xed to keep the electro-neutrality of the material. But sensitive techniques to the oxidation state, spin state, bond covalence, and the electronic structures such as EELS and XAS suggest that both oxygen and transition metal ions experience an electronic change, and both participate to the electrochemical reaction. The oxygen K-edge XAS for Li x -CoO 2 shows signicant change with respect to Li-ion content compared to the Co L II,III -edge XAS, suggesting that the electronic structure of the oxygen ions is more inuenced than Co ions by the Li + removal. In addition, the capacity fading upon cycling is explained with the change in the Co-O covalence that changes with the Li content as a result of the local structural distortion of CoO 6 octahedra that accompanies oxygen and cobalt orbitals re-hybridization. 13 Degradation has a signicant contribution to the performance decay of LiCoO 2 . Bulk structure degradation of LiCoO 2 is the main reason for capacity loss at high voltage. Jiang et al.
observed that by increasing the cutoff potential from 4.2 to 4.7 V, Li concentration was signicantly decreased in the bulk. 14 Besides, the samples with low Li content suffered from thermal instability when cycled to 4.7 V. The STEM-HAADF images, Fig. 4, show bright strips, where in the middle, the O1 phase is identied, and the density of these bright strips increases with increasing the cutoff voltage and the number of cycles. This reveals that there are uneven lithiation and de-lithiation behaviors in the bulk material. Besides structure degradation, interfacial degradation is one of the mechanisms assigned to the failure of layered cathode materials. It originates from the surface phase transformation layer and the cathode electrolyte interface (CEI) layer. The surface degradation is enhanced by increasing the cutoff potential. Fig. 5 shows that the thickness of the degradation layer aer 50 cycles is about four times higher in the samples cycled to a cutoff voltage of 4.7 V compared to that cycled to 4.2 V.
Tremendous efforts have been devoted to improving the functional performance of Li 1Àx CoO 2 at high voltage and elevated temperatures. Surface coating, doping with foreign  elements, and synthesis by wet chemistry methods such as solgel are traditional strategies that have been explored to improve the materials' properties. There are several methods for surface coating, including wet chemistry (sol-gel), chemical polymerization routes, and deposition techniques (sputtering, pulsed laser deposition, atomic layer deposition, chemical vapor deposition). 15 The properties of the coated material depend on the homogeneity and thickness of the coating layer. Insulator metals oxides, such as AlO 2 , ZrO 2 , and AlPO 4 , show outstanding improved performance in terms of capacity retention and rate capability at high voltage compared to the bare LiCoO 2 . The enhanced electrochemical performance is attributed to (1) the suppression of cobalt dissolution in the electrolyte, (2) the prevention of the direct contact between the cathode and the liquid electrolyte by the inactive coating layer, 16 and (3) cation diffusion to form Li-M-O (M ¼ Al, Co) on the surface during calcination that can suppress the phase transitions and hence decrease the capacity fading especially at high potentials. 17 Recently, Yano et al. 18 synthesized 0.5 wt% AlO 2 -coated LiCoO 2 by the sol-gel method. It shows electrochemical activity at potential > 4.2 V vs. Li/Li + and an initial capacity of 247 mA h g À1 at the rst cycle and preserved $82.6% of its initial capacity aer 20 cycles. The improved performance is attributed to the formation of a solid solution surface phase of LiAlO 2 and LiCoO 2 by the diffusion during the calcination aer the sol-gel coating. The coated LiCoO 2 shows very small phase transition upon cycling as Al 3+ is electrochemically inactive and shares the octahedral sites with Co ions in the MO 6 , resulting in a difficultly for both Li and O ions to leave the LiAlO 2 framework and thereby preserving the surface structure. . When x $ 0.5, the material decomposes into layered LiNiO 2 and spinel LiNi 2 O 4 (Fd3m) in the range between 180 and 250 C. In the composition range 0.5 > x $ 0.2, the material is completely transformed into spinel. Further increasing the temperature, the material turns into rock salt structure (Fm3m), with all transformations accompanied by oxygen evolution. It has been suggested that the exothermic reaction mechanism is driven by the phase transformations and oxygen evolution, which are originally due to the random mixing between Li and Ni ions. 33,34 Recently, Yoon et al. 27 successfully suppressed the H1 to H2 phase transition by preparing  a compact geometry of LiNiO 2 in which particles of 200 nm are packed tightly in a spherical particle of 10 microns. Such geometry minimized the contact between the electrolyte and the surface of the electrode resulting in better structure stability and capacity of 179 at 0.1 C rate, 95% of the capacity was retained aer 100 cycles when the cutoff voltage was 4.1 V. Similar to LiCoO 2 , better thermal, structural, and electrochemical stability is obtained by surface coating with ZrO 2 or SiO 2 , Co and (Mn, Co) as shown in Table 3.
LiCo 1Àx Ni x O 2 was suggested as an alternative to LiNiO 2 to merge the relatively better stability of LiCoO 2 and the high capacity of LiNiO 2 . The LiCo 1Àx Ni x O 2 solid solution can be formed in the whole x range. The lattice parameters and cell volume increase linearly while c/a decreases obeying Vegard's law when Ni ions are partially substituted for Co ions. 40 The higher the Co ions content in LiCo 1Àx Ni x O 2 , the more stable the structure is. Co substitution suppresses the non-stoichiometry of Li 1Àz Ni 1+z O 2 from z ¼ 0.04 to 0.01 when Co is increased from 0 to 0.2, and a pure stoichiometric 2D structure of LiCo 1Àx Ni x O 2 was observed when Co content reached 0.3. 41 The best performance of LiCo 1Àx Ni x O 2 was obtained for 0.1 $ x $ 0.3. 42 Differential scanning calorimetry for LiCo 1Àx Ni x O 2 proved relatively higher safety and thermal stability than LiNiO 2 at x ¼ 0.1 $ x $ 0.3 at cutoff voltages of 4.2 and 4.3 V vs. Li/Li + . 43 Upon increasing the Co content, the exothermic reaction shis toward higher temperatures, and its enthalpy decreases. In addition, when the cell is charged to higher potential, the exothermic event takes place at a lower temperature in agreement with the fact that the lower the Li + content, the lower the thermal stability of these oxide materials is. Recently, Xie et al. 44 doped Li(Co,Ni)O 2 with Mg ions to obtain highly structural, thermal, and electrochemical stable structure. Note that 80.1% of its initial capacity (214 mA h g À1 ) was maintained aer 500 cycles compared to 56.3% without Mg ions. The Mg 2+ prevented the Li + /Ni 2+ site mixing, suppressed the migration of Ni ions to the Li layers, and showed a pillar effect in the position of the Li ions, which improved the thermal stability signicantly even at such a high voltage as 4.7 V vs. Li + /Li. 45 Table 4 summarizes the thermal behavior of different transition metal oxides.

LiMnO 2
Manganese oxides are promising cathode materials due to their less toxicity and lower cost than cobalt and nickel oxides. Ion exchange is a successful method to synthesize layered LiMnO 2 by exchanging Na in NaMnO 2 with Li. 50 Truly layered LiMnO 2 with a-NaFeO 2 like structure is thermodynamically unstable because of the Jahn-Teller effect that distorts the structure to less symmetry orthorhombic (o-LiMnO 2 ) with the space group  Pmmn or monoclinic (m-LiMnO 2 ) with the space group C2/m. 51 Monoclinic LiMnO 2 is thermodynamically less stable than the spinel and orthorhombic lithium manganese oxides because of the antiferromagnetic interactions between Mn 3+ distorting the MnO 6 octahedra. 52 This results in a transition from the initial monoclinic to hexagonal at the rst charge then to spinel gradually upon cycling. 53 This transformation takes place through the migration of 25% of Mn ions from the octahedral sites in the transition metal layer to the lithium layer, and the Li ions move to tetrahedral sites. Stabilization of the layered structure can be obtained by partially substituting Mn ions by antiferromagnetic elements such as Al or Cr. 54 60 Li + prefers to occupy positions close to Mn 4+ ions than Ni 2+ in the transition metal layer. 58 Therefore, Li, Ni, and Mn ions are not randomly distributed in the transition metal layer. Single crystal selected area electron diffraction patterns of pristine LiNi 0.5 Mn 0.5 O 2 reveal the presence of superlattice reections indicating the long-range ordering of Li-rich and Li-decient sites in the transition metal layer. The superlattice reections are indexed in ffiffiffiffiffiffiffiffiffiffiffiffi 3a Hex: p Â ffiffiffiffiffiffiffiffiffiffiffiffi 3a Hex: p Â c Hex: and correspond to a trigonal structure with the space group P3 1 12. 61,62 Investigation of the local structure upon charging suggested the disappearance of Li + located in the transition metal layer very early during the rst charge. 63 Li + migrates to the tetrahedral sites that share faces with vacant octahedral sites in the transition metal layer. The presence of Li in the tetrahedral sites reects negatively on the electrochemical performance of the material. Because Li tet + is very energetically stable and requires high potential to be removed, which decreased the capacity obtained in the practical voltage range. 64 Regarding the Ni ions in the Li layer, by charging above 4.6 V, 75% of Ni Li ions migrate to vacant Li positions in the transition metal (TM) layer. 65 The migration of Ni ions to the TM layer is partly a reversible process and reects on the weaken or disappearance of the ffiffiffiffiffiffiffiffiffiffiffiffi  [70][71][72] This material has promoted many debates on the mechanism of the electrochemical behavior and Li + extraction. During the rst charge, the Li + extraction was found to be accompanied by oxygen evolution, leading to the removal of Li 2 O, resulting in electrochemical activation of Li 2 MnO 3 . 73 Robertson et al. suggested that there is a proton exchange with Li + to form Li 2Àx H x MnO 3 , 74 which can take place aer the production of hydrogen by electrolyte oxidation. The presence of H in the oxide layers forming O-H-O bond leads to a change to a more stable stacking sequence from O3(ABCABC) to P3(ABBCCA). Another proposed mechanism based on DFT calculations shows that the electrochemical behavior is attributed to the charge compensation by the oxidation of anion "oxygen". 75 However, the thermodynamically unstable localized holes at the beginning of the oxidation result in the evolution of oxygen. This study showed that the structural transformation to spinel is driven by the oxygen dimerization, which facilitates the Mn ions migration onto octahedral sites in the vacated lithium layers. Rana et al. suggested that both oxygen oxidation and proton exchange are responsible for the electrochemical behavior of Li 2 MnO 3 . 76 Oxygen and Li + are removed from the material, leaving layered Mn 2 O and the structure change from O3 to P3 due to the proton-Li + exchange. This structural change is reversible, and it turns to O3 during the following discharge, but there is an irreversible oxygen loss due to the material activation during the rst cycle. The repeated Li + -H + exchange, which leads to the change from O3 to P3 and the reverse upon cycling, was found to be responsible for the poor electrochemical performance of Li 2 MnO 3. The repeated shearing of oxygen layers results in permeant damage of the structure upon cycling. 4.7.2 xLi 2 MnO 3 $(1 À x)LiMO 2 . xLi 2 MnO 3 $(1 À x)LiMO 2 (M ¼ Ni, Co, Cr) are attractive materials due to their high-energy density and low cost. Li[Li 1/3À2x/3 Ni x Mn 2/3Àx/3 ]O 2 is a layered oxide composed of two phases: monoclinic xLi 2 MnO 3 and hexagonal (1 À x)LiMO 2 integrated into a single nanoparticle. 77,78 In the range 0 # x # 1/3, the Li and transition metals are ordered in the transition metal layer on a ffiffiffiffiffiffiffiffiffiffiffiffi 3a Hex: p Â ffiffiffiffiffiffiffiffiffiffiffiffi 3a Hex: p super structure. 79 It exhibits high voltage during charging (>4.5 V), high specic capacity (>250 mA h g À1 ), and high energy density. 70,80,81 The electrochemical behavior of Li-rich, layered oxides is divided into two stages. The rst is before 4.4 V when the Li + is extracted from LiMO 2 . The second starts at 4.5 V when the Li 2 O is extracted to activate Li 2 MnO 3 . The Li-rich, layered oxides suffer from some drawbacks such as an irreversible capacity loss due to the irreversible extraction of Li 2 O and the reaction with electrolyte at high potential, poor rate capability due to high charge transfer resistance and a voltage fade upon cycling. The origin of the voltage fade was suggested to be the transition from layered to spinel, which is accompanied by an irreversible transition metals migration to the Li-ions layers. [82][83][84][85] Despite the fact that the nanomaterials with the high surface area are known for their undesired side reactions with the electrolyte, reduction of particle size of Lirich materials to the nanoscale shows a better rate capability and a decrease in the diffusion path of Li ions. 86,87 Moreover, decreasing the interface area between the electrode and the electrolyte by controlling the morphology of the nanomaterials could suppress the erosion from the electrolyte and improve the cycling performance. Taking Li 1.2 Ni 0.2 Mn 0.6 O 2 as an example, hierarchical nanoplates structure shows outstanding rate capability as shown in Fig. 8, with specic capacity of $231, 216.5, 188, 163 and 142 mA h g À1 at as high rates as 1C, 2C, 5C, 10C, and 20C. 87 Furthermore, it shows high capacity retention of 95.5% and 86.6% aer 60 cycles for 1C and 2C rates, respectively. The classical voltage decay of Li-rich material was still observed due to the transition to spinel. An expected signicant decrease in capacity retention to 68% occurred when the hierarchical structure was damaged by ultrasonic treatment.

LiMn 2 O 4
Spinel LiMn 2 O 4 is one of the most attractive energy storage materials due to its low cost and reversible and fast intercalation and de-intercalation of Li ions. [89][90][91] The LiMn 2 O 4 spinel structure (space-group: Fd3m), consists of oxygen ions close-packed array occupy the 32e position, Mn ions occupy the octahedral site 16d, and Li + occupies the tetrahedral 8a site. In Li x Mn 2 O 4 (0 < x # 2), Mn 2 O 4 provides a 3-dimensional host network for Li + . When the voltage is $4 V, Li ions are hosted in the 8a site in the range 0 < x # 1. However, at 3 V, the inserted Li + ions occupy the octahedral 16c sites, which share faces with the 8a tetrahedral. The Li-ions at the 8a sites move to the empty 16c sites, resulting in a rst-order phase transition to Li 2 Mn 2 O 4 . In addition, when Mn 3+ concentration increases, Jahn-Teller distortion occurs to decrease the crystal symmetry from cubic to tetragonal Li 2 Mn 2 O 4 . Therefore, it is believed that the electrochemical performance of spinel LiMn 2 O 4 is better at 4 V than 3 V. However, it is still possible to observe the presence of Li 2 Mn 2 O 4 on the surface of LiMn 2 O 4 in a 4 V cell at the end of the discharge process as an over-discharge product. 92 Generally, LiMn 2 O 4 suffers from a capacity fading problem, especially at a temperature of 55 C due to the phase transitions discussed earlier as well as the chemical instability of the spinel LiMn 2 O 4 with the electrolyte, resulting in the dissolution of Mn ions. 94 The improvement of the electrode and electrolyte interface region by providing a protective coating on the material surface to decrease the direct contact with the electrolyte is one of the most extensively studied methods to avoid the previously mentioned problems. Electrochemically active materials coatings, such as LiNi 0.5 Mn 1.5 O 2 , LiCoO 2 , Li 4 Ti 5 O 12 , and metal oxide coating by B 2 O 3 , ZnO, or TiO 2 , show improved performance of the material at room temperature and 55 C. [95][96][97][98][99] Doping with Al, Co, Cr, or Ni improved the cyclability but at the expense of the Fig. 9 Polyhedral representation of the olivine structure. FeO 6 (light green), PO 4 (gray), and the Li + (orange) "by VESTA". The spinel type that exhibits Fd3m space group is a disordered phase that has a face centered cubic structure in which Li ions occupy the 8a sites, the transition metals Ni and Mn ions randomly occupy the 16d sites, and the ccp oxygen array occupies the 32e sites. The spinel type that exhibits P4 3 32 space group is an ordered phase that has a primitive cubic structure in which the Li ions occupy the 8c positions, Mn and Ni are regularly ordered in the 12d and 4a sites, respectively, while the oxygen ions occupy both the 24e and 8c positions. 104 The disordered structure exhibits better performance than the ordered spinel structure. 105 Despite its voltage is considered as one of the highest among other available materials, this voltage is outside the stability window of the electrolyte, which leads to the formation of SEI that affects the kinetics of the electrochemical insertion and de-insertion of Li ions. The stability of the SEI can be improved by surface modication via coating. 106 In addition, coating plays a benecial role in suppressing the dissolution of Mn ions from Li Ni 0.5 Mn 1.5 O 4 .

Phosphates-based cathode materials
Polyanionic olivine LiMPO 4 (M ¼ Fe, Mn, Co or Ni) are very attractive as cathode materials for LIBs. LiMPO 4 has an olivine structure, Fig. 9, and crystallizes in the orthorhombic Pnma space group in which the oxygen ions form a hexagonal closepacked framework, Li occupies the octahedral M1 site (4a) forming a tunnel along the direction [010], the transition metal (Fe, Ni, Mn, or Co) ion is located in the M2 site (4c) forming a zigzag chain of corner shared octahedra linked to the phosphate tetrahedra. The MO 6 octahedral and PO 4 tetrahedral share their edges. The strong P-O covalent bond provides high thermodynamic stability and prevents oxygen evolution at high temperatures compared to LiCoO 2 and LiNiO 2 . This strong PO 4 3À shows lower redox energy of 3d orbital of the metal than the Fermi level. Therefore, the P-O covalent bond is responsible for the lower potential in olivine than oxide materials. The cell parameters of LiMPO 4 (M ¼ Fe, Mn, Co, or Ni) are given in Table  5. 6.1.1 LiFePO 4 . LiFePO 4 is the most attractive commercialized cathode material because of its desirable safety features, high theoretical capacity (170 mA h g À1 ) at moderate current densities, stable Fe 3+ /Fe 4+ redox potential of 3.5 V vs. Li + /Li, at voltage plateau, thermal stability even at high temperatures, stable electrochemical and chemical properties, excellent cycling performance, low cost, and abundance of iron. During the insertion and removal of Li + , it was observed that there are two coexisting phases, LiFePO 4 and FePO 4 . Both phases have the same space group (Pnma). Padhi et al. suggested that the insertion and extraction of Li + is associated with the motion of the two-phase interface. 111 They suggested the shrinkage core model in which the Li ions move from the surface of a particle inward through the phase boundary. Upon lithiation, the surface area of the interface decreases until it reaches a critical surface area when the rate of Li insertion is not able to sustain the applied current. This makes LiFePO 4 a suitable material for low power applications as it inuences the performance of the material in terms of capacity and rate capability at high current density. Srinivasan et al. assumed a juxtaposition of the LiFePO 4 and FePO 4 phases and suggested a shrinking core model in which the shell of one phase covers the core of the other phase by considering that the Li-ion diffusion drives the phase boundary motion. 112 The battery cycling and calendar age inuence the cycle life, capacity, and power capability of LIB. The factors that inuence the performance can be a function of either the operation conditions, such as temperature or working potential window, showed that the high temperature to have a tremendous negative impact on the performance of the LiFePO 4 battery. 113 The battery lifetime at room temperature is about seven times higher than that at 55 C as presented in Fig. 10. The reason for capacity loss at room temperature is assigned to the irreversible loss of Li by the formation of a solid electrolyte interface (SEI) layer. Upon increasing the ambient temperature, the interplanar spacing between the crystallographic planes of the structure becomes larger, and the surface layer (SEL) becomes thicker, which is a sign of degradation. Also, the SEM images of the surface of the aged cathodes at temperatures higher than 45 C showed a rounded-shaped particle rich in uorine with the particle size increased from 0.7-0.8 mm to 1.3-1.5 mm with increasing the temperature, which is accompanied by the accelerated decomposition of the electrolyte.
There is a structural instability that usually occurs during the intercalation and de-intercalation process of Li, which results in fast capacity fading upon cycling. A study used in situ TEM and HRTEM analyses showed that the lithiation and de-lithiation in LiFePO 4 occur by two phase reactions (LiFePO 4 and FePO 4 ). 114 Upon applying potential to FePO 4 , Li ions inserted into the structure to form a layer of LiFePO 4 that gets thicker with time, which was evident via the migration of the step shape phase boundary along the b axis or [010] direction as shown in Fig. 11. This study proves that there is a lattice mismatch that induced dislocation at the phase boundary, where the transformation between the two phases produced an array of dislocations in the FePO 4 side. Also, the orientation of LiFePO 4 changed slightly This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 21662-21685 | 21673 compared to FePO 4 as a result of the elastic deformation that can accommodate the transformation strain. These dislocations accumulate during the cycling and can cause further defects/ cracks that cause degradation. Moreover, the ageing of C 6 / LiFePO 4 (LFP) batteries was studied theoretically by the electron-tunneling-based model for lithium immobilization. 115 The SEI formation model presumes the existence of a porous outer and dense inner SEI layers. Electron tunneling through the inner SEI layer is considered to be the rate-determined step in the SEI formation process and its initial thickness aer activation process controls the degradation rate. The outer SEI layer grows much faster than the inner layer. During cycling, cracks formation takes place, exposing the new surface of graphite to the electrolyte directly resulting in the formation of SEI, hence loss in capacity. The formed SEI was found to be dependent on the number of cycles. LiFePO 4 has a low intrinsic electronic conductivity of 10 À9 S cm À1 and low ionic conductivity. Several attempts were carried out to increase both its ionic and electronic conductivity and, consequently, its electrochemical performance. First, conductive surface coating by carbon was found to improve the electronic conductivity due to the improved electric contact between the electrode particles, as shown in Fig. 12a. 116 Several groups tailored the C/LiFePO 4 composite electrode and obtained specic capacity close to the theoretical limits of LiFePO 4 (170 mA h g À1 ). Furthermore, C-coating prevents particle growth during heat treatment resulting in the synthesis of smaller particle size and higher surface area LiFePO 4 . 117 Fig. 12b shows a comparison between the electrochemical prole of the bare and C-coated LiFePO 4 at C/12 rate; both have the same particle size (40 nm). The C-coated sample showed high capacity that is close to the theoretical limit, while bare LiFePO 4 showed only 55% of the theoretical capacity. Even at higher rate of C/4, C/LiFePO 4 shows excellent Li ions extraction kinetics and outstanding ability to maintain 100% of its initial capacity (160 mA h g À1 ) aer 120 cycles. However, the addition of additives such as C would increase the conductivity, but it decreases the energy density of the cathode because it is counted as electrochemical inactive material. Carbon-free LiFePO 4 small particles with narrow size distribution can shorten the ionic and electronic diffusion length and improve the conductivity of the material. 118,119 Delacourzt et al. prepared 100-200 nm LiFePO 4 , which showed a high reversible specic capacity of 145 mA h g À1 at C/2. Third, the electronic conductivity and Li ion diffusion in LiFePO 4 structure would be improved by doping at the anionic and/or cationic sites. Chung et al. reported that multivalent ions-doped LiFePO 4 showed outstanding electronic conductivity of eight orders of magnitude higher than pure LiFePO 4 . 120 However, the reason for the improved conductivity is questioned since Herle et al. 121 showed that the iron salt precursor in Chung's experiment was the source of C, which can reduce Fe and P to Fe 2 P and/or Fe 3 P (conductivity of LiFePO 4 + Fe 2 P is shown in Fig. 12a). Therefore, the metal-rich phosphides are believed to be responsible for the improved conductivity. A combination of C and Zn 4+ doping showed a synergetic effect that facilitated electron transport. 122 6.1.2 LiMnPO 4 . The low energy density of LiFePO 4 restricts their use in the automotive industry. The study and development of other members from the olivine family with higher discharge voltage and energy density, such as LiMnPO 4 , are critical. LiMnPO 4 showed higher potential than LiFePO 4 (4.1 V vs. Li 0 /Li + ) and a high energy density of $700 W h kg À1 . 124 However, cyclic stability and rate capability of LiMnPO 4 are hindered by the poor kinetics and low ionic and electronic conductivity. Furthermore, the presence of Mn 3+ ions in an unstable position in the lattice leads to cell distortion, which is controlled by the Jahn-Teller effect. 125 Similar to LiFePO 4 , several strategies were used to improve the conductivity of LiMnPO 4 , such as the synthesis of smaller particle size, carbon coating, and cation substitution by Fe, Mg, or Zn. [126][127][128] The solid solution of LiMnPO 4 and LiFePO 4 operates in the voltage range between 3.4 and 4 V vs. (Li/Li + ), which is not very high to prevent  spray pyrolysis and suggested that the performance was improved because the conductivity increased, and Mn dissolution decreased from 103.7 to 87.2 ppm. Seo et al. 131 studied the charge transfer resistance (R ct ) and the electrochemical performance of LiFe x Mn 1Àx PO 4 at x ¼ 0, 0.2, 0.4, 0.6, 0.8. The best electrochemical performance was observed when x ¼ 0.8 and 0.6, which corresponded to the highest discharge capacity of almost 154 and 149 mA h g À1 and retained 100% and 101% of its initial capacity aer 50 cycles, respectively. The impedance patterns of all compositions showed a single semicircle in the high frequency region as shown in Fig. 13. The semicircle at high frequencies is due to the combination between double layer capacitance and R ct . The R ct decreased by increasing the Fe content and reached its lowest value of 170 U when x ¼ 0.8. The Warburg impedance at low frequency is related to the diffusion of Li ions. The Li ions diffusion increased with decreasing the slope of the line and reached its highest value of 7.33 Â 10 À14 when x ¼ 0.8. Dai et al. reported that the co-substitution of Mn by Fe and Mg could signicantly improve the cycling performance, where the optimal composition was found to be LiMn 0.8 Fe 0.19 Mg 0.01 PO 4 . 132 Moreover, they studied the effect of C content on the electrochemical performance of this composition. The material with 10.5 wt% carbon showed the best rate capability and capacity of 143.1 mA h g À1 , 130 mA h g À1 , and 125.7 mA h g À1 at 1C, 5C, and 10C rates, respectively. Lower C content of 4.1 wt% showed a slight decrease in the rate capability. However, further decrease or increase in the C content to 2 wt% or 27.4 wt% resulted in fast capacity fade at high C rate. 133 This better performance is explained by the lowest charge transfer resistance of $27 U at 10 wt% C electrodes.  134 It is challenging to extract lithium from LiNiPO 4 within the stable voltage window of the available electrolytes. Moreover, both compounds showed much lower electronic conductivity than LFP and LMP. The cyclic voltammetry of LiCoPO 4 showed that there is a reduction peak at 4.6 V and an oxidation peak at 5.1 V, while there are no signicant reduction peaks for LiNiPO 4 in the range between 3.5 to 6 V vs. Li/Li + as shown in Fig. 14. 135

Silicate-based cathode materials
The replacement of phosphate by silicate was a crucial step in the search for new high energy cathode materials. Orthosilicate cathode materials (Li 2 MSiO 4 , M ¼ Fe, Mn) are of particular interest due to their superior electrochemical performance, safety, raw material abundance, and low cost. Li 2 MSiO 4 has a high theoretical specic capacity ($333 mA h g À1 ) if the two Li ions are fully extracted per formula unit. Several temperatures and pressure-dependent polymorphs of Li 2 MSiO 4 exist. 136 The framework consists of distorted hexagonal close backed oxygen array, and all cations occupy tetrahedral sites. The different possible cations ordering produce different polymorphs. These polymorphs can be divided into low or high temperature, which corresponds to b or g Li 3 PO 4 like structures, respectively. Each type could be explained in several space groups such as orthorhombic b I (Pbn2 1 , Pna2 1 ) and b II (Pmn2 1 ), monoclinic g s /g 0 (P2 1 /n), and orthorhombic g II (Pmna, Pmnb).
6.2.1 Li 2 FeSiO 4 . Li 2 FeSiO 4 exhibits a specic capacity of 166 mA h g À1 since only one Li + can be extracted due to the redox couple Fe 3+ /Fe 2+ at 3.1 V vs. Li + /Li. The lower electronegativity of Si compared to P results in a lowering of the Fe 3+ /Fe 2+ redox couple in Li 2 FeSiO 4 than in LiFePO 4 . The other redox couple Fe 4+ /Fe 3+ takes place at a high voltage of 4.85, which is outside the stability window of the electrolyte; therefore, it is not reached. Li 2 FeSiO 4 exhibits complex structural defects that make it difficult to analyze its structure. Nytén et al. 137 reported Li 2 FeSiO 4 in the orthorhombic system with Pmn2 1 space group and explained two additional peaks in its diffractogram as an unidentiable impurity. Its potential shis irreversibly from 3.1 V to 2.8 V in the second cycle. This voltage drop was  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 21662-21685 | 21677 suggested to be due to site mixing between Li ions and Fe ions since the Li : Fe ratio in the 4b site changed from 96 : 4 to 40 : 60 during the rst cycle. 138 Nishimura et al. 139 reported Li 2 FeSiO 4 in the monoclinic system and P2 1 space group. Sirisopanaporn et al. 140 synthesized a polymorph, which is an isostructural with Li 2 CdSiO 4 and is described by the Pmnb space group. Sirisopanaporn et al. 141 compared the electrochemical behavior of Li 2 FeSiO 4 polymorphs, which were explained in the Pmn2 1 , P2 1 /n, and Pmnb space groups by synthesis under hydrothermal conditions at 200, 700, and 900 C, respectively. The structural difference of these polymorphs is shown in Fig. 15a. Each polymorph showed different rst oxidation potential. The Pmnb polymorph exhibited the lowest potential (2.9 V vs. Li + /Li 0 ), P2 1 /n showed an intermediate potential of 2.97 vs. Li + /Li 0 , while Pmn2 1 showed the highest oxidation potential (3.1 V vs. Li + /Li 0 ). Derivative plots obtained from PITT of the three polymorphs showed that they all experience the voltage fade from 3 to 2.8 V upon cycling due to the transformation to Li-poor phase as shown in Fig. 15b-d. More precisely, it showed that the high temperature Pmnb and P2 1 /n, which have high degree of disorder, transformed faster than the low temperature Pmn2 1 . The rst discharge reduction potential of Pmnb and P2 1 /n was $2.76 V vs. Li + /Li 0 , while the discharge of Pmn2 1 showed two reduction peaks at $2.76 and $3.04 V vs. Li + / Li 0 indicating the presence of initial Li 2 FeSiO 4 . The hightemperature Pmnb and P2 1 /n experienced a phase transformation at the rst cycle faster than Pmn2 1 , which is fully transformed at the h discharge.
6.2.2 Li 2 MnSiO 4 . Li 2 MnSiO 4 is a more interesting material than Li 2 FeSiO 4 as it offers higher redox potential and specic capacity. Li 2 MnSiO 4 possesses a theoretical capacity of 333 mA h g À1 because of the contribution of two Li ions to the electrochemical reaction. This is because Mn ions can be reversibly oxidized and reduced from the +2 state to +4 state with the redox couples of Mn 2+ /Mn 3+ and Mn 3+ /Mn 4+ at 4.1 and 4.5 V vs. Li + /Li, respectively. At 0 K, the low-temperature b-Li 3 PO 4 is more thermodynamically stable than the hightemperature g-Li 3 PO 4 . The transformation from b to g structures is exothermic and can take place at higher temperatures. Also, the g to b transformation can happen under high-pressure conditions. 142 The separation of polymorphs is possible by applying the preferred conditions of each structure. P2 1 /n is obtained by synthesis at high temperatures, followed by fast quenching. The synthesis of Pmn2 1 polymorph is possible by controlling both the pressure and temperature at 2-8 GPa and 600-900 C, respectively. 142 Among these polymorphs, the lowtemperature orthorhombic Pmn2 1 is the best polymorph as it shows favorable thermodynamic stability and exhibits the lowest Li diffusion energy barrier. Li 2 MnSiO 4 suffers from poor performance and fast capacity fading problems upon cycling, as shown in Fig. 16. This is because of the low ionic and electronic conductivity ($5 Â 10 À16 S cm À1 at RT) 143 and the structural instability due to the presence of Mn ions in the tetrahedral sites. Several strategies were proposed to improve its structural stability upon cycling. The synthesis of new polymorphs in which the Mn ions are sitting in stable octahedral sites was tested under high pressure up to 15 GPa. However, it was not obtained because of the strong cationic repulsion in the dense structures. 144,145 The addition of conductive carbon, such as acetylene black, showed signicantly improved performance. 146 Nano C/Li 2 MnSiO 4 prepared by ionothermal method showed improved performance and capacity retention even at high C rates as shown in Fig. 17. Cation substitution by Mg, Ca, Cr, Ti, V, Ni, Na, Cu or P improved the intrinsic conductivity and increased the Li ion diffusion coefficient. [147][148][149][150][151] 7 Go green with Li-ion batteries Following the growth of the Li-ion batteries industry, the environmental impact of such a technology is critical and must be carefully considered. Cathode materials are the key components in the battery in terms of cost and active weight. Their toxic byproducts during the manufacturing process as well as inadequate disposal ways, require urgent green approaches to avoid their severe health and environmental problems. One of the proposed methods in literature is to switch from conventional to green synthesis routs. To avoid environmental pollution and avert additional costs associated with processing the harmful byproducts that contain anions such as Cl À , SO 4 2À , NO 3 À , and gases like N x O y , CO, and NH 3 . 154 The electrochemical performance is one of the main factors that direct the industry, Fig. 19 Comparison of two materials, one was sintered for 12 h at 750 C and the other was prepared by hydrothermal method then sintered for 12 h at 750 C. 159 Fig. 18 The waste management hierarchy and range of recycling options. 158 therefore, considering it side by side with the environmental issues is logical and crucial. As an example, Bolloju et al. obtained 165 mA g À1 at 0.1C from LiFePO 4 prepared using Fe powder as a precursor, which is comparable to that obtained by conventional chemical synthesis. 155 Fe powder is very advantageous over other precursors because it does not produce harmful anions. Also, there is a 100% atom economy delivered by the iron metal, which makes it cost-effective as well. Liu et al. prepared multi-doped LiFePO 4 /C with Mn, Co, and Ni using Fe powder and spent electroless nickel plating solution. 156 The product shows comparable performance to that obtained by the expensive ferrous oxalate or ferric oxide as well as to that prepared by conventional chemical routs. Waste management is a process that considers the protection of the environment with various strategies in a hierarchy order (disposal, recovery, recycling, reuse, and prevention), as illustrated in Fig. 18. Accordingly, reusing a battery is cleaner than recycling because recycling needs to spend energy on transportation and processing. Thus, it is possible, for instance, to collect the batteries that retain 85% to 80% of their initial capacity from electric vehicles to reuse in stationary grid applications. 157 Recycling can still help in decreasing pollution and satisfy the need for raw materials. Recycling relies rst on the recovery of different streams of raw materials followed by a pyrometallurgical or hydrometallurgical process to extract metals (Li, Ni, Mn, Al, Cu or Co). Pyrometallurgical is simple; the battery is smelted in a high-temperature furnace to reduce the active materials into metallic alloy form and also produce slag and gases. 158 This method requires lters to trap the emitted toxic gases.
On the other hand, the hydrometallurgy process involves the leaching of the metal from the cathode by the use of aqueous solutions, such as a mixture of acid and hydrogen peroxide. Each metal can be obtained by extraction, precipitation, or electrodeposition. Shi et al. combined hydrothermal and annealing method to efficiently obtain LiCoO 2 with controlled microstructure and composition and found that its rate capability is higher than that obtained by conventional solid-state synthesis (Fig. 19). 159 Transition metals oxides and hydroxides were collected from commercial spent Li-ion batteries and used as a potential material for supercapacitors. 160,161 The battery cathode tapes were recycled and used as a catalyst for the degradation of the toxic methylene blue dye. 162

Summary and outlook
The use of different cathode materials, such as layered oxides, olivines, and spinels, for Li-ion batteries is summarized and discussed. The classication of these cathodes materials is based on the Li ion diffusion pathway in different structures. The principle challenge for Li-ion batteries is the development of functional materials that can offer higher energy, power, and lifetime than the currently existing materials. In order to achieve this goal, huge efforts were explored, including the development of Li-rich oxides or mixed transition metal layered oxides. Moreover, the origin of the factors affecting the battery lifetime, such as degradation of the materials upon cycling and aging, are of great importance as outlined in this review. Improving the performance of existing materials is possible through different strategies, including the creation of defects in the material, manipulation of nanostructured complexes, substitution with different transition metals, doping with foreign elements, and smart coatings. Furthermore, continued progress to safer batteries, clean green materials, and lower cost is required. To this end, modelling is a great tool to identify new cathode materials that can satisfy the balance between the electrochemical performance and environmental requirements. Moreover, recycling Li-ion batteries and recovery of their components to construct new devices should be adapted and widely investigated.

Conflicts of interest
There are no conicts to declare.