Shuo
Sun†
a,
Jin
Li†
a,
Cuixia
Xu
b,
Teng
Zhai
*a and
Hui
Xia
*a
aHerbert Gleiter Institute of Nanoscience, School of Materials Science and Engineering, Nanjing University of Science and Technology, Nanjing, Jiangsu Province, China. E-mail: tengzhai@njust.edu.cn; xiahui@njust.edu.cn
bJiangsu Key Laboratory for Biofunctional Molecules, College of Life Science and Chemistry, Jiangsu Second Normal University, Nanjing 210013, P. R. China
First published on 20th May 2022
The ever-increasing demand for high-energy-density electrochemical energy storage has been driving research on the electrochemical degradation mechanisms of high-energy cathodes, among which manganese-based layered oxide (MLO) cathodes have attracted high attention thanks to their low cost and eco-friendliness. More importantly, MLO materials with large and tunable interlayer spacing are ideal candidates for the insertion of (monovalent, divalent, trivalent) alkaline ions, such as Li+, Na+, K+, Zn2+, Mg2+, and Al3+, enabling impressive electrochemical performance. Nevertheless, the local MnO6 octahedron distortion induced by the Jahn–Teller (J–T) effect can lead to irreversible phase transformation, dissolution/disproportionation reactions, interfacial degradation arising from Mn2+, and crack formation, which significantly impact the electrochemical stability of MLO materials. Hence, in this review, we discuss the various degradation processes caused by J–T distortion in MLO cathodes at the atomic level. Advances in the atomic-level structure and property optimizations of MLO materials and in-depth structure–function–property correlations are also systematically reviewed. Finally, we provide our perspectives on the future development of MLO materials. The integration of high-performance MLO cathodes in energy storage devices has great potential to address growing global energy demands.
Among the various cathode candidates, transition metal (TM) layered oxides provide many advantages thanks to their high capacities, simple structures, and ease of synthesis.11–13 Although different TMs, including Ni, Co, and Mn, have been utilized in layered cathode materials, only Mn-based layered cathode materials realize the ideal balance between sustainability and performance (Fig. 1a). Mn has the highest crust abundance among Ni, Co, and Mn (84 ppm, 25 ppm, and 950 ppm, respectively)14 and thus the lowest costs. Additionally, elemental Mn is nontoxic and eco-friendly, and the risks of metal pollution with Mn use are lower than those of Ni and Co.8,15,16 In terms of performance, owing to the high redox activity of Mn2+/Mn3+/Mn4+ and the corresponding high redox potential, as well as the relatively low atomic weight, Mn-based layered oxides (MLOs) deliver much higher energy densities than other cathode materials.4,14 Meanwhile, unlike Ni-based layered oxides, MLOs possess high thermal stability and are thus safer.17 Furthermore, the large and tunable interlayer spacing of MLO materials guarantees fast ionic diffusion kinetics, providing an ideal structure for (monovalent, divalent, trivalent) alkaline ions (e.g., Li+, Na+, K+, Zn2+, Mg2+, Ca2+, Al3+) insertion, representing a star family as cathode materials in various electrochemical energy storage devices for a long period (Fig. 1b).18–20 Apart from the commonly used metal anodes, such as Li, Na, and K, metal anode materials like Zn, Mg, and Al, which boast the merits of earth abundance, safety, good machinability, and ultrahigh volumetric energy density, are also attractive candidates for energy storage. Therefore, the development of MLO cathodes is beneficial to develop advanced energy storage devices and satisfy the future demands of large-scale energy storage.
The typical MLO materials are built by alternately layering TM–oxygen and alkaline metal ions. The TM is coordinated in an octahedral geometry by oxygen, and the alkaline metal ions are inserted in prismatic (P), tetrahedral (T), and octahedral (O) configurations, corresponding to the P2/P3, T2, and O2/O3 phases, respectively (Fig. 2a), in which the numbers 2 and 3 represent the repeat period of the TM–O layers.21,22 In addition, a special kind of MLO material, called birnessite, comprises a mixed-valent oxide surrounded by six oxygen atoms to constitute a basic unit [MnO6] octahedron with cations of various valences (e.g., K+, Na+, Zn2+) and crystal water filling the interlayers.23,24 Birnessite, with 2D interlayer space for ionic diffusion, is a hydrous Mn oxide. Layered oxides were first applied in energy storage by J. B. Goodenough in 1980;25 substantial efforts have since advanced the development of high-performance MLO materials for various energy storage purposes (Fig. 2b).26–30 Although remarkable progress has been made, the applications of MLO materials are still limited by irreversible phase transition, Mn dissolution/disproportionation reactions, interfacial degradation arising from Mn2+, and crack formation (Fig. 3), which lead to rapid electrochemical degradation. It is worth noting that the root of these detrimental issues is local MnO6 octahedron distortion caused by the Jahn–Teller (J–T) effect. The J–T theorem was proposed by H.A. Jahn and E. Teller in 1937;32 due to this effect, a nonlinear molecular system suffers from instability in the degenerate electron state. Mn has several oxidation states (+2, +3, +4, +7), and electrochemical redox reactions usually happen between the +4 and +2 oxidation states. The ligand field of Mn4+ is shown at the center of Fig. 3. The Mn ions are coordinated by six oxide ligands and are symmetrically arranged along the Cartesian axes. The electric field of the octahedrally arranged ligands separates the d orbitals of Mn into doubly degenerate eg (dz2 and dx2−y2) and triply degenerate t2g (dxy, dxz, and dyz) sets. The eg orbital set of Mn4+ is empty, thus possessing a cubically symmetrical electron density. This feature confers stability to Mn4+-derived layered materials. However, when Mn3+ is coordinated with oxygen, like in LiMnO2, the 3d4 configuration of Mn3+ exhibits an asymmetrical electron density distribution and a high-spin state due to the single-electron occupancy in the eg level.33–35 In this case, the two axial Mn–O bonds are elongated and the four equatorial bonds are shortened to relax this unfavorable state, i.e., the J–T distortion, transforming MnO6 from the cubic configuration into the tetragonal configuration. As such, this structural evolution at the atomic level detrimentally impacts the physicochemical properties of MLO materials.
Fig. 2 (a) Schematic illustrations of the crystal structures of the birnessite phase, reproduced with permission.31 Copyright 2015 American Chemical Society. Schematic illustrations of T, P, and O phases. Reproduced with permission.22 Copyright 2021 Wiley-VCH. (b) Main progress of MLOs in various energy storage systems, including Li, Na, K, Zn, Al, etc. batteries, as well as supercapacitors. |
In attempts to prevent J–T distortion, various strategies have been developed to optimize structural stability at the atomic level during the charging/discharging process.7,34,36–40 These strategies can be divided into four categories (Fig. 3): (i) elemental doping, (ii) interlayered structure manipulation, (iii) vacancy introduction, and (iv) novel structural design. Although previous reviews have discussed Mn-based and/or layered cathode materials, the relationships between the structural evolution of MLO materials at the atomic level and capacity fade during the ion deintercalation/intercalation processes have not been systematically reviewed.12,41–43 In particular, a comprehensive review of the atomic-level mechanisms and strategies that remedy the electrochemical degradation of MLO cathodes in various valence ion-based electrochemical energy storage devices is needed.
In this review, we discuss the relationships between the intrinsic structure, electrochemical behaviors, and degradation mechanisms from the perspective of the atomic level and summarize the effective strategies for structural degradation during ion deintercalation/intercalation. We begin by discussing the different electrochemical behaviors resulting from alkaline ion insertion into MLO cathodes, from monovalent ions to trivalent ions. Moreover, we describe the corresponding electrochemical degradation caused by irreversible phase transition, Mn dissolution/disproportionation reactions, interfacial degradation arising from Mn2+, and crack formation in various energy storage devices, as well as the relationships between these degradation mechanisms and J–T distortion. Finally, we outline the recent achievements in overcoming the aforementioned bottlenecks and provide perspectives on the future development of high-performance MLO materials.
Energy storage type | Material | Crystal structure | Interlayer space (Å) | Electrolyte type | Rate performance | Ref. |
---|---|---|---|---|---|---|
Supercapacitor | δ-MnO2 | Birnessite | 7.2 | Aqueous | 250 F g−1 at 0.2 A g−1 | 19 |
125 F g−1 at 10.0 A g−1 | ||||||
Li-ion battery | LiMnO2 | O3 | 5.4 | Organic | 286 mA h g−1 at 1C | 44 |
183 mA h g−1 at 12C | ||||||
Li-ion battery | Li2/3[Ni1/3Mn2/3]O2 | T2 | 5.0 | Organic | 180 mA h g−1 at 0.1C | 45 |
Li-ion battery | Li2MnO3 | O2 | 4.7 | Organic | 400 mA h g−1 at 0.01 A g−1 | 46 |
Na-ion battery | NaMnO2−y −δ(OH)2y | Birnessite | 7.2 | Organic | 211 mA h g−1 at 0.2C | 47 |
156 mA h g−1 at 50C | ||||||
Na-ion battery | NaxMnO2 | O3 (x = 0.9–1.0) | 5.7 | Organic | 185 mA h g−1 at 0.1C | 48 |
P2 (x = 0.6–0.8) | 194 mA h g−1 at C/30 | |||||
K-ion battery | K0·77MnO2 | Birnessite | 7.2 | Organic | 77 mA h g−1 at 1 A g−1 | 24 |
Zn-ion battery | Zn–MnO2 | Birnessite | 9.2 | Aqueous | 242 mA h g−1 at 3 A g−1 | 49 |
Ca-ion battery | CaMnO2 | Birnessite | 6.9 | Aqueous | 89 mA h g−1 at 2 A g−1 | 50 |
Mg-ion battery | MgMnO2 | Birnessite | 6.7 | Aqueous | 55 mA h g−1 at 8 A g−1 | 51 |
Mg-ion battery | MnO2 | Birnessite | 9.7 | Aqueous | 110 mA h g−1 at 1C | 52 |
50 mA h g−1 at 10C | ||||||
Al-ion battery | MnO2 | Birnessite | 7.0 | Aqueous | 59 mA h g−1 at 0.1 A g−1 | 53 |
Fig. 4 (a) The typical CV of δ-MnO2 showing nearly capacitive behavior at 10 mV s−1 and simulation of H2O and K+ intercalation into δ-MnO2. Reproduced with permission.19 Copyright 2021 Springer Nature. (b) CV curve of the spinel–layered LiMnO2 cathode at 0.1 mV s−1. Reproduced with permission.44 Copyright 2021 Springer Nature. (c) Discharge curves of NaMnO2 at a rate of 10 mA g−1. The 1st, 2nd, 5th, and 10th cycles are represented by the curves with black, red, blue, and green, respectively. Reproduced with permission.55 Copyright 2014 American Chemical Society. (d) Local atomic coordination around O with two Li–O–M and one Li–O–Li configurations in Li-excess layered oxides. Reproduced with permission.56 Copyright 2016 Springer Nature. (e) The metastable phase diagram suggests that the Mn would be oxidized to the +7 state in the form of Mn2O7 and LiMnO4 when the formation kinetics of oxygen–oxygen bonds is sluggish. (f) Comparison between the voltage curves for the Li2MnO3 component of composites and Li2MnO3. Reproduced with permission.57 Copyright 2019 Springer Nature. (g) Charge/discharge curves at different current densities in ZIBs. (h) Ragone plot of ZIB-based layered MnO2. Reproduced with permission.58 Copyright 2019 Royal Society of Chemistry. (i) Discharge curves at the current density of 2 mA cm−2 in the Al/MnO2 battery. Reproduced with permission.28 Copyright 2022 Elsevier. |
As for the P/O/T phases, such as layered LixMnO2 and NaxMnO2, both consist of edge-sharing MnO6 octahedra separated by an interlayer of Li+ or Na+. Generally, most AMO2 (A = Li, Na; M = transition metal) phases have α-NaFeO2-type structures.63,64 However, when M is Mn3+, the crystal structure is monoclinic (space group C2/m), not orthorhombic, due to the strong J–T effect of high-spin Mn3+, which distorts the local site symmetry near Mn3+ from a regular octahedron and changes the corresponding crystal structure.65–67 The practical use of layered LiMnO2 in batteries is greatly hindered by severe structural degradation and rapid capacity fading. The small difference in size between Mn3+ and Li+ makes it difficult to form the layered LiMnO2 structure; Mn migration into the Li layer is often observed.68,69 Thus, an irreversible transition from the layered structure to the spinel structure usually occurs during the first deintercalation. Specifically, the formation of Li vacancies during oxidation results in covalent MnO2 slab formation and an empty interlayer space with face-to-face O layers. Correspondingly, the Mn ions tend to migrate to the Li layers to homogenize the charge distribution.70 To realize this transfer process, the Mn ions must cross a tetrahedral interstitial site. Low-spin Co4+ (d5), low-spin Co3+ (d6), low-spin Ni4+ (d6), and low-spin Ni3+ (d6) have very strong stability in octahedral sites. However, high-spin Mn3+ ions are prone to disproportionate into Mn2+ and Mn4+.71,72 Mn2+ (d5) is much less stable in octahedral sites than in tetrahedral sites, promoting movement and thus fast capacity fade. As such, the crystal field stability determines whether the TM ions transfer.73 Generally, stable redox reactions of monoclinic layered LiMnO2 can be observed at ∼3.9 V in the cathodic scan and 3.8 V in the anodic scan (Fig. 4b).44,74,75 However, the sizes of Na+ and Mn ions differ, allowing NaMnO2 to form the O3-type layered structure. Meanwhile, there is a great increase in the covalency of the Mn–O bond due to the Na–O bond, which has a much stronger ionic character than the Li–O bond, improving the stability of the tetravalent ions.70,76,77 First-principles calculations suggest that monoclinic NaMnO2 is stable, whereas LiMnO2 prefers an orthorhombic structure.78,79 Nevertheless, with the decrease in Na content in NaxMnO2, the phase changes from O3-type (x = 0.9–1.0) to P2-type (x = 0.6–0.8) and P3-type (x = 0.5–0.6) oxygen packing.80 As the P3 and O3 structures share the same MnO6 octahedral orientation, the layer of O3-type NaMnO2 glides during the deintercalation of Na+ to form P3-type NaxMnO2. This process may be reversible; however, the stress/strain induced by phase transition significantly limits long-term cycling. In addition, NaxMnO2 materials exhibit different electrochemical behaviors; many voltage plateaus are observed during the deintercalation/intercalation process due to the ordering of Na+/vacancy (Fig. 4c).55,81
Li-rich MLO materials like O3-type Li2MnO3 have attracted extensive attention due to their large capacities, which significantly exceed those of conventional Li-containing cathode materials.82,83 As such, substantial efforts have focused on utilizing the Li+ in Li2MnO3 with 100% efficiency. The structure of Li2MnO3 is usually expressed as Li[Li1/3Mn2/3]O2, consisting of Li layers between MO2 (M = Li1/3Mn2/3) layers, in which one third of the Mn is replaced by Li in a specific honeycomb-like occupation pattern.84,85 This feature results in excess Li. It is worth noting that Li-rich Mn-based oxide (LRMO) cathode materials are usually synthesized by combining xLi2MnO3 and (1−x)LiTMO2 (TM = Ni, Co, Mn), when x = 0.5, in the form of Li1.2Mn0.54Co0.13Ni0.13O2.86,87 However, in addition to the J–T distortion in LiTMO2, the anionic redox reactions triggered by Li2MnO3 activation accelerate the migration of Mn ions. These behaviors imply unusual structural degradation,91–93 including large voltage hysteresis, a gradual decrease in the average voltage during cycling (voltage fade), and an “activation” plateau (4.5 V vs. Li/Li+) during the first charge.85,94 In the first step of the delithiation process, TM cations are oxidized. At a high voltage between 4.5 and 4.7 V, there is a voltage plateau, and most Li+ is deintercalated from the material. Meanwhile, in the bulk region, some oxygen anions are oxidized to O(2−x)−. The anionic redox is due to the existence of the Li–O–Li configuration, in which unhybridized O 2p states (Li–O–Li states) are generated.56,95,96 Because the energy of electrons in the Li–O–Li configuration is higher than in the other O 2p states (Fig. 4d), the oxygen in the Li–O–Li configuration is preferentially oxidized at a relatively low voltage of ∼4.3 V. It was recently proposed that the Mn can be oxidized from +4 to +7 through three-phase reactions, forming the Mn7+ compounds Mn2O7 and LiMnO4 and the Mn4+ compounds Li4Mn5O12 and MnO2.57 The phase diagrams in Fig. 4e facilitate the understanding of the electrochemical reaction mechanisms of Li-rich MLO materials. However, the detected voltage curves do not follow these three-phase reactions (Fig. 4f), which is mainly attributed to the slow kinetics of these reactions resulting from the long-range diffusion of Mn or O. Compared with Li, the relatively slow migration of Mn leads to the formation of a metastable Li2−xMnO3 compound. The oxygen-redox hypothesis also demonstrates the formation of this metastable phase where the oxygen anion valence states are −1.56,97
Zn-ion batteries possess high gravimetric and volumetric capacity (820 mA h g−1 and 5854 mA h cm−3). In addition, the excellent reversibility of Zn plating/stripping reactions, safety, and low toxicity make Zn-ion batteries a promising energy storage device. The insertion of Zn2+ into birnessite forms the layered δ-ZnxMnO2. Alfaruqi et al.27 demonstrated the high capacity of birnessite-based Zn-ion batteries (>250 mA h g−1). Despite the high capacity of layered δ-ZnxMnO2 compared with tunnel MnO2, δ-ZnxMnO2 still has low capacity retention during electrochemical processes. Recent research demonstrated that the capacity decay is due to the phase transition from the layered structure to the spinel structure and the dissolution of Mn ions in electrolytes.98 Wang et al.58 prepared a mesoporous layered MnO2 nanosphere with H2O and Zn2+ in the interlayers (Fig. 4g and h), which had a high specific capacity (358 mA h g−1 at 0.3 A g−1). In addition, in situ and ex situ characterizations have been used to uncover the structural changes of ZnxMnO2 during charge/discharge cycling. Energy storage mechanism that involves the co-insertion/extraction of H+ and Zn2+ along with the deposition/dissolution of zinc sulfate hydroxide hydrate on the electrode surface was discovered. In general, the specific structural degradation between Zn-batteries in aqueous solution and Li/Na-batteries in the organic solution is different due to their different electrolytes and inserted ions. Compared with the inserted Li+ or Na+ with monovalence, divalent Zn2+ has much higher electrostatic repulsion and atomic radius, which inevitably causes larger volume change of MLO materials during the electrochemical processes and thus results in more severe layered structure distortion than the insertion of Li+/Na+ into MLO materials.99,100 The electrolytes used in Zn batteries are always acidic, in which higher H+ concentration exists. As a result, MLO materials suffer from the more serious dissolution issues of Mn2+ originating from the disproportionation reaction than the organic electrolyte that has low H+ concentration used in Li/Na-batteries. In addition, Alfaruqi et al. proposed that layered MnO2 in the process of Zn-ion insertion undergoes a series of structural transformations: layered MnO2–layered ZnxMnO2–Zn-inserted phase–spinel ZnMn2O4.98 Although the phase transition during the discharging can be reverted to the layered MnO2 during the initial charging, these phase transformations may induce the huge change of layered structure and thus easily results in the structural degradation during the long-term cycles, which is demonstrated by the complete phase transformation from the layered structure into the spinel structure after prolonged cycles. Additionally, the co-intercalation mechanism of H+ and Zn2+ has been demonstrated.101 Unfortunately, the effect of co-intercalation of H+ and Zn2+ on structural degradation is not clear until now. But it must cause different degradation mechanisms compared with the single Li+/Na+ insertion in Li/Na-batteries.
Al-based batteries, which have three-electron redox states (from Al3+ to Al) and are low cost and easy to handle, provide an alternative to commercial Li/Na-ion batteries. Birnessite, with large interlayer spacing, enables the reversible intercalation and extraction of Al3+, which can deliver high capacity. However, the discharge rate is relatively low in this layered structure due to the sluggish diffusion kinetics of Al3+ in the interlayers. The strong electrostatic interaction between Al3+ and oxygen in layered birnessite can result in the formation of inactive AlxMnO2 and, subsequently, fast capacity decay. Therefore, it is crucial to prevent the direct insertion of Al3+ ions into the lattice. To address these issues, electrolyte engineering has been proposed to tune the solvation structure of Al3+. For example, Xu et al.28 developed an Al-based battery with high stability by tuning the electrolyte. When 1 M MnCl2 and 4.4 M AlCl3 were used, the intercalation of Al3+ was prohibited because of the large solvation structure of AlCl1.8(H2O)4.21.2+ that formed in this electrolyte.28 Correspondingly, the Al–MnO2 batteries delivered a large specific capacity of 493 mA h g−1 and stable cycling behaviors over 1000 cycles, demonstrating the potential of trivalent ion-based energy storage (Fig. 4i).
Fig. 5 (a) Schematic illustration of the phase transformation process of the layer into tunnel through layer buckling. Reproduced with permission.88 Copyright 2011 American Chemical Society. (b) Schematic illustration of the Li/Mn dumbbell configuration. Reproduced with permission.48 Copyright 2011. The Electrochemical Society. (c) Phase transition from the layered structure to the spinel of cycled layered LNMO after 100 cycles. Reproduced with permission.89 Copyright 2013 American Chemical Society. (d) Proposed Li1/2MnO3 structure showing the Li2MnO3 cathode material component at the end of the activation plateau. (e) The charge mechanisms in Li-rich MLO oxides. Reproduced with permission.57 Copyright 2019 Springer Nature. (f) Schematic illustration of the layered Na-containing oxides for P2-type, O2-type, O3-type, and P3-type stackings. Reproduced with permission.90 Copyright 2017 Wiley-VCH. |
Gu et al.89 reported that newly formed spinel domains show random orientation within the same particle during cycling. The extraction of Li+ ions together with the loss of oxygen from Li2MnO3 induces the displacement of TM cations. This change further destroys the lattice and creates numerous nanometer-scale spinel particles, leading to rapid capacity and voltage decay. Correspondingly, Li+ ions cannot intercalate in the transformed structure again, resulting in low initial coulombic efficiency. For example, in the first oxidation process, atomic columns in the Li layers can be observed in both the surface and bulk regions. This feature implies the migration of TM ions to the Li layer (Fig. 5c). Meanwhile, the development of pores and cracks among the layered nanoparticles results in lattice degradation and vacancy formation upon the extraction of Li+ ions, suggesting that the layered-to-spinel phase transition is a nucleation and growth mechanism. Recently, Radin et al.57 modeled the irreversible phase transition. As shown in Fig. 5d, a typical local configuration forms upon Li2MnO3 charging. Mn4+ remains in the [MnO6] octahedral sites, while Mn7+ tends to migrate to the tetrahedral sites in Li1/2MnO3. To form Li1/2MnO3, half of the Mn ions must migrate, with one quarter moving downward and one quarter moving upward (Fig. 5e). This behavior finally produces an unstable lattice structure.
As the large Na+ ions stabilize the prismatic sites in the P2 phase, MnO2 slabs tend to form octahedral sites, enabling the phase change from P2 to O2.90 For the O3-type phase, when a sufficient amount of Na+ is removed from the interlayers, MnO2 sheet gliding induces the formation of broad prismatic sites. The stacking of oxygen transits to AB, BC, CA, known as the P3 phase. In other words, the extraction of Na+ from both P2 and O3 phases generally causes the gliding of MnO2 sheets, promoting the phase transition (layer-to-layer). NaxMnO2 has a more complicated phase than its Li analogs due to the large size of Na+ coupled with the ordering arrangement of Na+ and vacancies. Because the MnO2 sheet gliding combined with the change in oxygen stacking brings severe structural degradation, cracks are often observed in NaxMnO2 during cycling. Therefore, reducing or suppressing these phase transformations is critical for improving the electrochemical performance.
Fig. 6 (a) The six equivalent Mn–O bonds (left) in MLOs are distorted and transformed into the four shorter equatorial bonds and two longer axial bonds (right). (b) H+ in the electrolyte attacks the axial oxide and transfers electrons from the manganese to oxygen. Reproduced with permission.103 Copyright 2020 American Association for the Advancement of Science. (c and d) Decomposition reaction of EC in electrolytes with/without interacting with metallic Mn, Mn2+, and Li+ before and after one-electron reduction. (e) The catalytic effect of Mn2+ on the decomposition of electrolytes. Reproduced with permission.104 Copyright 2019 Springer Nature. |
Specifically, the lower operation potential of conventional anodes like Li metal and graphite compared with the redox potential of Mn/Mn2+ results in the electroplating of soluble Mn2+ in the electrolyte on the anode. Simulations have indicated that the deposited Mn on the anode has high electrical conductivity or electrocatalytic properties, accelerating the decomposition of the electrolyte.112–114 The reconstructed SEI gradually covers the anode and affects the ionic diffusion kinetics. Recently, Wang et al.104 demonstrated that overlooking the effect of the solvation sheath on the Mn2+ ions is the main reason for the inaccurate conclusions that Mn is the final product of the anode. They proposed that the solvation sheaths around Mn2+ act as molecular-sized reaction vessels, in which the central Mn2+ activates its solvation member and guides the reaction process. This process explains why Mn atoms are rarely detected on anodes.115,116 If the solvation sheath begins to lose solvating members, Mn2+ ions can convert to Mn atoms, explaining the existence of Mn atoms on the surface of the anode. It was found that the Mn atoms (Fig. 6c) substantially decrease the activation energy of EC decomposition (from TS1 247.20 to TS3 33.69 kJ mol−1). In particular, when coordinated with Mn2+, the EC decomposition activation energy (TS7) is less than that of neat EC (TS1), which is only 16.57 kJ mol−1. After the EC anion receives an electron, the decomposition of EC (TS10) becomes much easier (Fig. 6d). With Mn2+, the reduced EC decomposition activation energy (TS21: 9 kJ mol−1) is higher than that of reduced EC (TS10), suggesting that the EC anion moves away from Mn2+ after receiving one electron. Nevertheless, the generated coordination vacancies around Mn2+ become occupied by other incoming EC molecules. Overall, the Mn2+ solvation sheath serves as a catalytic center on the electrode surface or in the electrolytes, which continuously accelerates the reduction of carbonate molecules, resulting in severe interfacial degradation, and cell performance fading (Fig. 6e).
Fig. 7 (a) The HAADF image of the pristine cathode material, forming a spinel-type grain boundary. (b) The generation of a large strain in the spinel-type grain boundary analyzed by Geometrical Phase Analysis (GPA). (c) Atomic resolution image of a nanocrack at the particle surface induced by a spinel-type grain boundary. (d) EDS maps collected from the nanocrack. Reproduced with permission.110 Copyright 2019 American Chemical Society. |
Material | Strategy | Synthesis method | Improved mechanism | Capacity | Capacity retention | Ref. |
---|---|---|---|---|---|---|
δ-MnO2 | Cr doping | Hydrothermal | Suppressing the elongation of the Mn–O bond | 250 F g−1 at 0.2 A g−1 | 69.1% (10000 cycles) vs.82.6% (30000 cycles) at 10 A g−1 | 124 |
Na0.65Mn0.75Ni0.25O2 | F doping | Solid-state reaction | Inhibiting P3–O1 phase transition | 160 mA h g−1 at 0.1C | 75.1% vs. 83.1% (100 cycles) at 0.2C | 125 |
Na0.833[Li0.25Mn0.75]O2 | Zn doping | Sol–gel method | Inhibiting P2–P2′ phase transition | 241 mA h g−1 at 0.2C | 41.3% vs. 96.9% (200 cycles) at 0.2C | 34 |
Na0.65Mn0.75Ni0.25O2 | F/B co-doping | Solid-state reaction | Suppressing P3–O1 phase transition | 164 mA h g−1 at 0.1C | 86% vs. 75% (100 cycles) at 0.2C | 125 |
Na0.5MnO2 | Na+ preintercalation | Hydrothermal | Stabilizing the interlayered structure | 366 F g−1 at 1A g−1 | 96% (10000 cycles) at 4 A g−1 | 126 |
δ-MnO2 | Polyaniline intercalation | Chemical oxidation and reduction | Strengthening the layered structure and eliminating layered−spinel phase transition | 280 mA h g−1 at 0.2 A g−1 | 89.2% (5000 cycles) at 2 A g−1 | 101 |
Na0.67Mn0.5Co0.5−xFexO2. | Fe3+ preintercalation | Solid-state reaction | Restraining the potential slab sliding | 160 mA h g−1 at 0.2C | 68.2% vs. 89.6% (100 cycles) at 10C | 127 |
Na2/3[Zn1/9Mn7/9□1/9]O2 | Mn vacancy | Pechini method | Shortening the Mn–O bond length | 204 mA h g−1 at 0.02 A g−1 | 43.4% vs. 25.6% (100 cycles) at 0.02 A g−1 | 128 |
Li1.22xNaxMn0.56Ni0.16Co0.08O2 | O vacancy | Solid-state reaction | Modulating the local Mn coordination environments | 250 mA h g−1 at 0.1C | 84% vs. 94% (100 cycles) at 0.5C | 40 |
Li1.2[Mn0.66Co0.17Ni0.17]0.8O2 | Mixed vacancies of O and TM | Solid-state reaction | Prohibiting the O3–P3 phase transition | 220 mA h g−1 at 0.02 A g−1 | 86% (250 cycles) at 0.125 A g−1 | 129 |
NaMnO2−y −δ(OH)2y | Novel monoclinic phase | Hydrothermal sodiation | Decreasing the volume change via a stable interlayer | 215 mA h g−1 at 0.2C | 94.6% (1000 cycles) at 10C | 47 |
Spinel–layered LiMnO2 | Heterostructure construction | Electrochemical conversion | Modulating interfacial orbital ordering | 286 mA h g−1 at 1C | 90.4% (2000 cycles) at 12C | 44 |
Li2Mn2/3Nb1/3O2F | Disordered structure | Ball-milling | Replacing the anionic reaction with a Mn reaction | 320 mA h g−1 at 0.01 A g−1 | 90% (25 cycles) at 0.02 A g−1 | 130 |
Li1.2Ni0.2Mn0.6O2 | Surface structure design | Chemical vapor deposition | Charge compensation | 238 mA h g−1 at 0.5C | 99% (100 cycles) at 1C | 131 |
Inferior structural stability and electronic conductivity have been demonstrated to limit the electrochemical performance of δ-MnO2. Zhao et al.124 proposed that incorporating Cr3+ into δ-MnO2 could boost its rate capability and cycling stability within the potential window of 0–1.2 V (vs. Ag/AgCl). δ-MnO2 without Cr doping undergoes severe J–T distortion, which increases the average Mn–O bond length and accelerates the migration and dissolution of Mn. Numerous dissolved Mn ions in the electrolyte generate cracks and pulverize the MnO2 nanosheet fragments to a moderate extent (Fig. 8a). In addition, the dissolved Mn ions form a precipitate in the electrolyte, impacting the electrochemical performance of the whole cell. Meanwhile, buckled Mn–O slabs are formed along with the J–T distortion, and these slabs dissociate and rearrange, leading to the irreversible phase transition from a layered structure to a 2 × 2 tunnel structure.88 Correspondingly, only 69.1% of the capacity of the pristine δ-MnO2 electrode is retained over 10000 cycles at the current density of 10 A g−1. The elongation of the Mn–O bond induced by the J–T distortion is effectively suppressed in Cr3+-doped δ-MnO2 (Fig. 8b), implying that the Mn ion migration barrier greatly increases. The reinforced birnessite shows significantly improved structural stability (82.6% capacity retention over 30000 cycles) in comparison with the pristine δ-MnO2, demonstrating the excellent effect of Cr doping in δ-MnO2 on suppressing the J-T effect and stabilizing the layered structure. Moreover, owing to the improved effect of Cr on the electronic conductivity of δ-MnO2, the Cr-doped δ-MnO2 showed a much higher capacity of 250 F g−1 than pristine δ-MnO2 with 178 F g−1 at the current density of 0.2 A g−1 with a high mass loading of ∼5 mg cm−2.
Fig. 8 (a) Mechanisms of Cr-doping for the stabilities of the δ-MnO2. (b) The shortened Mn–O bonds after Cr-doping. (c) Cycling performances of the pristine δ-MnO2 and Cr-doping δ-MnO2 electrodes at a current density of 10 A g−1. Reproduced with permission.124 Copyright 2020 Royal Society of Chemistry. (d) The PDOS of the Mn3+ sites in the layered Zn-doping Na[Li0.25Mn0.75]O2. (e) The two atomic models include 8 Na atoms in each unit cell. Reproduced with permission.34 Copyright 2019 Springer Nature. |
To further validate the role of elemental doping, Zhang et al.34 modeled the Zn-doped layered Na[Li0.25Mn0.75]O2 material to explore its effect on thermodynamics and the electronic structure. Zn doping enables the oxidation of the Mn3+ center and thus increases the Mn4+ content of the centers in the Zn-doped materials (Fig. 8d). As a result, the materials are more structurally stable and undergo less J–T distortion, leading to negligible capacity decay of the Zn-doped layered Na[Li0.25Mn0.75]O2 electrode (99% capacity retention over 200 cycles), which is much higher than that of the pristine Na[Li0.25Mn0.75]O2 electrode (41% capacity retention over 200 cycles). From the perspective of thermodynamics, the formation energies of mixing enthalpy (ΔHmix) have an important influence on phase stability. The calculated thermodynamic energy values can confirm the structural stability of materials according to charge carrier content. Fig. 8e shows that Na[Li0.25Mn0.75]O2 comprises five ground states, whereas Zn-doped Na[Li0.25Mn0.75]O2 has seven ground states. Specifically, in the first region (x = 0–0.25), Zn doping can decrease ΔHmix, indicating a reduced volumetric strain (Zn-doped structure: εV = −0.027; undoped structure: εV = −0.04). In the second region (x = 0.25–0.50), Zn-doped Na[Li0.25Mn0.75]O2 is illustrated by the red circles located on the tie line, which directly confirm that desodiation occurs in the same phase. In undoped Na[Li0.25Mn0.75]O2, a pseudo-ground state appears, suggesting the formation of a new phase during the desodiation process. In the third region (x = 0.50–0.88), Zn-doped Na[Li0.25Mn0.75]O2 has one pseudo-ground state as marked by the grey circle (x = 0.63), indicating that phase dissociation may occur within the narrow region from x = 0.5 to x = 0.75. The extra ground states in Zn-doped Na[Li0.25Mn0.75]O2 are beneficial for suppressing the inherent phase transition that occurs in Na[Li0.25Mn0.75]O2.132 In contrast to its doped counterpart, two pseudo-ground states of undoped Na[Li0.25Mn0.75]O2 are shown as grey circles above the tie line (at x = 0.63 and x = 0.75, respectively), reflecting the phase separation to this extent during desodiation.
In summary, elemental doping can restrain the elongated bonds and stabilize the metastable structure to inhibit J–T distortion by changing the atomic structures around Mn ions, giving rise to a more stable structure. Additionally, the average valence of Mn ions may be tuned to above +3.5, which is beneficial to suppress the J–T effect, such as the introduction of the low-valence ions. Although elemental doping has achieved tremendous progress, some concerns should be noted. Presently, annealing at high temperatures is the most popular method to realize doping, but this process involves high energy cost and has low efficiency. In addition, it is difficult to precisely control the doping content and doping sites. Hence, it is highly desirable to explore the powerful and facile method of accurate doping with low energy costs.
Fig. 9 (a) Schematic illustration of the preintercalation process during the electrochemical oxidation. (b) The ion-preintercalation effect on widening the potential windows of layered Na0.5MnO2. Reproduced with permission.126 Copyright 2017 Wiley-VCH. (c) A schematic of the expanded structure by the intercalation of polyaniline (PANI) into MnO2 nanolayers. Reproduced with permission.101 Copyright 2018 Springer Nature. (d) The typical in situ XRD patterns of the Na0.67Mn0.5Co0.5−xFexO2 electrode. (e) The variation of cell parameters c (green points), V (blue points), and a/b (pink points). (f) V variation of Na0.67Mn0.5Co0.5−xFexO2 and several typical sodium-storage cathode materials during the charge/discharge process. Reproduced with permission.127 Copyright 2021 Wiley-VCH. |
Aside from enlarging the interlayer spacing, some intercalated ions may have a pinning effect to suppress phase transitions and thus improve the structural stability. In Na-ion batteries, Na+ ions possess a large radius and cause huge lattice changes of more than 20% during the sodiation/desodiation process, inducing a severe phase transition, resulting in more than 30% capacity fading over 100 cycles at the current density of 2C.127 Chu et al.127 reported that intercalated Fe in the Na site acts as a pinning point to suppress slab sliding. The pinning point represents a structural stabilizer that significantly reduces the volume variation to appropriately 0.6% (Fig. 9d and e). This slight structural change confers outstanding structural stability and maintains a smooth channel for fast Na-ion transfer during the electrochemical processes. The ultra-low volume variation is even smaller than those of electrode materials like LiCoO2 and LiFePO4, as shown in Fig. 9f. Consequently, Na-ion batteries fabricated using this improved material can achieve high capacity retention of ∼89.1% over 100 cycles at the current density of 2C. Impressively, 71% capacity can be retained over 1000 cycles at the high current density of 10C.
On the basis of the above introduction, manipulating the interlayered structure via the preintercalation can greatly enlarge the interlayer space and thus facilitate the ions to diffuse in the bulk of the MLO materials, which also enables the MLO materials to provide more active sites to accommodate cations. In addition, the support effect of preintercalated ions/molecules on the interlayered structure can improve the structural stability, guaranteeing long-term cycling stability. Nevertheless, preintercalated ions may have the electrostatic repulsion effect with the inserted ions during the electrochemical processes. Moreover, the stability of supported ions/molecules should be considered. These preintercalated species suffer from issues of extraction of ions/molecules from the host interlayers during the electrochemical processes, which is unavoidable. Such issues eliminate the advantages of manipulating interlayered structures and induce severe structural degradation during the continuous intercalation/de-intercalation of cations. Excessive cation preintercalation would induce the phase transition where the structural reconstruction enables the loss of the fast ionic transport of layered pathways.100 Therefore, to avoid these disadvantages, careful screening of the preintercalated ions/molecules and optimizing their amount are necessary.
Fig. 10 (a) The evolution of bond lengths in the electrochemical processes. (b) The percentages of the low-spin and high-spin Mn3+ ions. Reproduced with permission.128 Copyright 2021 Elsevier. (c) Schematic illustration of the mechanism behind the prevention of the O3–P3 biphasic reaction. (d) Cycling performance of Li1.2[Mn0.66Co0.17Ni0.17]0.8O2 at 125 mA g−1. Reproduced with permission.129 Copyright 2021 Wiley-VCH. |
In addition to single vacancies, the introduction of mixed vacancies is also an effective way to suppress structural degradation due to the synergistic effects between anionic and cationic vacancies. In both P2-type and O3-type MLO materials, the irreversible layer gliding causes huge volume changes and significantly jeopardizes the structural stability.141,142 To solve this problem, reducing the net charge of the oxygen (i.e., creating oxygen vacancies) or enabling reversible TM migration during the charging/discharging process is critical. “Reversible” migration implies that TM can move between the TM layer and alkaline layer but does not need to return to the same site each time. Therefore, cationic vacancies can provide many more sites for TM migration. To demonstrate this concept, Xiao et al.129 prepared an Li- and Mn- rich layered cathode material, in which the formation of oxygen vacancies and Mn vacancies was simultaneous. The results of a combination of computational simulations and systematic characterizations demonstrated that the mixed vacancies enabled the TM to migrate to the tetrahedral sites of the Na layer during desodiation and move back during sodiation. The TM migration suppressed a complete O3–P3 phase transition by increasing the energy barriers and stabilized the performance. Fig. 10c shows the proposed energy evolution during the O3–P3 two-phase reaction by manipulating TM migrations. Such metastable states can immobilize the structure and inhibit the transition to global P stacking. Correspondingly, the Li- and Mn-rich layered cathode material delivered 86% capacity retention after 250 cycles (Fig. 10d). In addition, although the interlayer spacing decreased to 5.0 Å, the diffusion kinetics of Na+ ions was not affected, which was attributed to the formation of oxygen vacancies and the distortion of the metal oxide layer.
As described above, the incorporation of Mn and/or O ion vacancies in MLO materials promotes ion-reduction/oxidation and bond length changes, which significantly increase the percentage of low-spin Mn3+ and provide more sites for the reversible migration of Mn ions. It should be noted that the introduction of vacancies can alter the electronic structure and bonding energy in MLO materials. Too high amount of vacancies may sacrifice the conductivity of materials and structural stability. However, how to achieve precise and optimum vacancy concentration is a critical issue in the current synthesis process. Correspondingly, developing new synthesis methods to accurately control the vacancy concentration is necessary. Furthermore, advanced characterization methods to monitor the generation process of vacancies are also needed.
To address these issues, Xia et al. designed a novel monoclinic polymorph NaMnO2−y−δ(OH)2y with the coexistence of the O′3 and H′3.47 In sharp contrast to the monoclinic, O3-, and P2-NaxMnO2, the monoclinic polymorph NaMnO2−y−δ(OH)2y has a larger interlayer distance (∼7 Å), which greatly enhances the kinetics of Na+ ion diffusion. The large interlayer space is attributed to the simultaneous formation of O′3 and H′3 phases with six-coordinate octahedral sites and eight-coordinate hexahedral sites, respectively. It was found that the generated O vacancy in this material can reduce the bandgap, suggesting the enhanced electronic conductivity of NaMnO2−y−δ(OH)2y. Therefore, NaMnO2−y−δ(OH)2y with large interlayered spacing and high electronic conductivity accelerates the electron and ion transport in the electrode, which enables the NaMnO2−y−δ (OH)2y electrode to maintain a large capacity of 156 mA h g−1 even at the current density of 50C, outperforming the previously reported data in this field. In addition, the reduced distortion of Mn–O and the small β angle in NaMnO2−y−δ(OH)2y are beneficial to suppress the J–T distortion and minimize the volume change during electrochemical processes (Fig. 11a). These attributes explain why NaMnO2−y−δ(OH)2y can achieve superior structural stability and excellent cycle performance. As a result, the prepared NaMnO2−y−δ(OH)2y electrode achieved outstanding cycling stability (94.6% capacity retention at 10C over 1000 cycles) than the Birnessite NaMnO2−y−δ (OH)2y·0.61H2O (78% capacity retention at 10C over 1000 cycles).
Fig. 11 (a) Reversible crystal structure change of the monoclinic H′3/O′3 NaMnO2−y−δ(OH)2y in the electrochemical processes. Reproduced with permission.47 Copyright 2018 Springer Nature. (b) Alleviating Jahn–Teller distortion by interfacial structure manipulation in spinel–layered LiMnO2. (c) Bond lengths of MnO6 octahedra in spinel–layered LiMnO2, spinel LiMnO2, and monoclinic LiMnO2. Reproduced with permission.44 Copyright 2021 Springer Nature. (d) A high-magnification TEM image of polycrystalline Li2Mn2/3Nb1/3O2F (yellow circle). Reproduced with permission.130 Copyright 2018 Springer Nature. |
Designing novel interface structures is also effective in suppressing J–T distortion. Considering the orbital ordering in Mn3+-containing MLO materials, such as LiMnO2, which can generate the collinear J–T order, the strong cooperative J–T effect may cause a large volume change and promote structural degradation (Fig. 11b). To eliminate this cooperative J–T effect, Xia et al.44 designed a unique heterostructure of LiMnO2. The heterostructure comprised layered and spinel domains with various orientations for J–T distortion, which can effectively disrupt the long-range collinear J–T ordering for either the spinel phase or the layered phase in the heterostructure. Fig. 11c compares the axial Mn–O bond elongation in this heterostructure with those in spinel and monoclinic LiMnO2. The heterostructure increased the axial Mn–O bonds from 1.96 to 2.07 Å in spinel LiMnO2 and 1.98 to 2.03 Å in monoclinic LiMnO2, which corresponds to 5.5% and 2.5% distorted degrees, respectively. For layered LiMnO2 and spinel Li2Mn2O4,69,145 the distortion degrees were 18.0% and 15.9%, respectively, much higher than the corresponding phase in the heterostructure. Therefore, the layered–spinel heterostructure with controlled interfacial orbital ordering significantly suppressed J–T distortion. The electrochemical results demonstrated that the LiMnO2 heterostructure cathode delivered high specific capability ∼254.3 mA h g−1 with 97% coulombic efficiency. Furthermore, such a heterostructure achieved a high specific capacity of 183.6 mA h g−1 at 12C. Unprecedented cycling stability with ∼90.4% capacity retention over 2000 cycles was achieved, demonstrating the significant effect of interfacial engineering on mitigating J–T distortion.
Cation-disordered lithium-rich MLO materials are another example of a novel structure.146 Lee et al.130 prepared a new cation-disordered (disordered rock-salt structure) material Li2Mn2/3Nb1/3O2F by manipulating the layered phase of Li2MnO3 (Fig. 11d). The disordered rock-salt structure, originating from a layered structure, can generate more 0 TM channels.147 This diffusion channel has a low migration barrier, giving rise to fast Li+ diffusion kinetics. Furthermore, with cation and anion substitution, the disordered rock-salt Li-rich cathodes could deliver a high capacity (>300 mA h g−1), mainly from the Mn2+/Mn4+ double electron reaction (Fig. 11e). This Mn2+/Mn4+ double redox effectively suppresses the problems associated with O redox by reducing the anionic redox reaction content.
In short, novel structural design shows notable effectiveness in manipulating the localized structure to suppress the J–T effect, providing stable units in localized regions and thus effectively eliminating electrochemical degradation. Nevertheless, novel structural design faces great challenges, countering which needs an in-depth understanding of the structure–property relationships in MLO materials. Additionally, accurately manipulating the structure at the atomic level is difficult, which is not suited to large-scale applications in most cases.
Although various strategies have paved effective pathways to suppress electrochemical degradation, it is still difficult to thoroughly disentangle the issues raised above. Additional research efforts are needed to develop methods for stabilizing the crystal structure during electrochemical processes to suppress capacity degradation in the representative MLO systems. Our suggestions are as follows:
(1) Coupling various strategies to induce synergistic effects for high-performance MLO materials. Each strategy may have a limited effect on suppressing the degradation of MLO cathodes. In this case, integrating various strategies to exploit their synergistic effects may be worth exploring. For example, combining doping and interlayer manipulation may realize robust structural stability and fast diffusion kinetics.
(2) Developing novel MLO materials with stable structures. Presently, most MLO materials still have structural issues and are sensitive to air containing water vapor. It is critically necessary to innovate MLO materials with highly stable localized structures and microstructures by modulating the electronic states and tailoring the crystal domains. In addition, advanced machine learning and theoretical calculations can be applied to design and develop high-performance MLO materials. With improved computational analyses, it is possible to gain some quantitative comprehension of the effect of each adverse factor on structural stability using qualitative constraints in simulations.
(3) Realizing accurate MLO cathode characterization and comprehensively elucidating the structure–property correlations. Electrochemical energy storage devices store energy through electrochemical reactions. However, the understanding of the intrinsic electrochemical reaction mechanisms in MLO materials is still limited due to the complexity of each energy storage system and the effects of electrolytes, conductive carbon additives, and binders. In this regard, powerful characterization techniques, such as in situ transmission electron microscopy, in situ X-ray absorption spectroscopy, and electrochemical quartz crystal microbalance analysis, are indispensable to distinguish the dynamic structural evolution and electrochemical behaviors of MLO cathodes and comprehensively understand their structure–property correlations.
In conclusion, a comprehensive understanding of the electrochemical degradation mechanisms at the atomic level is essential for developing high-performance MLO cathodes, as is systematically outlined in this review. Although numerous challenges remain, the future is bright thanks to the tremendous endeavors to elucidate J–T distortion and the continual development of effective strategies, advanced characterization techniques, artificial intelligence, and theoretical calculations. We hope that the synergistic application of experiments and computational modeling will dramatically improve our understanding of the chemical degradation mechanisms and the electrochemical performance of energy storage systems based on MLO cathodes.
Footnote |
† These authors contributed equally to this work. |
This journal is © The Royal Society of Chemistry 2022 |