The quest for manganese-rich electrodes for lithium batteries: strategic design and electrochemical behavior

Abstract

Manganese oxides, notably γ-MnO₂ and modified derivatives, have played a major role in electrochemical energy storage for well over a century. They have been used as the positive electrode in primary (single discharge) Leclanché dry cells and alkaline cells, as well as in primary and secondary (rechargeable) lithium cells with non-aqueous electrolytes. Lithiated manganese oxides, such as LiMn₂O₄ (spinel) and layered lithium–nickel–manganese–cobalt (NMC) oxide systems, are playing an increasing role in the development of advanced rechargeable lithium-ion batteries. These manganese-rich electrodes have both cost and environmental advantages over their nickel counterpart, NiOOH, the dominant cathode material for rechargeable nickel–cadmium and nickel–metal hydride batteries, and their cobalt counterpart, LiCoO₂, the dominant cathode material in lithium-ion batteries that power cell phones. An additional benefit is that tetravalent manganese can be used as a redox-active and/or stabilizing ‘spectator’ ion in lithiated mixed-metal oxide electrodes. This paper provides an overview of the historical development of manganese-based oxide electrode materials and structures, leading to advanced systems for lithium-ion battery technology; it updates a twenty-year old review of manganese oxides for lithium batteries. The narrative emanates largely from strategies used to design manganese oxide electrode structures at the Council for Scientific and Industrial Research, South Africa (1980–1994), Oxford University, UK (1981–1982), and Argonne National Laboratory, USA (1994–2017); it highlights the worldwide evolution of ideas and recent trends to improve the design, stability, and electrochemical capacity of structurally integrated, manganese-rich electrode materials.

---

Historical background

Manganese oxides are a structurally versatile family of compounds, existing in a wide variety of polymeric forms¹ with atomic arrangements that provide an interstitial space of one-dimensional tunnels, two-dimensional layers, or three-dimensional intersecting channels. These compounds can act as a host framework for the intercalation and de-intercalation of guest ions, such as hydrogen or lithium ions. As such, they are of interest as insertion electrodes for aqueous and non-aqueous electrochemical cells and batteries.^2,3

In 1866, George Leclanché invented a 1.5 V Zn/MnO₂ ‘wet’ cell, which contained a zinc anode, manganese dioxide cathode with a carbon current collector, and an ammonium chloride electrolyte solution.⁴ In 1886, Carl Gassner developed a ‘dry’ cell in which the ammonium chloride was contained in a ‘plaster of paris’ paste.⁴ The manganese dioxide that was commonly known at that time was nsutite, a naturally occurring manganese dioxide, simply referred to as γ-MnO₂ (ref. 1) or ε-MnO₂, depending on the quality of its X-ray diffraction (XRD) pattern.² Nsutite electrodes were later replaced by higher quality synthetic γ-MnO₂ products prepared by electrolytic or chemical methods. Electrolytic manganese dioxide, rather than chemically prepared manganese dioxide, is the preferred cathode material for alkaline Zn/MnO₂ cells (also 1.5 V) introduced into the battery market in the late 1960s.⁵ An excellent historical account of the electrochemical reactions that occur during discharge of Zn/MnO₂ cells is provided in a review by Chabre and Pannetier.² By the 1960s, several groups had concluded that the monotonic drop in potential during discharge of the cells could be attributed to a H⁺-ion insertion reaction to form MnOOH (groutite),^2,6–9 rather than a displacement or conversion reaction. Specifically, in 1966, Kozawa and Powers¹⁰ proposed that the reaction could be described as ‘a homogeneous single phase, solid state process, following previously introduced thermodynamic concepts of non-stoichiometric oxides’ by Vetter in 1962.^2,11 This extensive characterization of γ-MnO₂ as an insertion electrode was the forerunner to intense research into lithium intercalation compounds in the mid-to-late 1970s, such as graphite^12–18 and the metal chalcogenides (e.g., TiS₂ (ref. 19–21)). Subsequently, the metal oxides LiCoO₂ (ref. 22 and 23) and LiMn₂O₄ (ref. 24–26) would spawn the non-aqueous carbon/lithium–metal-oxide battery industry in the early 1990s.

Attempts to develop and commercialize rechargeable alkaline Zn/MnO₂ cells have not been hugely successful; the cells suffer from a poor cycle life because the reaction products cannot be fully recharged,^27–29 and the zinc dendrites that form during charge can lead to internal short circuits.^30,31 However, a recent report has indicated that Zn/MnO₂ cells can be cycled with >80% energy efficiency for more than 3000 cycles if the depth of discharge is restricted to 10%.³²

The structural complexity of γ-MnO₂ materials is described in detail by Chabre et al.² Simplistically, γ-MnO₂ can be regarded as having a composite structure, depicted ideally in Fig. 1c, in which octahedral MnO₆ units and (‘1 × 1’) dimensional channels of β-MnO₂ (rutile-type structure, Fig. 1b) are intergrown with ramsdellite-MnO₂ units and channels of larger (‘2 × 1’) dimension (Fig. 1a). Synthetic β-MnO₂ tends to be anhydrous³³ whereas the (2 × 1) channels in ramsdellite- and γ-MnO₂ structures can be protonated or contain a minor amount of occluded water.^1,2 In practice, the arrangement of the β-MnO₂ and ramsdellite-MnO₂ building blocks in γ-MnO₂ is not ideal. The intergrown structures contain defects, stacking faults, and structural disorder that lead to XRD patterns with broad and sharp peaks, making accurate structural analysis of this family of compounds, and the differences between them, difficult.

Fig. 1 MnO₂ structure types: (a) ramsdellite-MnO₂, (b) β-MnO₂ (rutile), (c) γ-MnO₂, (d) α-MnO₂ (hollandite), (e) Li₂O-stabilized α-MnO₂, and (f) λ-MnO₂ (defect spinel).

In 1975, Ikeda disclosed a 3 V non-aqueous primary Li/γ-MnO₂ cell with a voltage double that of the aqueous Zn/MnO₂ cell,³⁴ which was commercialized the following year.³⁵ In this case, the cathode was heat-treated at 350–430 °C to remove occluded and surface water from the electrode prior to cell assembly, thereby significantly improving the electrochemical properties and life of the cell.³⁶ Lithium intercalation occurs predominantly in the ‘2 × 1’ channels of the heated γ-MnO₂ structure. The ‘1 × 1’ channels of the rutile component are small and energetically resistant to lithium uptake, consistent with the finding that, when lithiated chemically at 50 °C, β-MnO₂ transforms to spinel (LiMn₂O₄) to accommodate the additional lithium.³⁷

Rechargeable lithium cells with pre-lithiated and heat-treated γ-MnO₂ cathodes were subsequently developed for the battery market.³⁸ The cathodes were prepared by reacting LiOH (or LiNO₃) with γ-MnO₂, typically in a 3 : 7 molar ratio at 350–430 °C, to form an intergrown electrode structure, referred to as ‘composite dimensional manganese oxide’ (‘CDMO’) by Sanyo, which was initially believed to consist of structurally connected Li₂MnO₃ (Li : Mn = 2 : 1) and lithium-stabilized γ-MnO₂ components.³⁹ This interpretation was later changed because a Li : Mn ratio of 3 : 7 (1 : 2.33) would thermodynamically favor the formation of a lithium–manganese-oxide spinel component at ∼400 °C, such as Li₂Mn₄O₉ (Li : Mn = 1 : 2), rather than Li₂MnO₃, whose XRD patterns are similar.^40,41

These studies highlighted the concept of using composite electrode structures with electrochemically active and inactive components, the former contributing to capacity generation, and the latter providing structural stability. Subsequently, this approach was further exploited in designing advanced NMC electrode structures for lithium-ion batteries.

There also are MnO₂ structures with tunnels of larger dimension, e.g., ‘2 × 2’ tunnels (hollandite), ‘2 × 3’ tunnels (romanechite), and ‘3 × 3’ tunnels (todorokite).¹ The structure of hollandite, commonly known as α-MnO₂ (Fig. 1d), is particularly intriguing; it has the characteristic ‘2 × 1’ MnO₆ units of ramsdellite but also the characteristic (‘1 × 1’) channels of β-MnO₂. The α-MnO₂ framework contains stabilizing cations within the ‘2 × 2’ tunnels, for example, Na⁺, K⁺, Ca²⁺, or Ba²⁺ ions (Fig. 1e). The α-MnO₂ structure can also be stabilized by hydronium (H₃O⁺)⁴² or ammonium (NH₄⁺)⁴³ ions. In this case, negatively charged O²⁻ and N³⁻ ions are located at the center of the ‘2 × 2’ channels, occupying at least some of the vacant sites of a distorted close-packed oxygen array;⁴⁴ the protons provide compensating positive charge within the ‘2 × 2’ channels necessary to stabilize the structure. A recent report has provided evidence that hydrated α-MnO₂ is stabilized by both hydronium ions and water molecules.⁴⁵ The α-MnO₂ framework is zeolitic in nature; when this material is heated to 400 °C and cooled in a moist air, the water component can be reversibly removed from, and reincorporated, into the pores of the stable framework structure.⁴⁶ The protons in H₃O⁺-stabilized α-MnO₂ can be ion-exchanged with lithium to yield products that provide superior capacity and electrochemical stability relative to metal-stabilized analogues (Na⁺, K⁺, Ca²⁺, Ba²⁺, etc.) in rechargeable 3 V Li cells.^44,47 More recently, α-MnO₂ has been identified as a promising electrocatalyst for Li-oxygen cells.^46,48 This finding begs the question whether, in a hybrid Li-ion/Li–O₂ system in which both manganese and oxygen participate in the electrochemical redox reactions, the electrocatalytic activity of α-MnO₂ is associated with the removal of lithium and oxygen ions from the ‘2 × 2’ channels during charge, and re-accommodation during discharge.⁴⁹

The advent of lithium-ion technology

The advent of lithium-ion batteries in 1991, with the introduction of high potential lithium–metal oxide cathodes, coupled to a carbon anode, was a game changer in the battery industry. It opened the door to the exploration, identification, and exploitation of a vast number of cathode compositions and structures.^50–52 Remarkably, LiCoO₂ (ref. 20) is still the cathode material of choice for the ever-expanding cell phone industry. The high cost of cobalt and the political instability of cobalt-producing countries have intensified worldwide efforts to use lithium-ion batteries with alternative cathode materials, such as nickel- and manganese-based metal oxides, for large-scale, heavy-duty batteries for hybrid and all-electric vehicles, as well as stationary energy storage for telecommunications, and back-up for the electrical grid. Partial substitution of cobalt by nickel and aluminum has provided alternative nickel-rich cathode compositions, e.g., LiNi_0.8Co_0.15Al_0.05O₂ (NCA),^50,52 for high power applications such as electric vehicles. However, Ni-rich electrode materials such as NCA suffer from structural and chemical instabilities through oxygen loss above 4 V that can lead to a fast self-heating rate below 200 °C,⁵³ thereby compromising the safety of lithium-ion cells.⁵⁴ A concerted effort has therefore been made over the past two decades to replace or substitute cobalt and nickel in layered oxides with manganese, because lithium-ion cells with NMC cathodes offer competitive energy density and greater stability at high charging potentials (>4 V).⁵⁴

LiMn₂O₄ (spinel) and substituted derivatives

Fig. 2 presents a ‘compositional’ phase diagram for the Li–Mn–O system, constructed in the early 1990s to emphasize compositional and oxidation state differences that affect the electrochemical behavior of spinel-related electrodes.^3,55 The diagram highlights, in particular, (i) the stoichiometric spinel tie-line between LiMn₂O₄ and Li₄Mn₅O₁₂ represented by the Li_1+xMn_2−xO₄ system (0 ≤ x ≤ 0.33), (ii) the stoichiometric lithiated spinel (rock-salt) tie-line between LiMn₃O₄ and Li₇Mn₅O₁₂, and (iii) the LiMn₂O₄–Li₄Mn₅O₁₂–MnO₂ defect spinel tie-triangle, which contains, for example, λ-MnO₂ and Li₂Mn₄O₉.

Fig. 2 Compositional Li–Mn–O phase diagram with respect to spinel, rock salt, and layered compounds (adapted from Thackeray et al.⁵⁵).

Phase diagrams of lithium-ion electrode systems reported in the battery literature, e.g., Fig. 2, can reflect metastable materials formed by lithium insertion/extraction reactions at room temperature, each of which will have its own electrochemical signature. These diagrams, therefore, do not accurately indicate the stable compositions and structures that would co-exist under isothermal conditions, particularly at the high temperatures at which parent lithium metal oxide electrodes are synthesized. For example:

(1) In Fig. 2, the stoichiometric spinels LiMn₂O₄ and Li₄Mn₅O₁₂ are typically prepared in air at 850 °C and 400 °C, respectively, while the metastable, defect-spinel λ-MnO₂ (ref. 56) and the lithiated spinels Li₂[Mn₂]O₄ and Li₇[Mn₅]O₁₂ (ref. 25 and 55) can be prepared by chemical or electrochemical methods at room temperature. On heating in an inert atmosphere, the lithiated spinel Li₂[Mn₂]O₄ will transform to orthorhombic LiMnO₂ while, on heating in air, it will transform to Li₂MnO₃ and LiMn₂O₄. Note also that the composition LiMnO₂ can have several structure types depending on the method and temperature of synthesis, for example, layered (monoclinic),⁵⁷ lithiated spinel (tetragonal),²⁵ and staggered (orthorhombic)⁵⁸ configurations, or intergrowths between them.⁵⁷

(2) In the Li–Co–O system, layered LiCoO₂ is typically prepared at about 850 °C in air, whereas a metastable structural configuration approximating the lithiated spinel Li₂[Co₂]O₄ (also LiCoO₂) is generated at 400 °C; both structure types are preserved on cooling to room temperature but display significantly different electrochemical behavior.^59–61

Fig. 2, coupled to supporting experimental data,^25,55,62,63 underscores the following structure–electrochemical property relationships (at room temperature, ∼25 °C) in lithium–manganese-oxide spinel electrodes, notably those in the stoichiometric spinel Li_1+xMn_2−xO₄ (0 ≤ x ≤ 0.33) system:

(1) Lithium insertion invokes an immediate first-order transition during discharge from a spinel to a rock-salt-type structure by displacing cations from tetrahedral to octahedral sites.^25,55 This is an attractive electrochemical reaction because both spinel and rock-salt structures are found in nature and, therefore, represent intrinsically stable atomic configurations in the charged and discharged states of the electrode, respectively. In LiMn₂O₄ (x = 0), the spinel-to-rock-salt transition occurs by simultaneous lithium insertion and a two-phase conversion reaction, which results in a constant voltage discharge rather than a monotonically decreasing voltage expected for a homogeneous, single-phase (solid solution) reaction.

(2) The Li_1+xMn_2−xO₄ (0 ≤ x ≤ 0.33) spinel structures on the LiMn₂O₄ (x = 0)–Li₄Mn₅O₁₂ (x = 0.33) tie-line in Fig. 2 are cubic. Lithium insertion into these structures induces a Jahn–Teller (tetragonal) distortion when the average Mn oxidation state falls below 3.5+, i.e., when the Jahn–Teller active Mn³⁺ (d⁴) ion concentration reaches a critical value.⁵⁵ The magnitude of the crystallographic distortion depends on the Mn³⁺-ion concentration. For example, electrochemical lithiation of cubic LiMn₂O₄ (average Mn oxidation state = 3.5+; c/a = 1) immediately yields tetragonal Li₂Mn₂O₄ (average Mn oxidation state = 3.0+; c/a = 1.16).^25,62 On the other hand, Li₄Mn₅O₁₂ (average Mn oxidation state = 4.0+; c/a = 1) undergoes a cubic–cubic transition to yield Li_6.5Mn₅O₁₂ (average Mn oxidation state = 3.5+; c/a = 1) before the tetragonal phase Li₇Mn₅O₁₂ (average Mn oxidation state = 3.4+; c/a = 1.11) is formed;⁵⁵ in this case, the Jahn–Teller effect is not as pronounced as it is in Li₂Mn₂O₄. Note that this effect is also suppressed during the lithiation of the tetragonal spinel, Mn₃O₄ (c/a = 1.16); here, lithiation results in tetragonal LiMn₃O₄ (c/a = 1.05), which is close to cubic (c/a = 1.0), consistent with the reduction in the average Mn oxidation state from 2.67+ to 2.33+.²⁵

(3) Lithium is extracted electrochemically from tetrahedral sites of the Li_1+xMn_2−xO₄ spinel electrodes, such as LiMn₂O₄ (x = 0), at a potential of approximately 4 V vs. metallic lithium, whereas lithium is inserted into interstitial octahedral sites of the spinel structure at approximately 3 V to yield ordered rock-salt configurations.^24,62–66 Cubic–cubic transitions are reversible, whereas cubic–tetragonal transitions tend to be less so, particularly when lithium insertion induces large anisotropic changes to the unit cell parameters.⁵⁵

Despite the advantages of low cost and electrochemical stability (safety) that manganese offers over cobalt- and nickel-rich electrodes, LiMn₂O₄ and substituted derivatives (LMO), such as those containing excess lithium and a little aluminum to enhance structural and electrochemical stability,^67–69 provide inferior capacity and energy output relative to LiCoO₂ (LCO) and LiNi_0.8Co_0.15Al_0.05O₂ (NCA) electrodes. However, manganese dissolution from LMO electrodes, which has been attributed to the disproportionation reaction Mn³⁺ → Mn⁴⁺ and Mn²⁺ at the particle surface,^56,70,71 severely compromises the cycle life of carbon/LiMn₂O₄ lithium-ion cells.⁷¹ Nevertheless, significant progress has been made in enhancing surface stability, reducing manganese dissolution, and alleviating capacity fade through a wide variety of surface treatments.^72–77 Although LMO does not compete directly with LCO, NCA, and NMC cathodes in terms of capacity, energy output, and cycling stability, LMO has been blended with these cathode materials to offset the higher cost of cobalt and nickel.⁷⁸ The high-potential (4.7 V) Li_xMn_1.5Ni_0.5O₄ spinel (0 ≤ x ≤ 2) is receiving particular attention for next generation systems because it offers the possibility of significantly increasing the energy and power of a lithium-ion cell relative to an all-manganese (4.1 V) LMO cathode. Note that (1) in a standard Li_xMn₂O₄ electrode, the manganese ions are redox active over the full range of x (0 ≤ x ≤ 2), whereas, ideally, in well-ordered Li_xMn_1.5Ni_0.5O₄ electrodes, the Mn⁴⁺ ions are electrochemically inactive and play a spectator role that allows reversible Ni²⁺ to Ni⁴⁺ redox reactions to occur over the range 0 ≤ x ≤ 1; and (2), for the range 1 < x ≤ 2, the damaging Jahn–Teller (Mn³⁺) effect in Li_xMn_1.5Ni_0.5O₄ electrodes during lithium insertion at 3 V is not as pronounced as it is in Li_xMn₂O₄ electrodes, the average manganese oxidation state at the end of discharge of these two electrodes at x = 2 being 3.3+ and 3.0+, respectively.^25,79,80 However, to date, electrolyte instability at 4.7 V has precluded the commercialization of LiMn_1.5Ni_0.5O₄ spinel electrodes in Li-ion cells.

Li₂O-stabilized MnO₂ structures

Early studies of lithium–manganese-oxide spinel systems exposed the importance of lithia (Li₂O) as a stabilizing component in manganese oxide structures, in which the oxygen ions of the lithia and manganese oxide components form a close-packed oxygen array.³ For example:

(1) Li₂O-stabilized α-MnO₂ with 1-D channels provides significantly superior electrochemical properties relative to metal cation-substituted α-MnO₂ derivatives, such as K⁺-stabilized α-MnO₂.^40,44

(2) Li₂MnO₃ (Li₂O·MnO₂) has a Li₂O-stabilized, layered (2-D) MnO₂ rock-salt configuration, which is electrochemically inactive as an insertion electrode below ∼4.5 V vs. metallic lithium.^3,81

(3) The spinel Li₄Mn₅O₁₂, alternatively 2Li₂O·5MnO₂, with 3-D channels for Li⁺-ion transport, provides superior cycling stability to lithium insertion/extraction at 3 V relative to LiMn₂O₄ by mitigating the Jahn–Teller effect.⁵⁵

Lithia also stabilizes other transition metal oxide electrode materials. For example, V₂O₅ reacts with lithia to form LiV₃O₈ (Li₂O·V₂O₅),^82,83 which markedly improves the structural stability and electrochemical properties (e.g., energy and power) of the electrode system relative to unlithiated V₂O₅, notably in solid state lithium-polymer electrolyte cells that operate close to 100 °C.^84,85 The realization that lithia could be used effectively as a stabilizing component in metal oxides subsequently led to the design of structurally integrated electrodes and lithium-ion systems with enhanced electrochemical properties, particularly those with a Li₂MnO₃ (Li₂O·MnO₂) component.^86,87

Layered lithium manganese oxides and substituted derivatives

Layered manganese oxide structures, such as the minerals birnessite [Na,Ca,Mn(²⁺)]Mn₇O₁₄·2.8H₂O, lithiophorite LiAl₂[Mn(⁴⁺)₂Mn(³⁺)]O₆(OH)₆, and chalcophanite ZnMn(⁴⁺)₃O₇·3H₂O, occur in nature.¹ As the formulae indicate, their structures are stabilized by metal cations, water molecules, and/or hydroxide ions. Prior to the introduction of Li-ion batteries into the market in 1991, no anhydrous layered LiMnO₂ analogue of LiCoO₂ and LiNiO₂ was known.

Li₂MnO₃

In the late 1980s and early 1990s, Rossouw et al. synthesized a layered MnO₂ structure by digesting and leaching lithium from Li₂MnO₃ (Fig. 3a), which can be represented, alternatively, in layered notation as Li[Li_0.33Mn_0.67]O₂.^3,88,89 Acid treatment of Li₂MnO₃ is accompanied first by the exchange of Li⁺ by H⁺ in the lithium layer to yield H_1.5Li_0.5MnO₃ (alternatively, H[Li_0.33Mn_0.67]O₂); during this process, the close-packed oxygen planes (i.e., O3 stacking) shear to trigonal prismatic (P3) stacking.^89–91 Further acid treatment removes both lithium and oxygen from the structure to yield a layered product (H,Li)_2−xMnO_3−x/2 that contains both protons and residual Li⁺ ions.⁸⁹ Attempts to create an anhydrous, layered structure by heating (H,Li)_2−xMnO_3−x/2 have been unsuccessful. The structure transforms to γ-MnO₂ under prolonged acid treatment;⁹² the transformation also occurs when highly delithiated (H,Li)_2−xMnO_3−x/2 products are stored in air.⁹³

Fig. 3 Idealized close-packed structures of (a) layered Li₂MnO₃, (b) layered LiMO₂ (M = Co, Ni, Mn), and (c) cubic spinel LiMn₂O₄.

Chemical re-lithiation of (H,Li)_2−xMnO_3−x/2 with LiI in acetonitrile results in the exchange of protons by lithium ions with concomitant reduction of the manganese ions to yield the composition Li_1.1Mn_0.9O₂;⁸⁹ discounting any remaining protons, this formula can be written, alternatively, as 0.2Li₂MnO₃·0.8LiMnO₂.⁸⁶ Re-lithiation of (H,Li)_2−xMnO_3−x/2 with LiI regenerates the close-packed oxygen array of the parent Li₂MnO₃ structure.

The electrochemical extraction of lithium from Li₂MnO₃ at 4.5 V and above was first reported by Kalyani et al.⁹⁴ and subsequently by the Bruce group.^81,95 The activation of a supposedly ‘inactive’ electrode material was attributed to the loss of oxygen that accompanied lithium extraction at high potentials with a compositional net loss of ‘Li₂O’ (i.e., Li_2−xMnO_3−x/2), possibly facilitated by proton exchange. The loss of Li₂O from Li₂MnO₃ (Li₂O·MnO₂) increases the concentration of the MnO₂ component within the structure, thereby rendering the electrode electrochemically active to lithium insertion and extraction. Electrochemically activated Li₂MnO₃ electrodes do not cycle well, presumably because, like acid-treated Li₂MnO₃ electrodes, the layered structure transforms readily to a spinel-type configuration.⁸¹ The exploitation of oxygen redox reactions in high-capacity Li₂MnO₃-stabilized electrodes is discussed later in this paper.

The preparation of anhydrous layered LiMnO₂ from the isostructural sodium analogue, NaMnO₂, by ion exchange with Li in a non-aqueous solvent was reported in 1996 almost simultaneously by Delmas et al.⁹⁶ and Bruce et al.⁵⁷ Layered LiMnO₂ and substituted derivatives, such as LiMn_2−xCo_xO₂ (0 < x ≤ 0.5), are electrochemically unstable;⁹⁷ on cycling, manganese is displaced from the Mn-rich layers to adjacent lithium-depleted layers, thereby transforming to a structure with spinel-like character, consistent with the electrochemical behavior of Li₂MnO₃.⁸¹ The layered-to-spinel phase transition occurs via the disproportionation reaction:

2Mn³⁺ → Mn⁴⁺ + Mn²⁺

Trivalent Mn³⁺ ions in the octahedral sites of the manganese layer are displaced via tetrahedral sites, where they reside temporarily as Mn²⁺ (d⁵) ions, into octahedral sites of the lithium-depleted layer before being simultaneously re-oxidized to generate localized spinel-like arrangements within a residual layered matrix.^97,98 The displacement of the manganese ions, induced by the lithium extraction process at moderately high potentials, is disadvantageous because the uncontrolled formation of a structure with intermediate layered–spinel character leads to a pronounced hysteresis in the charge/discharge profile.⁹⁷

Layered LiMO₂ (M = Co, Ni, Mn)

The introduction of LiCoO₂ as a viable lithium-ion cathode material resulted in concerted efforts during the 1990s to synthesize layered mixed-metal oxide electrode structures,⁵⁰ such as lithium–cobalt–nickel oxides,^99,100 lithium–manganese–nickel oxides,^101,102 lithium–manganese–cobalt oxides,^103,104 and lithium–manganese–chromium oxides.^105,106 These studies were later expanded to include ternary and higher-order transition metal oxide systems, notably NMC electrodes.

Advances in designing layered electrode structures

At present, NMC materials, LiCoO₂, LiNi_0.80Co_0.15Al_0.05O₂, substituted derivatives of LiMn₂O₄, and olivine LiFePO₄ (LFP),^107–109 which was discovered in 1996, constitute the dominant cathode materials in the lithium-ion battery industry. Substituted Li_1+xMn_2−x−yM_yO₄ (e.g., M = Al) spinel and LFP electrode structures and compositions have been optimized by industry and do not compete with layered metal oxide cathodes from a cell capacity and energy standpoint. The best opportunity for advancing lithium-ion battery cathodes, therefore, rests with layered metal oxides and, in particular, NMC materials. Nickel-rich NMC electrode materials are currently in vogue because they offer high energy and power, but are compromised by thermal instability and safety concerns at high states of charge, while layered Mn-rich NMC electrodes suffer from structural instability and low power, yet remain attractive because of their potential energy, cost, and safety advantages.

Li₂MnO₃-stabilized electrode structures

Lithium extraction from layered LiMO₂ electrode structures, in general, reduces the binding energy between the oxygen layers, thereby destabilizing lithium-deficient materials. Therefore, it is not surprising that phase transitions can be induced by displacing transition metal ions from their layer to sites left vacant by the exiting lithium ions in adjacent layers, or by sliding the close-packed oxygen planes to reduce the free energy of the electrode system.^81,110,111 For Li_1−xCoO₂, these damaging transitions start to occur when x ≈ 0.5,¹¹² which corresponds to a capacity of approximately 140 mA h g⁻¹, whereas the transformation from layered LiMnO₂ to spinel LiMn₂O₄ occurs almost immediately with lithium extraction.⁸¹ More lithium (i.e., x > 0.5) can be extracted from Ni-substituted layered electrodes, such as Li_1−xNi_0.80Co_0.15Al_0.05O₂ and Li_1−xNi_0.0.33Mn_0.33Co_0.33O₂, in which Al and Mn play a stabilizing role, before major structural changes occur, thereby increasing the capacity and energy of the cell.⁵² One concept to suppress phase transitions in layered LiMO₂ cathodes, and thereby enhance cell capacity and energy, is to integrate a structurally compatible, and electrochemically inactive, layered Li₂M′O₃ (M′ = Mn, Ti, Zr) component (Li₂O·M′O₂) with the electrochemically active LiMO₂ component.^{86,87,113–116} These materials are commonly formulated as xLi₂M′O₃·(1 − x)LiMO₂ to emphasize their structural components and composition. At present, systems in which M′ = Mn and M = Ni, Mn, and Co are receiving the most attention. They are also referred to as ‘layered–layered’ electrodes, because it is easy to follow the compositional changes during electrochemical lithium extraction/reinsertion reactions on a xLi₂M′O₃–LiMO₂–M′′O₂ phase diagram (M′′ = MM′).⁸⁶ This design concept mimics the intergrowth structure of γ-MnO₂, as discussed earlier, in which the ramsdellite-MnO₂ component is electrochemically active to lithium insertion/extraction, while the β-MnO₂ component provides structural stability and is essentially inactive to lithium uptake. For this reason, the term ‘composite structure’ has been used to describe ‘layered–layered’ Li₂MnO₃-stabilized LiMO₂ electrode materials.⁸⁷ Note that the ‘layered–layered’ formula xLi₂MnO₃·(1 − x)LiMO₂ can be normalized to the standard layered (rock-salt) notation Li_1+(x/2+x)M′_1+(x/(2+x))O₂ in which M′ = Mn + M or, more simply, Li_1+yM′_1−yO₂ where y = (x/(2 + x)). Although the structures of these layered materials have also been referred to as ‘solid solutions,’ in reality they are inhomogeneous and highly complex.¹¹⁷

Structurally integrated cathode materials are being explored worldwide, particularly those using Li₂MnO₃ as a stabilizing component, integrated either with another structurally compatible layered component, LiMO₂ (M = Mn, Ni, Co), or a LiM′₂O₄ spinel component, or both. Numerous reports of the structural and electrochemical properties of these materials, both experimental and theoretical, have appeared in the literature over the past decade, many of which are cited in the references listed herein. Specifically, readers are referred to recent reviews of these materials by Croy et al.,¹¹⁸ Passerini et al.,¹¹⁹ Zheng et al.,¹²⁰ and Manthiram et al.¹²¹ and to computational studies by Benedek and Iddir,^122–124 Persson et al.,^125,126 and Wolverton et al.¹²⁷ These articles provide guidance in understanding the complex structural and electrochemical features of Li₂MnO₃-based systems.

Structural considerations

Fig. 3 illustrates the idealized and well-known structures of (a) layered Li₂MnO₃, (b) layered LiMO₂ (M = Co, Ni, and Mn), and (c) spinel LiMn₂O₄, i.e., the components under consideration for designing structurally integrated ‘layered–layered’, ‘layered–spinel’, and ‘layered–layered–spinel’ electrode materials. The LiMn₂O₄ spinel structure has an ideal cubic-close-packed (ccp) oxygen array, whereas the layered rock-salt structures Li₂MnO₃ (monoclinic) and LiMO₂ (M = Co,Ni – monoclinic, M = Mn – trigonal) deviate slightly from ideal ccp stacking. In an ideal LiMO₂ layered structure (Fig. 3b), trivalent M and monovalent Li ions occupy the octahedral sites in alternate layers in a 1 : 1 ratio. In Li₂MnO₃, or in layered LiMO₂ notation, Li[Li_1/3Mn_2/3]O₂ (Fig. 3a), lithium ions fully occupy the octahedral sites of one layer and share the octahedral sites in every alternate layer with tetravalent manganese ions in a 1 : 2 ratio. In LiMn₂O₄ (Fig. 3c), tetravalent and trivalent Mn ions (net oxidation state = 3.5+) occupy one-half of the octahedral sites; they are distributed in an ordered configuration in alternate layers in a 3 : 1 ratio, thus imparting some layered character to the spinel structure.

Fig. 4a shows the building block of Li₂MnO₃, which has been proposed to play a key role in stabilizing layered^86,87,116 and spinel^128–130 electrode structures. The lithium and manganese ions are each bonded to six oxygen ions in octahedral coordination and are ordered such that each lithium ion in the manganese layer has six nearest neighbor manganese ions, as depicted by the LiMn₆ configurations in Fig. 4b. Fig. 4c is an idealized representation of the cation distribution in the transition metal layers of a 0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂ electrode structure.¹³¹ These layers consist of an interconnected array of LiMn₆ and NiMn₆ units, which represent the Li₂MnO₃ and LiMn_0.5Ni_0.5O₂ components of the structure, respectively (Fig. 4c). Note that in 0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂, the manganese ions are all tetravalent while the nickel ions are divalent, as they are in layered LiMn_0.5Ni_0.5O₂ alone¹³² and in the spinel LiMn_1.5Ni_0.5O₄.¹³³ Below ∼4.5 V, all the electrochemical capacity of the electrode is derived from Ni²⁺ ↔ Ni⁴⁺ redox reactions, while the tetravalent manganese ions act as ‘spectator’ ions and stabilizing agents. By contrast, in 0.5Li₂MnO₃·0.5LiCoO₂, the Li₂MnO₃ and LiCoO₂ components tend to phase segregate during synthesis,¹³⁴ forming interconnected nanodomains in a complex mosaic pattern.¹³⁵

Fig. 4 (a) Characteristic LiMn₆ unit of Li₂MnO₃, (b, c) projections of atomic configurations in the transition-metal rich layers of ideal Li₂MnO₃ and 0.5Li₂MnO₃·0.5LiNi_0.5Mn_0.5O₂ structures, respectively.

‘Layered–layered’ electrodes: the Li₂MnO₃–LiMO₂–MO₂ system

A compositional phase diagram of the Li₂MnO₃–LiMO₂–MO₂ (M = Mn, Ni, Co) system is shown in Fig. 5. The Li₂MnO₃–LiMO₂ tie-line represents the compositions of structurally integrated ‘layered–layered’ xLi₂MnO₃·(1 − x)LiMO₂ electrode materials described in the preceding section. Lithium extraction from these materials occurs first with a concomitant oxidation of the transition metal (M) cations in the LiMO₂ component. During this reaction, the composition of the electrode follows the direction of the arrows away from the xLi₂MnO₃·(1 − x)LiMO₂ tie-line. For a standard LiCoO₂ electrode (i.e., x = 0, M = Co), lithium extraction follows the black arrow along the LiMO₂–MO₂ tie-line, which represents a continually changing electrode composition, e.g., Li_1−xCoO₂, as x increases. When charged to 4.3 V, approximately one-half of the lithium ions can be extracted reversibly from LiCoO₂, which corresponds to a capacity of ∼140 mA h g⁻¹, before the onset of structural perturbations and reactions with the electrolyte.^111,112 Slightly higher reversible capacities can be achieved from NMC electrodes when a small amount of additional lithium is added to the electrode composition.¹³⁶ These lithium-rich and manganese-containing electrodes can be represented by the notation xLi₂MnO₃·(1 − x)LiMO₂, because the excess lithium together with some manganese can constitute the Li₂MnO₃ component of the electrode structure.⁸⁷ Lithium extraction from slightly lithium-rich NMC electrodes follows the red reaction path on the Li₂MnO₃–LiMO₂–MO₂ phase diagram in Fig. 5; these electrodes yield a slightly higher capacity than LiCoO₂, typically ∼160 mA h g⁻¹ when charged to 4.3 V.¹³⁶ In this instance, the Li₂MnO₃ component is electrochemically inactive and acts as a spectator ion and stabilizing agent, allowing extra capacity (i.e., lithium) to be cycled during charge and discharge relative to stoichiometric (LiMO₂) NMC electrodes. Lithium-rich NMC electrodes demonstrate an enhanced rate capability, which is dependent on the relative amounts of nickel, manganese, and cobalt.^137,138

Fig. 5 Li₂MnO₃–LiMO₂–MO₂ compositional phase diagram (adapted from Thackeray et al.^86,87).

The amount of inactive Li₂MnO₃ in the xLi₂MnO₃·(1 − x)LiMO₂ electrode increases with x. Therefore, when charged to 4.3 V, electrodes with higher values of x provide a lower capacity as indicated by the short reaction path for x = 0.3, highlighted in green, for complete lithium extraction from the LiMO₂ component, relative to the LiMO₂-to-MO₂ reaction (i.e., x = 0), highlighted in black. However, for electrodes with x = 0.3, if charged above 4.5 V, lithium can be extracted from the Li₂MnO₃ component with a concomitant loss of oxygen (net loss Li₂O), which drives the composition of the electrode towards the MO₂ apex of the phase diagram, thereby increasing the concentration of an electrochemically active MnO₂ component in the structure, and hence the capacity of the electrode.^87,139–141 The reaction is complex, particularly during the early break-in cycles above 4.5 V. Not much gaseous oxygen is observed during lithium extraction at such high potentials,¹⁴² which implies that some undetected oxygen may react immediately with, and oxidize, the electrolyte at the particle surface.¹²¹ In studies of lithium-rich metal oxide electrodes at high potentials, hybridization of the transition metal d and oxygen 2p bands has been given as a reason for the oxygen redox activity.^121,143,144 Recently, Ceder et al. proposed that the oxygen redox capacity is also dependent on the need for particular Li–O–Li configurations within these structures.¹⁴⁵ The anomalously high capacity (∼300 mA h g⁻¹) that can be obtained from fully activated electrodes is significantly higher than the expected theoretical value for transition metal redox reactions alone (∼260 mA h g⁻¹).¹¹⁴ These high capacity reactions involving oxygen are partially reversible;^146,147 however, the capacity diminishes on extended cycling, consistent with progressive oxygen loss, to values that can be attributed solely to redox reactions on the transition metal ions.

Lithium- and manganese-rich ‘layered–layered’ electrode structures have high capacity but are unstable when cycled repeatedly to high potentials (>4.5 V).^{114,119–121,148,149} This inherent instability is not surprising, given that layered LiMnO₂ electrodes convert to spinel-like configurations, a process that involves the migration of manganese from the transition-metal layers to the lithium-depleted layers with the evolution of the characteristic 4 V and 3 V electrochemical signatures of a LiMn₂O₄ spinel electrode. Lithium- and manganese-rich xLi₂MnO₃·(1 − x)LiMO₂ electrodes, such as 0.5Li₂MnO₃·0.5LiMn_0.375Ni_0.375Co_0.250O₂, exhibit a similar ‘voltage fade’ towards 3 V on cycling (Fig. 6) – a result of interlayer metal migration that generates a structure with a lower Gibbs free energy relative to the parent layered material. When activated at high potentials, the electrochemical properties of these complex mixed-metal oxides are also highly dependent on the electrode composition, i.e., the Ni : Mn : Co ratio. For example, Long et al.¹³⁰ at Argonne National Laboratory have shown that the electrochemical properties of 0.25Li₂MnO₃·0.75LiMn_yNi_yCo_1−2yO₂ electrodes are strongly dependent on the value of y. The cycling protocol typically adopted at Argonne to evaluate ‘layered–layered’ compositions and structures is to subject lithium cells initially to one formation cycle between 4.6 and 2.0 V and, thereafter, to continuous cycling between ∼4.45 and ∼2.50 V.¹³⁰ Capacity vs. cycle number plots and corresponding dQ/dV plots of a Li/0.25Li₂MnO₃·0.75LiMn_yNi_yCo_1−2yO₂ cell are provided in Fig. 7a–c and d–f, respectively. The data clearly show enhanced cycling capacity and stability, as well as the suppression of voltage fade, when the electrode has a relatively high manganese and nickel content (y) and a low cobalt content (1 − 2y).

Fig. 6 (a) Voltage fade profiles and (b) corresponding dQ/dV plots during cycling of a Li/0.5Li₂MnO₃·0.5LiMn_0.375Ni_0.375Co_0.250O₂ cell.

Fig. 7 Electrochemical cycling stability (top) and dQ/dV plots (bottom) of ‘layered–layered’ 0.25 Li₂MnO₃·0.75LiMn_yNi_yCo_1−2yO₂ electrodes in lithium cells: (a) y = 0.125, (b) y = 0.250, and (c) y = 0.375 (reproduced with permission from J. Electrochem. Soc., Long et al.¹³⁰ Copyright 2014, The Electrochemical Society).

Although synthesizing high-quality xLi₂MnO₃·(1 − x)LiMO₂ electrode materials reproducibly can be difficult, Fig. 8 highlights the electrochemical stability of a Li/0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂ cell after activating the cathode at 4.6 V for several cycles to induce voltage fade.¹⁵⁰ In this instance, voltage vs. capacity plots of cycles 15 to 25 (i.e., after electrochemical activation) and corresponding dQ/dV plots of this cell (Fig. 8a and b, respectively), cycled continuously between 4.6 and 2.0 V, show that the charge and discharge reactions are remarkably reversible during these early cycles. However, (1) the Ni reduction peak at 3.7 V is significantly weaker than it is for the relatively stable 0.25Li₂MnO₃·0.75LiMn_yNi_yCo_1−2yO₂ electrode in Fig. 7c, which was charged to a slightly lower voltage limit (4.45 V), and (2) most of the discharge occurs below 3.5 V. This voltage fade, which causes asymmetry in the charge and discharge profiles because of transition metal migration during lithium insertion/extraction reactions, is disadvantageous for several reasons:^151–153

Fig. 8 Cycling stability of Li/0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂ cell (Mn : Ni ratio = 3 : 1): (a) voltage vs. capacity plots (cycles 15, 20, and 25) and (b) corresponding dQ/dV plot.

(1) it causes hysteresis in the charge/discharge process;

(2) it reduces the energy output of the cell;

(3) it increases cell impedance at both high and low states of charge, thereby lowering lithium diffusion rates (power); and

(4) it compromises the management of cells and batteries.

Despite these negative attributes, high-capacity ‘layered–layered’ xLi₂MnO₃·(1 − x)LiMO₂ electrodes continue to receive worldwide attention in attempts to find ways to mitigate their limitations, particularly the suppression or elimination of transition-metal migration during electrochemical reactions when the electrodes are charged to 4.5 V and above.

Configurational considerations

The complex interconnected nanodomains observed in 0.5Li₂MnO₃·0.5LiCoO₂ can be addressed in simple terms by using Pauling's principle of electrostatic valence. Consider the oxygen coordination in LiCoO₂ and Li₂MnO₃. In LiCoO₂, each oxygen is surrounded by three Li⁺ and three Co³⁺ ions. Given that all cations are octahedrally coordinated by six oxygen ions, and each oxygen ion is correspondingly coordinated by six cations, and following Pauling's principle, the Li⁺–O²⁻ and Co³⁺–O²⁻ bonds will have formal valences of and , respectively. For each oxygen, the lithium layer contributes summed bond valence while the Co layer contributes . This leads to a total formal bond valence of +2, which charge balances the formal valence of O²⁻. The situation with Li₂MnO₃ is similar except that the total LiMn₂ metal layer bond valence of is the sum of . This can be expressed in the rule that the sum of the formal valence of the cation triangles in the metal layer should equal 9 for charge balancing between nearest neighbor cations and anions.

To investigate cation ordering in the metal layer, a simulated annealing cation-swapping optimization algorithm, which minimizes the squared difference between 9 and the sum of every metal-triangle bond valence, was developed.¹⁵⁴ For a given chemical composition, cations are initially randomly placed on an extended two-dimensional hexagonal lattice. The cations are then swapped until a minimum is reached in the bond-valence cost function. For Li₂MnO₃, the signature honeycomb configuration is rapidly obtained (Fig. 4b). The results for 0.5Li₂MnO₃·0.5LiCoO₂ are presented in Fig. 9a, which shows nanodomain structures that are similar to those observed by Bareño et al.¹³⁵ Reducing the fraction of Li₂MnO₃ in xLi₂MnO₃·(1 − x)LiCoO₂ to x = 0.3 results in the formula Li(Li_0.1Mn_0.2Co_0.7)O₂; performing the cation-swapping optimization also produces segregated nanodomains (Fig. 9b). This segregation appears to be driven principally by the strong electrostatic preference for Li⁺ to be coordinated by two Mn⁴⁺ ions. Despite the simplicity of the cation-swapping algorithm, its ability to model 0.5Li₂MnO₃·0.5LiCoO₂ and both end members Li₂MnO₃ and LiCoO₂ suggests that it provides a reasonable description of the mesoscale structures of these materials and a tool to probe the configurational complexity of Li₂MnO₃-containing electrode materials.

Fig. 9 Theoretical potential topological configurations of the mesoscale structure of (a) 0.5Li₂MnO₃·0.5LiCoO₂ and (b) Li(Li_0.1Mn_0.2Co_0.7)O₂ indicating phase segregated regions of LiCoO₂ and Li₂MnO₃. Li⁺, Co²⁺ and Mn⁴⁺ ions are illustrated as yellow, blue and pink spheres, respectively. The larger transparent yellow and green spheres respectively indicate formal charge excess and charge deficiency of the oxygen ion associated with the cation triangle. All other oxygen ions above and below the metal layer have the correct formal charge.

‘Layered–layered–spinel’ (LLS) electrodes: the Li₂MnO₃–LiMO₂–LiM′₂O₄ system

As stated above, the structural instability and voltage fade of ‘layered–layered’ xLi₂MnO₃·(1 − x)LiMO₂ electrodes, when charged to potentials greater than ∼4.5 V vs. Li⁰, have been attributed predominantly to the migration of transition metals into the lithium-depleted layers during electrochemical cycling. In an attempt to address this limitation, a strategy has been adopted to fabricate xLi₂MnO₃·(1 − x)LiMO₂ electrodes in which voltage fade is built into the electrode structures during synthesis, i.e., by designing an electrode with stabilizing transition metal ions in the lithium layers of the parent material.^130,150 A seemingly logical way to achieve this is to add a spinel component to the composite ‘layered–layered’ structure because, unlike the two layered components that have discrete layers of lithium and transition metal ions, the transition metal cations in a LiM₂O₄ spinel component are arranged in a 3 : 1 ratio in alternate layers between ccp oxygen planes. The overall objective of the strategy, therefore, is to embed a spinel component to stabilize a ‘layered–layered’ structure, which is simply an extension of the concept to use a Li₂MnO₃ component to stabilize the LiMO₂ component in an xLi₂MnO₃·(1 − x)LiMO₂ electrode system.^86,129

Electrochemical delithiation of layered LiMnO₂ causes an internal transformation to a spinel-like structure^57,81,98 such that the spinel and layered components are structurally integrated with one another in what has been defined colloquially as a ‘sp layered’ configuration.¹⁵⁵ Note that such configurations have also recently been observed by high-resolution transmission electron microscopy (TEM) when LiMn₂O₄ spinel electrodes are charged to high anodic potentials.¹⁵⁶ In earlier studies of composite lithium–manganese-oxide materials, Rossouw et al. highlighted the close similarity of the XRD patterns of Li₂MnO₃ and lithia-stabilized MnO₂ materials in the Li₂O·yMnO₂ (y ≥ 2.5) system, such as Li₄Mn₅O₁₂ (y = 2.5) and Li₂Mn₄O₉ (y = 4).^40,157 This finding indicates that the interlayer spacing of the ccp oxygen planes is comparable in Li₂MnO₃ and spinel structures in which all the manganese ions are tetravalent, suggesting that these materials would be conducive to structural integration. Indeed, a subsequent high-resolution TEM study of the Li₂MnO₃–Li_1+xMn_2−xO₄ ‘layered–spinel’ system revealed the remarkable structural intergrowth of Li₂MnO₃ and Li₄Mn₅O₁₂ (x = 0.33) nanodomains in a 0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂ sample synthesized at 400 °C, and a seemingly perfect alignment of the close-packed 001 and 111 oxygen planes of the layered and spinel components, respectively (Fig. 10a).¹²⁸ XRD patterns showed that heating xLi₂MnO₃·(1 − x)Li₄Mn₅O₁₂ materials to high temperatures in air drives the composition of the spinel component, by oxygen loss, towards LiMn₂O₄ (x = 0) in the Li–Mn–O phase diagram (Fig. 2). During this process, the cubic lattice parameter of the spinel component, a, expands with increasing Mn³⁺ concentration, as expected (Fig. 10b). The changing Li : Mn ratio in the spinel component along the Li₄Mn₅O₁₂–LiMn₂O₄ tie-line is accommodated during heat treatment by adjusting the amount of Li₂MnO₃ in the final composite structure.¹²⁸ Electrochemical charge and discharge profiles of a Li/0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂ half-cell confirmed the dual layered-spinel character of the cathode material. When cycled between 5.0 and 2.0 V, the capacity delivered during the initial discharge is 270 mA h g⁻¹ (Fig. 10c).¹²⁸ On subsequent cycling, the electrode continued to provide more than 250 mA h g⁻¹, but lost capacity steadily, an unsurprising result given the high cutoff voltage. Note that the charge capacity on the initial cycle (252 mA h g⁻¹ in Fig. 10c), which includes some electrolyte oxidation at high potentials, is smaller than the discharge capacity (270 mA h g⁻¹) because the Li₄Mn₅O₁₂ component in the parent electrode is in a fully charged state and, therefore, can accommodate lithium from the metallic lithium anode to yield the fully lithiated spinel composition (Li₇Mn₅O₁₂) during discharge at ∼3 V. In a full cell configuration, C₆/0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂, lithium from the Li₂MnO₃ component can, therefore, serve to load the graphite anode during the electrochemical activation process, thereby countering or even eliminating first-cycle irreversible capacity losses.¹⁵⁸

Fig. 10 (a) TEM image of 0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂ prepared at 400 °C, (b) XRD patterns of 0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂ (Li : Mn = 1.2 : 1) as a function of synthesis temperature, and (c) voltage profile of a Li/0.7Li₂MnO₃·0.3Li₄Mn₅O₁₂ (750 °C) cell on initial charge–discharge cycle (reproduced with permission from Electrochemical Communications, Johnson et al.¹²⁸).

A compositional phase diagram of a three-component Li₂MnO₃–LiMO₂–LiM′₂O₄ (‘layered–layered–spinel’) system shown in Fig. 11 highlights the wide compositional space that exists to exploit structurally integrated layered and spinel electrode materials. The Li₂MnO₃–LiMO₂ tie-line reflects the composition of ‘layered–layered’ electrodes described in the preceding sections. Reducing the lithium content in the parent Li₂MnO₃–LiMO₂ electrode material necessarily drives the composition of the electrode within the tie-triangle towards the spinel apex of the triangle. This strategy has been exploited to induce the formation of a stabilizing spinel component within layered or ‘layered–layered’ electrode structures during synthesis at elevated temperature.^129,130

Fig. 11 Conceptual design space of a Li₂MnO₃ (layered)–LiMO₂ (layered)–LiM′₂O₄ (spinel) compositional phase diagram highlighting the ‘layered–layered–spinel’ tie-line (in red) between 0.5Li₂MnO₃·0.5LiMO₂ and LiM′₂O₄ and a region within the phase diagram representing low concentrations of stabilizing Li₂MnO₃ and LiM′₂O₄ components (in blue).

For example, reducing the amount of lithium required to synthesize Li₂MnO₃ and heating lithium-deficient precursors, Li_2−xMnO_δ (0 ≤ x ≤ 1.5), at 850 °C in air induces the formation of LiMn₂O₄, which requires some oxygen loss (δ) for charge compensation, as shown by XRD patterns (Fig. 12a) and the electrochemical profiles of Li/Li_2−xMnO_δ cells for x = 0, 0.5, 1.0, and 1.5 (Fig. 12b)¹⁵⁹ (note that this system can be represented, alternatively, as yLi₂MnO₃·(1 − y)LiMn₂O₄ for y = 0, 0.2, 0.5, and 1.0, respectively). For x = 1 (i.e., 0.5Li₂MnO₃·0.5LiMn₂O₄), a TEM image of the product shows structurally integrated Li₂MnO₃ and LiMn₂O₄ domains, as expected, with the spinel component residing predominantly at the surface of the particles (Fig. 12c). Unlike the TEM image in Fig. 10a, which shows a Li₂MnO₃–Li₄Mn₅O₁₂ composite structure synthesized at 400 °C with near perfect alignment of the close-packed planes of the two components, the image of the Li₂MnO₃–LiMn₂O₄ product in Fig. 12c shows a convoluted structure, which can likely be attributed to a mismatch in the lattice parameters and d-spacings of the layered Li₂MnO₃ and spinel LiMn₂O₄ components and to stacking fault disorder.

Fig. 12 Effects of reducing the Li content in Li_2−xMnO_δ materials after heating to 850 °C: (a) XRD data, (b) electrochemical data, and (c) TEM image of Li_2−xMnO_δ (x = 1.0).

Reducing the lithium content in the ‘layered–layered’ composition, 0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂ (in normalized notation Li_1.5Mn_0.75Ni_0.25O_2.5, or Li_1.2Mn_0.6Ni_0.2O₂), shows similar layered–spinel behavior as the Li_2−xMnO_δ system, described above. Increasing the lithium deficiency drives the composition of the electrode during high temperature synthesis towards the spinel composition LiMn_1.5Ni_0.5O₄ because the Mn : Ni ratio (3 : 1) remains constant on the 0.5Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂–LiMn_1.5Ni_0.5O₄ tie-line.¹²⁹ X-ray diffraction data and electrochemical profiles confirm increasing spinel-like character as the Li : transition metal ratio in the parent ‘layered–layered’ electrode is reduced from x = 1 to x = 0 in x(Li₂MnO₃·0.5LiMn_0.5Ni_0.5O₂)·(1 − x)LiMn_1.5Ni_0.5O₄.¹²⁹

A study by Long et al. of the compositional, structural, and electrochemical features of spinel-stabilized ‘layered–layered’ xLi₂MnO₃·(1 − x)LiMO₂ (M = Mn, Ni, Co) electrodes has revealed intriguing features of ‘layered–layered–spinel’ electrode materials.¹³⁰ Compounds in the xLi₂MnO_3·(1 − x)LiMn_yNi_yCo_1−2yO₂ system for x = 0.25 and x = 0.33 were selected for this study. The cobalt content, 1 − 2y, was varied over the range 0.125 ≤ y ≤ 0.375 while keeping a constant Mn : Ni ratio of 1 : 1. ‘Layered–layered–spinel’ products with a targeted spinel content of 6%, 15%, and 25% were synthesized from metal oxalate precursors by reducing the Li content in a parent ‘layered–layered’ composition by the appropriate amount. Fig. 13 shows, for x = 0.25 and y = 0.375, the impact of reducing the lithium content on the electrochemical capacity (in mA h g⁻¹) after an initial ‘activation’ charge to 4.6 V and subsequent discharge to 2.0 V. Also shown is the first-cycle efficiency. The capacity of the parent 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ electrode (in normalized notation, Li_1.25Mn_0.531Ni_0.281Co_0.187O_2.25) increases from 208 to 212 mA h g⁻¹ when the amount of lithium in the parent electrode is reduced to target a 6% spinel content. Increasing the targeted spinel content results in a decline in capacity, but it increases the first-cycle efficiency because, unlike a discharged layered LiMO₂ component, a charged spinel component (e.g., Li_1+xMn_2−xO₄) in the parent electrode can accommodate one lithium ion per formula unit. This electrochemical behavior is consistent with reports of other lithium- and manganese-rich ‘layered–layered–spinel’ electrode compositions.^129,130,160

Fig. 13 The effect of varying the lithium content in 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂, i.e., x in Li_xMn_0.531Ni_0.281Co_0.188O_δ (reproduced with permission from J. Electrochem. Soc., Long et al.¹³⁰ Copyright 2014, The Electrochemical Society).

With the above-mentioned electrochemical data in mind, the current strategy at Argonne National Laboratory is to use a bottom-up strategy to develop ‘layered–layered–spinel’ materials that can deliver a stable capacity of at least 200 mA h g⁻¹ when cycled below 4.5 V, which would make them competitive with layered Ni-rich systems.¹⁵⁰ Approaches are being sought to overcome electrode surface and electrolyte instabilities and capacity fade (for example, by using effective surface coatings), which would allow ‘layered–layered–spinel’ electrodes, when charged repeatedly to 4.6 V, to deliver close to their theoretical capacity based on the redox reactions of the transition metal ions alone (∼250 mA h g⁻¹).

The impact of lowering the lithium content in ‘layered–layered’ electrodes on their capacity, cycling stability, and rate performance is highlighted in Fig. 14.¹⁵⁰ For these experiments, lithium-deficient 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ materials were prepared at Argonne's Materials Engineering and Research Facility to target a spinel content of 2, 5, 10, and 15% in the ‘layered–layered–spinel’ products. Lithium coin cells containing these materials were activated on the initial charge/discharge cycle (4.6–2.0 V) and subsequently cycled between 4.45 and 2.0 V at a low rate (15 mA g⁻¹; ∼C/12). Fig. 14a shows that a maximum capacity is obtained when the targeted spinel content is 5–10%, consistent with the capacity vs. composition plot of cells with 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ electrodes (Fig. 13). Fig. 14a also highlights the large decline in capacity for electrodes with 15% spinel content, and the poor capacity retention on cycling for electrodes with 2% spinel content. Of particular significance is that electrodes with 5–10% spinel are also more tolerant to higher rates than the 2% spinel electrode (Fig. 14b). This finding suggests that the spinel component may reside predominantly at the surface of the ‘layered–layered–spinel’ electrode particles, thereby stabilizing the electrode/electrolyte interface and providing good capacity retention on long-term cycling (275 cycles, Fig. 14c).

Fig. 14 (a) Cycling stability (40 cycles) and (b) rate performance of ‘layered–layered–spinel’ electrodes derived from a 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ parent compound by reducing the lithium content to target a spinel content of 2%, 5%, 10%, and 15%; (c) cycling stability (275 cycles) of an electrode with a targeted 5% spinel content.

As shown in Fig. 15a and b, high-resolution (synchrotron) XRD and TEM data of a ‘layered–layered–spinel’ product with a targeted 15% spinel content, derived from a lithium-deficient 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ parent composition, provided unequivocal evidence of a spinel component embedded within the layered structure (Fig. 15a and b).¹³⁰ Of particular significance is that elemental mapping of the sample (Fig. 15c) by electron energy loss spectroscopy (EELS) indicated that regions rich in Mn and Ni adopt a spinel configuration, while regions rich in cobalt have a layered configuration. The data emphasize the complexity and inhomogeneity of these materials and the tendency to form thermodynamically favored configurations or to phase segregate on cooling the samples from the high temperature at which they are synthesized (typically 850–900 °C) to room temperature, consistent with data reported by Bareno et al.¹³⁵ and Long et al.¹³⁴ for ‘layered–layered’ electrodes in the Li₂MnO₃–LiCoO₂ system. A question that remains to be answered is: when synthesizing ‘layered–layered–spinel’ structures, at what lithium concentration do the transition metal ions start diffusing into the lithium-rich layers to provide a spinel character to the electrode structure, rather than maintain a layered configuration in which the lithium deficiency is compensated by oxidation state changes on the transition metal ions? In this respect, recent structural refinements of lithium-deficient 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ electrodes with synchrotron XRD data have revealed that reducing the lithium content does not immediately induce spinel formation. The lattice parameter of the cubic spinel component that ultimately forms is 8.145 Å, which is coincidentally close to that expected for lithium–manganese-oxide spinels, such as Li₄Mn₅O₁₂,⁵⁵ and probably cobalt- and/or nickel-substituted derivatives.¹⁶¹

Fig. 15 (a) High-resolution XRD pattern, (b) TEM image, and (c) EELS data of a lithium-deficient 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ electrode targeting 15% spinel content (reproduced with permission from J. Electrochem. Soc., Long et al.¹³⁰ Copyright 2014, The Electrochemical Society).

Coatings and surface treatments

The formation of protective solid-electrolyte interphase (SEI) layers at the surface of lithium-ion battery electrodes that function outside the stability window of organic electrolyte solvents, particularly lithiated graphite (LiC₆) anodes that operate at a potential of ∼100 mV above metallic lithium, is a well-known and studied phenomenon;¹⁶² without it, lithium-ion cells would not work. Unprotected lithium–metal-oxide cathodes that are electrochemically active above the upper stability limit of the electrolyte (∼4 V vs. Li⁰) also tend to suffer from interfacial electrode/electrolyte side reactions. These reactions can lead to structural decay, metal dissolution, oxygen loss, lower capacity, a slower electrochemical reaction rate, and reduced operating life, all of which can severely compromise the energy and power output of a lithium-ion cell. Numerous materials have been tried and tested as protective coatings for lithium–metal-oxide cathodes,¹⁶³ many of which, for example, Al₂O₃,^73,164 AlF₃,¹⁶⁵ ZrO₂,¹⁶⁶ Li₂ZrO₃,¹⁶⁷ AlPO₄,^168,169 LiNiPO₄,¹⁷⁰ and graphene,⁷⁷ can counter the above limitations to various extents.

‘Layered–layered–spinel’ electrodes derived from a parent composition 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂, when activated between 4.6 and 2 V, and subsequently cycled between 4.45 and 2.5 V at 15 mA g⁻¹, also benefit from surface treatment. Fig. 16 shows the electrochemical charge and discharge profiles of the first 10 cycles (top) and the 40^th to 50^th cycles (bottom) of lithium half cells with (a) a layered–layered 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ electrode, (b) an untreated ‘layered–layered–spinel’ electrode with a targeted spinel content of 10% (based on inductively coupled plasma atomic emission spectroscopy (ICP-AES) analysis of the Li content), and (c) the ‘layered–layered–spinel’ electrode in (b) after a proprietary surface treatment. Despite several formation cycles required to stabilize the electrode reaction (Fig. 16a–c, top), all cells cycle thereafter with excellent capacity retention, with maximum capacity (∼215 mA h g⁻¹) being generated by the surface-treated ‘layered–layered–spinel’ electrode (Fig. 16c, bottom). Note the distinct discontinuity (arrows) in the electrochemical profiles of cells containing the ‘layered–layered–spinel’ electrodes, tentatively attributed to structural and/or polarization effects, which could serve as an early end-of-life indicator for the cell.

Fig. 16 Electrochemical cycling stability of untreated and surface-treated ‘layered–layered’ and ‘layered–layered–spinel’ electrodes derived from a parent composition 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂.

Corresponding dQ/dV plots (Fig. 17) of the voltage–capacity data in Fig. 16 confirm the excellent cycling stability of untreated and surface-treated ‘layered–layered’ and ‘layered–layered–spinel’ electrodes derived from 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂. What is noteworthy about the dQ/dV plots is:

Fig. 17 Corresponding dQ/dV plots of the voltage–capacity plots in Fig. 16, highlighting the electrochemical cycling stability of untreated and surface-treated ‘layered–layered’ and ‘layered–layered–spinel’ electrodes derived from 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂.

(1) there is significantly less hysteresis and voltage fade than for the cell containing a ‘layered–layered’ cathode with higher lithium and manganese content, 0.5Li₂MnO₃·0.5LiMn_0.375Ni_0.375Co_0.250O₂ (Fig. 6);

(2) the Ni redox peak at approximately 3.7 V remains stable; and

(3) the small amount of capacity generated by voltage fade during the early cycles is delivered over a wide voltage range (approximately 3.4 to 2.7 V), implying that lithium insertion takes place in a highly complex ‘layered–spinel’ structure in which the interstitial site energies vary widely.

Nickel-rich layered electrode materials

Nickel-rich, layered electrodes, such as NCA (LiNi_0.8Co_0.15Al_0.05O₂) and NMC analogues LiNi_0.8Mn_0.1Co_0.1O₂ (‘811’) and LiNi_0.6Mn_0.2Co_0.2O₂ (‘622’), are currently considered next generation lithium-ion cathode materials.^54,171 Gradient cathode designs, in which a high concentration of nickel decreases from the bulk to the surface of NMC particles are also being developed.¹⁷² However, the instability of tetravalent nickel and the possible release of oxygen in these electrodes at the top of charge introduce a safety risk, which can be reduced by increasing the manganese content. The electrochemical performance of a commercial, moderately nickel-rich NMC electrode, LiNi_0.5Mn_0.3Co_0.2O₂ (‘532’), was therefore compared with a moderately manganese-rich LLS ‘MNC’ electrode to assess their relative electrochemical behavior, the latter product being derived from a ‘layered–layered’ composition 0.25Li₂MnO₃·0.75LiMn_0.375Ni_0.375Co_0.250O₂ with a targeted 10% spinel content in which the Mn : Ni : Co ratio is 5.3 : 2.8 : 1.9, i.e., close to ‘532’ (Fig. 18).¹⁷³ In these tests, the nickel-rich Li/NMC cells were charged and discharged continuously between 4.45 and 2.5 V, whereas the manganese-rich Li/MNC cells (referenced as LLS in Fig. 18) were subjected to one activation cycle between 4.6 and 2.0 V, before continuous cycling between 4.45 and 2.5 V. Cells were cycled at a 15 mA g⁻¹ rate and at 30 °C.

Fig. 18 Electrochemical profiles and properties of a commercial, nickel-rich layered ‘532’ NMC electrode with a manganese-rich ‘layered–layered–spinel’ MNC (532) electrode with targeted spinel content of ∼8%.

Fig. 18a shows the first-cycle electrochemical profiles of the Li/NMC and Li/MNC (LLS) cells described above. The voltage profile of the NMC cell is consistent with a layered nickel-rich cathode structure, delivering approximately 190 mA h g⁻¹ on the initial discharge. By contrast, the initial discharge capacity of the manganese-rich LLS cathode is considerably higher (238 mA h g⁻¹), the additional capacity below 3.5 V being generated by the spinel-like component within the composite electrode structure. The electrochemical profiles of the Li/NMC and Li/MNC for cycles 2 and 50 are shown in Fig. 18b and c, respectively. Despite the more pronounced voltage fade of the manganese-rich LLS electrode, Li/MNC cells outperformed their nickel-rich NMC counterpart, both in terms of capacity and energy output. More specifically, the manganese-rich LLS electrodes delivered more than 200 mA h g⁻¹ and an energy output of more than 750 W h kg⁻¹ (based on the mass of the cathode alone) after 50 cycles, whereas the layered nickel-rich cathodes provided approximately 180 mA h g⁻¹ and less than 700 W h kg⁻¹, respectively. This result augurs well for further progress in exploiting manganese-rich LLS technology.

New approaches and future directions

Exploiting lithiated spinels as stabilizers for Li₂MnO₃–LiMO₂ electrodes

Despite the advances that have been made in developing manganese-rich LLS electrode materials, Fig. 18 indicates that a LLS MNC ‘532’ electrode with 50% Mn still suffers from some structural decay and voltage fade on long term cycling. These limitations are thought to arise because MNC ‘532’ electrodes are not as effective in arresting transition metal migration as their layered nickel- and cobalt-rich counterparts. A recent strategy that has been adopted at Argonne to address this challenge is to use a cobalt-based lithiated spinel composition Li₂[Co_2−2xM_2x]O₄ (e.g., M = Ni, Al), alternatively LiCo_1−xM_xO₂, rather than a stoichiometric, manganese-rich spinel, to stabilize high-capacity ‘layered–layered’ xLi₂MnO₃·(1 − x)LiMO₂ electrodes.¹⁷⁴ This approach is attractive for several reasons:

(1) Like xLi₂MnO₃·(1 − x)LiMO₂ materials, lithiated spinels Li₂[Co_2−2xM_2x]O₄ have close-packed structures with a rock-salt stoichiometry, making them compositionally, and potentially structurally, compatible with one another.

(2) Relative to manganese and nickel, cobalt has a lower propensity to migrate in a ccp oxygen lattice,^103,175 thereby offering the possibility of mitigating voltage fade.

(3) Lithium extraction from a lithiated cobalt-rich spinel component, Li_2−δ[Co_2−2xM_2x]O₄, occurs at a significantly higher potential (∼3.6 V) than a lithiated manganese-oxide spinel analogue, Li_2−δMn₂O₄ (∼2.9 V).

Limited research has been conducted on lithiated cobalt spinel materials, and substituted derivatives, since their discovery in the early 1990s.^59,176 They are synthesized at a relatively low temperature (∼400 °C). The lithiated spinel Li₂[Co₂]O₄ (or simply LiCoO₂) has cubic symmetry (Fd3̄m) while its layered analogue has trigonal symmetry, R3̄m (Fig. 19). Some of these early structural analyses of Li₂[Co₂]O₄ were misleading because it had not been recognized that the atomic vector space of a cubic lithiated spinel structure is identical to that of a (hypothetical) layered LiCoO₂ structure in which the oxygen array is ideally cubic-close-packed.^60,61,177 In such an instance, the XRD patterns of a cubic lithiated spinel, Li₂[Co₂]O₄, and a cubic layered LiCoO₂ structure would be indistinguishable from one another, as shown in Fig. 19b and c.⁶⁰ In practice, however, layered LiCoO₂ deviates slightly from ideal cubic close packing, yielding a c/a ratio of 4.99, whereas cubic Li₂[Co₂]O₄ has the ideal c/a ratio of 4.90 for a ccp structure;¹⁷⁷ this difference separates the trigonal and cubic diffraction peaks from one another. Recent synchrotron XRD data and analyses have, however, confirmed the earlier conclusions that LiCoO₂ prepared at 400 °C (LT-LiCoO₂) has both layered- and lithiated-spinel character, and that 10% nickel substitution for cobalt (LT-LiCo_0.9Ni_0.1O₂) essentially eliminates the layered LiCoO₂ component from the structure (Fig. 20).¹⁷⁴ Recent studies at Argonne have shown that other substituents such as Al can improve the cycling stability of these materials considerably; it has also been shown that Ni-substituted spinels can react with Li₂MnO₃ to form a series of xLi₂MnO₃·(1 − x)LiCo_1−yNi_yO₂ compounds (0 ≤ x ≤ 0.2; 0 ≤ y ≤ 0.2), and that their electrochemical cycling stability is highly dependent on the values of x and y (Fig. 21).¹⁷⁸ These cobalt-rich materials, in which the surplus lithium in the transition metal layer of the Li₂MnO₃ component and the cobalt in the lithium-rich layer of the lithiated spinel component play a stabilizing role, are under investigation as possible agents for minimizing the voltage fade and capacity loss of high-capacity, manganese-rich xLi₂MnO₃·(1 − x)LiMO₂ (MNC) electrodes to an industrially acceptable level.

Fig. 19 Layered–spinel LiCoO₂ structural anomaly: simulated XRD patterns of (a) layered-LiCoO₂ structure (R3̄m, c/a = 4.99), (b) hypothetical layered structure with ideal cubic-close-packed oxygen array (R3̄m, c/a = 4.90), and (c) a cubic lithiated spinel, Li₂[Co₂]O₄ (Fd3̄m), showing identical patterns to those of (b) and (c) (reproduced with permission from ACS Appl. Mater. Interfaces, Lee et al.¹⁷⁴).

Fig. 20 Synchrotron XRD patterns of LT-LiCo_1−xNi_xO₂ (x = 0 and 0.1) (reproduced with permission from ACS Appl. Mater. Interfaces, Lee et al.¹⁷⁴).

Fig. 21 Voltage profiles of ‘layered–spinel’ Li/xLi₂MnO₃·(1 − x)LiCo_1−yNi_yO₂ cells (0 ≤ x ≤ 0.2; 0 ≤ y ≤ 0.2) cycled between 2.5 and 4.2 V vs. Li at a current rate of 15 mA g⁻¹.

Exploiting oxygen redox reactions in high-capacity Li₂MnO₃-stabilized electrodes

It is now well known that anomalously high capacities can be derived from lithium- and manganese-rich electrodes by activating the electrodes through oxygen loss at potentials above 4.5 V, as described earlier.^139–142 Intensive studies have been undertaken in recent years to understand and exploit electrochemically induced oxygen redox reactions in lithium metal oxides, such as Li₂MO₃ (Li₂O·MO₂) (M = Ru, Ir),^{143,179–181} Li₅FeO₄ (5Li₂O·Fe₂O₃),^182,183 and xLi₂MnO₃·(1 − x)LiMO₂ (alternatively, x[Li₂O·MnO₂]·(1 − x)LiMO₂) materials,^123,124,184 all of which can be regarded as having Li₂O-stabilized structures. Although some reversible (O₂ⁿ⁻) redox behavior has been observed in these materials, it can push the structural stability of the electrode to the limit such that there is competition between the O²⁻/O₂²⁻ anionic redox reaction and oxygen loss (O₂), particularly at the electrode surface. It has been reported that oxygen loss is preceded by a shortening of the O–O bond in Li₂IrO₃,¹⁴³ while first principles molecular dynamics simulations predict that O–O dimerization occurs in xLi₂MnO₃·(1 − x)LiMO₂ systems.¹²⁴ Both processes would result in a displacement of the oxygen ions from their close-packed positions in the parent electrode structure. In this respect, a recent XRD refinement of a 0.4Li₂MnO₃·0.6LiMn_0.5Ni_0.5O₂ electrode using synchrotron data (Fig. 22a) revealed that, on the initial charge of a Li/0.4Li₂MnO₃·0.6LiMn_0.5Ni_0.5O₂ cell to 4.7 V, the oxygen site occupancy decreased from 0.99 after ∼200 mA h g⁻¹ of capacity had been withdrawn from the electrode (Fig. 22b) to 0.91 at the top of charge (∼280 mA h g⁻¹), and that on the subsequent discharge to 2.5 V the site occupancy increased from 0.91 to ∼0.95.¹⁴⁷ Similar behavior was observed on the second cycle (Fig. 22c) with improved coulombic efficiency. The study also revealed that diffraction peaks characteristic of the Li₂MnO₃ component disappeared during the initial charge and reappeared during discharge, consistent with changes in Mn–O correlations (observed by X-ray absorption spectroscopy) and a partially reversible lithium and oxygen displacement process. The data provide some hope that, with time, control of these anionic redox processes may become possible and lead to the successful exploitation of reversible oxygen electrochemistry, even if only to a limited extent, in lithium-ion cathode technology.

Fig. 22 (a) Synchrotron XRD data and (b, c) variation of the oxygen site occupancy in a 0.4Li₂MnO₃·0.6LiMn_0.5Ni_0.5O₂ electrode during the first two cycles of a lithium cell when charged and discharged between 4.7 and 2.5 V (reproduced with permission from Nano Energy, Yoon et al.¹⁴⁷).

Concluding remarks

Today's lithium-ion battery cathodes are dominated by cobalt- and nickel-rich materials, such as LiCoO₂ (LCO) and LiNi_0.8Co_0.15Al_0.05O₂ (NCA), whereas manganese-based systems, such as the spinel LiMn₂O₄ (LMO) and LiNi_0.33Mn_0.33Co_0.0.33O₂ (NMC) play a relatively minor role in the market. Advances in cathode performance and cell energy density are likely to be made incrementally by improving the composition and structural design of lithium–metal-oxide materials and stabilized surfaces. Nickel-rich cathodes, such as NCA and NMC (e.g., ‘811’ and ‘622’), are currently in vogue but suffer from high cost and have potential safety concerns relative to manganese-rich materials. On the other hand, manganese-based cathodes such as LiMn₂O₄ and NMC (‘333’) will continue to serve the market, notably in the transportation sector for all-electric and hybrid-electric vehicles. The need to buffer cost when the prices of cobalt and nickel fluctuate to high values motivates and justifies further R&D of the manganese-rich high-voltage spinel, LiMn_1.5Ni_0.5O₄, and lithium- and manganese-rich (Li₂MnO₃-stabilized) materials such as the ‘layered–layered’ and ‘layered–layered–spinel’ systems discussed in this paper.

Statement of contributions

M. M. Thackeray – main author of this review and contributor to all aspects herein. J. R. Croy – co-author of text and author/contributor to data presented in Fig. 4, 7, 8, 13, 14–20. E. Lee – co-author of text and author/contributor to data presented in Fig. 1, 3, 6, 12,19–21. A. Gutierrez – co-author and contributor/author to data presented in Fig. 16–18. M. He – co-author and contributor/author to data presented in Fig. 16–18. J.-S. Park − co-author and contributor/author to data presented in Fig. 6, 7, 11, 13–15, and 19. B. T. Yonemoto – co-author and contributor/author to data presented in Fig. 11, 14, 16, and 17. B. R. Long – co-author and contributor/author to data presented in Fig. 3, 7, 13–15. J. D. Blauwkamp – co-author and contributor/author to data presented in Fig. 12, 19–21. C. S. Johnson – co-author and contributor/author to data presented in Fig. 1, 2, 5, and 10. Y. Shin – co-author and contributor/author to data presented in Fig. 14, 16, and 18. W. I. F. David – co-author and sole author of the work presented in Fig. 9.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

Support from the Office of Vehicle Technologies of the U.S. Department of Energy, particularly from Tien Duong, David Howell, and Peter Faguy is gratefully acknowledged. The submitted manuscript has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up, nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government.

References

1. J. E. Post, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 3447–3454.
2. Y. Chabre and J. Pannetier, Prog. Solid State Chem., 1995, 23, 1–130.
3. M. M. Thackeray, Prog. Solid State Chem., 1997, 25, 1–71.
4. D. Linden and T. B. Reddy, Handbook of Batteries, McGraw-Hill, New York, 3rd edn, 2001.
5. P. A. Marsal, K. Kordesch and L. F. Urry, US Pat., 2960558, 1960.
6. J. Brenet and P. Malessan, C. R. Acad. Sci., 1956, 242, 3064–3066.
7. K. Neumann and W. Fink, Z. Elektrochem., 1958, 62, 114–122.
8. W. C. Vosburgh, J. Electrochem. Soc., 1959, 106, 839.
9. P. Brouillet, A. Grund, F. Jolas and R. Mellet, C. R. Acad. Sci., 1963, 257, 3390.
10. A. Kozawa and R. A. Powers, J. Electrochem. Soc., 1966, 113, 870–878.
11. K. J. Vetter, Z. Elektrochem., 1962, 66, 577–586.
12. J. O. Besenhard, Carbon, 1976, 14, 111–115.
13. M. Zanini, S. Basu and J. E. Fischer, Carbon, 1978, 16, 211–212.
14. S. Basu, C. Zeller, P. J. Flanders, C. D. Fuerst, W. D. Johnson and J. E. Fischer, Mater. Sci. Eng., 1979, 38, 275–283.
15. M. S. Dresselhaus and G. Dresselhaus, Adv. Phys., 1981, 30, 139–326.
16. H. Ikeda, K. Narukawa and H. Nakashima, Jpn. Pat., 57-208079, 1982.
17. S. Basu, US Pat., 4304825, 1981.
18. S. Basu, US Pat., 4423125, 1983.
19. M. S. Whittingham, US Pat., 4009052, 1977.
20. M. S. Whittingham, Science, 1976, 192, 1126–1127.
21. M. S. Whittingham, Prog. Solid State Chem., 1978, 12, 41–99.
22. J. B. Goodenough and K. Mizuchima, US Pat., 4302518, 1981.
23. K. Mizushima, P. C. Jones, P. J. Wiseman and J. B. Goodenough, Mater. Res. Bull., 1980, 15, 783–789.
24. M. M. Thackeray and J. B. Goodenough, US Pat., 4507371, 1983.
25. M. M. Thackeray, W. I. F. David, P. G. Bruce and J. B. Goodenough, Mater. Res. Bull., 1983, 18, 461–472.
26. M. M. Thackeray, P. J. Johnson, L. A. de Picciotto, P. G. Bruce and J. B. Goodenough, Mater. Res. Bull., 1984, 19, 179–187.
27. K. Kordesch, J. Gsellmann, M. Peri, K. Tomantschger and R. Chemelli, Electrochim. Acta, 1981, 26, 1495–1504.
28. R. Patrice, B. Gerand, J. B. Leriche, L. Seguin, E. Wang, R. Moses, K. Brandt and J. M. Tarascon, J. Electrochem. Soc., 2001, 148, A448–A455.
29. M. A. Dzieciuch, N. Gupta and H. S. Wroblowa, J. Electrochem. Soc., 1988, 135, 2415–2418.
30. F. R. McLarnon and E. J. Cairns, J. Electrochem. Soc., 1991, 138, 645–664.
31. T. C. Adler, F. R. Mclarnon and E. J. Cairns, J. Electrochem. Soc., 1993, 140, 289–294.
32. N. D. Ingale, J. W. Gallaway, M. Nyce, A. Couzis and S. Banerjee, J. Power Sources, 2015, 276, 7–18.
33. T. Hirai and I. Tari, Prog. Batteries Sol. Cells, 1980, 3, 157.
34. H. Ikeda, T. Saito and H. Tamura, Proc. Manganese Dioxide Symposium, 1975, vol. 1, p. 384.
35. R. J. Brodd, A. Kozawa and K. V. Kordesch, J. Electrochem. Soc., 1978, 125, C271–C283.
36. H. Ikeda, M. Hara and S. Narukawa, US Pat., 4133856, 1979.
37. W. I. F. David, M. M. Thackeray, P. G. Bruce and J. B. Goodenough, Mater. Res. Bull., 1984, 19, 99–106.
38. T. Nohma, T. Saito, N. Furukawa and H. Ikeda, J. Power Sources, 1989, 26, 389–396.
39. T. Nohma, Y. Yamamoto, K. Nishio, I. Nakane and N. Furukawa, J. Power Sources, 1990, 32, 373–379.
40. M. M. Thackeray, M. H. Rossouw, A. de Kock, A. P. de la Harpe, R. J. Gummow, K. Pearce and D. C. Liles, J. Power Sources, 1993, 43–44, 289–300.
41. E. Levi, E. Zinigrad, H. Teller, M. D. Levi, D. Aurbach, E. Mengeritsky, E. Elster, P. Dan, E. Granot and H. Yamin, J. Electrochem. Soc., 1997, 144, 4133–4141.
42. M. H. Rossouw, D. C. Liles, M. M. Thackeray, W. I. F. David and S. Hull, Mater. Res. Bull., 1992, 27, 221–230.
43. P. Botkovitz, P. Deniard, M. Tournoux and R. Brec, J. Power Sources, 1993, 44, 657–665.
44. C. S. Johnson, D. W. Dees, M. F. Mansuetto, M. M. Thackeray, D. R. Vissers, D. Argyriou, C. K. Loong and L. Christensen, J. Power Sources, 1997, 68, 570–577.
45. Z. Z. Yang, D. C. Ford, J. S. Park, Y. Ren, S. Kim, H. Kim, T. T. Fister, M. K. Y. Chan and M. M. Thackeray, Chem. Mater., 2017, 29, 1507–1517.
46. Z. Z. Yang, L. Trahey, Y. Ren, M. K. Y. Chan, C. K. Lin, J. Okasinski and M. M. Thackeray, J. Mater. Chem. A, 2015, 3, 7389–7398.
47. M. H. Rossouw and M. M. Thackeray, US Pat., 5166012, 1992.
48. A. Debart, A. J. Paterson, J. Bao and P. G. Bruce, Angew. Chem., Int. Ed., 2008, 47, 4521–4524.
49. M. M. Thackeray, M. K. Y. Chan, L. Trahey, S. Kirklin and C. Wolverton, J. Phys. Chem. Lett., 2013, 4, 3607–3611.
50. M. S. Whittingham, Chem. Rev., 2004, 104, 4271–4301.
51. J. B. Goodenough and Y. Kim, Chem. Mater., 2010, 22, 587–603.
52. N. Nitta, F. X. Wu, J. T. Lee and G. Yushin, Mater. Today, 2015, 18, 252–264.
53. M. Brand, S. Glaser, J. Geder, S. Menacher, S. Obpacher, A. Jossen and D. Quinger, 2013 World Electric Vehicle Symposium and Exhibition (EVS27), Barcelona, Spain, 2013.
54. A. Manthiram, B. Song and W. Li, Energy Storage Materials, 2017, 6, 125–139.
55. M. M. Thackeray, A. Dekock, M. H. Rossouw, D. Liles, R. Bittihn and D. Hoge, J. Electrochem. Soc., 1992, 139, 363–366.
56. J. C. Hunter, J. Solid State Chem., 1981, 39, 142–147.
57. A. R. Armstrong and P. G. Bruce, Nature, 1996, 381, 499–500.
58. R. J. Gummow, D. C. Liles and M. M. Thackeray, Mater. Res. Bull., 1993, 28, 1249–1256.
59. R. J. Gummow and M. M. Thackeray, Solid State Ionics, 1992, 53, 681–687.
60. E. Rossen, J. N. Reimers and J. R. Dahn, Solid State Ionics, 1993, 62, 53–60.
61. R. J. Gummow, D. C. Liles and M. M. Thackeray, Mater. Res. Bull., 1993, 28, 235–246.
62. J. B. Goodenough, M. M. Thackeray, W. I. F. David and P. G. Bruce, Rev. Chim. Miner., 1984, 21, 435–455.
63. T. Ohzuku, M. Kitagawa and T. Hirai, J. Electrochem. Soc., 1990, 137, 769–775.
64. C. Masquelier, M. Tabuchi, K. Ado, R. Kanno, Y. Kobayashi, Y. Maki, O. Nakamura and J. B. Goodenough, J. Solid State Chem., 1996, 123, 255–266.
65. J. M. Tarascon, E. Wang, F. K. Shokoohi, W. R. McKinnon and S. Colson, J. Electrochem. Soc., 1991, 138, 2859–2864.
66. J. M. Tarascon and D. Guyomard, J. Electrochem. Soc., 1991, 138, 2864–2868.
67. M. M. Thackeray and R. J. Gummow, US Pat., 5316877, 1994.
68. R. J. Gummow, A. de Kock and M. M. Thackeray, Solid State Ionics, 1994, 69, 59–67.
69. Y. S. Lee, N. Kumada and M. Yoshio, J. Power Sources, 2001, 96, 376–384.
70. R. Benedek and M. M. Thackeray, Electrochem. Solid-State Lett., 2006, 9, A265–A267.
71. C. Zhan, J. Lu, A. J. Kropf, T. P. Wu, A. N. Jansen, Y. K. Sun, X. P. Qiu and K. Amine, Nat. Commun., 2013, 4, 2437.
72. T. F. Yi, Y. R. Zhu, X. D. Zhu, J. Shu, C. B. Yue and A. N. Zhou, Ionics, 2009, 15, 779–784.
73. J.-S. Kim, C. S. Johnson, J. T. Vaughey, S. A. Hackney, K. A. Walz, W. A. Zeltner, M. A. Anderson and M. M. Thackeray, J. Electrochem. Soc., 2004, 151, A1755–A1761.
74. S. W. Lee, K. S. Kim, H. S. Moon, H. J. Kim, B. W. Cho, W. I. Cho, J. B. Ju and J. W. Park, J. Power Sources, 2004, 126, 150–155.
75. S. Lee, M. Jeong and J. Cho, Adv. Energy Mater., 2013, 3, 1623–1629.
76. L. Jaber-Ansari, K. P. Puntambekar, S. Kim, M. Aykol, L. L. Luo, J. S. Wu, B. D. Myers, H. Iddir, J. T. Russell, S. J. Saldana, R. Kumar, M. M. Thackeray, L. A. Curtiss, V. P. Dravid, C. Wolverton and M. C. Hersam, Adv. Energy Mater., 2015, 5, 1500646.
77. K. S. Chen, R. Xu, N. S. Luu, E. B. Secor, K. Hamamoto, Q. Q. Li, S. Kim, V. K. Sangwan, I. Balla, L. M. Guiney, J. W. T. Seo, X. K. Yu, W. W. Liu, J. S. Wu, C. Wolverton, V. P. Dravid, S. A. Barnett, J. Lu, K. Amine and M. C. Hersam, Nano Lett., 2017, 17, 2539–2546.
78. S. B. Chikkannanavar, D. M. Bernardi and L. Y. Liu, J. Power Sources, 2014, 248, 91–100.
79. K. Amine, H. Tukamoto, H. Yasuda and Y. Fujita, J. Electrochem. Soc., 1996, 143, 1607–1613.
80. E.-S. Lee, K.-W. Nam, E. Hu and A. Manthiram, Chem. Mater., 2012, 24, 3610–3620.
81. A. D. Robertson and P. G. Bruce, Chem. Mater., 2003, 15, 1984–1992.
82. L. A. de Picciotto, K. T. Adendorff, D. C. Liles and M. M. Thackeray, Solid State Ionics, 1993, 62, 297–307.
83. G. Pistoia, S. Panero, M. Tocci, R. V. Moshtev and V. Manev, Solid State Ionics, 1984, 13, 311–318.
84. R. Benedek, M. M. Thackeray and L. H. Yang, J. Power Sources, 1999, 81, 487–490.
85. P. Hovington, M. Lagace, A. Guerfi, P. Bouchard, A. Manger, C. M. Julien, M. Armand and K. Zaghib, Nano Lett., 2015, 15, 2671–2678.
86. M. M. Thackeray, C. S. Johnson, J. T. Vaughey, N. Li and S. A. Hackney, J. Mater. Chem., 2005, 15, 2257–2267.
87. M. M. Thackeray, S. H. Kang, C. S. Johnson, J. T. Vaughey, R. Benedek and S. A. Hackney, J. Mater. Chem., 2007, 17, 3112–3125.
88. M. M. Thackeray and M. H. Rossouw, US Pat., 5153081, 1992.
89. M. H. Rossouw, D. C. Liles and M. M. Thackeray, J. Solid State Chem., 1993, 104, 464–466.
90. Y. Paik, C. P. Grey, C. S. Johnson, J. S. Kim and M. M. Thackeray, Chem. Mater., 2002, 14, 5109–5115.
91. W. Tang, H. Kanoh, X. Yang and K. Ooi, Chem. Mater., 2000, 12, 3271–3279.
92. M. H. Rossouw and M. M. Thackeray, Mater. Res. Bull., 1991, 26, 463–473.
93. C. S. Johnson, S. D. Korte, J. T. Vaughey, M. M. Thackeray, T. E. Bofinger, Y. Shao-Horn and S. A. Hackney, J. Power Sources, 1999, 81, 491–495.
94. P. Kalyani, S. Chitra, T. Mohan and S. Gopukumar, J. Power Sources, 1999, 80, 103–106.
95. A. R. Armstrong, A. D. Robertson and P. G. Bruce, J. Power Sources, 2005, 146, 275–280.
96. F. Capitaine, P. Gravereau and C. Delmas, Solid State Ionics, 1996, 89, 197–202.
97. P. G. Bruce, A. R. Armstrong and R. L. Gitzendanner, J. Mater. Chem., 1999, 9, 193–198.
98. J. Reed, G. Ceder and A. Van der Ven, Electrochem. Solid-State Lett., 2001, 4, A78–A81.
99. C. Delmas, I. Saadoune and A. Rougier, J. Power Sources, 1993, 44, 595–602.
100. C. Delmas, M. Menetrier, L. Croguennec, I. Saadoune, A. Rougier, C. Pouillerie, G. Prado, M. Grune and L. Fournes, Electrochim. Acta, 1999, 45, 243–253.
101. E. Rossen, C. D. W. Jones and J. R. Dahn, Solid State Ionics, 1992, 57, 311–318.
102. M. E. Spahr, P. Novak, B. Schnyder, O. Haas and R. Nesper, J. Electrochem. Soc., 1998, 145, 1113–1121.
103. A. R. Armstrong, A. D. Robertson and P. G. Bruce, Electrochim. Acta, 1999, 45, 285–294.
104. A. D. Robertson, A. R. Armstrong and P. G. Bruce, Chem. Mater., 2001, 13, 2380–2386.
105. C. Storey, I. Kargina, Y. Grincourt, I. J. Davidson, Y. C. Yoo and D. Y. Seung, J. Power Sources, 2001, 97–8, 541–544.
106. B. Ammundsen, J. Paulsen, I. Davidson, R. S. Liu, C. H. Shen, J. M. Chen, L. Y. Jang and J. F. Lee, J. Electrochem. Soc., 2002, 149, A431–A436.
107. J. B. Goodenough, A. K. Padhi, K. S. Nanjundaswamy and C. Masquelier, US Pat., 5910382, 1999.
108. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194.
109. A. K. Padhi, K. S. Nanjundaswamy, C. Masquelier, S. Okada and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1609–1613.
110. L. A. de Picciotto and M. M. Thackeray, Solid State Ionics, 1986, 18–9, 773–777.
111. G. G. Amatucci, J. M. Tarascon and L. C. Klein, J. Electrochem. Soc., 1996, 143, 1114–1123.
112. J. N. Reimers and J. R. Dahn, J. Electrochem. Soc., 1992, 139, 2091–2097.
113. M. M. Thackeray, C. S. Johnson, K. Amine and J. Kim, US Pat., 6677082, 2004.
114. M. M. Thackeray, C. S. Johnson, K. Amine and J. Kim, US Pat., 6680143, 2004.
115. J. S. Kim, C. S. Johnson and M. M. Thackeray, Electrochem. Commun., 2002, 4, 205–209.
116. J.-S. Kim, C. S. Johnson, J. T. Vaughey, M. M. Thackeray, S. A. Hackney, W. Yoon and C. P. Grey, Chem. Mater., 2004, 16, 1996–2006.
117. M. M. Thackeray, S. H. Kang, C. S. Johnson, J. T. Vaughey and S. A. Hackney, Electrochem. Commun., 2006, 8, 1531–1538.
118. J. R. Croy, M. Balasubramanian, K. G. Gallagher and A. K. Burrell, Acc. Chem. Res., 2015, 48, 2813–2821.
119. J. Wang, X. He, E. Paillard, N. Laszczynski, J. Li and S. Passerini, Adv. Energy Mater., 2016, 6, 1600906.
120. J. M. Zheng, S. J. Myeong, W. R. Cho, P. F. Yan, J. Xiao, C. M. Wang, J. Cho and J. G. Zhang, Adv. Energy Mater., 2017, 7, 1601284.
121. W. D. Li, B. H. Song and A. Manthiram, Chem. Soc. Rev., 2017, 46, 3006–3059.
122. H. Iddir and R. Benedek, Chem. Mater., 2014, 26, 2407–2413.
123. H. Iddir, J. Bareño and R. Benedek, J. Electrochem. Soc., 2016, 163, A1784–A1789.
124. R. Benedek and H. Iddir, J. Phys. Chem. C, 2017, 121, 6492–6499.
125. E. Lee and K. A. Persson, Adv. Energy Mater., 2014, 4, 1400498.
126. Y. Shin, H. Ding and K. A. Persson, Chem. Mater., 2016, 28, 2081–2088.
127. S. Kim, M. Aykol, V. I. Hegde, Z. Lu, S. Kirklin, J. R. Croy, M. M. Thackeray and C. Wolverton, Energy Environ. Sci., 2017, 10, 2201–2211.
128. C. S. Johnson, N. Li, J. T. Vaughey, S. A. Hackney and M. M. Thackeray, Electrochem. Commun., 2005, 7, 528–536.
129. S. H. Park, S. H. Kang, C. S. Johnson, K. Amine and M. M. Thackeray, Electrochem. Commun., 2007, 9, 262–268.
130. B. R. Long, J. R. Croy, J. S. Park, J. G. Wen, D. J. Miller and M. M. Thackeray, J. Electrochem. Soc., 2014, 161, A2160–A2167.
131. K. A. Jarvis, Z. Deng, L. F. Allard, A. Manthiram and P. J. Ferreira, Chem. Mater., 2011, 23, 3614–3621.
132. J. Reed and G. Ceder, Electrochem. Solid-State Lett., 2002, 5, A145–A148.
133. Y. Gao, K. Myrtle, M. J. Zhang, J. N. Reimers and J. R. Dahn, Phys. Rev. B, 1996, 54, 16670–16675.
134. B. R. Long, J. R. Croy, F. Dogan, M. R. Suchomel, B. Key, J. Wen, D. J. Miller, M. M. Thackeray and M. Balasubramanian, Chem. Mater., 2014, 26, 3565–3572.
135. J. Bareño, M. Balasubramanian, S. H. Kang, J. G. Wen, C. H. Lei, S. V. Pol, I. Petrov and D. P. Abraham, Chem. Mater., 2011, 23, 2039–2050.
136. J. Choi and A. Manthiram, J. Electrochem. Soc., 2005, 152, A1714–A1718.
137. K. Shizuka, T. Kobayashi, K. Okahara, K. Okamoto, S. Kanzaki and R. Kanno, J. Power Sources, 2005, 146, 589–593.
138. S. H. Park, S. H. Kang, I. Belharouak, Y. K. Sun and K. Amine, J. Power Sources, 2008, 177, 177–183.
139. Z. H. Lu, L. Y. Beaulieu, R. A. Donaberger, C. L. Thomas and J. R. Dahn, J. Electrochem. Soc., 2002, 149, A778–A791.
140. Z. H. Lu and J. R. Dahn, J. Electrochem. Soc., 2002, 149, A815–A822.
141. C. S. Johnson, J. S. Kim, C. Lefief, N. Li, J. T. Vaughey and M. M. Thackeray, Electrochem. Commun., 2004, 6, 1085–1091.
142. A. R. Armstrong, M. Holzapfel, P. Novak, C. S. Johnson, S. H. Kang, M. M. Thackeray and P. G. Bruce, J. Am. Chem. Soc., 2006, 128, 8694–8698.
143. E. McCalla, A. M. Abakumov, M. Saubanère, D. Foix, E. J. Berg, G. Rousse, M.-L. Doublet, D. Gonbeau, P. Novák, G. Van Tendeloo, R. Dominko and J.-M. Tarascon, Science, 2015, 350, 1516–1521.
144. A. Grimaud, W. T. Hong, Y. Shao-Horn and J. M. Tarascon, Nat. Mater., 2016, 15, 121–126.
145. D. H. Seo, J. Lee, A. Urban, R. Malik, S. Kang and G. Ceder, Nat. Chem., 2016, 8, 692–697.
146. C. S. Johnson, N. Li, C. Lefief and M. M. Thackeray, Electrochem. Commun., 2007, 9, 787–795.
147. S. Muhammad, H. Kim, Y. Kim, D. Kim, J. H. Song, J. Yoon, J.-H. Park, S.-J. Ahn, S.-H. Kang, M. M. Thackeray and W.-S. Yoon, Nano Energy, 2016, 21, 172–184.
148. P. K. Nayak, J. Grinblat, M. Levi, E. Levi, S. Kim, J. W. Choi and D. Aurbach, Adv. Energy Mater., 2016, 6, 1502398.
149. P. K. Nayak, J. Grinblat, E. Levi, T. R. Penki, M. Levi, Y.-K. Sun, B. Markovsky and D. Aurbach, ACS Appl. Mater. Interfaces, 2017, 9, 4309–4319.
150. J. R. Croy, J. S. Park, Y. Shin, B. T. Yonemoto, M. Balasubramanian, B. R. Long, Y. Ren and M. M. Thackeray, J. Power Sources, 2016, 334, 213–220.
151. M. M. Thackeray, C. Wolverton and E. D. Isaacs, Energy Environ. Sci., 2012, 5, 7854–7863.
152. K. G. Gallagher, J. R. Croy, M. Balasubramanian, M. Bettge, D. P. Abraham, A. K. Burrell and M. M. Thackeray, Electrochem. Commun., 2013, 33, 96–98.
153. J. R. Croy, K. G. Gallagher, M. Balasubramanian, Z. Chen, Y. Ren, D. Kim, S.-H. Kang, D. W. Dees and M. M. Thackeray, J. Phys. Chem. C, 2013, 117, 6525–6536.
154. W. I. F. David, unpublished work.
155. G.-A. Nazri and G. Pistoia, Lithium Batteries: Science and Technology, Kluwer Academic Publishers, Massachusetts, 2003.
156. N. Leifer, F. Schipper, E. M. Erickson, C. Ghanty, M. Talianker, J. Grinblat, C. M. Julien, B. Markovsky and D. Aurbach, J. Phys. Chem. C, 2017, 121, 9120–9130.
157. M. H. Rossouw, A. Dekock, L. A. Depicciotto and M. M. Thackeray, Mater. Res. Bull., 1990, 25, 173–182.
158. C. S. Johnson, S. H. Kang, J. T. Vaughey, S. V. Pol, M. Balasubramanian and M. M. Thackeray, Chem. Mater., 2010, 22, 1263–1270.
159. E. Lee, J. S. Park, H. Kim, Q. Li, J. Wu, V. P. Dravid, J. Blauwkamp, J. Croy and M. M. Thackeray, DOE Vehicle Technologies Annual Merit Review, ES049, 2017.
160. D. Kim, G. Sandi, J. R. Croy, K. G. Gallagher, S. H. Kang, E. Lee, M. D. Slater, C. S. Johnson and M. M. Thackeray, J. Electrochem. Soc., 2013, 160, A31–A38.
161. E. Lee, D. Ahn, J. S. Park, A. Gutierrez, M. He, D. Mohanty, A. Huq, Y. Ren, J. R. Croy and M. M. Thackeray, DOE Vehicle Technologies Annual Merit Review, BAT235 & BAT049, 2018.
162. E. Peled and S. Menkin, J. Electrochem. Soc., 2017, 164, A1703–A1719.
163. C. Li, H. P. Zhang, L. J. Fu, H. Liu, Y. P. Wu, E. Rahm, R. Holze and H. Q. Wu, Electrochim. Acta, 2006, 51, 3872–3883.
164. J. Liu and A. Manthiram, J. Mater. Chem., 2010, 20, 3961–3967.
165. Y.-K. Sun, M.-J. Lee, C. S. Yoon, J. Hassoun, K. Amine and B. Scrosati, Adv. Mater., 2012, 24, 1192–1196.
166. H. M. Wu, I. Belharouak, A. Abouimrane, Y. K. Sun and K. Amine, J. Power Sources, 2010, 195, 2909–2913.
167. M. M. Thackeray, C. S. Johnson, J. S. Kim, K. C. Lauzze, J. T. Vaughey, N. Dietz, D. Abraham, S. A. Hackney, W. Zeltner and M. A. Anderson, Electrochem. Commun., 2003, 5, 752–758.
168. J. Cho, T.-J. Kim, J. Kim, M. Noh and B. Park, J. Electrochem. Soc., 2004, 151, A1899–A1904.
169. Q. Y. Wang, J. Liu, A. V. Murugan and A. Manthiram, J. Mater. Chem., 2009, 19, 4965–4972.
170. S. H. Kang and M. M. Thackeray, Electrochem. Commun., 2009, 11, 748–751.
171. Visual Capitalist, http://www.visualcapitalist.com/nickel-secret-driver-battery-revolution, accessed January 2018.
172. Y.-K. Sun, Z. Chen, H.-J. Noh, D.-J. Lee, H.-G. Jung, Y. Ren, S. Wang, C. S. Yoon, S.-T. Myung and K. Amine, Nat. Mater., 2012, 11, 942–947.
173. J. R. Croy, A. Gutierrez, E. Lee and M. M. Thackeray, Manganese-Rich Alternatives to Nickel-Rich NMC Cathodes, in preparation.
174. E. Lee, J. Blauwkamp, F. C. Castro, J. Wu, V. P. Dravid, P. Yan, C. Wang, S. Kim, C. Wolverton, R. Benedek, F. Dogan, J. S. Park, J. R. Croy and M. M. Thackeray, ACS Appl. Mater. Interfaces, 2016, 8, 27720–27729.
175. J. Reed and G. Ceder, Chem. Rev., 2004, 104, 4513–4533.
176. R. J. Gummow, M. M. Thackeray, W. I. F. David and S. Hull, Mater. Res. Bull., 1992, 27, 327–337.
177. R. J. Gummow, D. C. Liles, M. M. Thackeray and W. I. F. David, Mater. Res. Bull., 1993, 28, 1177–1184.
178. E. Lee, J. Blauwkamp, A. Gutierrez, M. He, J. Croy and M. M. Thackeray, Stabilizing Low Temperature LiCoO₂ by Li₂MnO₃ and Ni Integration, in preparation.
179. M. Sathiya, K. Ramesha, G. Rousse, D. Foix, D. Gonbeau, A. S. Prakash, M. L. Doublet, K. Hemalatha and J. M. Tarascon, Chem. Mater., 2013, 25, 1121–1131.
180. M. Sathiya, G. Rousse, K. Ramesha, C. P. Laisa, H. Vezin, M. T. Sougrati, M. L. Doublet, D. Foix, D. Gonbeau, W. Walker, A. S. Prakash, M. Ben Hassine, L. Dupont and J. M. Tarascon, Nat. Mater., 2013, 12, 827–835.
181. P. E. Pearce, A. J. Perez, G. Rousse, M. Saubanere, D. Batuk, D. Foix, E. McCalla, A. M. Abakumov, G. Van Tendeloo, M. L. Doublet and J. M. Tarascon, Nat. Mater., 2017, 16, 580–586.
182. C. S. Johnson, S. H. Kang, J. T. Vaughey, S. V. Pol, M. Balasubramanian and M. M. Thackeray, Chem. Mater., 2010, 22, 1263–1270.
183. C. Zhan, Z. Yao, J. Lu, L. Ma, V. A. Maroni, L. Li, E. Lee, E. E. Alp, T. Wu, J. Wen, Y. Ren, C. Johnson, M. M. Thackeray, M. K. Y. Chan, C. Wolverton and K. Amine, Nat. Energy, 2017, 2, 963–971.
184. K. Luo, M. R. Roberts, N. Guerrini, N. Tapia-Ruiz, R. Hao, F. Massel, D. M. Pickup, S. Ramos, Y.-S. Liu, J. Guo, A. V. Chadwick, L. C. Duda and P. G. Bruce, J. Am. Chem. Soc., 2016, 138, 11211–11218.

---

Footnotes

† Current addresses: SAFT America, Cockeysville, Maryland 21030, USA.

‡ Current addresses: Microvast Power Solutions, Inc., Orlando, Florida 32826, USA.

§ Current addresses: BASF, Beachwood, Ohio 44122, USA.

¶ Current addresses: Challenge Manufacturing, Grand Rapids, MI 49544.

---