Layered electrode materials for non-aqueous multivalent metal batteries

Abstract

The global transition to clean energy production accelerates the necessity for efficient energy storage solutions. Lithium-ion batteries (LIBs), developed considerably over the past three decades, have been widely applied in portable electronics and electric vehicles. Nevertheless, the application of LIBs in large-scale energy storage applications is restricted by the Li resource scarcity, high-cost raw materials, and severe safety issues, which thus triggers the development of new battery chemistries beyond Li⁺. In this regard, multivalent metal batteries (MVMBs, e.g., Zn, Mg, and Al batteries) are promising alternatives owing to the advantages brought by the direct use of corresponding metals as anodes, such as high elemental abundance, low anode redox potential, multielectron redox capability, and facile metal stripping and plating chemistry. However, MVMBs suffer from the lack of available cathodes for efficiently accommodating multivalent metal ions (i.e., Zn²⁺, Mg²⁺, and Al³⁺), which originates from the strong electrostatic interactions between charge carrier ions and cathodes. Promising cathode candidates for MVMBs to address this challenge are layered electrode materials, whose structures can be engineered with versatile approaches to regulate the charge-storage behaviour. Here, layered electrode materials used for non-aqueous MVMB cathodes are thoroughly reviewed. We first introduce the cell configurations and the thus-far developed anode–electrolyte–cathode chemistries for non-aqueous MVMBs. Recent progress in the exploration of layered materials for non-aqueous MVMBs is subsequently summarized. Emphasis is put on examining the employed structure engineering strategies and their effects on both the intrinsic properties and electrochemical behaviours of layered electrode materials. Finally, perspectives on the challenges and future directions in this research field are provided with aspects to the cathode structure engineering, performance assessment, and device demonstration.

---

Ahiud Morag

Ahiud Morag received his PhD degree in interdisciplinary studies from Ben Gurion University of the Negev, Israel, in June 2019. In October 2019, he joined Prof Xinliang Feng's group at Technische Universität Dresden (TU Dresden) as a postdoctoral research associate. His research mainly focuses on the development of novel cathodes and electrolytes for non-aqueous magnesium and aluminum multivalent metal batteries.

Minghao Yu

Minghao Yu received his PhD degree in Material Physics and Chemistry from Sun Yat-sen University in June 2017. In March 2019, he became a research group leader of the Chair for Molecular Functional Materials at Technische Universität Dresden. His research interests focus on the development of advanced functional materials for applications of energy storage (supercapacitors and metal-ion batteries) and conversion (electrocatalysis and metal-air batteries).

---

1. Introduction

Global warming is pushing towards the transition from traditional fossil fuels to clean energy resources like solar, wind, and tidal energies.^1,2 However, the intermittent nature of these clean and cheap energy resources restricts their direct utilization, which imposes a strong demand for high-performance energy storage technologies.^3–5 Among existing electrochemical energy storage technologies, lithium-ion batteries (LIBs) represent the most commercially successful one, and they have been implemented in diverse applications, such as portable electronics and electric vehicles.^6,7 However, the severe safety issues of LIBs, limited Li resource in the earth crust (20 ppm, 0.002%), and the increasing demand for cheap energy storage solutions motivate the development of new battery chemistries relying on cheap and abundant elements. In this context, multivalent metal batteries (MVMBs) are proposed as promising alternatives, which directly employ resource-abundant and low-cost multivalent metals (e.g., Zn, Mg, and Al) as anodes.^8–16 Typically, multivalent metal anodes present low stripping/plating potentials (−0.76, −2.37, and −1.66 V vs. standard hydrogen electrode (SHE) for Zn, Mg, and Al, respectively), which are beneficial for the construction of high-voltage and high-energy energy storage devices (Fig. 1). Importantly, Zn, Mg, and Al are the 24th, 8th, and 3rd most abundant elements in the Earth's crust, respectively. These metal anodes can deliver high specific capacities due to their multielectron redox capability (820, 2206, and 2981 mA h g⁻¹ for Zn, Mg, and Al, respectively). Meanwhile, the high mass densities of multivalent metals empower the corresponding anodes with impressive volumetric capacities (5849, 3834, and 8047 Ah L⁻¹ for Zn, Mg, and Al, respectively). Interestingly, the partial compatibility of multivalent metals (especially Zn) with water-based electrolytes inspired researchers to develop aqueous MVMBs. Aqueous electrolytes enable the fabrication of low-cost and safe MVMBs with fast cation-storage kinetics of the cathodes.^17,18 However, aqueous MVMBs suffer from low voltage windows (normally less than 1.8 V) due to the narrow stable potential windows of water-based electrolytes. Moreover, water in electrolytes would lead to the formation of passivation layers on metal anodes, which can inhibit the stripping/plating of Mg and Al and accelerate the dendrite growth of Zn. By contrast, multivalent metals show negligible passivation or dendrite growth in non-aqueous electrolytes, providing the metal stripping/plating process with high coulombic efficiency.^19–24 Additionally, non-aqueous electrolyte displays substantially higher potential window than aqueous electrolytes, enabling the use of many cathode materials with high redox potentials. All these features allow non-aqueous MVMBs to be promising candidates to abreast with LIBs as leading energy storage technologies.

Fig. 1 Specific capacities, volumetric capacities, ionic radii, and stripping/plating potentials of various metal ions.

However, the construction of practical non-aqueous MVMBs is hindered by severe challenges induced by the multivalent metal ions as charge carriers for cathodes. Multivalent metal ions display large charge densities, resulting from the multivalent nature and the small ionic radii (0.74, 0.72, and 0.53 Å for Zn²⁺, Mg²⁺, and Al³⁺ compared with 0.76 Å for Li⁺). It further leads to the strong electrostatic interactions between charge carrier ions and cathodes during the charge/discharge of non-aqueous MVMBs. Consequently, most well-recognized metal oxide cathodes for LIBs stand inappropriate for non-aqueous MVMB cathodes.^25–31 In this sense, cathode structures for non-aqueous MVMBs should be re-designed at the atomic level to enable facile Zn²⁺, Mg²⁺, and Al³⁺ storage.

As promising cathode candidates for non-aqueous MVMBs, natural layered materials are a class of materials with strong atom bonding in the basal plane and weak van der Waals (vdW) interaction between layers. These materials are equipped with versatile physical, chemical, electronic properties, as well as broad structural diversity.^32–35 Particularly, layered materials are appealing for energy storage, as they exhibit some intrinsic advantages over non-layered materials. For example, compared with non-layered materials, layered materials depicted higher accessibility of exposed active sites, enabling higher specific capacities and better ion diffusion kinetics.^36–38 However, some layered materials suffer from phase transition during electrochemical processes and poor stability in the ambient environment.³⁹ Thus far, a variety of layered materials have been explored as cathodes for MVMBs, including layered transition metal oxides (TMOs),^32,40 transition metal dichalcogenides (TMDs),^41–43 graphite,^44,45 and two-dimensional (2D) covalent organic frameworks (COFs).⁴⁶ Pristine layered materials as non-aqueous MVMB cathodes generally exhibit limited electrochemical performance.⁴⁷ Importantly, the weak vdW interaction between the stacked layers enables layered materials with diverse possibilities for rational structure engineering, such as exfoliation into 2D nanoflakes, interlayer expansion with guest molecules, and hybrid structure construction (Fig. 2).⁴⁸ These structure engineering strategies are highly desired for layered materials to tailor their intrinsic properties (e.g., electronic structure, conductivity, and redox capability) and electrochemical behaviours (e.g., ion desolvation energy, solid-state ion diffusion kinetics, charge-storage mechanism) for multivalent metal ion storage.

Fig. 2 Schematic illustration of the review content, including the cell configuration of non-aqueous MVMBs, layered cathode materials, and structure engineering methods.

In this review, the latest development of layered materials for non-aqueous MVMB cathodes is reviewed. We first summarize the typical configurations of non-aqueous MVMBs in terms of the developed electrolytes, typical anode–electrolyte–cathode chemistries, and possible charge carrier ions. Next, a comprehensive overview of state-of-the-art layered materials used for non-aqueous MVMBs is carried out by emphasizing the strategies applied to tailor the structures of layered materials. Emphatic efforts are devoted to analysing the induced changes in the intrinsic properties and electrochemical behaviours of layered materials upon the structure regulation. Finally, perspectives on the current challenges and future directions of layered materials for non-aqueous MVMBs are presented.

2. Cell configurations of non-aqueous MVMBs

Contrasting to Li-metal anodes, Zn-, Mg-, and Al-metal anodes can be easily processed and show no aggressive dendrite growth in non-aqueous electrolytes during battery operation.⁴⁹ Although multivalent metals can be directly used as anodes, their stripping/plating reactions are not the simple M^x+ + xe⁻ → M reaction. Exact reaction mechanisms on multivalent metal anodes are introduced in this section. Additionally, according to the different cathode chemistry, non-aqueous MVMBs can be briefly categorized into cation ‘rocking-chair’ cells and dual-ion cells. In ‘rocking-chair’ cells, cation released from cathodes and metal plating on anodes occur during the charge of cells, while cation storage in cathodes and metal stripping on anodes occur during the discharge of cells. The electrolyte concentration in the ‘rocking-chair’ cells keeps constant during the cell operation. In dual-ion cells, anions serve as charge carrier ions for cathodes, which are stored/released in/from cathodes during the charge/discharge process of cells. The charge process would consume ions in electrolytes, and high-concentration or large amounts of electrolytes are preferred for dual-ion cells.⁵⁰ In this section, we also discuss the cathode chemistries of MVMBs with a particular focus on the different charge carrier species.

2.1 Zn-metal batteries

The Zn²⁺/Zn couple has a sufficiently high redox potential of −0.76 V vs. SHE, which enables the Zn stripping/plating reaction to be feasible in either alkaline or close-to-neutral aqueous electrolytes. Moreover, the use of water-based electrolytes brings the advantages of low cost, high safety, high ionic conductivity, and easy-to-manufacture feature.⁵¹ Thereby, dominant researches of Zn-metal batteries (ZMBs) are focused on aqueous systems. Aqueous ZMBs are not covered in this review, and interested readers are directed to recent comprehensive reviews of this field.^52–57 It is worth noting that aqueous ZMBs are restricted by several intrinsic drawbacks arising from the use of aqueous electrolytes. Cathode materials like MnO₂ and VOPO₄ suffer from dissolution in water, resulting in fast capacity degradation during cycling.^58,59 Additionally, Zn-metal anodes in aqueous electrolytes exhibit severe side reactions (e.g., hydrogen evolution, ZnO formation) and Zn dendrite growth, which further lead to the electrolyte decomposition, irreversible Zn consumption, low stripping/plating coulombic efficiency, and safety issues.⁶⁰

Non-aqueous electrolytes based on ionic liquid, acetonitrile (AN) or carbonate solvents with Zn salts (e.g., zinc bis(trifluoromethylsulfonyl)imide (Zn(TFSI)₂), Zn(CF₃SO₃)₂ (Zn(OTf)₂), Zn(ClO₄)₂, Zn(BF₄)₂, and Zn(PF₆)₂) were demonstrated feasible for reversible Zn stripping and plating.^24,61–66 These electrolytes offer much wider voltage windows than aqueous electrolytes. For example, Han et al.⁶⁴ showed that both AN and propylene carbonate (PC), dissolving different Zn salts (e.g., Zn(TFSI)₂, Zn(OTf)₂, Zn(BF₄)₂, and Zn(PF₆)₂), demonstrated high anodic stability of 3.8 V vs. Zn²⁺/Zn. Wang et al.²⁴ developed a Li⁺-containing hybrid electrolyte, exhibiting high anodic stability of 4 V vs. Zn²⁺/Zn. The high anodic stability was achieved by adding LiPF₆ to Zn(TFSI)₂ in ethyl methyl carbonate (EMC). The wide voltage windows of non-aqueous Zn electrolytes are crucial for the fabrication of high-voltage and high-energy ZMBs.

According to the cathode chemistries, two possible charge storage mechanisms for non-aqueous ZMBs are presented in Fig. 3. The first one relies on the Zn²⁺ ‘rocking-chair’ mechanism. The high energy barrier of Zn²⁺ desolvation on the cathode surface accounts for the large charge-transfer resistance of these ZMBs.^61,66,67 The second mechanism is based on dual-ion cells, involving Zn plating/stripping on the anode and anion storage/release on the cathode. The charge-transfer resistance of anion storage is considerably lower than that of Zn²⁺ storage, enabling the excellent rate performance of ZMBs.^24,68 So far, several cathode materials with the Zn²⁺-storage mechanism were reported for non-aqueous ZMBs, including MnO₂,⁶⁶ vanadium oxide,^61,65 and VOPO₄.⁶⁹ On the other hand, graphite-based cathodes, exhibiting anion intercalation/de-intercalation, were used for constructing dual-ion ZMBs.^23,24,68 Anions, such as PF₆⁻, TFSI⁻, and OTf⁻ can be used as charge carriers for cathodes, providing ZMBs with high operation voltages and high energy densities. Overall, compared with aqueous ZMBs, non-aqueous ZMBs benefit from high anodic stability and dendrite-free Zn stripping and plating.

Fig. 3 Schematic illustration showing the charge storage mechanism of ZMBs.

2.2 Mg-metal batteries

The Mg metal surface would be passivated in most solvents by an Mg²⁺ non-conducting layer, preventing the electrochemical Mg plating. In detail, electrolytes comprised of conventional solvents and Mg salts, such as Mg(TFSI)₂ or Mg(ClO₄)₂ in AN or carbonate solvents, are not suitable for Mg-metal batteries (MMBs),⁷⁰ because these electrolytes tend to form ionic-insulating passivation layers (e.g., MgO, MgCl₂, and Mg_xClO_y) on Mg metal. Grignard reagents (e.g., EtMgBr, and Mg(BR₄)₂, where R = butyl, phenyl) dissolved in ethers were presented as the first suitable electrolytes for the facile Mg stripping/plating.^70–72 However, Grignard reagent-based electrolytes show low anodic stability (<1.5 V vs. Mg²⁺/Mg).^72,73 In 2000, Aurbach with his colleagues reported the MMB electrolyte, Mg(AlCl₂BuEt)₂, with high anodic stability.⁷⁴ The electrolyte was prepared by mixing MgR₂ (R is butyl or ethyl) and AlCl₂R/AlCl₃ in tetrahydrofuran (THF). The reaction between RMgCl and R_xAlCl_3−x produced R_xAlCl_4−x⁻ and active Mg-containing cations (e.g., MgCl⁺ and Mg₂Cl₃⁺).^75–78 Importantly, the Mg(AlCl₂BuEt)₂ electrolyte exhibited a large electrochemical window of 2.5 V vs. Mg²⁺/Mg, superior to electrolytes without the addition of AlCl₃ (e.g., Mg(BPh₂Bu₂)₂ and BuMgCl). Moreover, all phenyl complex (APC) electrolytes, the most commonly used electrolytes for MMBs, can be obtained by substituting the alkyl group of RMgCl with the phenyl group. These APC electrolytes can demonstrate wide stable potential windows of more than 3 V vs. Mg²⁺/Mg.⁷⁹ Recent studies revealed that reducing the R_xAlCl_4−x⁻ concentration and increasing the AlCl₄⁻ concentration further improved the anodic stability of electrolytes. The low HOMO level of AlCl₄⁻ played an important role in improving the potential window. Zhao et al.⁸⁰ further improved the anodic stability of APC electrolytes by adding 1-butyl-1-methylpiperidinium chloride (PP₁₄Cl). The Cl⁻ ions reacted with R_xAlCl_4−x⁻ to produce AlCl₄⁻, consequently increasing the anodic stability from 2.9 to 3.1 V vs. Mg²⁺/Mg.

Magnesium aluminium chloride complex (MACC) dissolved in ether solvents (e.g., dimethoxyethane (DME)) represents an important type of MMB electrolytes with all-inorganic salts (e.g., AlCl₃ and MgCl₂).^81–83 During the preparation of MACC electrolytes, different acid–base reactions occur depending on the ratio between AlCl₃ and MgCl₂. For MgCl₂ : AlCl₃ ratio of 2 : 1 and 1 : 1, the dominating Mg_xCl_y^2x−y species in the electrolyte are Mg₂Cl₃⁺ and MgCl⁺ through eqn (1) and (2). In these cases, Mg undergoes facile electrochemical stripping and plating due to the presence of Mg_xCl_y^2x−y species. Increasing the MgCl₂ : AlCl₃ ratio to 1 : 2 inhibits Mg plating due to the low concentration of Mg_xCl_y^2x−y species.⁸² Moreover, when the AlCl₃ : MgCl₂ ratio is higher than 2 : 1, the formation of AlCl₂⁺ would cause severe Mg metal corrosion.⁸⁴ Typical MACC electrolytes require the pre-cycling process to reduce the overpotential and improve the electrochemical stripping/plating coulombic efficiency.⁸⁵ The pre-cycling process is termed conditioning, and would complicate the production of MMBs. Conditioning-free MACC electrolytes can be obtained by adding additional salts like Mg(TFSI)₂ and Mg(hexamethyldisilazide)₂ (Mg(HMDS)₂).^84,86 Besides, these Mg-salt additives can improve the water-resistance of MACC electrolytes, allowing excellent Mg plating at a water content of 2000 ppm. Electrochemical Mg plating is also feasible by completely replacing AlCl₃ in MACC electrolytes with either Mg(TFSI)₂ or Mg(HMDS)₂.^87,88 In the MgCl₂/Mg(TFSI)₂ electrolyte, salts and solvents must have high purity by removing water residue to realize the facile Mg stripping and plating. By contrast, purification is not necessary for the Mg(HMDS)₂/MgCl₂ electrolyte to achieve reversible Mg stripping and plating.

AlCl₃ + 2MgCl₂ → AlCl₄⁻ + Mg₂Cl₃⁺ (1)

AlCl₃ + MgCl₂ → AlCl₄⁻ + MgCl⁺ (2)

Chloride presence in MMB electrolytes accounts for the corrosive nature of electrolytes for common current collectors (e.g., Al and stainless steel), thus encouraging the development of chlorine-free Mg electrolytes. Mohtadi et al.⁹⁰ presented a chlorine-free electrolyte based on Mg(BH₄)₂ in DME. Nevertheless, the electrolyte displayed low coulombic efficiency (67%), and the decomposition of BH₄⁻ lead to the low anodic stability (2.3 V vs. Mg²⁺/Mg) of the electrolyte. Inspired by Mohtadi's work, Tutusaus and co-workers⁹¹ reported a halogen-free electrolyte with dodecaborate dianions for successful Mg stripping and plating. Impressively, the electrolyte demonstrated outstanding anodic stability of 3.8 V vs. Mg²⁺/Mg.⁹¹ Following this strategy, Luo et al.⁸⁹ synthesized Mg perfluorinated pinacolatoborate (FPB) for MMB electrolytes, which depicted outstanding anodic stability of 4.0 V vs. Mg²⁺/Mg. However, it should be pointed out that the synthesis of boron-based Mg salts generally requires harsh conditions under inert environments. Complicated synthetic routes impose a potential risk for the practical application of these boron-based electrolytes. A summary of the main features of developed Mg electrolytes is presented in Table 1.

Table 1 Comparison of developed MMB electrolytes

Electrolyte name	APC^79	Mg-HMDS^88	MACC^81,82	Mg(BRn)2 (ref. 89)
Main precursors	PhMgCl, AlCl3	Mg(HMDS)2, MgCl2	MgCl2, AlCl3	Mg(BH4)2, C6H2F12O2
Cations and anions	MgxCly^+ and PhxAlCl4−x^−	Mg2Cl3^+ and HMDSAlCl3^−	Mg2Cl3^+ and AlCl4^−	Mg^2+ and B(O2C2(CF3)4)2^−
Anodic stability	2.5 V vs. Mg^2+/Mg	2.8 V vs. Mg^2+/Mg	3.5 V vs. Mg^2+/Mg	4.0 V vs. Mg^2+/Mg
Coulombic efficiency	∼100%	99%	99%	95%
Advantages	High coulombic efficiency	Good cathode compatibility	Simple synthesis, high concentration	High anodic stability
Drawbacks	Low anodic stability	Low anodic stability	Conditioning required	Complicated synthetic process

---

With the demonstrated anode–electrolyte chemistries, searching compatible cathodes represents the main task to establish suitable full-device chemistries. The large charge density of Mg²⁺ and its strong binding to Cl⁻ and solvent molecules rule out many recognized cathode materials in other battery systems. The first demonstrated MMB cathode with decent performance is the Chevrel-phase Mo₆S₈.⁷⁴ Theoretical studies suggest that the Chevrel phase surface can significantly accelerate the desolvation and dissociation of Mg₂Cl₃⁺ in THF, reducing the dissociation energy of Mg₂Cl₃⁺ from 3.0 eV to 0.2 eV.⁹² For this reason, the Chevrel phase is one of a few inorganic materials that can be used as cathodes of MMBs without structure modification or special conditions (e.g., elevated temperature).

Besides, TMOs are considered promising materials for energy storage due to their high theoretical capacities and operation potentials. Apart from classic Mg²⁺ intercalation, TMO can also store charge through the conversion mechanism. The charge storage mechanism is determined by the type of TMOs and the nature of electrolytes.^93–95 However, TMO cathodes for MMBs suffer from large voltage hysteresis and sluggish intercalation kinetics. These issues can be assigned to the strong electrostatic interactions between O atoms in TMOs and Mg²⁺.⁹⁶ In addition, the compatibility of TMOs with chloride-containing electrolytes is poor, making the fabrication of MMBs with TMO cathodes a great challenge.⁹⁴ Another class of MMB cathodes is TMDs, which display better Mg²⁺-intercalation kinetics than TMOs in light of the weak electrostatic interactions between chalcogenide atoms and Mg²⁺. Unlike TMOs, TMDs are compatible with chloride-containing electrolytes. However, the main issue associated with TMDs is their much lower theoretical capacities and redox potentials than those of TMOs. The charge storage mechanism was shown to be either intercalation reaction or intercalation/conversion reaction with Mg²⁺, MgCl⁺, or [Mg(DME)₃]²⁺ as charge-carrier species (Fig. 4).

Fig. 4 Schematic illustration showing the charge storage mechanism of MMBs.

2.3 Al-metal batteries

Trivalent Al³⁺ and the high abundance of Al resources make Al-metal batteries (AMBs) to be promising candidates for large-scale energy storage. However, electrochemical Al stripping and plating is challenging in both organic and ionic liquid electrolytes. The large charge density of Al³⁺ induces the strong coulombic anion–cation interactions in Al salts (e.g., Al(TFSI)₃ and Al(OTf)₃), which further results in the low solubility of Al salts in common carbonate solvents. On the other hand, although ether-based solvents (e.g., DME and diglyme) can dissolve Al salts, these solvents exhibit strong interactions with cations, leading to high solvation energies.^97,98 To achieve facile Al stripping/plating, the commonly used chemistry takes use of Al₂X₇⁻ (X is Cl or Br), as shown in eqn (3).^99–102 The Al plating chemistry was first discovered in high-temperature (140–180 °C) molten salts, such as AlCl₃–LiCl and AlCl₃–NaCl–KCl.^100,101 However, the requirement of high temperature would greatly restrict the practical application of AMBs in energy storage devices. To this end, room-temperature ionic liquids with low viscosity are employed for AMBs. The most commonly used ionic liquids are mixtures of AlCl₃ and 1-ethyl-3-methylimidazolium chloride (EMIMCl). To achieve Al plating, the AlCl₃ : EMIMCl molar ratio should be higher than 1 to ensure the presence of Al₂Cl₇⁻ through the reactions of eqn (4) and (5). The standard reduction potential of Al₂Cl₇⁻ is −0.7 V vs. SHE (eqn (3)), which is almost 1.0 V higher than the direct reduction of Al³⁺ (−1.66 V vs. SHE).¹⁰³

4Al₂X₇⁻ + 3e⁻ ↔ Al + 7AlX₄⁻ (3)

AlCl₃ + EMIMCl ↔ AlCl₄⁻ + EMIM⁺ (4)

AlCl₃ + AlCl₄⁻ ↔ Al₂Cl₇⁻ (5)

Besides, mixtures of AlCl₃ with amides (e.g., urea and acetamide) can serve as economical alternatives for ionic liquid-based electrolytes.^104,105 In these cases, AlCl₄⁻, Al₂Cl₇⁻, and AlCl₂⁺ are formed through the reactions in eqn (5) and (6). Thermodynamically, Al plating can also be achieved through the reduction of AlCl₂⁺. However, calculations showed that the large energy barrier associated with the dissociation of Al-amide bonds restricted this reaction (Fig. 5).¹⁰⁴ This fact could explain why electrochemical Al striping/plating occurs only with the presence of Al₂Cl₇⁻. While amide/AlCl₃-based electrolytes are appealing for replacing ionic liquid-based electrolytes from a cost perspective, they suffer from high viscosity and low ionic conductivity. These drawbacks result in the inferior performance of AMBs with amide/AlCl₃ electrolytes, compared with AMBs with EMIMCl/AlCl₃ electrolytes.¹⁰⁵

2AlCl₃ + 2urea ↔ AlCl₄⁻ + AlCl₂(urea)₂⁺ (6)

Fig. 5 Energy profiles of the dissociation reactions of (a) Al₂Cl₇⁻ and (b) [AlCl₂(acetamide)₂]⁺. Reprinted with permission from ref. 104. Copyright 2018 Elsevier.

Intercalation of large-charge-density Al³⁺ usually leads AMB cathodes to poor cyclability and low-rate capability. For example, inorganic cathode materials (e.g., TiO₂,²⁹ TiS₂,¹⁰⁶ Co₉S₈,¹⁰⁷ and V₂O₅ (ref. 108)), with charge storage involving Al³⁺, demonstrate significant capacity fading during the first few cycles. In addition to the sluggish Al³⁺ intercalation, the high dissociation energy of Al₂Cl₇⁻ to Al³⁺ (eqn (7)) is another important reason for the low rate capability and large charge/discharge voltage hysteresis of AMBs with Al³⁺ storage. Recently, Yang et al.¹⁰⁹ reported that the dissociation rate of Al₂Cl₆Br⁻, prepared by mixing AlCl₃ with EMIMBr, was 15-fold faster than that of Al₂Cl₇⁻. However, the voltage window of this Al₂Cl₆Br⁻-based electrolyte exhibited a narrow potential window, which was only suitable for low-potential cathodes, such as TiS₂ (ref. 106) and TiO₂.²⁹ Other Al-based species, such as AlCl₂⁺ and AlCl₄⁻, generally act as favourable charge carriers for AMB cathodes (Fig. 6).^21,110,111 In detail, AlCl₄⁻ can serve as the charge carrier for graphite and amine compounds.^21,112,113 On the other hand, n-type organic electrodes (e.g., pyrazine,¹¹⁴ polyaniline,¹¹⁵ and phenanthrenequinone¹¹⁰) can store charges through redox reactions with AlCl₂⁺. Compared with Al³⁺, AlCl₂⁺ as the charge carrier allows cathodes with higher charge/discharge rates and better cyclability. The intercalation of Al³⁺ was disclosed to be challenging in most materials, pushing towards the development of new n-type and p-type organic materials for AlCl₂⁺/AlCl₄⁻ storage.

4Al₂Cl₇⁻ → 4AlCl₄⁻ + 4AlCl₃ → 7AlCl₄⁻ + Al³⁺ (7)

Fig. 6 Schematic illustration showing the charge storage mechanism of AMBs.

3. Layered cathode materials for MVMBs

Layered materials present 2D topologies with flat planes, forming a bulk structure through the weak vdW interaction. They have been widely used for both cation and anion storage due to their potentially high theoretical capacities and favourable 2D ion transport channels.^41,44,46 Specifically, efficient cation/anion storage in layered materials can be promoted by large interlayer spacing, weak interactions between intercalating ions and host lattices, and the electronic conductivity of layered materials.^116,117 Therefore, both intrinsic properties of layered materials and the applied structure modification have a prominent effect on the charge storage capability. In this section, we present the current state-of-the-art layered cathode materials used for non-aqueous MVMBs. The materials in this section are divided into four major groups, layered TMOs, TMDs, graphite-based materials, and 2D COFs. We analyse the proposed structure modification approaches and their influence on both the intrinsic physicochemical properties and the electrochemical performance enhancement of layered materials.

3.1 Layered transition metal oxides

TMOs include a broad class of materials with diverse structural frameworks, which exhibit excellent stability, environmental friendliness, and simple synthetic routes.^117–120 The multiple valence states of transition metal elements arise from the coordination of the transition metal with highly electronegative oxygen, empowering TMOs with the superior charge storage capability (e.g., superior specific capacity). In addition, the high ionic feature of M–O bonds results in relatively high electrochemical redox potentials.³⁷ However, the oxygen atoms of TMOs impose strong interactions with charge carrier ions (e.g., Zn²⁺, Mg²⁺, and Al³⁺), which leads to the limited ion-storage kinetics and poor structural stability of TMOs used in MVMBs.^25,121 In this sense, layered TMOs allow solid-state ion diffusion through the 2D interlayer space, providing efficient ion transport pathways.¹²² Furthermore, structure modification of layered TMOs aims at weakening the interaction between intercalating ions and oxygen atoms. So far, several layered TMOs have been explored for MVMB cathodes, including manganese oxide, vanadium oxide, and vanadyl phosphate (VOPO₄).

3.1.1 Manganese oxides. Manganese oxides have two layer-structure phases, namely, ε- and δ-phase MnO₂. The δ-phase MnO₂ (birnessite) is believed to provide more efficient pathways for cation diffusion than the ε-phase MnO₂.¹²³ Both MnO₂ phases are composed of Mn⁴⁺ ions surrounded by six oxygen atoms, creating [MnO₆] octahedral units. The [MnO₆] units are linked via edge corners to form the layer structure.^124–127 Birnessite MnO₂ exhibits an interlayer distance of 7.24 Å, with a crystal water layer between two [MnO₆] octahedral layers. It is promising battery cathode material in light of the high specific capacity (308 mA h g⁻¹ for Mn³⁺/Mn⁴⁺ transition), nontoxicity, and low cost.

Due to the high specific capacity of MnO₂,⁵⁸ extensive researches have been conducted to use MnO₂ as cathodes for aqueous ZMBs. Recently, Han et al.⁶⁶ studied the charge storage performance of the δ-MnO₂ cathode using Zn(TFSI)₂ in AN as the electrolyte. δ-MnO₂ was prepared by the reaction between KMnO₄ and MnSO₄, displaying a nanofloret-like morphology with an interlayer spacing of 7 Å. The Zn²⁺-intercalation mechanism of δ-MnO₂ nanoflorets was studied using ex situ X-ray diffraction (XRD) and scanning transmission electron microscopy energy dispersive X-ray spectroscopy (STEM-EDS). After the first discharge, the (001) and (002) peak intensity of δ-MnO₂ was significantly decreased, indicating the loss of long-range order in the layer direction. This was caused by the structural transformation of δ-MnO₂ to the non-layered polymorph. The structural change caused the specific capacity increase to 123 mA h g⁻¹ at 12.3 mA g⁻¹ during the first 20 cycles. After 30 cycles, the specific capacity started to decay, reaching 55 mA h g⁻¹ after 125 cycles. The decrease was attributed to the severe electrolyte decomposition and the formation of the passivation layer on the anode.

MnO₂ was also studied as the cathodes of MMBs in light of its high redox potential vs. Mg²⁺/Mg. However, Mg²⁺ intercalation into layered MnO₂ was not feasible with the electrolyte consisting of anhydrous Mg(ClO₄)₂ in AN.¹²⁸ Therefore, the water-containing electrolyte was prepared using Mg(ClO₄)₂·xH₂O with the H₂O : Mg²⁺ ratio of 6. After the addition of water, Mg²⁺ intercalation was improved from a ratio of <0.2 Mg²⁺ per Mn for anhydrous Mg(ClO₄)₂ to a ratio of 0.7 Mg²⁺ per Mn (Fig. 7a). Interestingly, cycling with the water-containing electrolyte acted as an activation step of MnO₂. The activated MnO₂ maintained the good Mg²⁺-storage performance in the anhydrous electrolyte. The authors claimed that, during the discharge/charge cycling in the water-containing electrolyte, a slight structure re-orientation occurred, causing the activation of MnO₂. However, due to the low crystallinity of the sample, the exact phase transition was not detected.

Fig. 7 (a) Inductively coupled plasma-atomic emission spectroscopy Mg : Mn ratios of MnO₂ nanowire electrodes discharged at −0.4 V vs. Ag/AgCl for 15 minutes in electrolytes with various H₂O/Mg²⁺ ratios and expected Mg : Mn ratios which were calculated based on the total accumulated charges. Error bars indicate 95% confidence intervals. Reproduced from ref. 128 with permission from the PCCP Owner Societies. (b) Birnessite crystal structure showing a water monolayer between the MnO₂ sheets. (c) The water content of the electrolyte as determined by Karl Fischer titration, showing the full water release after 20 cycles. (d) Cycling performance of MnO₂·0.9H₂O at C/10 and C/5. (e) Mn 2p_3/2 XPS spectra of the pristine, discharged, and charged Mg-birnessite/carbon cloth electrodes; fits are shown in colour as labelled. Reprinted with permission from ref. 95. Copyright 2016 American Chemical Society.

To further understand the role of water in MnO₂, Sun et al.⁹⁵ investigated Mg²⁺ intercalation into Mg_0.15MnO₂·0.9H₂O (Fig. 7b). Using Karl Fischer titration, it was disclosed that water molecules in the Mg_0.15MnO₂·0.9H₂O structure were removed during the first 20 cycles. Remarkably, the first 20 cycles were also the cycles required for the full conditioning of the Mg_0.15MnO₂·0.9H₂O cathode (Fig. 7c and d). The activation was also necessary for the water-containing electrolyte, suggesting that the activation is related to the structural change of Mg_0.15MnO₂·0.9H₂O. X-ray photoelectron spectroscopy (XPS) measurements detected the exitance of MnOOH, MnO, and Mg(OH)₂ in the electrode at different charged states (Fig. 7e). In detail, the original electrode contained 77% Mn⁴⁺, 20% Mn³⁺, and 3% Mn²⁺, which was indicative of the chemical formula Mg_0.15MnO₂·0.9H₂O. At the discharged state, peaks corresponding to MnOOH and MnO were observed. At the charged state, MnO was still detected, indicating that the electrochemical process was partially reversible. These results indicated that the magnesiation process of Mg_0.15MnO₂·0.9H₂O was accompanied by the conversion reaction, as shown in eqn (8). Moreover, the small specific capacity of 135 mA h g⁻¹ after 20 cycles suggested that the electrochemical charge storage reaction was limited to the surface of Mg_0.15MnO₂·0.9H₂O.

Overall, water molecules in the electrolyte or the crystal structure of MnO₂ play a vital role in promoting the initial intercalation of Mg²⁺ into MnO₂. After the pre-cycling in the water-containing electrolyte, MnO₂ undergoes structural changes, and the water-containing electrolyte can be replaced by the anhydrous electrolyte. Since Mg metal is not stable in the water environment, water removal is an important step for the full Mg//MnO₂ cell.

Mg_0.15MnO₂ + H₂O + xMg²⁺ + 2xe⁻ ↔ (2x − 0.7)MnO + (1.7 − 2x)MnOOH + (x + 0.15)Mg(OH)₂ (8)

3.1.2 Vanadium oxides. Orthorhombic vanadium pentoxide (V₂O₅) is a typical layered material with an interlayer spacing of 4.37 Å.¹²⁹ The layer is composed of [VO₅] pyramids connected via edge-sharing. [VO₅] pyramids show two types of bonds, namely, the O=V double bond and the V–O–V bond.¹³⁰ The V⁵⁺/V⁴⁺ redox couple empowers V₂O₅ with a high theoretical specific capacity of 295 mA h g⁻¹. In addition, V₂O₅ displays a high redox potential of ∼3.2 V vs. Li⁺/Li, making V₂O₅ a promising cathode material for MVMB cathodes.¹³¹

Novák and Desilvestro,¹³³ for the first time, demonstrated the intercalation of Mg²⁺ into V₂O₅ by adding water to the Mg(ClO₄)₂ in AN electrolyte. The intercalation kinetics of Mg²⁺ was significantly improved through the shielding effect of water. Inspired by this study, several studies showed that the incorporation of water into the V₂O₅ structure could attain a similar shielding effect. For example, An et al.¹³² synthesized the V₂O₅·1.4H₂O nanowire/graphene hybrid (denoted VOG), which presented greatly improved conductivity and Mg²⁺-intercalation kinetics (Fig. 8a and b). VOG delivered a large specific capacity of 330 mA h g⁻¹ at 50 mA g⁻¹ and high capacity retention of 80% after 200 cycles. For comparison, VOG without crystal water was subsequently prepared via annealing. The annealed VOG exhibited a very limited Mg²⁺-intercalation performance with a small specific capacity of <75 mA h g⁻¹ at 100 mA g⁻¹ (Fig. 8c). Remarkably, XRD patterns showed only a small difference between the interlayer distance of VOG (11 Å) and the annealed VOG (10 Å). It implied that, compared with the expanded interlayer spacing, the shielding effect of water molecules played the dominant role in promoting the Mg²⁺-storage kinetics (Fig. 8d). Significantly, water molecules were located at the interstitial sites of V₂O₅·1.4H₂O, and kept confined within the V₂O₅ lattice during charge/discharge, even at elevated temperatures.^132,134 It thereby explained the excellent capacity retention of VOG over the prolonged cycling (100 mA h g⁻¹ after 200 cycles at 1 A g⁻¹).

Fig. 8 (a) Shielding effect of water in V₂O₅·xH₂O. The strong polarization of divalent Mg²⁺ could be significantly reduced by solvating with crystal water molecules. (b) Schematic illustration of the V₂O₅·xH₂O/graphene nanocomposite with bi-continuous electron/ion transport pathways, large area of electrode–electrolyte interface, and facile strain relaxation during Mg²⁺ insertion/extraction. (c) Charge/discharge curves of hydrate V₂O₅·xH₂O (VOG-1), V₂O₅·xH₂O after heating in Ar (VOG-2) and in air (VOG-3) at 350 °C for 2 h. (d) XRD spectra of V₂O₅·1.35H₂O before and after annealing in Ar. Reprinted with permission from ref. 132. Copyright 2015 Elsevier.

Metal atom incorporation into V₂O₅ interlayers (e.g., Mg, Mn, or Na)^135–139 represents another effective strategy to promote the stability of layered V₂O₅. For instance, Mg incorporation was achieved through electrochemically discharging V₂O₅ nanowires to 0.2 V vs. Mg²⁺/Mg.¹³⁶ Near edge X-ray absorption fine structure (NEXAFS) of V K-edge revealed that the edge position of the Mg-incorporated V₂O₅ was close to that of VO₂ standard, indicating a stoichiometry of MgV₂O₅ for the Mg-incorporated V₂O₅. Moreover, the decreased intensity of the pre-edge peak suggested a local reordering of the V environment upon Mg incorporation. Significantly, the obtained MgV₂O₅ electrode exhibited a high specific capacity of 160 mA h g⁻¹ at 20 mA g⁻¹. Nevertheless, the electrochemical performance of pristine V₂O₅ was not presented, making it difficult to quantify the contribution of Mg incorporation. Mg can also be incorporated into the crystal structure of V₂O₅ during the material synthesis. Xu et al.¹³⁵ synthesized Mg_0.3V₂O₅·1.1H₂O nanowires through the reaction between V₂O₅ and C₄H₆O₄Mg·4H₂O. The [MgO₆] octahedra species acted as pillars and enhanced the structural stability of V₂O₅, while water molecules improved the Mg²⁺-intercalation kinetics (Fig. 9a). The Mg_0.3V₂O₅·1.1H₂O electrode delivered a high specific capacity of 164 mA h g⁻¹ (Fig. 9b), which is substantially higher than those of V₂O₅·1.1H₂O (114 mA h g⁻¹) and Mg_0.3V₂O₅ (91 mA h g⁻¹). Impressively, the Mg_0.3V₂O₅·1.1H₂O electrode also showed a superior cycling life with 80% capacity retained after 10 000 cycles at a current density of 1 A g⁻¹, which remarkably outclassed V₂O₅·1.1H₂O and Mg_0.3V₂O₅ (fast capacity decay over 200 cycles). The incorporation of Mg into the structure was uncovered to improve the conductivity of Mg_0.3V₂O₅·1.1H₂O (Fig. 9c and d), which contributed to the superior rate performance. XRD studies further observed the slight decrease by 0.5 Å in the interlayer spacing of Mg_0.3V₂O₅·1.1H₂O after charging. This result indicated that the [MgO₆] sites as pillars alleviated the electrode structure change during the charge/discharge process, further accounting for the excellent cycle stability of Mg_0.3V₂O₅·1.1H₂O.^135,137

Fig. 9 (a) Schematic showing the crystal structure of Mg_0.3V₂O₅·1.1H₂O. (b) Charge–discharge curves of Mg_0.3V₂O₅·1.1H₂O, V₂O₅·xH₂O, and Mg_0.3V₂O₅. (c) Scanning electron microscope images of the Mg_0.3V₂O₅·1.1H₂O single-nanowire devices. (d) I–V curves of Mg_0.3V₂O₅·1.1H₂O, V₂O₅·xH₂O, and Mg_0.3V₂O₅. Reprinted with permission from ref. 135. Copyright 2019 Elsevier.

Besides, ammonium¹⁴⁰ or poly(ethylene oxide) (PEO)¹⁴¹ incorporated between vanadium oxide layers can also promote the Mg²⁺-intercalation kinetics. Impressively, NH₄V₄O₁₀ exhibited the initial specific capacity of 175 mA h g⁻¹ in the first cycle, and the specific capacity increased to ∼250 mA h g⁻¹ in the second cycle. The boosted specific capacity was assigned to the irreversible de-ammoniation process occurring at the first charge, which created additional sites for Mg²⁺ storage. This de-ammoniation process was identified by the irreversible oxidation peak in the first cyclic voltammetry cycle and the disappearance of the N 1s XPS peak after the first charge. Similarly, PEO-incorporated V₂O₅·1.5H₂O displayed a superior specific capacity of 100 mA h g⁻¹, which substantially outweighed the low specific capacity of V₂O₅·1.5H₂O (20 mA h g⁻¹ at 10 mA g⁻¹).¹⁴¹ In addition, the specific capacity of PEO-incorporated V₂O₅·1.5H₂O was stabilized at ∼90 mA h g⁻¹ after 20 cycles, in contrast to the fast capacity decay of V₂O₅·1.5H₂O (4.3 mA g⁻¹ after 35 cycles). The improved specific capacity of PEO-incorporated V₂O₅·1.5H₂O was assigned to the interlayer spacing expansion from 11.6 Å for V₂O₅·1.5H₂O to 12.6 Å for PEO-incorporated V₂O₅·1.5H₂O.

The Mg²⁺-intercalation kinetics of pristine vanadium oxide can also be improved by the charge/discharge at elevated temperatures. Rastgoo-Deylami et al.¹⁴² showed that the capacity of V₃O₇·H₂O was boosted from 80 mA h g⁻¹ to 231 mA h g⁻¹ when the temperature was increased from 25 °C to 60 °C. Moreover, initial cycling of the electrode at the elevated temperature was also explored as an activation step for further cycling at room temperature.⁹³ To activate the cathode, commercial α-V₂O₅ was discharged/charged at 110 °C with the electrolyte of 0.5 M Mg(TFSI)₂ in 1-butyl-1-methyl-pyrrolidinium TFSI (Py₁₄TFSI). EDS and STEM analysis were used to evaluate the effect of the activation step on the structure of α-V₂O₅. First, the α-V₂O₅ cathode after discharge at 110 °C exhibited the uniform distribution of V and Mg, implying that the charge storage mechanism of α-V₂O₅ was high-kinetics Mg²⁺ intercalation. STEM images of α-V₂O₅ at the discharged state revealed the considerable delamination of α-V₂O₅, which lead to the greatly reduced domain size (3.5 nm). The specific capacity of α-V₂O₅ without activation was only 16 mA h g⁻¹. Interestingly, after the temperature was increased to 110 °C, the specific capacity increased to 295 mA h g⁻¹. Afterwards, the specific capacity could be stabilized at 96 mA h g⁻¹ after the temperature recovered to room temperature. It should be noted that the activation step of the cathode would be a high-cost step in the battery fabrication process, thus restricting the real-life application of the assembled devices.

Although preliminary studies demonstrated the feasible Mg²⁺ storage of layered vanadium oxides in electrolytes like Mg(TFSI)₂ or Mg(ClO₄)₂ in AN and ionic liquids, these electrolytes are generally not compatible with Mg metal anodes. To enable full Mg//V₂O₅ cell, more efforts are needed to develop electrolytes compatible with both V₂O₅ cathodes and Mg metal anodes.

Unlike MMB electrolytes, metal anode-compatible AMB electrolytes (e.g., AlCl₃/EMIMCl) are also suitable for oxide cathodes. AMBs composed of Al metal, V₂O₅ nanowires, and AlCl₃/EMIMCl (molar ratio: 1.1) were assembled.¹⁴³ As revealed, the V₂O₅ cathode delivered a high specific capacity of 305 mA h g⁻¹ in the first cycle with a voltage plateau of around 0.5 V. The specific capacity kept 273 mA h g⁻¹ after 20 cycles, suggesting decent capacity retention of the V₂O₅ cathode. Recent studies have indicated that stainless steel exhibited nonnegligible electrochemical activity in AlCl₃/EMIMCl electrolyte.¹⁴⁴ This electrochemical activity could provide false interpretation of the obtained results. To further understand the intercalation of Al³⁺ into V₂O₅, Gu and co-workers¹⁰⁸ conducted the high-resolution TEM (HRTEM) analysis on V₂O₅ nanowires at different charge/discharge stages. They found that the intercalation of Al³⁺ led to the formation of 10 nm amorphous layers on the outer shell of V₂O₅ nanowires. After de-intercalation of Al³⁺, a new V₂O₅ phase appeared between the outer amorphous layer and the core of nanowires (Fig. 10a–c). The structural evolution suggested that the charge storage of V₂O₅ nanowires occurred only at the surface, and the charge storage mechanism was a combination of intercalation and phase transition. Furthermore, ex situ XPS measurements revealed that the oxidation state of V was not fully converted back to V⁵⁺ during the charging process. In addition, XPS also indicated the presence of Al after charging. These XPS results were consistent in several studies of AMBs with V₂O₅ cathodes.^108,145 Due to the irreversible electrochemical process, the specific capacity of V₂O₅ decayed fast during cycling, retaining only 40 mA h g⁻¹ after 10 cycles (Fig. 10d). The compatibility of the AMB electrolytes with V₂O₅ will motivate further investigations to improve both the Al³⁺-intercalation kinetics and the cycling stability of V₂O₅via diverse structure engineering strategies.

Fig. 10 (a and b) HRTEM images of the V₂O₅ nanowire (a) before cycling, and (b) at the fully discharged state (insets: fast Fourier transform images). (c) Schematic diagram showing the electrochemical Al³⁺ insertion/extraction of crystallized V₂O₅ nanowires. (d) Galvanostatic charge/discharge profiles of the V₂O₅ nanowire cathode. Reprinted with permission from ref. 108. Copyright 2017 Elsevier.

3.1.3 Vanadyl phosphate. Layer-structure VOPO₄ has several phases, such as α₁-, α₂-, γ-, δ-, and ω-VOPO₄. The common ones for electrochemical intercalation are α₁-VOPO₄ and α₂-VOPO₄ with a P4/n space group.¹⁴⁶ Both phases are based on the corner-sharing [PO₄] tetrahedra and strongly distorted vanadium octahedra. The difference between α₁-VOPO₄ and α₂-VOPO₄ is the position of oxygen atoms in vanadyl groups. The oxygen atoms of α₁-VOPO₄ are located within the layer, while the oxygen atoms of α₂-VOPO₄ are not in the same plane as V (Fig. 11).¹⁴⁷ VOPO₄ has attracted significant attention as a possible cathode material due to the large specific capacity, high redox potential, and stable cyclability.¹⁴⁸ Impressively, VOPO₄ can deliver a high theoretical specific capacity of 166 mA h g⁻¹ for the one-electron redox couple V⁴⁺/V⁵⁺ and 331 mA h g⁻¹ for the two-electron redox couple V³⁺/V⁵⁺.¹⁴⁹ Due to the enhanced ionic character of V–O bonds when (PO₄)³⁻ anions are introduced, VOPO₄ possesses a much higher redox potential (3.95 V vs. Li⁺/Li) than that of V₂O₅ (2–3.4 V vs. Li⁺/Li).^37,150

Fig. 11 Representation of (a) the α₁-VOPO₄ structure (C2/m, a = 8.73187(14) Å, b = 6.20548(5) Å, c = 6.20902(4) Å and β = 104.448(3)° at 280 °C) and (b) the α₂-VOPO₄ structure (P4/n, a = 6.0156(3) Å, c = 4.4375(3) Å at RT). Reproduced from ref. 147 with permission from The Royal Society of Chemistry.

VOPO₄ suffers from significant voltage decay during cycling in aqueous Zn electrolytes, which can be assigned to the decomposition of VOPO₄ into VO_x. Several studies suggested that low amounts of water additive would not cause the decomposition of VOPO₄.^59,151 Recently, Wang et al.¹⁵¹ compared the Zn²⁺-storage performance of VOPO₄ and VOPO₄·2H₂O in both water-free and water-containing electrolytes. VOPO₄·2H₂O was synthesized by the reaction between V₂O₅ and concentrated H₃PO₄ under the reflux condition. Subsequently, VOPO₄ was prepared through the calcination of VOPO₄·2H₂O at 550 °C in the air. The electrochemical performance was evaluated using 0.1 M Zn(OTf)₂ in AN with/without water. When the water content in the electrolyte was 0.5%, the performance of VOPO₄ showed only slight improvement. By contrast, the specific capacity of VOPO₄ was increased by 3-fold, when the water content reached 1%. The performance improvement was explained by the chemical co-intercalation of free water molecules and the formation of VOPO₄·xH₂O (Fig. 12a), which was supported by the XRD pattern (Fig. 12b). Besides, the electrochemical performance of VOPO₄·2H₂O was evaluated in AN with and without 1% water. Interestingly, the addition of water to the electrolyte can apparently improve the performance of VOPO₄·2H₂O. The specific capacities in the water-free electrolyte are 10 and 80 mA h g⁻¹ at 20 mA g⁻¹ for VOPO₄ and VOPO₄·2H₂O, respectively. When 1% water was added, the specific capacities were increased to 122 and 135 mA h g⁻¹ at 20 mA g⁻¹ for VOPO₄ and VOPO₄·2H₂O, respectively (Fig. 12c). XRD analysis revealed that water molecules were extracted from VOPO₄·2H₂O to the electrolyte, thus restricting the solid-state diffusion of Zn²⁺ (Fig. 12d and e). Apart from the increase of specific capacity, VOPO₄·2H₂O in the water-containing electrolyte exhibited an increase of the discharge voltage by ∼0.2 V and superior cycling capabilities with 100 mA h g⁻¹ retained after 25 cycles at 20 mA g⁻¹.

Fig. 12 (a) Schematic illustrations of VOPO₄ in 0.1 M Zn(OTf)₂ dissolved in AN without water and with 1% H₂O. (b) XRD patterns of VOPO₄ soaked in different electrolytes for 24 h. (c) The charge/discharge profiles of VOPO₄ and VOPO₄·2H₂O between 0.2 V and 1.9 V at 20 mA g⁻¹. (d) Schematic illustrations of VOPO₄·2H₂O in 0.1 M Zn(OTf)₂ dissolved in AN without water and with 1% H₂O. (e) XRD patterns of the pristine VOPO₄·2H₂O and VOPO₄·2H₂O soaked in the electrolyte of 0.1 M Zn(OTf)₂ dissolved in AN for 24 h. Reprinted with permission from ref. 151. Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

To further improve the cyclability of VOPO₄, Verma et al.⁵⁹ synthesized a polypyrrole-incorporated VOPO₄ (PPy-VOPO₄). Polypyrrole between VOPO₄ layers induced the partial reduction of V⁵⁺ to V⁴⁺, thus reducing the charge transfer resistance. The capacity retention of PPy-VOPO₄ was increased from 25% for the pristine VOPO₄ to 90% after 30 cycles at 30 mA g⁻¹. In addition, the cyclability was further improved by adding 10% volume of water to the electrolyte, thus enabling VOPO₄·2H₂O with pronouncedly enhanced capacity retention (60 mA h g⁻¹ after 350 cycles at 100 mA g⁻¹). The improved stability was ascribed to the presence of polypyrrole, which prevented the distortion of the lattice structure. In contrast, VOPO₄·2H₂O without polypyrrole suffered from the fast capacity decay caused by the extraction of crystal water.

VOPO₄ was also studied as the Mg²⁺-intercalation cathodes, and their performance showed strong dependence on both the interlayer spacing of VOPO₄ and the water content in the electrolyte. Ji et al.¹⁵² studied the effect of water molecules in both the VOPO₄ structure and the electrolyte on the Mg²⁺-intercalation performance of VOPO₄. Interestingly, only when the VOPO₄·H₂O cathode with water-containing electrolyte was used, the significant capacity enhancement could be observed. Theoretical calculation indicated that the energy barrier of Mg²⁺-intercalation was reduced from 1.56 eV for VOPO₄ to 0.48 eV for VOPO₄·H₂O. Such a pronounced difference resulted in that the diffusion coefficient of VOPO₄·H₂O was 1.2 × 10¹⁸ times higher than that of VOPO₄. In addition, the calculated formation energies of different [Mg(PC)_n(H₂O)_6−n]²⁺ (n ≤ 6) species in the PC electrolyte without water were much more negative than those in the water-containing PC electrolyte (Fig. 13a). The difference in formation energies revealed the immense desolvation energy of [Mg(PC)₆]²⁺. Additionally, the improved Mg²⁺-intercalation kinetics contributed to reducing the charge/discharge mid-voltage hysteresis from 1.15 V for VOPO₄ in the water-free electrolyte to 0.49 V for VOPO₄·2H₂O in the water-containing electrolyte (Fig. 13b). The average equilibrium voltages in electrolytes with various water contents were also calculated for the V⁵⁺/V⁴⁺ redox stage (Mg_0.5VOPO₄) and the V⁴⁺/V³⁺ redox stage (MgVOPO₄). The V⁴⁺/V³⁺ stage displayed a pronounced dependence on the water content in electrolytes compared with the V⁵⁺/V⁴⁺ stage. The voltage plateau improved from 1.79 V to 2.19 V, when the water activity was increased from 10⁻⁶ to 10⁻². To clarify the influence of water, Fig. 13c illustrates how Mg²⁺ intercalation into VOPO₄ and VOPO₄·2H₂O in both water-free and water-containing electrolytes.

Fig. 13 (a) Formation energies of [Mg(PC)_n]²⁺ and [Mg(PC)_n(H₂O)_6−n]²⁺ (n ≤ 6). (b) Voltage profiles of VOPO₄ and VOPO₄·2H₂O in the water-free (0.1 M Mg(ClO₄)₂ in PC) and water-containing (0.1 M Mg(ClO₄)₂·6H₂O in PC) electrolytes at 5 mA g⁻¹ in a three-electrode cell with Ag/AgCl and active carbon as the reference and counter electrode, respectively. (c) Schematic illustrating the charge storage mechanism of VOPO₄·nH₂O in the water-free and water-containing electrolytes. The green curves indicate the activation energy barriers, and the white dash lines represent the electrode/electrolyte interface. Reprinted with permission from ref. 152. Copyright 2018 American Chemical Society.

As mentioned earlier, water-containing electrolytes are not compatible with the Mg metal anode owing to the formation of the passivation layer on the Mg surface. By incorporating phenylamine molecules between VOPO₄ layers (denoted PA-VOPO₄), efficient Mg²⁺ intercalation in a water-free electrolyte was achieved.¹⁵³ PA-VOPO₄ was prepared by the exfoliation and self-assemble of VOPO₄·2H₂O nanosheets in a phenylamine solution (Fig. 14a). Phenylamine enabled VOPO₄ with apparent layer expansion (interlayer distance: 7.4 vs. 14.2 Å). The large interlayer distance of PA-VOPO₄ reduced the electrostatic interaction between the intercalating Mg²⁺ and VOPO₄. Meanwhile, the interlayer expansion also boosted the charge-storage kinetics of VOPO₄ by promoting the intercalation of MgCl⁺ (Fig. 14b–d).^153,154 Impressively, PA-VOPO₄ kept 87% of the initial specific capacity after 500 cycles at 100 mA g⁻¹, which was superior to that of VOPO₄·2H₂O (40% after 150 cycles). Moreover, PA-VOPO₄ showed a decent rate performance with high specific capacities of 275 and 109 mA h g⁻¹ at 50 and 2000 mA g⁻¹, respectively (Fig. 14e). The excellent cycling stability of PA-VOPO₄ is rather unique, given the APC electrolyte is known to be incompatible with oxides. Recently, Hu et al.¹⁵⁵ suggested that the incorporated PA molecules could improve the stability of VOPO₄ in aqueous ZMBs due to the increased hydrophobicity of PA-VOPO₄. It is highly desirable to understand the stability of PA-VOPO₄ in the APC electrolyte with future efforts. In addition, the low charge/discharge voltages and the higher specific capacities of PA-VOPO₄ than the theoretical value should be further clarified.

Fig. 14 (a) Schematic illustration of the experimental approach and proposed mechanism of PA-VOPO₄ nanosheets as the Mg-storage material. (b) Mg 1s and Cl 2p XPS spectra of PA-VOPO₄ nanosheets at the fully charged/discharged states. Mass change of the electrodes upon discharge when (c) PA-VOPO₄ nanosheets and (d) VOPO₄·2H₂O bulk served as cathode materials. (e) Rate capability of VOPO₄·2H₂O bulk and PA-VOPO₄ nanosheets. Reprinted with permission from ref. 153. Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

3.2 Transition metal dichalcogenides

Transition metal dichalcogenides, sharing the formula MX₂, are made of transition metal (M = Ti, V, Nb, Mo, W, etc.) and chalcogen (X = S, Se, and Te) atoms. The TMD layer has a three-atom-thick configuration of X–M–X. Stacked layers are connected by weak vdW interactions, allowing the exfoliation of bulk TMD into single-layer and few-layer TMD nanoflakes.⁴² TMDs have three polytypes, namely, trigonal prismatic, octahedral, and distorted octahedral, which differ in the metal atom coordination configuration and stacking orders. The electronic properties of TMDs (metallic, semiconducting, and insulating) are determined by the different polytypes and chalcogen types.^41,156 The weaker electrostatic interactions between chalcogenide anions and the intercalating cations account for the better multivalent metal ion-intercalation kinetics of TMD cathodes than that of TMO cathodes.¹⁵⁷ Generally, TMDs exhibit the intercalation mechanism at potentials above ∼1.0 V vs. Li⁺/Li. Under a low cut-off voltage (below ∼1.0 V vs. Li⁺/Li), the specific capacities of TMDs are substantially enhanced due to the involved conversion reactions.

3.2.1 Ti-based TMDs. Titanium disulphide (TiS₂) and titanium diselenide (TiSe₂) show tetragonal symmetry and metallic properties, exhibiting layer spacings of 5.65 Å and 6.00 Å, respectively. TiS₂ depicts a theoretical specific capacity of 239 mA h g⁻¹ for the one-electron Ti³⁺/Ti⁴⁺ reaction, as shown in eqn (9). Compared with TiS₂, TiSe₂ has a lower theoretical specific capacity of 130 mA h g⁻¹, but better cation-intercalation kinetics.¹⁵⁷ The superior cation-intercalation kinetics of TiSe₂ originates from the unique hybridization between Ti 3d orbitals and Se 4p orbitals, which boosts the electron delocalization and reduces the electrostatic interactions between the intercalating cations and TiSe₂ (Fig. 15a and b).^157,158 In addition, TiSe₂ displays the lowest charge rehybridization upon Mg²⁺ intercalation in comparison with TiS₂ and TiO₂, further implying the lowest Mg²⁺-diffusion barrier of TiSe₂ (Fig. 15c).¹⁵⁷

Fig. 15 (a) Energy diagram of atomic orbitals. Red circles, blue squares and orange triangles represent the energy levels of 3d-, 4d- and 5d-orbitals of each transition metal, respectively. The horizontal lines display the energy level position of O 2p-orbital, S 3p-orbital and Se 4p-orbital, respectively. (b) A schematic illustration of charge distribution in the electronic state with strong d–p hybridization. Electrons are accommodated in the delocalized state, which extends over transition metal and ligand atoms as schematically shown by the red framework. Reprinted with permission from ref. 158. Copyright 2015, Springer Nature. (c) Charge rehybridization upon the diffusion of Mg²⁺ between layers of TiO₂, TiS₂, and TiSe₂ with layer spacing (unit Å). Charge accumulation is shown in yellow, while depletion is shown in blue. Oxygen atoms are shown as small red spheres, sulfur as small yellow spheres, selenium as small kelly spheres, Ti as large sapphire spheres, and Mg at the center as an orange sphere. The isovalue number used for displaying the differential charge density is 0.003. Reprinted with permission from ref. 157. Copyright 2019 American Chemical Society.

Several strategies were applied to improve the Mg²⁺-intercalation performance of Ti-based TMDs, such as particle size optimization, temperature increase, and interlayer expansion. Tao et al.¹⁵⁹ studied the Mg²⁺-intercalation behaviours of the TiS₂ nanotubes and large-flake TiS₂ (20 μm). TiS₂ nanotubes (234 mA h g⁻¹) delivered more than double the capacity of the large-flake TiS₂ (96 mA h g⁻¹) at 10 mA g⁻¹ (Fig. 16). In addition to the particle size, the Mg²⁺-intercalation kinetics of TiS₂ was shown to be highly temperature-depended. For example, Sun et al.¹⁶⁰ demonstrated efficient Mg²⁺ intercalation into TiS₂ using the APC electrolyte at 60 °C. The TiS₂ electrode at 60 °C exhibited the initial specific capacity of 270 mA h g⁻¹ at 12 mA g⁻¹ and a reversible capacity of 160 mA h g⁻¹ in the following cycles. Interestingly, a fast decay of the specific capacity (to ∼110 mA h g⁻¹ after 40 cycles) suggested that neither TiS₂ nor the Mg anode was stable during cycling at high temperatures. To understand the Mg²⁺-intercalation mechanism of TiS₂, in situ XRD revealed the appearance of several new phases during the discharge process of TiS₂. At the initial discharge stage, partially irreversible interlayer expansion (c parameter increase) occurred, followed by the reversible increase in the a parameter. The increase in the c parameter was explained by Mg²⁺ intercalation into the octahedral sites of TiS₂, while the change in the a parameter indicated Mg²⁺ intercalation into the tetrahedral sites.

Fig. 16 (a) Cyclic voltammograms at 0.5 mV s⁻¹ of TiS₂ nanotube (solid) and polycrystalline powder (dots) electrodes at 20 °C. (b) Discharge curves of TiS₂ nanotubes (solid) and polycrystalline TiS₂ (dots) at various current densities and 20 °C. Reproduced from ref. 159 with permission from The Royal Society of Chemistry.

Yoo et al.,¹⁵⁴ reported the electrochemical intercalation of organic Py₁₄⁺ cations into TiS₂ interlayers, which remarkably improved the Mg²⁺-intercalation kinetics and promoted Mg–Cl⁺ intercalation. In this case, Py₁₄Cl ionic liquid was added to the APC electrolyte. Py₁₄⁺ cations were intercalated into TiS₂ at a low rate (5 mA g⁻¹), causing the irreversible interlayer expansion of TiS₂ (Fig. 17a). As a result, the specific capacity of TiS₂ was remarkably increased from ∼25 mA h g⁻¹ to 239 mA h g⁻¹ at 24 mA g⁻¹ (Fig. 17b). Moreover, the specific capacity of 239 mA h g⁻¹ was boosted to 400 mA h g⁻¹ at a high temperature of 60 °C, corresponding to the intercalation of two Mg–Cl⁺ species per Ti atom. N 1s XPS spectra revealed that only Py₁₄⁺ cations were intercalated into TiS₂ at the beginning of the first discharge. Cl 2p and Mg 2s XPS signals appeared at a later stage of the discharge, which indicated the intercalation of MgCl⁺ after the expansion of the interlayer distance by Py₁₄⁺. Besides, Mg K-edge NEXAFS of Mg²⁺-intercalated TiS₂ displayed a similar onset energy with tetra-coordinated [Mg₂Cl₂·4THF]²⁺, implying that the intercalated Mg²⁺ was tetra-coordinated with one Cl atom and three S atoms. MgCl⁺ acting as charge carriers improved the intercalation kinetics due to its lower charge density than that of Mg²⁺. Additionally, MgCl⁺ avoided the Mg–Cl dissociation step, and exhibited weaker binding to S than Mg²⁺ (Fig. 17c and d).

Fig. 17 (a) A schematic illustrating the structural evolution of TiS₂ at different stages of intercalation. (b) Galvanostatic voltage profiles of the exfoliated TiS₂ electrode at 24 and 240 mA g⁻¹ at 25 °C. The number of MgCl⁺ intercalation per exfoliated TiS₂ is also shown in the top axis. (c and d) Energy diagrams for the intercalation and diffusion of Mg²⁺ and MgCl⁺. (c) Typical intercalation of Mg²⁺ involves the scission of MgCl⁺ ions into Mg²⁺ and Cl⁻, which requires substantial activation energy of 3 eV at least. Subsequent diffusion of divalent Mg²⁺ also has a high-migration energy barrier of 1.06 eV, which results in the limited level of intercalation at room temperature. (d) Intercalation of MgCl⁺ bypasses the sluggish scission of the Mg–Cl bond at the electrolyte–cathode interface. Afterwards, MgCl⁺ diffuses fast in the expanded interlayers due to the fairly low-migration energy barrier of 0.18 eV. Mg and Cl atoms are shown as purple and green spheres, respectively. Reprinted with permission from ref. 154. Copyright 2017, Springer Nature.

Besides, the Al³⁺ intercalation kinetics into TiS₂ shows a strong dependency on the temperature and the particle size of TiS₂. Geng et al.¹⁰⁶ compared the Al³⁺ intercalation into TiS₂ at room temperature and 50 °C. As revealed, the specific capacity of TiS₂ at room temperature achieved 50 mA h g⁻¹ in the initial cycle and decayed to 30 mA h g⁻¹ after 50 cycles at 5 mA g⁻¹. By contrast, the specific capacity of TiS₂ at 50 °C reached 45 mA h g⁻¹ and increased to 70 mA h g⁻¹ after 50 cycles. The authors claimed that cycling at high temperatures altered the crystal structure of TiS₂, thus facilitating Al³⁺ intercalation. Hawkins et al.¹⁶¹ showed that TiS₂ nanobelts, cycled at 50 °C, displayed a superior specific capacity of 150 mA h g⁻¹ even at a high current density of 240 mA g⁻¹. It was suggested that both Al³⁺ and AlCl₄⁻ could be intercalated into TiS₂ nanobelts, which also accounted for the improved intercalation kinetics.^{30,108,162–164}

3.2.2 V-based TMDs. Vanadium disulphide (VS₂) and vanadium diselenide (VSe₂) exhibit metallic properties, tetragonal symmetry, and layer spacings of 5.74 Å and 6.10 Å, respectively.¹⁶⁵ The one-electron V⁴⁺/V³⁺ redox reaction enables VS₂ and VSe₂ with theoretical specific capacities of 233 mA h g⁻¹ and 128 mA h g⁻¹, respectively. Nevertheless, the strong orbital hybridization in VSe₂ accounts for the superior cation-intercalation kinetics of VSe₂ to that of VS₂. Therefore, pristine VSe₂ exhibited higher specific capacity than pristine VS₂ (Fig. 18).¹⁵⁷ Apart from intercalation, reversible conversion mechanism was also observed for VS₂ and VSe₂ in alkali metal-ion batteries, empowering them with high specific capacities (>700 mA h g⁻¹).^41,166

Fig. 18 Mg²⁺ intercalation into (a) pristine VS₂ and (b) pristine VSe₂ at 5 mA g⁻¹. Reprinted with permission from ref. 157. Copyright 2019 American Chemical Society.

To improve the Mg²⁺-intercalation kinetics of VS₂, several interlayer expansion strategies were developed for VS₂. For example, Xue et al.¹⁶⁷ synthesized VS₂ through a solvothermal reaction employing 2-ethylhexylamine as the solvent. Consequently, 2-ethylhexylamine molecules were incorporated into the VS₂ structure during the synthesis, obtaining interlayer-expanded VS₂ nanoflowers (Fig. 19a). The single-step preparation of the expanded VS₂ is vital for practical application, as the industry could not rely on elaborated post-synthesis steps. The combined XRD and Fourier transform infrared results verified that 2-ethylhexylamine molecules were located between VS₂ interlayers, rather than on the surface of VS₂. Impressively, the fabricated VS₂ exhibited a large interlayer spacing of 9.93 Å, which contrasted with the small interlayer distance of the annealed VS₂ (5.73 Å). The expanded VS₂ showed superior electrochemical performance with a high specific capacity of 245 mA h g⁻¹ at 100 mA g⁻¹ and 77% capacity retention after 100 cycles. Moreover, an excellent rate capability of the expanded VS₂ was evidenced by the high specific capacities of 140 and 102 mA h g⁻¹ at current densities of 1 and 2 A g⁻¹, respectively (Fig. 19b and c). XPS spectra of the expanded VS₂ electrode at different discharge stages presented both Mg 2s and Cl 2p signals with an Mg : Cl ratio >1, implying both Mg²⁺ and MgCl⁺ intercalation. The diffusion coefficient of Mg²⁺/MgCl⁺ in the expanded VS₂ was calculated to be in the range of 7.58 × 10⁻¹¹ to 6.03 × 10⁻¹³ cm² s⁻¹, which was substantially higher than the Mg²⁺-diffusion coefficient of pristine VS₂ (4.20 × 10⁻²³ cm² s⁻¹). The ex situ XRD and HRTEM measurements showed that a conversion reaction occurred at a low voltage range (<0.4 V vs. Mg²⁺/Mg), forming MgS and V (Fig. 19d and e). MgS and V were not observed when the expanded VS₂ was charged back to 2.2 V, indicating the good reversibility of the conversion reaction.

Fig. 19 (a) Schematic illustration of the one-step synthesis and in situ intercalation process of the expanded VS₂ nanoflowers and annealed VS₂ nanoflowers. (b) Galvanostatic discharge/charge profiles at various current densities. (c) Rate capabilities of the expanded VS₂ and annealed VS₂ electrodes at the current densities from 100 to 2000 mA g⁻¹. (d) HRTEM images of the expanded VS₂ nanoflowers at different discharge/charge states (from left to right): discharged to 0.6 V, discharged to 0.4 V, fully discharged to 0.2 V, and fully charged to 2.2 V. (e) Schematic of the reversible storage mechanism of active Mg species in the expanded VS₂ nanoflowers. Reprinted with permission from ref. 167. Copyright 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Another method to increase the interlayer distance of VS₂ relies on the irreversible electrochemical intercalation of ionic liquid cations.^80,154 Recently, Zhao et al.⁸⁰ used PP₁₄Cl as an electrolyte additive to enable PP₁₄⁺ intercalation between VS₂ layers. During the electrochemical activation at 20 mA g⁻¹, PP₁₄⁺ was intercalated between VS₂ nanosheets. Ex situ XPS analysis of the VS₂ electrode showed no change in the N 1s signal after the first discharge, indicating the irreversibility of PP₁₄⁺ intercalation. After the activation, VS₂ nanosheets exhibited high specific capacities of 299 and 214 mA h g⁻¹ at 50 and 2000 mA g⁻¹, respectively. Moreover, a high energy density of 152 W h kg⁻¹ at a power density of 1600 W kg⁻¹ was achieved by VS₂ nanosheets. It should be pointed out that a significant amount of capacity was detected at the low voltage range (e.g., 150 mA h g⁻¹ between 0.5–0.01 V at 20 mA g⁻¹). The contribution of the capacity at the low voltage range was considerably low, and the low voltage was shown to affect the cyclability of TMDs in alkali metal batteries due to the occurring conversion reaction.^168,169

Compared to Mg²⁺ intercalation, Li⁺ intercalation generally shows higher kinetics owing to the low charge density.^157,165 In this regard, Sun et al.¹⁷⁰ reported hybrid devices combining Mg stripping/plating anodes and Li⁺-intercalation VS₂ cathodes. Such Mg–Li hybrid device exploited both the stable Mg stripping/plating on the anode, and the efficient Li⁺-intercalation into VS₂ cathode. Impressively, the specific capacity of the VS₂ cathode was increased by 10 folds (from <25 to 250 mA h g⁻¹ at 0.5C) by adding LiCl to the APC electrolyte. EDS mapping of VS₂ nanosheets after discharge showed no Mg signal, confirming the negligible intercalation of Mg²⁺ into VS₂. The galvanostatic intermittent titration technique (GITT) verified that the diffusion coefficient of Li⁺ in VS₂ was around 10⁻¹³ cm s⁻¹ even at a high Li content (e.g., Li₂VS₂). By contrast, the diffusion coefficient of Mg²⁺ in VS₂ dropped from 10⁻¹⁴ to 10⁻¹⁵ cm s⁻¹ along with the slight change of Mg²⁺ content from Mg_0.025VS₂ to Mg_0.04VS₂. Thereby, the specific capacity improvement was thus ascribed to the higher diffusivity of Li⁺ in VS₂ than that of Mg²⁺. This study opens up an interesting direction for Mg–alkali metal hybrid batteries, which can be extended to constructing other energy storage systems, such as Mg–Na, Mg–K hybrid batteries. Still, alkali ions were consumed during the discharge process of VS₂. Such a fact requires the use of a large volume of electrolyte to assemble the device, consequently reducing the overall device performance.

V-based TMDs were also used as cathode materials for Al³⁺ intercalation. Wu et al.¹⁷¹ investigated the use of VS₂ and graphene-composited VS₂ (G-VS₂) as cathodes for AMBs. VS₂ was prepared using a hydrothermal method, and G-VS₂ was prepared by sonicating graphene together with VS₂. The G-VS₂ and VS₂ electrodes achieved specific capacities of 186 and 145 mA h g⁻¹ in the initial cycle at 100 mA g⁻¹, and maintained 50 and 25 mA h g⁻¹ after 50 cycles, respectively. The reversible Al³⁺ intercalation was identified by the in situ XRD analysis of G-VS₂, in which the intensity of VS₂ peaks decreased and increased reversibly during discharge and charge. The improved performance of the G-VS₂ cathode was associated with the reduced charge transfer resistance of G-VS₂ in comparison with VS₂. Furthermore, Lei et al.¹⁷² investigated single-crystal VSe₂ as the cathodes for AMBs. The XPS analysis indicated that AlCl₄⁻ served as the dominant charge carrier for VSe₂. The VSe₂ cathode exhibited the initial capacity of 650 mA h g⁻¹ at 100 mA g⁻¹, which decayed to 50 mA h g⁻¹ after 100 cycles. The rapid decay in the specific capacity of VS₂ and VSe₂ needs to be further clarified in the future.

3.2.3 Mo-based TMDs. Both molybdenum disulphide (MoS₂) and molybdenum diselenide (MoSe₂) show the hexagonal symmetry with two possible stacking configurations, namely, the 2H and 3R phases. The interlayer distances of the 2H phases are 6.15 and 6.7 Å for MoS₂ and MoSe₂, respectively. The 3R MoS₂ and MoSe₂ phases exhibit larger interlayer distances of 6.12 and 6.7 Å, respectively. In addition, the most common monolayer polytypes are semiconducting trigonal prismatic (2H) and metallic octahedral (1T). The 2H phase is thermodynamically favourable, and phase transition from 2H to 1T can occur during the electrochemical/chemical processes, such as intercalation and exfoliation.^173,174 The theoretical capacities associated with the Mo³⁺/Mo⁴⁺ transition are 168 and 106 mA h g⁻¹ for MoS₂ and MoSe₂, respectively.¹⁷⁵ Nevertheless, below ∼1.1 V vs. Li⁺/Li, both MoS₂ and MoSe₂ can undergo reversible conversion reactions, allowing high specific capacities of >700 mA h g⁻¹.

Due to the large interlayer spacing, MoS₂ can enable efficient Mg²⁺ intercalation. Liang et al.¹⁷⁶ assembled devices composed of Mg nanoparticles as the anodes and highly exfoliated graphene-like MoS₂ (G-MoS₂) as the cathodes. G-MoS₂ exhibited an enlarged interlayer spacing of 6.5–7 Å in comparison with bulk MoS₂ (B-MoS₂, 6.3 Å). The reduced particle size, together with the enlarged lattice spacing, increased the Mg²⁺-intercalation kinetics of G-MoS₂. As a result, G-MoS₂ exhibited a superior specific capacity of 170 mA h g⁻¹ at 20 mA g⁻¹, which was significantly higher than the specific capacity of B-MoS₂ (71 mA h g⁻¹, Fig. 20a) and comparable with the calculated theoretical specific capacity of 223.2 mA h g⁻¹ (Fig. 20b). Interestingly, the excellent specific capacity highly depended on the anode configuration as well. In specific, with the bulk Mg anode, the specific capacity of the G-MoS₂ cathode was two-time lower (90 mA h g⁻¹) than the specific capacity of G-MoS₂ with the Mg nanoparticle anode (170 mA h g⁻¹). The authors proposed that the formation of a thin passivation layer on the Mg nanoparticle anode promoted the Mg²⁺ diffusion across the particle surface. Yet, no experimental evidence was provided to identify the formation of the passivation film and the amount of electrolyte that is consumed during the film formation. In addition, the formation of passivation layers is usually avoided in the chlorine-containing electrolyte.

Fig. 20 (a) Typical cycling behaviour of the cells fabricated with B-MoS₂ or G-MoS₂ cathodes and bulk- or nanoparticles-Mg anodes with a discharge rate of 20 mA g⁻¹. In the inset table, the relative capacity at a certain cycle refers to the ratio of the discharge capacity of the cells at the corresponding cycle to that at the first cycle. (b) Graphical illustrations of theoretically modelled Mg adsorption on MoS₂ single-layered nanoribbon. Reprinted with permission from ref. 176. Copyright 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

To improve the cyclability of the MoS₂ cathodes for MMBs, polyvinylpyrrolidone (PVP) was incorporated between MoS₂ layers during the hydrothermal synthesis of MoS₂.¹⁷⁷ The interlayer spacings of the expanded MoS₂ with and without PVP were determined to be 9.7 Å and 9.4 Å, respectively. Additionally, PVP-incorporated MoS₂ exhibited a new XRD peak at 18.4°, implying the new lamellar phase constructed from PVP and MoS₂ monolayers, with an interlayer spacing of 4.8 Å. An initial specific capacity of 143.3 mA h g⁻¹ at the first discharge and 92% capacity retention after 100 cycles were reached by PVP-incorporated MoS₂. In contrast, expanded MoS₂ without PVP delivered a specific capacity of 131.9 mA h g⁻¹, but retained only 52% of the initial capacity after 100 cycles.

In addition to the interlayer expansion, altering the Mg-cation intercalating species was used to improve the electrochemical performance of MoS₂. Li et al.¹⁷⁴ showed that [Mg(DME)₃]²⁺ intercalation into porous 2H-MoS₂@C exhibited superior kinetics to the Mg²⁺-intercalation kinetics. The mixture of Mg(BH₄)₂ and hexafluoroisopropanol in DME was employed as the electrolyte (denoted MgBOR/DME), in which [Mg(DME)₃]²⁺ cations acted as the charge carrier ion. [Mg(DME)₃]²⁺ intercalation was confirmed using the STEM-EDS analysis, which uncovered similar distribution of O and Mg in the outer layer of 2H-MoS₂@C. Moreover, the XPS peak analysis revealed that 2H-MoS₂ underwent a phase transition to 1T-MoS₂ during the initial activation process (20 cycles at 20 mA g⁻¹). HRTEM image of the 2H-MoS₂@C cathode after 30th discharge cycles showed an amorphous outer layer, which indicated the fragmentation and structural distortion of MoS₂ during extended [Mg(DME)₃]²⁺ intercalation/de-intercalation. Fig. 21 illustrates the charge storage process of 2H-MoS₂@C, including the intercalation of large [Mg(DME)₃]²⁺ ions, 2H–1T phase transition, and the fragmentation and structural distortion. The effect of [Mg(DME)₃]²⁺ on the electrochemical performance of 2H-MoS₂@C was assessed in different Mg electrolytes, such as MgBOR/DME, APC, and Mg(HMDS)₂/MgCl₂/AlCl₃. The MgBOR/DME electrolyte enabled 2H-MoS₂@C with the highest specific capacity of 95 mA h g⁻¹ at 50 mA g⁻¹, in comparison with the APC (85 mA h g⁻¹) and Mg(HMDS)₂/MgCl₂/AlCl₃ (45 mA h g⁻¹) electrolytes. This work highlighted a new electrochemical method for boosting the efficient Mg²⁺ intercalation in MoS₂.

Fig. 21 Schematic illustration of the [Mg(DME)₃]²⁺ storage mechanism in MoS₂ structures with the MgBOR/DME electrolyte. Reprinted with permission from ref. 174. Copyright 2018, Springer Nature.

Several studies explored MoS₂ as cathodes for AMBs, where Al³⁺ served as the intercalating species with the AlCl₃/EMIMCl electrolyte.^111,178,179 For example, Li et al.¹¹¹ reported that MoS₂ spheres delivered a high specific capacity of 254 mA h g⁻¹ at 20 mA g⁻¹ in the initial cycle, but exhibited poor cyclicality with 67 mA h g⁻¹ retained after 100 cycles at 40 mA g⁻¹. XRD spectra revealed that the intercalation of Al³⁺ into MoS₂ spheres resulted in apparent interlayer expansion from 6.2 Å to 7.3 Å and obvious lattice stripe distortion (Fig. 22a and b). The XPS analysis showed that partial Al³⁺ intercalation into the MoS₂ spheres was irreversible, which also accounted for the inferior cycling performance.

Fig. 22 TEM images of MoS₂ microspheres (a) before and (b) after cycling. Reprinted with permission from ref. 111. Copyright 2018 American Chemical Society. (c) The cycling performance and corresponding coulombic efficiency of MoSe₂@C at 1 A g⁻¹. Reprinted with permission from ref. 180. Copyright 2019 American Chemical Society.

To further improve the electrochemical performance, MoS₂ and MoSe₂ were hybridized with carbon materials.^178,180 Yang et al.¹⁷⁸ fabricated MoS₂ nanostructure-incorporated free-standing carbon fibres through an electrospinning method. The highly conductive carbon fibres provided efficient encapsulation for MoS₂ nanostructures. The specific capacity of the carbon fiber-MoS₂ electrode achieved 293 mA h g⁻¹ in the initial cycle, and maintained 125 mA h g⁻¹ after 200 cycles at 100 mA g⁻¹. Moreover, Zhao et al.¹⁸⁰ prepared MoSe₂@C using a multistep synthetic route. First, dopamine hydrochloride as the carbon precursor was mixed with ammonium molybdate to create a homogenous mixture. Afterwards, the mixture was calcinated and selenizated to obtain MoSe₂@C. The MoSe₂@C cathode for AMBs showed a high specific capacity of 267 mA h g⁻¹ at 100 mA g⁻¹ with no noticeable capacity fading for up to 5000 cycles at 1 A g⁻¹ (Fig. 22c). Interestingly, the specific capacities of both MoS₂ and MoSe₂ were higher than the theoretical value based on the intercalation mechanism. This phenomenon suggested that the conversion mechanism was involved in the charge-storage process.

3.3 Graphite-based materials

Graphite-based cathode materials are appealing due to the abundant resource, low cost, and environmental friendliness. The unique redox amphoteric feature allows graphite to host both cations (e.g., Li⁺, K⁺, and Py₁₄⁺) and anions (e.g., Br⁻, Cl⁻, BF₄⁻, FSI⁻, PF₆⁻, TFSI⁻, TOf⁻, and AlCl₄⁻). In particular, anion intercalation into graphite occurs at a high potential (∼1.75 V vs. SHE), which is beneficial for the construction of high-voltage energy storage devices.^44,45,181 Anion intercalation of graphite-based materials is based on a staging mechanism. The stage number represents the number of graphite layers between intercalating anions (Fig. 23).¹⁸² Different intercalation stages can be detected by distinctive voltage plateau during charge/discharge and the spectroscopy analysis (e.g., XRD and Raman). For example, the intercalation stage can be extracted from the peak position ratio of the two most dominant peaks, (00n + 1) and (00n + 2). Anion-intercalation graphite-based materials are widely employed to construct dual-ion MVMBs. During charging, metal cations are deposited on the anodes, and anions are intercalated into the graphite-based cathodes. Owing to the consumption of electrolyte ions, dual-ion batteries with graphite cathodes generally require the use of large-amount or high-concertation electrolytes.⁵⁰

Fig. 23 Schematic illustration of the staging mechanism of intercalant guest species into graphite: I_c = periodic repeat distance; d_i = intercalant gallery height; Δd = gallery expansion. Reprinted with permission from ref. 182. Copyright 2013 Elsevier.

In contrast to ZMBs constructed with commonly used cathodes (e.g., V₂O₅ and MnO₂), ZMBs based on the graphite cathodes (denoted graphite-ZMBs) can depict high average voltages of above 2 V. However, most organic electrolytes of ZMBs (e.g., Zn(TfO)₂ and Zn(TFSI)₂ in ionic liquid) have low anodic stable potential windows (<2.6 V vs. Zn²⁺/Zn), which cannot fulfil the high potential requirement of anion intercalation into graphite.^23,24,68,183 For example, graphite-ZMBs charged to a cut-off voltage of 2.6 V vs. Zn²⁺/Zn only demonstrated a specific capacity of 50 mA h g⁻¹.⁶⁸ Recently, Wang et al.²⁴ showed that adding LiPF₆ into the Zn electrolyte composed of Zn(TFSI)₂ in AN can greatly enhance the anodic stability. It was shown that 1 M Zn(TFSI)₂ in AN started to decompose at ∼2.3 V vs. Zn²⁺/Zn. The presence of LiPF₆ efficiently suppressed the anodic dissolution and the decomposition of Zn(TFSI)₂ electrolyte. The mixed electrolyte (0.5 M Zn(TFSI)₂ + 2 M LiPF₆) depicted a high stable potential of more than 4 V vs. Zn²⁺/Zn (Fig. 24a). The floating test at 3 V exhibited negligible leakage currents (<10⁻³ mA cm⁻²), which contrasted with the high leakage currents of the 1 M Zn(TFSI)₂ electrolyte (>30 mA cm⁻², Fig. 24b). ¹⁹F NMR identified the co-intercalation of PF₆⁻ and TFSI⁻ into graphite during charging. Density functional theory (DFT) calculations showed that PF₆⁻ had a lower diffusion energy barrier than TFSI⁻, thus suggesting the higher diffusion rate of PF₆⁻ (Fig. 24c and d). Graphite-ZMB with the LiPF₆ additive could be charged to a high voltage of 2.8 V vs. Zn²⁺/Zn, depicting large specific capacities of 105 and 97 mA h g⁻¹ at 100 and 2000 mA g⁻¹, respectively (Fig. 24e). In addition, the superior cyclability of graphite-ZMB was demonstrated with nearly 100% retention after 2000 cycles at 1000 mA g⁻¹. Recently, Wang et al.²³ demonstrated that Zn metal would react with PF₆⁻, forming dissolved Zn²⁺ in the electrolyte. The reaction between PF₆⁻ and Zn also led to the formation of a solid electrolyte interface on Zn, which was composed of ZnF₂ and LiF. With an electrolyte of 2.5 M LiPF₆ in EMC, graphite-ZMB exhibited a high specific capacity of 95 mA h g⁻¹ at 200 mA g⁻¹. It is highly desirable to find alternative metal-PF₆ salts and avoid the use of Li salts.

Fig. 24 (a) Linear sweep voltammetry curves at 2 mV s⁻¹ of different electrolytes on stainless steel electrode within Zn//stainless steel cells. (b) Floating test of electrolytes at 3 V. Sensitivity is 0.01 A V⁻¹. (c) The optimized anion diffusion path in the graphite layers. (d) The optimized anion diffusion energy barriers in the graphite layers. (e) Rate performance of Zn//graphite dual-ion battery. Reprinted with permission from ref. 24. Copyright 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

AMBs are the most studied battery system with anion-intercalation graphite cathodes. Although the Al stripping and plating electrochemistry in the ionic liquid mixture containing Al₂Cl₇⁻ anions has been well known for several decades,⁹⁹ Dai et al.²¹ was the first to show AMBs constructed with the AlCl₄⁻ intercalation graphite cathodes in 2015. The optimal AMBs were assembled in the electrolyte with an AlCl₄⁻/Al₂Cl₇⁻ ratio of ∼2.33, which was prepared by mixing AlCl₃ and EMIMCl in a molar ratio of 1.3. The AlCl₄⁻-intercalation chemistry allowed pyrolytic graphite with a specific capacity of 60 mA h g⁻¹ at 66 mA g⁻¹ and an average cell voltage of 2 V. However, the large size of AlCl₄⁻ anions led to the unsatisfactory rate capability (only 20 mA h g⁻¹ retained at 264 mA g⁻¹) of the pyrolytic graphite cathode. To address these issues, the graphitic foam was used to replace pyrolytic graphite, delivering a specific capacity of 60 mA h g⁻¹ at a high current density of 5 A g⁻¹. More impressively, the pouch-cell batteries assembled with the graphitic foam cathodes achieved high capacity retention of 100% after 7500 cycles with high coulombic efficiencies (>97%). Ex situ XRD measurements provided significant insights into the AlCl₄⁻-intercalation mechanism of graphite. In detail, the (002) peak of graphite vanished in the fully charged cathode, while two new peaks, assigned to lattice spacings of 3.15 and 3.77 Å, appeared. In the fully discharged cathode, the graphite (002) peak reappeared with a broad shoulder, which was indicative of an irreversible change in the stacking of graphite layers. AlCl₄⁻ intercalation into graphite was determined to be a stage-4 process. Besides, the same group also found that free-standing natural graphite film employed as the AMB cathode depicted superior cyclability (∼100% retention after 6000 cycles), good coulombic efficiencies (∼98%), and importantly considerably enhanced specific capacity (110 mA h g⁻¹ at 99 mA g⁻¹).¹¹³

To further improve the rate performance, AlCl₄⁻-intercalated graphite was subjected to rapid thermal expansion at 1000 °C and subsequently transferred to water for electrolysis.¹¹² The electrolysis process produced massive amounts of hydrogen gas, which further introduced large porosity into the expanded graphite. The obtained porous graphite was comprised of microparticles (∼1 μm) with 4–5 graphene layers, presenting superior rate performance with a specific capacity of 60 mA h g⁻¹ at 12 A g⁻¹ (18 seconds charge). Moreover, Zhang et al.¹⁸⁴ studied the AlCl₄⁻-intercalation behaviours of four different graphitic materials, namely, large-size graphite (L-graphite) and graphene (L-graphene), small-size graphite (S-graphite) and graphene (S-graphene). At a low current density of 60 mA g⁻¹, the specific capacities of L-graphite and L-graphene were both ∼85 mA h g⁻¹ with two apparent voltage plateaus. In contrast, S-graphite and S-graphene showed a relatively low specific capacity of ∼72 mA h g⁻¹ without distinctive voltage plateaus. In addition, L-graphene showed the best rate capability among four samples, retaining 90% of the initial capacity at a large current density of 4.8 A g⁻¹ (Fig. 25a). The excellent rate capability originated from both the high conductivity and structural flexibility of L-graphene, which could well endure the structural stress during the repeat AlCl₄⁻ intercalation/deintercalation (Fig. 25b). Interestingly, the good crystallinity of graphitic materials contributed to the high specific capacity of L-graphite. The carboxyl and hydroxyl groups on the edge of S-graphene caused the repulsive interaction with the intercalating AlCl₄⁻, thus imposing a specific activation energy for AlCl₄⁻-intercalation.^184,185 In addition, slight performance degradation was detected for L-graphene during the continuous charge/discharge cycles, which significantly contrasted with the notable performance degradation of S-graphene caused by the apparent structural change.¹⁸⁴

Fig. 25 (a) Rate capability from 60 to 4800 mA h g⁻¹ of L-graphene (black), S-graphene (red), L-graphite (blue), and S-graphite (orange). (b) Electrochemical impedance spectra (circle represents high voltage and square represents low voltage). Reprinted with permission from ref. 184. Copyright 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

High-performance cathodes relying on simple and robust preparation methods are highly pursued for practical applications. In this regard, Hu et al.¹⁸⁶ reported a simple acid treatment strategy, which greatly improved the rate performance of commercial carbon nanofibers (CNFs). The acid treatment cleaved the outer graphitic layer of CNFs, creating edge-rich graphitic nanoribbons interconnected by the nanofiber core (Fig. 26a). In the original CNFs, the intercalation of AlCl₄⁻ was blocked due to the dense outer wrapping graphite layer. By contrast, the inner graphitic carbon was sufficiently exposed after the cleavage of the outer layer, enabling the superior AlCl₄⁻-intercalation kinetics (Fig. 26b). In addition, the cleaved nanoribbons increased the charge-transport capability of the electrode by serving as conductive ties to link the stacked nanofibers. Raman spectra confirmed that the graphitic structure of the cleaved nanoribbons was not destroyed by the acid treatment. Additionally, the acid treatment increased the specific surface area from 23.2 m² g⁻¹ for the pristine CNFs to 55.3 m² g⁻¹ for the cleaved CNFs. DFT calculations suggested that the edge-rich graphene structure (2.45 eV) had a stronger binding with AlCl₄⁻ than the edge-less graphene structure (2.21 eV). Furthermore, the AlCl₄⁻ binding energies in the bilayer structures are −3.14 eV, −1.66 eV, and −1.17 eV for the edge-less graphene, the centre of edge-rich graphene, and the edge of edge-rich graphene, respectively (Fig. 26c–h). These calculation results supported that the intercalation of AlCl₄⁻ was preferred with edge-rich graphene. This conclusion was also consistent with a recent study,¹⁸⁷ disclosing that the voids in few-layers graphene promoted the stage-2 and stage-1 AlCl₄⁻ intercalation. After the acid treatment, the cleaved CNFs presented a high specific capacity of 126 mA h g⁻¹ at 1 A g⁻¹, which significantly contrasted with the pristine CNFs (15 mA h g⁻¹). Besides, excellent rate performance and cyclability were demonstrated for the cleaved CNFs with high specific capacities of 95 mA h g⁻¹ at 50 A g⁻¹ and 105 mA h g⁻¹ (∼100% retention) after 5000 cycles at 10 A g⁻¹ (Fig. 26i and j).

Fig. 26 (a) Schematic comparison of anion intercalation/deintercalation in commercial CNFs and cleaved CNFs. (b) Representative charge/discharge curves of CNFs and cleaved CNFs. (c–h) Calculated charge density differences of the AlCl₄⁻ in different graphene structures: the (c) top and (d) side views of the bilayer edge-less graphene with one AlCl₄⁻ ion placed between the two layers; the (e) top and (f) side views of the bilayer edge-rich graphene with one AlCl₄⁻ ion placed in the middle area; and the (j) top and (h) side views of bilayer edge-rich graphene with one AlCl₄⁻ ion placed near the edge, between two layers. An isosurface level of 0.0007 eÅ⁻³ has been used in all images. (i) Charge/discharge curves of cleaved-CNFs at current densities from 2 to 50 A g⁻¹. (j) Long-term stability of cleaved-CNFs with coulombic efficiency and discharge specific capacity versus the cycle number at a current density of 10 A g⁻¹. The inset in d shows a flexible cleaved-CNFs electrode and a flexible AIB lighting a light emitting diode. Reproduced from ref. 186 with permission from The Royal Society of Chemistry.

As discussed earlier, low-cost amides or amines can be used to replace the ionic liquid in the AMB electrolytes. For example, mixing AlCl₃ with urea,¹⁰⁵ triethylamine hydrochloride (Et₃NCl),¹⁸⁸ or acetamide¹⁰⁴ can form AlCl₄⁻, AlCl₂⁺, and Al₂Cl₇⁻ in electrolytes. The charge/discharge electrode reaction in the AlCl₃/urea electrolyte is the same as in the AlCl₃/EMIMCl electrolyte, involving the Al₂Cl₇⁻ reaction (eqn (3)) at the anode side and AlCl₄⁻ intercalation at the cathode side.¹⁰⁴ However, the high viscosity of AlCl₃/urea led the graphite cathode in the AlCl₃/urea electrolyte to exhibit poor performance.¹⁰⁵ The specific capacity decreased from 73 mA h g⁻¹ at 100 mA g⁻¹ to 50 mA h g⁻¹ at 200 mA g⁻¹. In addition, the average discharge voltage of graphite in the AlCl₃/urea electrolyte was 1.73 V, which was 0.27 V lower than the graphite cathode in the AlCl₃/EMIMCl electrolyte. To overcome the viscosity issue, Xu et al.¹⁸⁸ showed a device combining the Et₃NAlCl₄ electrolyte and the graphene aerogel cathode. The cathode was fabricated through freeze-drying and subsequent annealing of graphene oxide. In spite of the high viscosity of the electrolyte, the favourable porosity of the graphene aerogel cathode enabled an excellent rate performance. A specific capacity of nearly 110 mA h g⁻¹ was reached at a current density of 5 A g⁻¹. Besides, the graphene aerogel was charged to a high voltage due to the high cut-off voltage of the Et₃NAlCl₄ electrolyte (2.62 V). Charging the graphene aerogel to 2.51 V resulted in a high energy density of ∼260 W h kg⁻¹ at a power density of 3 kW kg⁻¹. Nevertheless, Al-based dual-ion batteries require further electrolyte development to improve the anodic stability, as no commercial current collector can be used for the cathodes.

Anion intercalation provides a good strategy to avoid the drawback associated with the intercalation of multivalent Zn²⁺, Mg²⁺, and Al³⁺. In addition, the anion-intercalation chemistry has been extensively studied in Li⁺-based systems, which accumulate insightful experience for constructing MVMB devices. In this direction, the development of suitable electrolytes with wide stable potential windows and efficient metal stripping and plating will be highly desired for MVMBs.

3.4 Two-dimensional covalent organic frameworks

Two-dimensional covalent organic frameworks are a class of crystalline and porous 2D polymers, which are constructed with dynamic covalent bonds in the layer and stacked by noncovalent aromatic π-interactions. Interestingly, 2D COFs are equipped with one-dimensional channels along the c-crystallographic direction, allowing efficient mass transport through the material. Recently, 2D COFs have attracted intensive research attention as a group of multifunctional materials.^189–192 Importantly, the regular porosity, large specific surface area, and the tailorable chemistries/topologies empower 2D COFs with versatile opportunities for energy storage by periodically organizing redox-active sites into porous frameworks.¹⁹³ According to the redox potentials, many organic groups are potentially suitable for constructing redox-active 2D COFs, such as quinone,^194–200 phenazine,^201,202 triphenylamine,^203,204 cyano,^205,206 bipyridine,^114,207 pyridinic nitrogen,²⁰⁸ and phenanthrenequinone.^20,110,209 At present, the energy-storage investigation on 2D COFs is still at the primary stage, and basic electrochemistry understanding is accumulating with dominant efforts devoted to the exploration of 2D COFs for LIBs. It should be noted that the electrochemical behaviours of organic groups are expected to be quite different for the applications of LIBs and MVMBs due to the change of charge carrier ions. For example, hexaazatrinaphthalene was demonstrated to exhibit a high specific capacity of ∼400 mA h g⁻¹ for Li⁺ storage based on the accommodation of two Li⁺ ions in each bipyridine site. In the case of Mg²⁺ storage, two bipyridine sites would accommodate only one Mg²⁺, displaying inferior capacity and rate performance.¹¹⁴

Recently, some pioneering studies attempted to demonstrate the application of 2D COFs in MVMBs. Distinct from inorganic cathodes for MVMBs, 2D COFs provide flexible Mg²⁺ diffusion pathways and superior electrochemical reaction kinetics. Recently, Sun et al.²⁰⁸ synthesized a triazine-based 2D COF for Mg²⁺ storage (Fig. 27a). The 2D COF was synthesized through the polymerization of 1,4-dicyanobenzene by annealing with ZnCl₂ at 400 °C. An impressive rate performance was exemplified with a specific capacity of 110 mA h g⁻¹ and more than 50% capacity retention when the current density was increased from 50 to 1300 mA g⁻¹ (Fig. 27b). In addition, the 2D COF presented impressive cyclability, retaining a specific capacity of ∼40 mA h g⁻¹ after 5000 cycles at 1300 mA g⁻¹ (Fig. 27c). Based on the proposed structure in Fig. 27a, the high density of active triazine sites enabled the 2D COF with a high theoretical specific capacity of 419 mA h g⁻¹, considerably higher than the measured value (110 mA h g⁻¹). XPS measurements detected changes in the signal of pyridinic-N in triazine during charge/discharge, indicating that charge storage occurred through interaction between Mg²⁺ and –C=N– in triazine rings. However, pyrrolic-N and graphitic-N could not contribute to the charge storage, which explained the low specific capacity.

Fig. 27 (a) Chemical structure and possible electrochemical redox mechanism of the triazine-based COF. (b) Discharge/charge curves of electrodes at different rates. (c) Long-term cycling performance of the 2D COF at 5C. Reprinted with permission from ref. 208. Copyright 2019 American Chemical Society.

Apart from cation-storage 2D COFs, anion-storage 2D COFs were also demonstrated for MVMBs, which delivered high voltage and large specific capacity. Lu et al.²¹⁰ synthesized a bipyridine-containing COF (denoted TpBpy-COF) through the reaction between 5,5′-diamino-2,2′-bipyridine and triformylphloroglucinol (Fig. 28a). Importantly, TpBpy-COF exhibited a high crystallinity with a large specific surface area of 1794 m² g⁻¹ and regular porosity with pore sizes of ∼2.1 nm. The TpBpy-COF cathode for AMBs exhibited an exceptionally high specific capacity of 307 mA h g⁻¹ at 125 mA g⁻¹ (Fig. 28b), close to its theoretical value of 369.7 mA h g⁻¹. Additionally, a high specific capacity of 150 mA h g⁻¹ was retained after 13 000 cycles at a high current density of 2 A g⁻¹. Ex situ XRD reflected that TpBpy-COF well retained its crystalline structure during repeat charge/discharge cycles, accounting for the excellent cyclability. Furthermore, Fourier-transform infrared measurements suggested that AlCl₄⁻ storage altered the environment of nitrogen atoms (ν_C=N and ν_C–N signals) in TpBpy-COF upon charging. N 1s peak related to the secondary N (–NH–) shifted from 399.2 eV to higher binding energy of 400.5 eV upon the charge, and shifted back to 399.3 eV after the discharge. Meanwhile, the ratio between secondary N and pyridinic N changed from 1 : 1 to 8 : 1 upon the charge and recovered to 2 : 1 after the discharge (Fig. 28c and d). These XPS results indicated that the C–N and C=N groups were involved in the p-type oxidation with AlCl₄⁻.²¹⁰

Fig. 28 (a) Synthetic reaction and structure illustration of 2D COF fabricated from 5,5′-diamino-2,2′-bipyridine and triformylphloroglucinol. (b) Galvanostatic discharge/charge curves under different current densities in the potential range 0.01–2.3 V. N 1s XPS spectra of electrodes (c) charged to 2.3 V and (d) discharged to 0.1 V at 2 A g⁻¹. Reprinted with permission from ref. 210. Copyright 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

4. Summary and outlook

To sum up, MVMBs, using highly available multivalent metal as anodes, have emerged as promising alternatives for LIBs, particularly in large-scale and stationary energy storage applications. Currently, MVMBs are still at the preliminary research stage, and the whole community focuses on exploiting suitable “anode–electrolyte–cathode” systems for potential practical applications. In this review, we briefly introduced the cell configuration and the so-far developed “anode–electrolyte” chemistries for non-aqueous MVMBs, including ZMBs, MMBs, and AMBs. Particularly, we put the emphasis on discussing the recent progress in the development of layered cathode materials (including layered TMOs, TMDs, graphite, and 2D COFs) for non-aqueous MVMBs (also summarized in Tables 2–4). It should be noted that Ca-metal batteries are also promising alternative MVMBs. However, the development of Ca metal-compatible electrolytes is still the main challenge of Ca-metal batteries, resulting in rare studies on the Ca²⁺ intercalation of layered materials.²¹¹ Therefore, Ca-metal batteries are not discussed in this review. Among layered materials, layered TMOs (e.g., MnO₂ and V₂O₅) demonstrate auspicious specific capacities and redox potentials, but display sluggish multivalent metal ion-storage kinetics (e.g., Zn²⁺ and Al³⁺) and compatibility issues with MVMB electrolytes (e.g., MMB electrolytes). In comparison with TMOs, TMDs show good electrolyte compatibility with efficient cation intercalation and superior rate capabilities. Nevertheless, inferior discharge voltages lead to the low energy densities of MVMBs constructed with TMD cathodes. Graphite-based materials with the anion-intercalation chemistry exhibit high redox potentials (e.g., ∼1.75 V vs. SHE), avoiding the drawbacks related to multivalent metal ion intercalation. However, graphite-based materials are limited by the low theoretical capacity, as well as the insufficient electrolyte studies for ZMBs and MMBs. In addition, 2D COFs provide a unique material platform for exceptional active site engineering at the molecular level, providing many opportunities for both cation and anion storage. Fig. 29 compares the average discharge voltages and specific capacities of recently developed layered cathodes for ZMBs, MMBs, and AMBs. It provides an unambiguous picture about the advantages and limitations of different layered materials for different MVMBs. In light of the unique vdW interaction between layers, the structure of layered materials can be easily engineered via versatile strategies to tailor their intrinsic properties and ion-storage kinetics. In this sense, interlayer expansion strategies represent the most employed one to facilitate the efficient ion diffusion and promote the intercalation of new charge carrier species (e.g., MgCl⁺ in MMBs). Besides, guest species incorporation strategies can enhance the intrinsic conductivity and structural stability of layered materials, while imposing a shielding effect on the interaction of charge carrier ions and cathodes. Given the remarkable progress achieved in this research direction, there are many challenges to be addressed in the near future, which are highlighted below.

Table 2 Summary of layered cathode materials for non-aqueous ZMBs^a

Material	Electrolyte	Specific capacity	Average discharge voltage	Energy density based on the cathode	Cyclability	Ref.
a TMS – tetramethylene sulfone.
Bilayered hydrated V2O5	0.5 M Zn(TFSI)2 in AN	196 mA h g^−1 at 14.4 mA g^−1	0.9 V	176 W h kg^−1	170 mA h g^−1 after 120 cycles at 14.4 mA g^−1	65
V3O7·H2O nanofibers	0.25 M Zn(OTf)2 in AN	175 mA h g^−1 at 5 mA g^−1	0.75 V	131 W h kg^−1	175 mA h g^−1 after 50 cycles at 5 mA g^−1	61
δ-MnO2 nanoflorets	0.5 M Zn(TFSI)2 in AN with 10% water	123 mA h g^−1 at 12.3 mA g^−1	1.37 V	169 W h kg^−1	55 mA h g^−1 after 125 cycles at 12.3 mA g^−1	66
PPy-intercalated VOPO4	1 M Zn(OTf)2 in AN	67 mA h g^−1 at 30 mA g^−1	1.3 V	87 W h kg^−1	60 mA h g^−1 after 350 cycles at 100 mA g^−1	59
Natural graphite	0.2 M Zn(OTf)2 in EMImOTf	33.7 mA h g^−1 at 200 mA g^−1	2.0 V	65.1 W h kg^−1	20 mA h g^−1 after 100 cycles at 200 mA g^−1	183
Graphite powder	2 M LiPF6 + 0.5 M Zn(TFSI)2 in EMC	105 mA h g^−1 at 100 mA g^−1	2.2 V	231 W h kg^−1	96 mA h g^−1 after 2000 cycles at 1 A g^−1	24
Natural graphite	2.5 M LiPF6 in EMC/TMS (4 : 1)	98 mA h g^−1 at 100 mA g^−1	2.2 V	216 W h kg^−1	82 mA h g^−1 after 1200 cycles at 300 mA g^−1	23

---

Table 3 Summary of layered cathode materials for MMBs^a

Material	Electrolyte	Specific capacity	Average discharge voltage	Energy density based on the cathode	Cyclability	Ref.
a NFs – nanofibers, NW – nanowire, AC – active carbon, NB – nanobelts, NS – nanosheet, NR – nanorod.
V2O5 on CNFs	1 M Mg(ClO4)2 in AN	160 mA h g^−1 at 20 mA g^−1	0.6 V	96 W h kg^−1	120 mA h g^−1 after 50 cycles at 30 mA g^−1	136
α-V2O5	0.5 M Mg(TFSI)2 in PY14TFSI	295 mA h g^−1 at 15 mA g^−1 (110 °C)	0.0 V vs. AC (2.0 V vs. Mg^2+/Mg)	N/A	200 mA h g^−1 after 50 cycles at 59 mA g^−1 (110 °C)	93
V2O5-PEO nanocomposites	0.5 M Mg(ClO4)2 in AN	100 mA h g^−1 at 10 mA g^−1	∼1.4 V	∼140 W h kg^−1	90 mA h g^−1 after 35 cycles at 10 mA g^−1	141
Graphene-decorated V2O5	0.5 M Mg(TFSI)2 in AN	320 mA h g^−1 at 50 mA g^−1	−0.4 V vs. AC (2.0 V vs. Mg^2+/Mg)	N/A	95 mA h g^−1 after 200 cycles at 1 A g^−1	132
V3O7·H2O NWs	0.5 M Mg(ClO4)2 in AN	231 mA h g^−1 at 10 mA g^−1 (at 60 °C)	−0.75 V vs. AC (1.75 V vs. Mg^2+/Mg)	N/A	132 mA h g^−1 after 100 cycles at 40 mA g^−1 (60 °C)	142
NH4V4O10	0.5 M Mg(ClO4)2 in AN	174.8 mA h g^−1 at 42 mA g^−1	−0.19 V vs. AC (2.19 V vs. Mg^2+/Mg)	N/A	36.8 mA h g^−1 after 100 cycles at 210 mA g^−1	140
Mn0.04V2O5·1.17H2O NBs	0.3 M Mg(TFSI)2 in AN	140 mA h g^−1 at 50 mA g^−1	−0.2 V vs. AC (2.2 V vs. Mg^2+/Mg)	N/A	70 mA h g^−1 after 10 000 cycles at 2 A g^−1	138
Mg0.3V2O5·1.1H2O NWs	0.3 M Mg(TFSI)2 in AN	162 mA h g^−1 at 100 mA g^−1	−0.25 V vs. AC (2.25 V vs. Mg^2+/Mg)	N/A	108 mA h g^−1 after 10 000 cycles at 1 A g^−1	135
MgxV5O12·nH2O NFs	0.3 M Mg(TFSI)2 in AN	160 mA h g^−1 at 50 mA g^−1	0.0 V vs. AC (2.0 V vs. Mg^2+/Mg)	N/A	68 mA h g^−1 after 10 000 cycles at 2 A g^−1	137
Na2V6O16·1.63H2O NWs	0.5 M Mg(TFSI)2 in DME	175 mA h g^−1 at 50 mA g^−1	−0.4 V vs. AC (2.0 V vs. Mg^2+/Mg)	N/A	46 mA h g^−1 after 450 cycles at 200 mA g^−1	139
Free-standing MnO2 NWs film	0.1 M Mg(ClO4)2 in PC	120 mA h g^−1 at 246 mA g^−1	N/A	N/A	92 mA h g^−1 after 100 cycles at 246 mA g^−1	128
MnO2	0.25 M Mg(TFSI)2 in diglyme	135 mA h g^−1 at 25 mA g^−1	1.4 V	N/A	90 mA h g^−1 after 100 cycles at 125 mA g^−1	95
MoO3	0.5 M Mg(ClO4)2 in AN	220 mA h g^−1 at N/A	1.8 V	N/A	N/A	221
PA-VOPO4	APC in THF	275 mA h g^−1 at 100 mA g^−1	1 V	275 W h kg^−1	192 mA h g^−1 after 500 cycles at 100 mA g^−1	153
VOPO4·2H2O	0.1 M Mg(ClO4)2·6H2O in PC	91.7 mA h g^−1 at 5 mA g^−1	0.1 V vs. Ag/AgCl (2.7 V vs. Mg^2+/Mg)	N/A	N/A	152
TiS2 nanotubes	1 M Mg(ClO4)2 in AN	236 mA h g^−1 at 10 mA g^−1	1.2 V	N/A	184 mA h g^−1 after 80 cycles at 10 mA g^−1	159
TiSe2 nanocrystal	0.25 M Mg(AlCl2EtBu)2 in THF	110 mA h g^−1 at 5 mA g^−1	0.9 V	99 W h kg^−1	∼100 mA h g^−1 after 50 cycles at 5 mA g^−1	158
TiS2	APC in tetraglyme	160 mA h g^−1 at 12.5 mA g^−1 (60 °C)	∼0.7 V	∼112 W h kg^−1	∼115 mA h g^−1 after 40 cycles at 25 mA g^−1 (60 °C)	160
Layer-expanded TiS2	APC in THF with 0.25 M Py14Cl	239 mA h g^−1 at 24 mA g^−1	0.7 V	176 W h kg^−1	120 mA h g^−1 after 500 cycles at 240 mA g^−1	154
Pristine TiSe2	APC in THF	127 mA h g^−1 at 5 mA g^−1	1.25 V	159 W h kg^−1	80 mA h g^−1 after 40 cycles at 5 mA g^−1	157
Pristine VSe2	APC in THF	115 mA h g^−1 at 5 mA g^−1	1.2 V	138 W h kg^−1	86 mA h g^−1 after 40 cycles at 5 mA g^−1	157
2-Ethylhexylamine pillared VS2 nanoflowers	Mg(HMDS)2–4MgCl2 in THF with PP14TFSI	249 mA h g^−1 at 100 mA g^−1	0.7 V	174 W h kg^−1	90 mA h g^−1 after 600 cycles at 1 A g^−1	167
VS2 NSs	APC in THF with 0.2 M PP14Cl	348 mA h g^−1 at 100 mA g^−1	0.7 V	244 W h kg^−1	200 mA h g^−1 after 300 cycles at 1 A g^−1	80
Graphene-like MoS2	Mg(AlCl3Bu)2 in THF	170 mA h g^−1 at 20 mA g^−1	1.9 V	323 W h kg^−1	162 mA h g^−1 after 50 cycles at 20 mA g^−1	176
MoS2@C porous NRs	Fluorinated Mg alkoxyborate	120 mA h g^−1 at 10 mA g^−1	0.65 V	80 W h kg^−1	56 mA h g^−1 after 200 cycles at 100 mA g^−1	174
PVP-incorporated MoS2	APC in THF	143.4 mA h g^−1 at 20 mA g^−1	∼0.75 V	∼108 W h kg^−1	131.9 mA h g^−1 after 100 cycles at 20 mA g^−1	177
Expanded graphite	0.5 M Mg(TFSI)2 in Py14TFSI	93 mA h g^−1 at 100 mA g^−1	1.83 V	174 W h kg^−1	62 mA h g^−1 after 500 cycles at 300 mA g^−1	212
Triazine-based porous COF	0.5 M Mg(TFSI)2 in DME	102 mA h g^−1 at 57 mA g^−1	1.45 V	146 W h kg^−1	30 mA h g^−1 after 3000 cycles at 570 mA g^−1	208

---

Table 4 Summary of layered cathode materials for AMBs^a

Material	Electrolyte	Specific capacity	Average discharge voltage	Energy density based on the cathode	Cyclability	Ref.
a NFs – nanofibers, TMS – tetramethylene sulfone, NW – nanowire, AC – active carbon, NB – nanobelts, NS – nanosheet, NR – nanorod, BMIM – 1-butyl-3-methylimidazolium, PMIM – 1-methyl-3-propylimidazolium chlorides, TpBpy – 1,3,5-triformylphloroglucinol + 2,2′-bipyridine-5,5′-diamine.
V2O5 NWs	AlCl3 : [EMIm]Cl molar ratio 1.1 : 1	305 mA h g^−1 at 125 mA g^−1	0.55 V	168 W h kg^−1	273 mA h g^−1 after 20 cycles at 125 mA g^−1	143
Binder-free V2O5	AlCl3 : [BMIm]Cl molar ratio 1.1 : 1	239 mA h g^−1 at 44.2 mA g^−1	0.6 V	143 W h kg^−1	N/A	213
V2O5 NWs	AlCl3 : [BMIm]Cl molar ratio 1.1 : 1	107 mA h g^−1 at N/A	0.5 V	54 W h kg^−1	40 mA h g^−1 after 10 cycles	108
α-MoO3	AlCl3 : [EMIm]Cl molar ratio 1.1 : 1	100 mA h g^−1 at 3 mA g^−1	0.9 V	90 W h kg^−1	12.5 mA h g^−1 after 25 cycles at 10 mA g^−1	214
TiS2	AlCl3 : [BMIm]Cl molar ratio 1.5 : 1	70 mA h g^−1 at 5 mA g^−1 (50 °C)	0.75 V	53 W h kg^−1	70 mA h g^−1 after 50 cycles at 5 mA g^−1 (50 °C)	106
TiS2 NB	AlCl3 : [EMIm]Cl molar ratio 1.5 : 1	200 mA h g^−1 at 240 mA g^−1 (at 50 °C)	0.4 V	80 W h kg^−1	150 mA h g^−1 after 90 cycles at 240 mA g^−1	161
Graphene-VS2 NS	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	186 mA h g^−1 at 100 mA g^−1	0.6 V	112 W h kg^−1	50 mA h g^−1 after 50 cycles at 100 mA g^−1	171
VSe2	AlCl3 : [EMIm]Cl	419 mA h g^−1 at 100 mA g^−1	1.2 V	503 W h kg^−1	50 mA h g^−1 after 100 cycles at 100 mA g^−1	172
MoS2 microspheres	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	253.6 mA h g^−1 at 20 mA g^−1	0.6 V	152 W h kg^−1	66.7 mA h g^−1 after 100 cycles at 40 mA g^−1	111
Free-standing MoS2/carbon NF	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	293.2 mA h g^−1 at 100 mA g^−1	∼0.5 V	∼147 W h kg^−1	127 mA h g^−1 after 200 cycles at 100 mA g^−1	178
Flower-like MoS2 microspheres	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	154 mA h g^−1 at 50 mA g^−1	∼0.4 V	∼61 W h kg^−1	112 mA h g^−1 after 100 cycles at 50 mA g^−1	179
Carbon paper	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	90 mA h g^−1 at 50 mA g^−1	1.8 V	162 W h kg^−1	70 mA h g^−1 after 100 cycles at 100 mA g^−1	215
Graphitic foam	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	60 mA h g^−1 at 5 A g^−1	1.8 V	108 W h kg^−1	60 mA h g^−1 after 7500 cycles at 4 A g^−1	21
Aligned graphene sheets	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	60 mA h g^−1 at 12 A g^−1	1.8 V	108 W h kg^−1	60 mA h g^−1 after 4000 cycles at 12 A g^−1	112
Few-layer graphene	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	110 mA h g^−1 at 100 mA g^−1	2 V	220 W h kg^−1	75 mA h g^−1 after 1000 cycles at 3 A g^−1	216
Natural graphite	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	132 mA h g^−1 at 100 mA g^−1	2 V	264 W h kg^−1	132 mA h g^−1 after 100 cycles at 100 mA g^−1	217
Graphite powder	AlCl3 : urea molar ratio 1.3 : 1	73 mA h g^−1 at 100 mA g^−1	1.73 V	126 W h kg^−1	∼73 mA h g^−1 after 200 cycles at 100 mA g^−1	105
Natural graphite	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	110 mA h g^−1 at 99 mA g^−1	∼1.9 V	209 W h kg^−1	60 mA h g^−1 after 6000 cycles at 600 mA g^−1	113
Large-size few-layer graphene	AlCl3 : [PMIm]Cl molar ratio 1.3 : 1	85 mA h g^−1 at 60 mA g^−1	1.8 V	153 W h kg^−1	80 mA h g^−1 after 200 cycles at 60 mA g^−1	184
Carbon nanoscrolls	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	104 mA h g^−1 at 1 A g^−1	1.5 V	156 W h kg^−1	101 mA h g^−1 after 55 000 cycles at 50 A g^−1	218
Graphite powder	AlCl3 : [EMIm]Cl molar ratio 1.7 : 1	∼110 mA h g^−1 at 50 mA g^−1 (at −10 °C)	∼2 V (−10 °C)	∼220 W h kg^−1 (at −10 °C)	85 mA h g^−1 after 1200 cycles at 100 mA g^−1 (−20 °C)	219
Edge-rich graphene	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	128 mA h g^−1 at 2 A g^−1	1.8 V	230 W h kg^−1	90 mA h g^−1 after 20 000 cycles at 8 A g^−1	220
Graphene aerogel	AlCl3/Et3NHCl molar ratio 1.5 : 1	112 mA h g^−1 at 5 A g^−1	∼1.7 V	190 W h kg^−1	109 mA h g^−1 after 30 000 cycles at 5 A g^−1	188
Edge-rich graphitic nanoribbons	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	126 mA h g^−1 at 1 A g^−1	1.91 V	241 W h kg^−1	105 mA h g^−1 after 20 000 cycles at 10 A g^−1	186
Surface-perforated graphene	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	197 mA h g^−1 at 2 A g^−1	∼1.75 V	∼345 W h kg^−1	147 mA h g^−1 after 1000 cycles at 5 A g^−1	187
TpBpy-COF	AlCl3 : [EMIm]Cl molar ratio 1.3 : 1	307 mA h g^−1 at 100 mA g^−1	∼1.15 V	∼353 W h kg^−1	150 mA h g^−1 after 13 000 cycles at 2 A g^−1	210

---

Fig. 29 Average discharge voltage vs. specific capacity of recently developed layered cathodes for (a) ZMBs,^{23,24,59,61,65,66,183} (b) MMBs,^{80,136,141,153,154,157,158,160,167,174,176,177,208,212} and (c) AMBs.^{21,105,106,108,111–113,143,161,171,172,178,179,184,186–188,210,213–220} The black (100 mA h g⁻¹), red (200 mA h g⁻¹), green (300 mA h g⁻¹), and blue (400 mA h g⁻¹) dash lines correspond to energy density at the same values.

Although many material engineering strategies have been reported to modify the structures of layered materials, their side effects on the electrochemical performance were not adequately addressed in the previous studies. For example, guest species incorporation would bring the extraction of molecules (e.g., water) into the electrolyte, inducing the formation of the passivation layer on metal anodes. The incorporated molecules would also induce the partial reduction of active sites, thus lessening the theoretical specific capacity. Moreover, interlayer expansion strategies would also lead to low discharge voltages compared with the theoretical values for pristine materials. New material structure engineering strategies at the molecule level are highly desired. In this sense, many approaches, which have been well demonstrated in alkali metal ion batteries (e.g., defect engineering, artifice electrode/electrolyte interface construction), could be considered in future researches.

The sluggish charge-storage kinetics of MVMB cathodes associated with the multivalent metal ions as charge carriers represents one of the dominant factors restricting the performance. To this end, employing anion-storage cathodes (e.g., graphite and p-type organic redox compounds) for MVMBs represents a feasible way to avoid the kinetics issue, because the low-charge-density anion-storage (e.g., PF₆⁻, TFSI⁻, and AlCl₄⁻) chemistry allows cathode with high rate capability and high redox potentials. Anion-storage chemistries have been extensively explored in AMBs and preliminarily demonstrated in ZMBs. More future efforts are desired to further expand anion-storage chemistries to more MVMB devices, which will require the development of high-capacity anion-storage sites and wide-potential-window electrolytes. Besides, rationally constructing 2D COFs with dense p-type organic redox groups could be an effective approach to obtain superior anion-storage cathodes.

Furthermore, the dominant motivation to develop MVMBs comes from the advantages brought by the direct use of multivalent metal anodes (e.g., superior volumetric capacity and dendrite-free stripping/plating). However, in many studies, the developed cathodes were evaluated in electrolytes that are incompatible with the metal anode chemistry (e.g., the Mg(ClO₄)₂ in AN electrolyte for MMBs). In this regard, the acquired electrochemical data would not be the suitable reference for assembling MVMBs. In addition, various device parameters (e.g., mass loading, electrode preparation method, and electrolyte) for the performance assessment of cathodes could significantly affect the electrochemical performance. Therefore, the comparison between cathodes evaluated in different device systems is somehow unjustified. Additionally, the cathode analysis focused primarily on the specific capacity and rate capability. Important parameters, such as energy density, energy efficiency, and power density, are often overlooked. Especially, in some cases, the specific capacities were measured at very low voltages, resulting in the negligible contribution to the energy. Therefore, it is essential for the community to push forwards the standardization of electrode evaluation for MVMBs, which take under consideration the important parameters measured at agreed conditions.

Although the research on MVMBs is still at the primary research stage, commercialization is always the ultimate goal for new battery chemistries. Thus, we summarize some promising research directions of non-aqueous MVMBs in views of practical application as below. Zn metal stability in aqueous electrolytes is restricted by the severe dendrite growth, thus motivating the development of non-aqueous ZMBs with reversible and efficient anode electrochemistry. However, the intercalation kinetics of Zn²⁺ into layered TMO cathodes is sluggish in non-aqueous electrolytes. Learned from aqueous ZMBs, trace water additive holds the promise to greatly improve the performance of TMO cathodes for non-aqueous ZMBs. In this sense, additional efforts are required to explore the effect of the water additive on the Zn anode and fine-tune the amount of the water additive in organic solvents. As Mg has the lowest stripping/plating potential among the presented multivalent metals, it has great potential for constructing batteries with high energy and power densities. Among the possible cathode materials for MMBs, TMOs exhibit the most promising performance with large theoretical capacities, high redox potentials, and thus large theoretical energy densities. In addition, the demonstration of effective Mg²⁺ intercalation into TMOs and decent cyclability imply their potential as promising cathodes for MMBs. However, before the implementation of TMOs-based MMB devices can be made, considerable efforts should be devoted to developing compatible electrolytes for both the TMO cathodes and the Mg metal anode. Unlike TMOs, TMDs are fully compatible with currently developed Mg electrolytes. Nevertheless, large amounts of their capacities are at the low operation voltage range (<0.5 V), which is not feasible for practical applications. For AMBs, the intercalation of Al³⁺ causes the fast capacity degradation of most layered cathodes. Recent studies suggest other Al species, such as AlCl₂⁺, AlCl²⁺, and AlCl₄⁻, to be promising charge carriers for AMB cathodes. In this regard, carbon-based/carbon-rich materials, such as graphite, 2D COFs, and 2D conjugated metal–organic frameworks, are considered as the potential alternative cathodes for AMBs, as they enable new electrochemistries, high specific capacities, long-term cyclability, and high operation voltage that other layered materials are lacking (e.g., TMOs and TMDs).

Finally, the assessment of the cathode performance alone could not promote the transition from research to practical application. Most reported studies focus only on evaluating cathodes with flood electrolytes and much over-capacity metal anodes, and the full-device demonstration for MVMBs is often missing in previous studies. When assembling full devices, the anode–electrolyte–cathode ratios could have a prominent effect on the performance. In addition, some MVMB electrolytes (e.g., APC for MMBs and AlCl₃/EMIMCl for AMBs) used in previous studies are not compatible with the commonly used current collectors (e.g., Al, Cu, and stainless steel) due to the strong corrosive effect. In this sense, the fabrication of promising cathodes should go side-by-side with the development of suitable electrolytes. Furthermore, effort should also be directed to the development of simple procedures for cathodes material synthesis. Since most pristine layered materials display inefficient multivalent metal ion storage, post structure engineering (e.g., layer expansion) is commonly required to promote the ion-storage kinetics. However, such structure engineering steps could hinder their commercialization due to the high material processing cost.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by European Union’s Horizon 2020 research and innovation programme (GrapheneCore3 881603), M-ERA.NET and Sächsisches Staatsministerium für Wissenschaft und Kunst (HYSUCAP 100478697), German Research Foundation (DFG) within the Cluster of Excellence, CRC 1415 (grant no. 417590517), and Polymer-based Batteries (SPP 2248, RACOF-MMIS). The authors appreciate the helpful discussion with Prof. Xinliang Feng at TU Dresden.

References

1. S. C. Pryor and R. J. Barthelmie, Nat. Energy, 2021, 6, 268–276.
2. C. Mitchell, Nat. Energy, 2016, 1, 15030.
3. M. Yu and X. Feng, Joule, 2019, 3, 338–360.
4. P. Zhang, F. Wang, S. Yang, G. Wang, M. Yu and X. Feng, Energy Storage Mater., 2020, 28, 160–187.
5. M. Wang, R. Dong and X. Feng, Chem. Soc. Rev., 2021, 50, 2764–2793.
6. J. Jaguemont, L. Boulon and Y. Dubé, Appl. Energy, 2016, 164, 99–114.
7. Y. Wang, B. Liu, Q. Li, S. Cartmell, S. Ferrara, Z. D. Deng and J. Xiao, J. Power Sources, 2015, 286, 330–345.
8. J. Muldoon, C. B. Bucur and T. Gregory, Chem. Rev., 2014, 114, 11683–11720.
9. M. Mao, T. Gao, S. Hou and C. Wang, Chem. Soc. Rev., 2018, 47, 8804–8841.
10. M. M. Huie, D. C. Bock, E. S. Takeuchi, A. C. Marschilok and K. J. Takeuchi, Coord. Chem. Rev., 2015, 287, 15–27.
11. A. Ponrouch, J. Bitenc, R. Dominko, N. Lindahl, P. Johansson and M. R. Palacin, Energy Storage Mater., 2019, 20, 253–262.
12. J. Xie and Q. Zhang, Small, 2019, 15, 1805061.
13. R. Dominko, J. Bitenc, R. Berthelot, M. Gauthier, G. Pagot and V. Di Noto, J. Power Sources, 2020, 478, 229027.
14. H. Li, L. Ma, C. Han, Z. Wang, Z. Liu, Z. Tang and C. Zhi, Nano Energy, 2019, 62, 550–587.
15. A. Ponrouch, J. Bitenc, R. Dominko, N. Lindahl, P. Johansson and M. R. R. Palacin, Energy Storage Mater., 2019, 20, 253–262.
16. Z. Hu, H. Zhang, H. Wang, F. Zhang, Q. Li and H. Li, ACS Mater. Lett., 2020, 2, 887–904.
17. Y. Liu, G. He, H. Jiang, I. P. Parkin, P. R. Shearing and D. J. L. Brett, Adv. Funct. Mater., 2021, 31, 2010445.
18. C. Wu, H. Tan, W. Huang, C. Liu, W. Wei, L. Chen and Q. Yan, Mater. Today Energy, 2021, 19, 100595.
19. T. Watkins, A. Kumar and D. A. Buttry, J. Am. Chem. Soc., 2016, 138, 641–650.
20. H. Dong, O. Tutusaus, Y. Liang, Y. Zhang, Z. Lebens-Higgins, W. Yang, R. Mohtadi and Y. Yao, Nat. Energy, 2020, 5, 1043–1050.
21. M. C. Lin, M. Gong, B. Lu, Y. Wu, D. Y. Wang, M. Guan, M. Angell, C. Chen, J. Yang, B. J. Hwang and H. Dai, Nature, 2015, 520, 324–328.
22. J. Zheng, D. C. Bock, T. Tang, Q. Zhao, J. Yin, K. R. Tallman, G. Wheeler, X. Liu, Y. Deng, S. Jin, A. C. Marschilok, E. S. Takeuchi, K. J. Takeuchi and L. A. Archer, Nat. Energy, 2021, 6, 398–406.
23. Y. Wang, L. Zhang, F. Zhang, X. Ding, K. Shin and Y. Tang, J. Energy Chem., 2021, 58, 602–609.
24. G. Wang, B. Kohn, U. Scheler, F. Wang, S. Oswald, M. Löffler, D. Tan, P. Zhang, J. Zhang and X. Feng, Adv. Mater., 2020, 32, 1905681.
25. W. Chen, X. Zhan, B. Luo, Z. Ou, P. C. Shih, L. Yao, S. Pidaparthy, A. Patra, H. An, P. V. Braun, R. M. Stephens, H. Yang, J.-M. Zuo and Q. Chen, Nano Lett., 2019, 19, 4712–4720.
26. R. D. Bayliss, B. Key, G. Sai Gautam, P. Canepa, B. J. Kwon, S. H. Lapidus, F. Dogan, A. A. Adil, A. S. Lipton, P. J. Baker, G. Ceder, J. T. Vaughey and J. Cabana, Chem. Mater., 2020, 32, 663–670.
27. C. Kim, P. J. Phillips, B. Key, T. Yi, D. Nordlund, Y. S. Yu, R. D. Bayliss, S. D. Han, M. He, Z. Zhang, A. K. Burrell, R. F. Klie and J. Cabana, Adv. Mater., 2015, 27, 3377–3384.
28. Q. D. Truong, M. Kempaiah Devaraju, P. D. Tran, Y. Gambe, K. Nayuki, Y. Sasaki and I. Honma, Chem. Mater., 2017, 29, 6245–6251.
29. T. Koketsu, J. Ma, B. J. Morgan, M. Body, C. Legein, W. Dachraoui, M. Giannini, A. Demortière, M. Salanne, F. Dardoize, H. Groult, O. J. Borkiewicz, K. W. Chapman, P. Strasser and D. Dambournet, Nat. Mater., 2017, 16, 1142–1148.
30. C. Legein, B. J. Morgan, F. Fayon, T. Koketsu, J. Ma, M. Body, V. Sarou-Kanian, X. Wei, M. Heggen, O. J. Borkiewicz, P. Strasser and D. Dambournet, Angew. Chem., Int. Ed., 2020, 59, 19247–19253.
31. T. D. Gregory, R. J. Hoffman and R. C. Winterton, J. Electrochem. Soc., 1990, 137, 775–780.
32. K. Kalantar-zadeh, J. Z. Ou, T. Daeneke, A. Mitchell, T. Sasaki and M. S. Fuhrer, Appl. Mater. Today, 2016, 5, 73–89.
33. G. Fiori, F. Bonaccorso, G. Iannaccone, T. Palacios, D. Neumaier, A. Seabaugh, S. K. Banerjee and L. Colombo, Nat. Nanotechnol., 2014, 9, 768–779.
34. A. B. Kaul, J. Mater. Res., 2014, 29, 348–361.
35. D. L. Duong, S. J. Yun and Y. H. Lee, ACS Nano, 2017, 11, 11803–11830.
36. A. Manthiram, ACS Cent. Sci., 2017, 3, 1063–1069.
37. C. Liu, Z. G. Neale and G. Cao, Mater. Today, 2016, 19, 109–123.
38. F. Wei, Q. Zhang, P. Zhang, W. Tian, K. Dai, L. Zhang, J. Mao and G. Shao, J. Electrochem. Soc., 2021, 168, 050524.
39. Y. Xiao, N. M. Abbasi, Y. Zhu, S. Li, S. Tan, W. Ling, L. Peng, T. Yang, L. Wang, X. Guo, Y. Yin, H. Zhang and Y. Guo, Adv. Funct. Mater., 2020, 30, 2001334.
40. Q. Liu, Z. Hu, M. Chen, C. Zou, H. Jin, S. Wang, S. Chou, Y. Liu and S. Dou, Adv. Funct. Mater., 2020, 30, 1909530.
41. B. Chen, D. Chao, E. Liu, M. Jaroniec, N. Zhao and S. Z. Qiao, Energy Environ. Sci., 2020, 13, 1096–1131.
42. W. S. V. Lee, T. Xiong, X. Wang and J. Xue, Small Methods, 2021, 5, 2000815.
43. S. Manzeli, D. Ovchinnikov, D. Pasquier, O. V. Yazyev and A. Kis, Nat. Rev. Mater., 2017, 2, 17033.
44. G. Wang, M. Yu and X. Feng, Chem. Soc. Rev., 2021, 50, 2388–2443.
45. Y. Li, Y. Lu, P. Adelhelm, M. M. Titirici and Y. S. Hu, Chem. Soc. Rev., 2019, 48, 4655–4687.
46. M. Yu, R. Dong and X. Feng, J. Am. Chem. Soc., 2020, 142, 12903–12915.
47. Y. Cao, Y. Lin, J. Wu, X. Huang, Z. Pei, J. Zhou and G. Wang, ChemSusChem, 2020, 13, 1392–1408.
48. Y. Shen, Y. Wang, Y. Miao, M. Yang, X. Zhao and X. Shen, Adv. Mater., 2020, 32, 1905524.
49. Y. Liang, H. Dong, D. Aurbach and Y. Yao, Nat. Energy, 2020, 5, 646–656.
50. T. Placke, A. Heckmann, R. Schmuch, P. Meister, K. Beltrop and M. Winter, Joule, 2018, 2, 2528–2550.
51. Z. Yi, G. Chen, F. Hou, L. Wang and J. Liang, Adv. Energy Mater., 2021, 11, 2003065.
52. B. Tang, L. Shan, S. Liang and J. Zhou, Energy Environ. Sci., 2019, 12, 3288–3304.
53. M. Song, H. Tan, D. Chao and H. J. Fan, Adv. Funct. Mater., 2018, 28, 1802564.
54. D. Selvakumaran, A. Pan, S. Liang and G. Cao, J. Mater. Chem. A, 2019, 7, 18209–18236.
55. S. Huang, J. Zhu, J. Tian and Z. Niu, Chem.–Eur. J., 2019, 25, 14480–14494.
56. N. Zhang, X. Chen, M. Yu, Z. Niu, F. Cheng and J. Chen, Chem. Soc. Rev., 2020, 49, 4203–4219.
57. P. Yu, Y. Zeng, H. Zhang, M. Yu, Y. Tong and X. Lu, Small, 2019, 15, 1804760.
58. H. Pan, Y. Shao, P. Yan, Y. Cheng, K. S. Han, Z. Nie, C. Wang, J. Yang, X. Li, P. Bhattacharya, K. T. Mueller and J. Liu, Nat. Energy, 2016, 1, 16039.
59. V. Verma, S. Kumar, W. Manalastas, J. Zhao, R. Chua, S. Meng, P. Kidkhunthod and M. Srinivasan, ACS Appl. Energy Mater., 2019, 2, 8667–8674.
60. Q. Yang, Q. Li, Z. Liu, D. Wang, Y. Guo, X. Li, Y. Tang, H. Li, B. Dong and C. Zhi, Adv. Mater., 2020, 32, 2001854.
61. D. Kundu, S. Hosseini Vajargah, L. Wan, B. Adams, D. Prendergast and L. F. Nazar, Energy Environ. Sci., 2018, 11, 881–892.
62. C. Pan, R. Zhang, R. G. Nuzzo and A. A. Gewirth, Adv. Energy Mater., 2018, 8, 1800589.
63. A. Guerfi, J. Trottier, I. Boyano, I. De Meatza, J. A. Blazquez, S. Brewer, K. S. Ryder, A. Vijh and K. Zaghib, J. Power Sources, 2014, 248, 1099–1104.
64. S. D. Han, N. N. Rajput, X. Qu, B. Pan, M. He, M. S. Ferrandon, C. Liao, K. A. Persson and A. K. Burrell, ACS Appl. Mater. Interfaces, 2016, 8, 3021–3031.
65. P. Senguttuvan, S. D. Han, S. Kim, A. L. Lipson, S. Tepavcevic, T. T. Fister, I. D. Bloom, A. K. Burrell and C. S. Johnson, Adv. Energy Mater., 2016, 6, 1600826.
66. S. D. Han, S. Kim, D. Li, V. Petkov, H. D. Yoo, P. J. Phillips, H. Wang, J. J. Kim, K. L. More, B. Key, R. F. Klie, J. Cabana, V. R. Stamenkovic, T. T. Fister, N. M. Markovic, A. K. Burrell, S. Tepavcevic and J. T. Vaughey, Chem. Mater., 2017, 29, 4874–4884.
67. Z. Chen, F. Mo, T. Wang, Q. Yang, Z. Huang, D. Wang, G. Liang, A. Chen, Q. Li, Y. Guo, X. Li, J. Fan and C. Zhi, Energy Environ. Sci., 2021, 14, 2441–2450.
68. B. Ji, W. Yao and Y. Tang, Sustainable Energy Fuels, 2020, 4, 101–107.
69. M. J. Park and A. Manthiram, ACS Appl. Energy Mater., 2020, 3, 5015–5023.
70. Z. Lu, A. Schechter, M. Moshkovich and D. Aurbach, J. Electroanal. Chem., 1999, 466, 203–217.
71. L. W. Gaddum and H. E. French, J. Am. Chem. Soc., 1927, 49, 1295–1299.
72. C. Liebenow, J. Appl. Electrochem., 1997, 27, 221–225.
73. T. D. Gregory, R. J. Hoffman and R. C. Winterton, J. Electrochem. Soc., 1990, 137, 775–780.
74. D. Aurbach, Z. Lu, A. Schechter, Y. Gofer, H. Gizbar, R. Turgeman, Y. Cohen, M. Moshkovich and E. Levi, Nature, 2000, 407, 724–727.
75. S. Sakamoto, T. Imamoto and K. Yamaguchi, Org. Lett., 2001, 3, 1793–1795.
76. H. S. Kim, T. S. Arthur, G. D. Allred, J. Zajicek, J. G. Newman, A. E. Rodnyansky, A. G. Oliver, W. C. Boggess and J. Muldoon, Nat. Commun., 2011, 2, 427.
77. A. K. Lautar, J. Bitenc, R. Dominko and J. S. Filhol, ACS Appl. Mater. Interfaces, 2021, 13, 8263–8273.
78. Z. Zhao-Karger, J. E. Mueller, X. Zhao, O. Fuhr, T. Jacob and M. Fichtner, RSC Adv., 2014, 4, 26924–26927.
79. N. Pour, Y. Gofer, D. T. Major and D. Aurbach, J. Am. Chem. Soc., 2011, 133, 6270–6278.
80. Y. Zhao, D. Wang, D. Yang, L. Wei, B. Liu, X. Wang, G. Chen and Y. Wei, Energy Storage Mater., 2019, 23, 749–756.
81. R. E. Doe, R. Han, J. Hwang, A. J. Gmitter, I. Shterenberg, H. D. Yoo, N. Pour and D. Aurbach, Chem. Commun., 2014, 50, 243–245.
82. S. He, J. Luo and T. L. Liu, J. Mater. Chem. A, 2017, 5, 12718–12722.
83. T. Liu, Y. Shao, G. Li, M. Gu, J. Hu, S. Xu, Z. Nie, X. Chen, C. Wang and J. Liu, J. Mater. Chem. A, 2014, 2, 3430–3438.
84. G. Bieker, M. Salama, M. Kolek, Y. Gofer, P. Bieker, D. Aurbach and M. Winter, ACS Appl. Mater. Interfaces, 2019, 11, 24057–24066.
85. S. J. Kang, H. Kim, S. Hwang, M. Jo, M. Jang, C. Park, S. T. Hong and H. Lee, ACS Appl. Mater. Interfaces, 2019, 11, 517–524.
86. Y. He, Q. Li, L. Yang, C. Yang and D. Xu, Angew. Chem., Int. Ed., 2019, 58, 7615–7619.
87. J. Bitenc, K. Pirnat, E. Žagar, A. Randon-Vitanova and R. Dominko, J. Power Sources, 2019, 430, 90–94.
88. C. Liao, N. Sa, B. Key, A. K. Burrell, L. Cheng, L. A. Curtiss, J. T. Vaughey, J. J. Woo, L. Hu, B. Pan and Z. Zhang, J. Mater. Chem. A, 2015, 3, 6082–6087.
89. J. Luo, Y. Bi, L. Zhang, X. Zhang and T. L. Liu, Angew. Chem., Int. Ed., 2019, 58, 6967–6971.
90. R. Mohtadi, M. Matsui, T. S. Arthur and S. J. Hwang, Angew. Chem., Int. Ed., 2012, 51, 9780–9783.
91. O. Tutusaus, R. Mohtadi, T. S. Arthur, F. Mizuno, E. G. Nelson and Y. V. Sevryugina, Angew. Chem., Int. Ed., 2015, 54, 7900–7904.
92. L. F. Wan, B. R. Perdue, C. A. Apblett and D. Prendergast, Chem. Mater., 2015, 27, 5932–5940.
93. H. D. Yoo, J. R. Jokisaari, Y. S. Yu, B. J. Kwon, L. Hu, S. Kim, S. D. Han, M. Lopez, S. H. Lapidus, G. M. Nolis, B. J. Ingram, I. Bolotin, S. Ahmed, R. F. Klie, J. T. Vaughey, T. T. Fister and J. Cabana, ACS Energy Lett., 2019, 4, 1528–1534.
94. L. Hu, J. R. Jokisaari, B. J. Kwon, L. Yin, S. Kim, H. Park, S. H. Lapidus, R. F. Klie, B. Key, P. Zapol, B. J. Ingram, J. T. Vaughey and J. Cabana, ACS Energy Lett., 2020, 5, 2721–2727.
95. X. Sun, V. Duffort, B. L. Mehdi, N. D. Browning and L. F. Nazar, Chem. Mater., 2016, 28, 534–542.
96. B. J. Kwon, L. Yin, H. Park, P. Parajuli, K. Kumar, S. Kim, M. Yang, M. Murphy, P. Zapol, C. Liao, T. T. Fister, R. F. Klie, J. Cabana, J. T. Vaughey, S. H. Lapidus and B. Key, Chem. Mater., 2020, 32, 6577–6587.
97. H. Yang, H. Li, J. Li, Z. Sun, K. He, H. M. Cheng and F. Li, Angew. Chem., Int. Ed., 2019, 58, 11978–11996.
98. L. D. Reed, A. Arteaga and E. J. Menke, J. Phys. Chem. B, 2015, 119, 12677–12681.
99. T. J. Melton, J. Joyce, J. T. Maloy, J. A. Boon and J. S. Wilkes, J. Electrochem. Soc., 1990, 137, 3865–3869.
100. M. Jafarian, M. G. Mahjani, F. Gobal and I. Danaee, J. Electroanal. Chem., 2006, 588, 190–196.
101. P. Rolland and G. Mamantov, J. Electrochem. Soc., 1976, 123, 1299–1303.
102. T. Tsuda, I. Kokubo, M. Kawabata, M. Yamagata, M. Ishikawa, S. Kusumoto, A. Imanishi and S. Kuwabata, J. Electrochem. Soc., 2014, 161, A908–A914.
103. K. V. Kravchyk and M. V. Kovalenko, Commun. Chem., 2020, 3, 120.
104. W. Chu, X. Zhang, J. Wang, S. Zhao, S. Liu and H. Yu, Energy Storage Mater., 2019, 22, 418–423.
105. M. Angell, C. J. Pan, Y. Rong, C. Yuan, M. C. Lin, B. Hwang and H. Dai, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 834–839.
106. L. Geng, J. P. Scheifers, C. Fu, J. Zhang, B. P. T. Fokwa and J. Guo, ACS Appl. Mater. Interfaces, 2017, 9, 21251–21257.
107. Z. Hu, K. Zhi, Q. Li, Z. Zhao, H. Liang, X. Liu, J. Huang, C. Zhang, H. Li and X. Guo, J. Power Sources, 2019, 440, 227147.
108. S. Gu, H. Wang, C. Wu, Y. Bai, H. Li and F. Wu, Energy Storage Mater., 2017, 6, 9–17.
109. H. Yang, L. Yin, J. Liang, Z. Sun, Y. Wang, H. Li, K. He, L. Ma, Z. Peng, S. Qiu, C. Sun, H. M. Cheng and F. Li, Angew. Chem., Int. Ed., 2018, 57, 1898–1902.
110. D. J. Kim, D. J. Yoo, M. T. Otley, A. Prokofjevs, C. Pezzato, M. Owczarek, S. J. Lee, J. W. Choi and J. F. Stoddart, Nat. Energy, 2019, 4, 51–59.
111. Z. Li, B. Niu, J. Liu, J. Li and F. Kang, ACS Appl. Mater. Interfaces, 2018, 10, 9451–9459.
112. Y. Wu, M. Gong, M. C. Lin, C. Yuan, M. Angell, L. Huang, D. Y. Wang, X. Zhang, J. Yang, B. J. Hwang and H. Dai, Adv. Mater., 2016, 28, 9218–9222.
113. D.-Y. Wang, C. Y. Wei, M. C. Lin, C. J. Pan, H. L. Chou, H. A. Chen, M. Gong, Y. Wu, C. Yuan, M. Angell, Y. J. Hsieh, Y. H. Chen, C. Y. Wen, C. W. Chen, B. J. Hwang, C. C. Chen and H. Dai, Nat. Commun., 2017, 8, 14283.
114. M. Mao, C. Luo, T. P. Pollard, S. Hou, T. Gao, X. Fan, C. Cui, J. Yue, Y. Tong, G. Yang, T. Deng, M. Zhang, J. Ma, L. Suo, O. Borodin and C. Wang, Angew. Chem., Int. Ed., 2019, 58, 17820–17826.
115. S. Wang, S. Huang, M. Yao, Y. Zhang and Z. Niu, Angew. Chem., 2020, 132, 11898–11905.
116. Y. Zhang, E. H. Ang, Y. Yang, M. Ye, W. Du and C. C. Li, Adv. Funct. Mater., 2021, 31, 2007358.
117. A. Chakraborty, S. Kunnikuruvan, S. Kumar, B. Markovsky, D. Aurbach, M. Dixit and D. T. Major, Chem. Mater., 2020, 32, 915–952.
118. G. M. Kanyolo, T. Masese, N. Matsubara, C. Y. Chen, J. Rizell, Z. D. Huang, Y. Sassa, M. Månsson, H. Senoh and H. Matsumoto, Chem. Soc. Rev., 2021, 50, 3990–4030.
119. K. Kubota, Electrochemistry, 2020, 88, 507–514.
120. J. Zhou, X. Lu and M. Yu, Mater. Chem. Front., 2021, 5, 2996–3020.
121. J. S. Kim, W. S. Chang, R. H. Kim, D. Y. Kim, D. W. Han, K. H. Lee, S. S. Lee and S. G. Doo, J. Power Sources, 2015, 273, 210–215.
122. W. Liu, P. Oh, X. Liu, M. J. Lee, W. Cho, S. Chae, Y. Kim and J. Cho, Angew. Chem., Int. Ed., 2015, 54, 4440–4457.
123. S. Komaba, N. Kumagai and S. Chiba, Electrochim. Acta, 2000, 46, 31–37.
124. M. H. Alfaruqi, J. Gim, S. Kim, J. Song, D. T. Pham, J. Jo, Z. Xiu, V. Mathew and J. Kim, Electrochem. Commun., 2015, 60, 121–125.
125. J. E. Post, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 3447–3454.
126. S. Devaraj and N. Munichandraiah, J. Phys. Chem. C, 2008, 112, 4406–4417.
127. C. H. Kim, Z. Akase, L. Zhang, A. H. Heuer, A. E. Newman and P. J. Hughes, J. Solid State Chem., 2006, 179, 753–774.
128. J. Song, M. Noked, E. Gillette, J. Duay, G. Rubloff and S. B. Lee, Phys. Chem. Chem. Phys., 2015, 17, 5256–5264.
129. A. Chakrabarti, K. Hermann, R. Druzinic, M. Witko, F. Wagner and M. Petersen, Phys. Rev. B, 1999, 59, 10583–10590.
130. M. Hibino, in Encyclopedia of Electrochemical Power Sources, Elsevier, 2009, pp. 40–50.
131. Q. An, P. Zhang, Q. Wei, L. He, F. Xiong, J. Sheng, Q. Wang and L. Mai, J. Mater. Chem. A, 2014, 2, 3297–3302.
132. Q. An, Y. Li, H. Deog Yoo, S. Chen, Q. Ru, L. Mai and Y. Yao, Nano Energy, 2015, 18, 265–272.
133. P. Novák and J. Desilvestro, J. Electrochem. Soc., 1993, 140, 140–144.
134. J. Song, L. Wang, Y. Lu, J. Liu, B. Guo, P. Xiao, J.-J. Lee, X.-Q. Yang, G. Henkelman and J. B. Goodenough, J. Am. Chem. Soc., 2015, 137, 2658–2664.
135. Y. Xu, X. Deng, Q. Li, G. Zhang, F. Xiong, S. Tan, Q. Wei, J. Lu, J. Li, Q. An and L. Mai, Chem, 2019, 5, 1194–1209.
136. S. Tepavcevic, Y. Liu, D. Zhou, B. Lai, J. Maser, X. Zuo, H. Chan, P. Král, C. S. Johnson, V. Stamenkovic, N. M. Markovic and T. Rajh, ACS Nano, 2015, 9, 8194–8205.
137. Y. Zhu, G. Huang, J. Yin, Y. Lei, A. Emwas, X. Yu, O. F. Mohammed and H. N. Alshareef, Adv. Energy Mater., 2020, 10, 2002128.
138. X. Deng, Y. Xu, Q. An, F. Xiong, S. Tan, L. Wu and L. Mai, J. Mater. Chem. A, 2019, 7, 10644–10650.
139. R. Sun, X. Ji, C. Luo, S. Hou, P. Hu, X. Pu, L. Cao, L. Mai and C. Wang, Small, 2020, 16, 2000741.
140. E. A. Esparcia, M. S. Chae, J. D. Ocon and S. T. Hong, Chem. Mater., 2018, 30, 3690–3696.
141. S. D. Perera, R. B. Archer, C. A. Damin, R. Mendoza-Cruz and C. P. Rhodes, J. Power Sources, 2017, 343, 580–591.
142. M. Rastgoo-Deylami, M. S. Chae and S. T. Hong, Chem. Mater., 2018, 30, 7464–7472.
143. N. Jayaprakash, S. K. Das and L. A. Archer, Chem. Commun., 2011, 47, 12610–12612.
144. L. D. Reed and E. Menke, J. Electrochem. Soc., 2013, 160, A915–A917.
145. M. Chiku, H. Takeda, S. Matsumura, E. Higuchi and H. Inoue, ACS Appl. Mater. Interfaces, 2015, 7, 24385–24389.
146. L. Beneš, K. Melánová, J. Svoboda and V. Zima, J. Inclusion Phenom. Macrocyclic Chem., 2012, 73, 33–53.
147. R. Gautier, R. Gautier, O. Hernandez, N. Audebrand, T. Bataille, C. Roiland, E. Elkaïm, L. Le Pollès, E. Furet and E. Le Fur, Dalton Trans., 2013, 42, 8124–8131.
148. Y. Fang, Q. Liu, L. Xiao, Y. Rong, Y. Liu, Z. Chen, X. Ai, Y. Cao, H. Yang, J. Xie, C. Sun, X. Zhang, B. Aoun, X. Xing, X. Xiao and Y. Ren, Chem, 2018, 4, 1167–1180.
149. G. He, W. H. Kan and A. Manthiram, Chem. Mater., 2016, 28, 682–688.
150. C. Wu, X. Lu, L. Peng, K. Xu, X. Peng, J. Huang, G. Yu and Y. Xie, Nat. Commun., 2013, 4, 2431.
151. F. Wang, W. Sun, Z. Shadike, E. Hu, X. Ji, T. Gao, X. Q. Yang, K. Xu and C. Wang, Angew. Chem., Int. Ed., 2018, 57, 11978–11981.
152. X. Ji, J. Chen, F. Wang, W. Sun, Y. Ruan, L. Miao, J. Jiang and C. Wang, Nano Lett., 2018, 18, 6441–6448.
153. L. Zhou, Q. Liu, Z. Zhang, K. Zhang, F. Xiong, S. Tan, Q. An, Y.-M. Kang, Z. Zhou and L. Mai, Adv. Mater., 2018, 30, 1801984.
154. H. D. Yoo, Y. Liang, H. Dong, J. Lin, H. Wang, Y. Liu, L. Ma, T. Wu, Y. Li, Q. Ru, Y. Jing, Q. An, W. Zhou, J. Guo, J. Lu, S. T. Pantelides, X. Qian and Y. Yao, Nat. Commun., 2017, 8, 339.
155. L. Hu, Z. Wu, C. Lu, F. Ye, Q. Liu and Z. Sun, Energy Environ. Sci. 10.1039/D1EE01158H.
156. S. Niu, J. Cai and G. Wang, Nano Res., 2021, 14, 1985–2002.
157. M. Mao, X. Ji, S. Hou, T. Gao, F. Wang, L. Chen, X. Fan, J. Chen, J. Ma and C. Wang, Chem. Mater., 2019, 31, 3183–3191.
158. Y. Gu, Y. Katsura, T. Yoshino, H. Takagi and K. Taniguchi, Sci. Rep., 2015, 5, 12486.
159. Z. L. Tao, L. N. Xu, X. L. Gou, J. Chen and H. T. Yuan, Chem. Commun., 2004, 2080–2081.
160. X. Sun, P. Bonnick and L. F. Nazar, ACS Energy Lett., 2016, 1, 297–301.
161. C. G. Hawkins, A. Verma, W. Horbinski, R. Weeks, P. P. Mukherjee and L. Whittaker-Brooks, ACS Appl. Mater. Interfaces, 2020, 12, 21788–21798.
162. L. Geng, G. Lv, X. Xing and J. Guo, Chem. Mater., 2015, 27, 4926–4929.
163. Q. Yan, Y. Shen, Y. Miao, M. Wang, M. Yang and X. Zhao, J. Alloys Compd., 2019, 806, 1109–1115.
164. Z. Zhao, Z. Hu, Q. Li, H. Li, X. Zhang, Y. Zhuang, F. Wang and G. Yu, Nano Today, 2020, 32, 100870.
165. Y. Jing, Z. Zhou, C. R. Cabrera and Z. Chen, J. Phys. Chem. C, 2013, 117, 25409–25413.
166. J. Zhou, L. Wang, M. Yang, J. Wu, F. Chen, W. Huang, N. Han, H. Ye, F. Zhao, Y. Li and Y. Li, Adv. Mater., 2017, 29, 1702061.
167. X. Xue, R. Chen, C. Yan, P. Zhao, Y. Hu, W. Kong, H. Lin, L. Wang and Z. Jin, Adv. Energy Mater., 2019, 9, 1900145.
168. H. Tao, M. Zhou, R. Wang, K. Wang, S. Cheng and K. Jiang, Adv. Sci., 2018, 5, 1801021.
169. L. Zhang, D. Sun, J. Kang, H.-T. Wang, S.-H. Hsieh, W.-F. Pong, H. A. Bechtel, J. Feng, L.-W. Wang, E. J. Cairns and J. Guo, Nano Lett., 2018, 18, 4506–4515.
170. R. Sun, C. Pei, J. Sheng, D. Wang, L. Wu, S. Liu, Q. An and L. Mai, Energy Storage Mater., 2018, 12, 61–68.
171. L. Wu, R. Sun, F. Xiong, C. Pei, K. Han, C. Peng, Y. Fan, W. Yang, Q. An and L. Mai, Phys. Chem. Chem. Phys., 2018, 20, 22563–22568.
172. H. Lei, M. Wang, J. Tu and S. Jiao, Sustainable Energy Fuels, 2019, 3, 2717–2724.
173. D. Voiry, A. Goswami, R. Kappera, C. D. C. C. E. Silva, D. Kaplan, T. Fujita, M. Chen, T. Asefa and M. Chhowalla, Nat. Chem., 2015, 7, 45–49.
174. Z. Li, X. Mu, Z. Zhao-Karger, T. Diemant, R. J. Behm, C. Kübel and M. Fichtner, Nat. Commun., 2018, 9, 5115.
175. T. Stephenson, Z. Li, B. Olsen and D. Mitlin, Energy Environ. Sci., 2014, 7, 209–231.
176. Y. Liang, R. Feng, S. Yang, H. Ma, J. Liang and J. Chen, Adv. Mater., 2011, 23, 640–643.
177. C. Wu, G. Zhao, S. Gong, N. Zhang and K. Sun, J. Mater. Chem. A, 2019, 7, 4426–4430.
178. W. Yang, H. Lu, Y. Cao, B. Xu, Y. Deng and W. Cai, ACS Sustainable Chem. Eng., 2019, 7, 4861–4867.
179. J. Tu, X. Xiao, M. Wang and S. Jiao, J. Phys. Chem. C, 2019, 123, 26794–26802.
180. Z. Zhao, Z. Hu, H. Liang, S. Li, H. Wang, F. Gao, X. Sang and H. Li, ACS Appl. Mater. Interfaces, 2019, 11, 44333–44341.
181. X. Zhou, Q. Liu, C. Jiang, B. Ji, X. Ji, Y. Tang and H. M. Cheng, Angew. Chem., Int. Ed., 2020, 59, 3802–3832.
182. G. Schmuelling, T. Placke, R. Kloepsch, O. Fromm, H. W. Meyer, S. Passerini and M. Winter, J. Power Sources, 2013, 239, 563–571.
183. J. Fan, Q. Xiao, Y. Fang, L. Li and W. Yuan, Ionics, 2019, 25, 1303–1313.
184. L. Zhang, L. Chen, H. Luo, X. Zhou and Z. Liu, Adv. Energy Mater., 2017, 7, 1700034.
185. H. Huang, F. Zhou, X. Shi, J. Qin, Z. Zhang, X. Bao and Z. S. Wu, Energy Storage Mater., 2019, 23, 664–669.
186. Y. Hu, S. Debnath, H. Hu, B. Luo, X. Zhu, S. Wang, M. Hankel, D. J. Searles and L. Wang, J. Mater. Chem. A, 2019, 7, 15123–15130.
187. Y. Kong, C. Tang, X. Huang, A. K. Nanjundan, J. Zou, A. Du and C. Yu, Adv. Funct. Mater., 2021, 31, 2010569.
188. H. Xu, T. Bai, H. Chen, F. Guo, J. Xi, T. Huang, S. Cai, X. Chu, J. Ling, W. Gao, Z. Xu and C. Gao, Energy Storage Mater., 2019, 17, 38–45.
189. G. Wang, N. Chandrasekhar, B. P. Biswal, D. Becker, S. Paasch, E. Brunner, M. Addicoat, M. Yu, R. Berger and X. Feng, Adv. Mater., 2019, 31, 1901478.
190. M. Yu, N. Chandrasekhar, R. K. M. Raghupathy, K. H. Ly, H. Zhang, E. Dmitrieva, C. Liang, X. Lu, T. D. Kühne, H. Mirhosseini, I. M. Weidinger and X. Feng, J. Am. Chem. Soc., 2020, 142, 19570–19578.
191. Z. Lei, Q. Yang, Y. Xu, S. Guo, W. Sun, H. Liu, L.-P. Lv, Y. Zhang and Y. Wang, Nat. Commun., 2018, 9, 576.
192. F. Wang, Z. Liu, C. Yang, H. Zhong, G. Nam, P. Zhang, R. Dong, Y. Wu, J. Cho, J. Zhang and X. Feng, Adv. Mater., 2020, 32, 1905361.
193. R. Shi, L. Liu, Y. Lu, C. Wang, Y. Li, L. Li, Z. Yan and J. Chen, Nat. Commun., 2020, 11, 178.
194. B. Pan, J. Huang, Z. Feng, L. Zeng, M. He, L. Zhang, J. T. Vaughey, M. J. Bedzyk, P. Fenter, Z. Zhang, A. K. Burrell and C. Liao, Adv. Energy Mater., 2016, 6, 1600140.
195. A. M. Khayum, M. Ghosh, V. Vijayakumar, A. Halder, M. Nurhuda, S. Kumar, M. Addicoat, S. Kurungot and R. Banerjee, Chem. Sci., 2019, 10, 8889–8894.
196. B. Pan, D. Zhou, J. Huang, L. Zhang, A. K. Burrell, J. T. Vaughey, Z. Zhang and C. Liao, J. Electrochem. Soc., 2016, 163, A580–A583.
197. J. Bitenc, N. Lindahl, A. Vizintin, M. E. Abdelhamid, R. Dominko and P. Johansson, Energy Storage Mater., 2020, 24, 379–383.
198. J. Bitenc, K. Pirnat, E. Žagar, A. Randon-Vitanova and R. Dominko, J. Power Sources, 2019, 430, 90–94.
199. J. Bitenc, K. Pirnat, T. Bančič, M. Gaberšček, B. Genorio, A. Randon-Vitanova and R. Dominko, ChemSusChem, 2015, 8, 4128–4132.
200. L. Ma, S. Chen, X. Li, A. Chen, B. Dong and C. Zhi, Angew. Chem., 2020, 132, 24044–24052.
201. Q. Wang, Y. Liu and P. Chen, J. Power Sources, 2020, 468, 228401.
202. S. Zhang, S. Long, H. Li and Q. Xu, Chem. Eng. J., 2020, 400, 125898.
203. H. Glatz, E. Lizundia, F. Pacifico and D. Kundu, ACS Appl. Energy Mater., 2019, 2, 1288–1294.
204. D. Lu, H. Liu, T. Huang, Z. Xu, L. Ma, P. Yang, P. Qiang, F. Zhang and D. Wu, J. Mater. Chem. A, 2018, 6, 17297–17302.
205. F. Guo, Z. Huang, M. Wang, W. L. Song, A. Lv, X. Han, J. Tu and S. Jiao, Energy Storage Mater., 2020, 33, 250–257.
206. Q. Wang, X. Xu, G. Yang, Y. Liu and X. Yao, Chem. Commun., 2020, 56, 11859–11862.
207. W. Wang, V. S. Kale, Z. Cao, S. Kandambeth, W. Zhang, J. Ming, P. T. Parvatkar, E. Abou-Hamad, O. Shekhah, L. Cavallo, M. Eddaoudi and H. N. Alshareef, ACS Energy Lett., 2020, 5, 2256–2264.
208. R. Sun, S. Hou, C. Luo, X. Ji, L. Wang, L. Mai and C. Wang, Nano Lett., 2020, 20, 3880–3888.
209. K. W. Nam, H. Kim, Y. Beldjoudi, T. Kwon, D. J. Kim and J. F. Stoddart, J. Am. Chem. Soc., 2020, 142, 2541–2548.
210. H. Lu, F. Ning, R. Jin, C. Teng, Y. Wang, K. Xi, D. Zhou and G. Xue, ChemSusChem, 2020, 13, 3447–3454.
211. B. Ji, H. He, W. Yao and Y. Tang, Adv. Mater., 2021, 33, 2005501.
212. R. Yang, F. Zhang, X. Lei, Y. Zheng, G. Zhao, Y. Tang and C. S. Lee, ACS Appl. Mater. Interfaces, 2020, 12, 47539–47547.
213. H. Wang, Y. Bai, S. Chen, X. Luo, C. Wu, F. Wu, J. Lu and K. Amine, ACS Appl. Mater. Interfaces, 2015, 7, 80–84.
214. F. Nacimiento, M. Cabello, R. Alcántara, C. Pérez-Vicente, P. Lavela and J. L. Tirado, J. Electrochem. Soc., 2018, 165, A2994–A2999.
215. H. Sun, W. Wang, Z. Yu, Y. Yuan, S. Wang and S. Jiao, Chem. Commun., 2015, 51, 11892–11895.
216. A. S. Childress, P. Parajuli, J. Zhu, R. Podila and A. M. Rao, Nano Energy, 2017, 39, 69–76.
217. K. V. Kravchyk, S. Wang, L. Piveteau and M. V. Kovalenko, Chem. Mater., 2017, 29, 4484–4492.
218. Z. Liu, J. Wang, H. Ding, S. Chen, X. Yu and B. Lu, ACS Nano, 2018, 12, 8456–8466.
219. C. J. Pan, C. Yuan, G. Zhu, Q. Zhang, C. J. Huang, M. C. Lin, M. Angell, B. J. Hwang, P. Kaghazchi and H. Dai, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 5670–5675.
220. Q. Zhang, L. Wang, J. Wang, C. Xing, J. Ge, L. Fan, Z. Liu, X. Lu, M. Wu, X. Yu, H. Zhang and B. Lu, Energy Storage Mater., 2018, 15, 361–367.
221. G. Gershinsky, H. D. Yoo, Y. Gofer and D. Aurbach, Langmuir, 2013, 29, 10964–10972.

---