Progress and Challenges of Vanadium Oxide Cathodes for Rechargeable Magnesium Batteries

Abstract

Among the challenges related to rechargeable magnesium batteries (RMBs) still not resolved are positive electrode materials with sufficient charge storage and rate capability as well as stability and raw material resources. Out of the materials proposed and studied so far, vanadium oxides stand out for these requirements, but significant further improvements are expected and required. They will be based on new materials and an improved understanding of their mode of operation. This report provides a critical review focused on this material, which is embedded in a brief overview on the general subject. It starts with the main strategic ways to design layered vanadium oxides cathodes for RMBs. Taking these examples in more detail, the typical issues and challenges often missed in broader overviews and reviews are discussed. In particular, issues related to the electrochemistry of intercalation processes in layered vanadium oxides; advantageous strategies for the development of vanadium oxide composite cathodes; their mechanism in aqueous, “wet”, and dry non-aqueous aprotic systems; and the possibility of co-intercalation processes involving protons and magnesium ions are considered. The perspectives for future development of vanadium oxide-based cathode materials are finally discussed and summarized.

1. Introduction

Magnesium ion batteries are among the promising power sources that can provide high energy density combined with increased environmental and operational safety at lower device costs mostly because of the high magnesium content in the Earth’s crust [1,2,3,4,5,6]. The Mg/Mg²⁺ system provides metallic anodes offering a high theoretical specific capacity (3834 mAh·cm⁻³), which is higher than that of metallic lithium (2061 mAh·cm⁻³). It also provides an attractively low electrode potential (−2.37 V) and forms metallic deposits without dendrites [1,4,5,6].

Due to these advantages, rechargeable magnesium ion batteries have attracted considerable and growing research interest in the field of post-lithium secondary batteries [7]. For comparison, Table 1 collects pertinent data on relevant negative electrode materials.

Systems with both non-aqueous and aqueous electrolytes are considered, in addition to the extended window of electrochemical stability available with water-in-salt electrolytes [8] has been discussed and extended into a quasi-solid state concept [9]. A first pouch cell demonstration has been reported [10].

Challenges for developing a practical magnesium full cell are related to all components, including in particular Mg²⁺-containing electrolytes that would remain electrochemically stable with both a Mg anode and cathodes. Nonaqueous rechargeable magnesium batteries also suffer from the complicated and moisture-sensitive electrolyte chemistry at magnesium electrode. Practical realization of a RMB is, in particular, handicapped by the absence of high-performance electrode materials due to the intrinsically slow Mg²⁺-ion diffusion in solids.

Table 1. Selected data of relevant negative electrode materials for metal ion batteries [11].

Element	Atomic Mass	E0, SHE/V	Gravimetric Capacity/mAh·g−1	Volumetric Capacity/mAh·cm−3
Li	6.94	−3.040	3860	2061
Na	23.0	−2.713	1165	1129
K	39.1	−2.924	685	610
Mg	24.31	−2.356	2206	3834
Ca	40.08	−2.840	1337	2072
Zn	65.41	−0.763	820	5855
Al	26.98	−1.676	2980	8046

Table 1. Selected data of relevant negative electrode materials for metal ion batteries [11].

Element	Atomic Mass	E0, SHE/V	Gravimetric Capacity/mAh·g−1	Volumetric Capacity/mAh·cm−3
Li	6.94	−3.040	3860	2061
Na	23.0	−2.713	1165	1129
K	39.1	−2.924	685	610
Mg	24.31	−2.356	2206	3834
Ca	40.08	−2.840	1337	2072
Zn	65.41	−0.763	820	5855
Al	26.98	−1.676	2980	8046

Critical characteristics, such as specific capacity, rate capability, and cycling stability of RMBs, are strongly affected by the intrinsic electrochemical properties of the cathode, i.e., the positive electrode (following an earlier recommendation by Huggins [12,13]), and, in particular, the materials they are made of. There is continuous progress in the improvements of their properties.

Several earlier reviews [14,15,16,17] as well as selective overviews [10,18,19,20,21,22,23] on RMBs are available. They address cathode-related issues like electrode reaction kinetics particularly relevant for rate capability and mass utilization [24] and detrimental effects of particular properties attributed to Mg²⁺-ions [25], highlight potential advantages of layered compounds [26], and provide mostly brief considerations and discussions of aspects concerning different intercalation-type cathode materials (like transition metal oxides, phosphates, chalcogenides, and Prussian blue analogues), but they are usually very short on considerations of their properties and mostly without an in-depth analysis of the reasons for the differences in functional properties (like specific capacity and power) and mechanism of intercalation processes regarding the latter class of materials. A focal issue is the high charge density of the magnesium ion with associated problems of strong interactions with a host material used in the positive electrode and associated slow diffusion. This also affects interfacial processes at the electrode/solution interface, and possibly even the electrolyte solution.

The number of investigations on different cathode materials for RMBs is rapidly increasing every year. Several reviews are published each year summarizing experimental progress on various cathode materials suitable for RMBs (transition metal oxides, sulfides, selenides, and other layered materials) [10,19,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40]. However, given the large number of material classes and research areas, they are discussed too superficially, without more detailed consideration of the existing challenges.

Some more specialized reviews discuss and analyze reaction mechanisms in cathodes structure–kinetics correlations, strategies for improving the electrode kinetics [24], the role of the electrolyte composition in the intercalation processes of Mg²⁺-ions [41], cathode materials and electrolytes for aqueous magnesium batteries [42], nanostructured cathode materials [43], and employing graphene and graphene-based materials to improve cathode materials [44].

Some reviews on RMBs have been dedicated to vanadium oxides solely. The structural characteristics and electrochemical performance of vanadium-based materials as RMB cathodes are discussed in [6,45,46]. The role of structural water molecules in vanadium oxide was analyzed in [47], the developments in application of vanadium oxide bronzes in metal ion batteries were discussed in [48], and the intercalation mechanisms in vanadium oxide were discussed in [49,50].

In this review, we will focus on the recent investigations of one of the promising and widely studied groups of cathode materials—layered vanadium oxides, including a systematic discussion of the achievements in their development and their possible improvement, the existing opinions on the mechanisms of charge–discharge processes in aqueous, “wet”, and dry non-aqueous electrolytes, and strategies to optimize functional properties of these cathode materials. We believe that this specialized and detailed review will help researchers to optimize the functional properties of these cathode materials.

2. Vanadium Oxide-Based Cathodes

Why did we select vanadium oxides as cathode materials for RMBs for this review?

Although the radius of Mg²⁺-ions (0.72 Å) is similar to that of Li⁺-ions (0.76 Å), the intercalation of Mg²⁺-ions into the same host materials is more difficult due to the difference in charge. The large charge/radius ratio of Mg²⁺-ions results in their strong electrostatic interactions with cathode materials, making intercalation and diffusion processes sluggish. As a result, most cathode materials suitable for RMBs exhibit a low degree of magnesiation, large voltage hysteresis, and low rate capabilities. Therefore, it is important to develop high-performance cathode materials for RMBs.

Vanadium-based materials have high theoretical capacity and energy density resulting from the multiple valence states of vanadium. Many vanadium-based materials have a layered structure, and the interlayer can be enlarged and adjusted to favor the intercalation and diffusion of Mg²⁺-ions into the cathode material.

Among vanadium-based materials, vanadium oxides are the most common and have numerous advantages for multivalent metal ion batteries. Vanadium oxides are cheap and abundant. Vanadium oxides with different structures have a variety of valence states (V⁺³/V⁺⁴/V⁺⁵) and can be easily synthesized by a variety of methods. Vanadium oxide polymorphs with different lattice symmetries have different electronic properties and thus different metal ion insertion thermodynamics and kinetics. Different synthetic strategies offer possibilities to synthesize different polymorphic structures and to fit the obtained layered structures to different interlayer distances and to weaken (i.e., screen) the strong electrostatic interaction of Mg²⁺-ions with the host lattice.

Among other cathode materials, the vanadium pentoxide family has been considered promising for RMBs due to the high theoretical capacity of V₂O₅ (294 mAh·g⁻¹, considering 1 mol of Mg²⁺ per mol of V₂O₅) and high working voltage of ≈2.4 V, resulting in a high specific energy (>600 Wh·kg⁻¹) [6,51,52,53]. Orthorhombic V₂O₅ was the first of the vanadium oxides to be investigated for Mg²⁺ intercalation.

In recent years, researchers have focused on other types of vanadium oxides like V₂O₅·nH₂O, M_xV₂O₅·nH₂O (with M: intercalated foreign metal ion) [54], VO₂ [55,56], and others, and also specially designed mixed-valence composites, such as V₃O₇/VO₂ [57].

In the thermodynamically most stable α-phase, V₂O₅ has a layered structure consisting of alternating edge- and corner-sharing distorted VO₅ trigonal-bipyramidal coordination polyhedrons stacked along the c-axis with an interlayer spacing of 4.37 Å (Figure 1a). Vanadium oxide gels (V₂O₅·nH₂O) have bilayer crystal structures comprised of layered square-pyramidal VO₅ polyhedra in which V⁵⁺ is coordinated with five oxygen atoms. Water molecules are bound between the layers of V₂O₅ (Figure 1b). The reported water content n in V₂O₅·nH₂O is usually about 1.6–2.0 moles per mole of V₂O₅ [58]. The layered structure provides channels for ion insertion and de-insertion (Figure 1b).

Brookite-phase vanadium dioxide VO₂(B) (Figure 1c) has a stable open framework. The crystal structure of VO₂(B) consists of corner- and edge-sharing VO₆ octahedra that provide three plausible migration channels along the c- and b-axes. It has a monoclinic symmetry with a space group of C2/m. Because the b-axis is the shortest unit cell dimension, crystals grow predominantly along the b-axis, forming one-dimensional morphologies such as nanowires, nanorods, and nanobelts. VO₂(B) is a promising candidate for RMBs due to its high theoretical capacity of 323 mAh·g⁻¹ per one electron transfer for the V⁴⁺/V³⁺ redox coupe [56].

Hydrated vanadium oxide (H₂V₃O₈, V₃O₇·H₂O, HVO) (Figure 1d) has an orthorhombic crystal structure and comprises V₃O₈ layers, consisting of edge- and corner-shared VO₅ square pyramids and VO₆ octahedra. In contrast to van der Waals forces in V₂O₅, vanadium oxide layers in HVO are linked by hydrogen bonds, which decrease the volume expansion/contraction during cycling.

The crystal structure of V₆O₁₃ (Figure 1e) is composed of corner- and edge-sharing distorted VO₆ octahedra, which form alternating single and double vanadium oxide layers connected by shared corners. V₆O₁₃ has high electrical conductivity at room temperature and has a high theoretical capacity of 417 mAh·g⁻¹ when high-valent vanadium (V⁵⁺/V⁴⁺) is all reduced to V³⁺.

The vanadium trioxide V₂O₃ (Figure 1f) possesses a tunnel-like 3D structure, which determines its good ion intercalation performance. The theoretical capacity of V₂O₃ is 357.5 mAh g⁻¹ and 715.1 mAh g⁻¹ for one- and two-electron redox processes.

Consideration of this class of insertion-type compounds allows to highlight the main strategic ways to design layered metal oxide cathodes and to consider, on these examples, the typical issues and mechanistic details often missing within broader overviews and reviews.

In particular, these issues should be focused on (i) the electrochemistry of intercalation processes in layered oxides (vanadium oxide V₂O₅ and its modifications), their mechanism in aqueous, wet, and dry non-aqueous aprotic systems (PC, AN), the possibility of co-intercalation processes involving protons and magnesium ions; and (ii) the dependence of electrochemical properties on the modification of the interlayer space due to the inclusion of foreign metal ions, organic molecules (polymeric and others), and water molecules.

The basis for the selection of a suitable inorganic host material in metal ion battery technology is the reversible electrochemical intercalation processes associated with electrochemical redox processes with the recharging of metal ions and the accompanying reversible processes of insertion/removal of electrolyte cations into/from the host electrode material (Figure 1g). The driving force of these processes is the gradient of electrochemical potentials realized in the host materials. The three main factors affecting the intercalation and diffusion of metal ions are (i) steric factors, depending on the geometric size of intercalated metal ions and crystal structures providing the interstitial voids along ion hopping trajectories, (ii) desolvation of ions at injection, and (iii) electrostatic and other interactions of intercalated ions with the host materials.

Several innovative approaches have been proposed for layered oxide cathodes to successfully overcome all these challenges in the case of Mg²⁺-intercalation. Among them is the strategy based on a significant increase in the interlayer spacing in layered materials by the synthesis of layered materials pre-intercalated/doped with foreign metal ions or organic molecules, which could facilitate Mg²⁺-intercalation and diffusion in the host material. Another important feature of the increased interlayer spacing is the possible intercalation of solvated magnesium ions, avoiding the sluggish desolvation process at the interface.

Additional attention is given to the retention of water molecules in the host structures and the co-insertion of water molecules in the case of aqueous or wet non-aqueous electrolytes. The degree of reversible Mg²⁺-insertion strongly depends on the “lubrication” of the layered channels. The selected strategic approaches are usually combined with more common measures, such as nanostructuring of materials, which results in shortening the solid state diffusion path for highly polarized Mg²⁺ in the host and increasing the surface area of the active grains. In the following, we discuss these issues in more detail, based on experimental examples from published original works.

2.1. Orthorhombic Vanadium Pentoxide V₂O₅

Orthorhombic vanadium pentoxide (α-V₂O₅) was the first compound in the family of vanadium oxides to be proposed for use as cathode in RMBs. Early works showed that its typical layered crystal structure with weak van der Waals forces is suitable for tuning the interlayer space, making it accessible for intercalation of Mg²⁺-ions [29,59,60,61].

The electrochemical intercalation of Mg²⁺-ions into orthorhombic V₂O₅ in a solid state RMB composed of a metallic Mg anode, a solid Mg-montmorillonite electrolyte, and a V₂O₅ cathode has been confirmed [59]. The quantitative estimation of magnesium intercalation was reported [60], where the authors reported that 0.66 mol of Mg can be intercalated into 1 mol of V₂O₅ by chemical intercalation in the dibutylmagnesium/heptane solution, which corresponds to a high specific capacity of 194 mAh·g⁻¹. This value is consistent with the capacity determined electrochemically in a 1 M Mg(ClO₄)₂/THF electrolyte.

A stronger indication that V₂O₅ is a potential cathode material for RMBs was the first publication on the ability of α-V₂O₅ to reversibly electrochemically intercalate Mg²⁺ in 1 M Mg(ClO₄)₂/AN electrolytes with water added to AN [61]. It was shown that the insertion of Mg²⁺ into V₂O₅ depends on the ratio between the amounts of H₂O and Mg²⁺ and on the absolute amount of H₂O in the electrolyte solution. The preferential solvation of Mg²⁺-ions by water molecules facilitated the insertion process. The highest capacities of up to 170 Ah·kg⁻¹ were obtained in acetonitrile solutions containing 1 M Mg(ClO₄)₂ + 1 M H₂O; however, the electrode had low cycling stability—only about 50 Ah·kg⁻¹ was retained after 20 cycles.

To optimize the diffusion pathways of Mg²⁺ in V₂O₅ during the charge–discharge, orthorhombic V₂O₅ with particle size in the range of 20 to 50 nm obtained via a combustion flame-chemical vapor condensation process was suggested [62]. The reversible capacity of the V₂O₅ in 0.5 M Mg(ClO₄)₂/PC was 180 mAh·g⁻¹ at a current density 7.6 mA·g⁻¹.

During the discharge process, Mg²⁺-ions are gradually inserted into the interlayer of V₂O₅ and become coordinated with the VO₅ pyramids, and the oxidation state of vanadium decreases from +5 to +4. The V⁴⁺/V⁵⁺ redox pair endows V₂O₅ with a high potential of 2.66 V vs. Mg/Mg²⁺. The overall reaction of the electrochemical insertion of Mg²⁺ ions into V₂O₅ as a RMB cathode material can be written as

$${\mathrm{V}}_2{\mathrm{O}}_5{+\mathrm{x}\mathrm{Mg}}^{2+}{+2\mathrm{x}\mathrm{e}}^{-}\underset{}{\rightleftarrows }{\mathrm{Mg}}_{\mathrm{x}}{\mathrm{V}}_2{\mathrm{O}}_5$$

(1)

V₂O₅ has a high theoretical specific capacity of 294 mAh·g⁻¹ when 1 mol of Mg²⁺ ion is inserted (x = 1) [28].

V₂O₅ forms metastable polymorphs (β-, γ-, δ-, ε-, and ζ-V₂O₅) in addition to the thermodynamically stable phase (α-V₂O₅). Some of them have been the subject of theoretical and experimental studies concerning RMBs.

First-principles calculations were performed to investigate Mg intercalation in the α- and δ-polymorphs of V₂O₅ [63]. It was supposed that the conversion of fully demagnesiated stable α-V₂O₅ into a two-phase structure consisting of fully magnesiated δ-V₂O₅ and fully demagnesiated α-V₂O₅ phases during the discharge process results in slow diffusion kinetics in V₂O₅ electrodes. It was shown that since the calculated α-phase migration barriers indicate poor Mg mobility, reversible Mg intercalation is only possible at very low rates and in small particles, and the δ-V₂O₅ polymorph exhibits superior Mg mobility, provided that the δ-V₂O₅ host structure can remain stable or metastable over a wide Mg concentration range. The first-principles calculations have also shown that the Mg²⁺-ion diffusion barrier in δ-V₂O₅ (~0.6–0.8 eV) is lower than that in α-V₂O₅ (~0.975–1.1 eV) [64].

The DFT calculations have also shown that V₂O₅ undergoes a structural transformation from the α-phase to the ε-phase (ε-Mg_0.5V₂O₅) and δ-phase (δ-MgV₂O₅) as the concentration of intercalated Mg²⁺-ions is increased [65].

The DTF calculations have shown a significant reduction in barriers for Mg diffusion in the β-V₂O₅ phase (0.65 eV) compared to α-V₂O₅ (>1 eV), leading to possibly fast charge–discharge rates of β-V₂O₅ as a cathode material for Mg²⁺-ion batteries [66]. β-V₂O₅ synthesized at high pressure demonstrated large voltage hysteresis in a dry (<20 ppm of H₂O) 0.1 M Mg(ClO₄)₂/AN electrolyte and a maximum discharge capacity of 361 mAh·g⁻¹ in GCD tests at C/25 [67].

The migration barriers for Mg²⁺ in ζ-V₂O₅ calculated by DFT were 0.62–0.86 eV [68], also suggesting its applicability in RMB cathodes. The metastable ζ-V₂O₅ obtained by topochemical extraction of the Ag⁺ ion from ζ-Ag_0.33V₂O₅ also demonstrated large voltage hysteresis in Mg(TFSI)₂/AN electrolyte, and displayed a specific capacity of 90 mAh g⁻¹ with insertion of up to 0.33 Mg²⁺ per formula unit [69]. Nanosized ζ-V₂O₅ had a discharge capacity of 130 mAh·g⁻¹ and lower voltage hysteresis [70].

The experimental investigations also indicated possible multi-stage transformation during the Mg²⁺-intercalation. Mg ion intercalation into nanoscale films of V₂O₅ deposited on Pt was studied [71]. It was shown that highly reversible Mg insertion/de-insertion is possible within V₂O₅ thin films. In the 0.5 M Mg(ClO₄)₂/AN electrolyte, the V₂O₅ thin film was cycled over a potential range of 2.2–3.0 V vs. Mg/Mg²⁺, and had an initial capacity of 180 mAh·g⁻¹ at a 0.15 mV s⁻¹ scan rate, the capacity retention after 36 cycles was ~83%. The electrodes had 100% coulombic efficiency. The differential capacity plots (dQ/dV − V) obtained for low current density galvanostatic cycling revealed four different Mg²⁺-insertion stages or processes with different thermodynamic and kinetic characteristics.

In order to check the inherent ability of α-V₂O₅ to intercalate Mg²⁺, electrochemical tests were conducted in an ionic liquid electrolyte in [72]. The α-V₂O₅ cathode prepared from commercial vanadium oxide powder had a low reversible capacity of only 16 mAh·g⁻¹ at 25 °C, which increased to 295 mAh·g⁻¹ (at C/5) at 110 °C. This capacity corresponds to reversible intercalation of 1 mol Mg²⁺ per unit formula. After 1 cycle at 110 °C, the electrochemical activity of α-V₂O₅ at room temperature was significantly enhanced (specific capacity of about 95 mAh·g⁻¹ at C/5). The composition of the cathode after the thermal activation was Mg_0.2V₂O₅; accordingly, the activation resulted in an expanded interlayer spacing with Mg-pillaring, facilitating fast Mg²⁺-migration. The reversible intercalation of Mg²⁺ into α-V₂O₅ was also observed during the characterization of the material composition, crystal structure, and redox state.

Thin films of orthorhombic V₂O₅ were grown on fluorine-doped tin oxide (FTO) glass electrodes using AACVD, and studied in aqueous solutions of MgCl₂ [73]. The material showed a specific discharge capacity as high as 427 mAh·g⁻¹ at a high current density of 5.9 A·g⁻¹ with 82% capacity retention after 2000 cycles. At a current density of 2.4 A·g⁻¹, the specific capacity was 170 mAh·g⁻¹, and when the current was returned to 5.9 A·g⁻¹, the cathode delivered 425 mAh·g⁻¹, corresponding to a capacity retention of 99%. The increase in the specific discharge capacity with increasing specific current was explained by the fast diffusion kinetics of Mg²⁺ in the metal oxide framework.

Several recent studies of V₂O₅ nanoclusters have shown that the use of vanadium-based nanoscale materials can improve the diffusion of Mg²⁺-ions and provide high reversibility of intercalation processes. Highly dispersed vanadium oxide nanoclusters supported on porous carbon frameworks were synthesized using the ambient hydrolysis deposition method in order to improve the electrical conductivity of V₂O₅ [74]. The composite with 45 wt.% V₂O₅ had a surface area and pore volume of 593 m²·g⁻¹ and 2.8 cm³·g⁻¹, respectively. Reversible intercalation of Mg²⁺ with an initial capacity of 350 mAh·g⁻¹ (per V₂O₅) or 180 mAh·g⁻¹ (per composite weight) at 40 mA·g⁻¹ within the voltage range of 0.5–2.8 V was observed in 0.2 M [Mg2(µ-Cl)₂(DME)₄][AlCl₄]₂ in DME. During the subsequent five cycles, the capacity of the composite electrode decreased to 160 mAh·g⁻¹ (per V₂O₅) or 90 mAh·g⁻¹ (per composite).

2.2. Nanostructured Vanadium Oxides

The findings confirming slow diffusion kinetics in V₂O₅ and low capacities pointed at the necessity of modifying the structure of V₂O₅ for Mg battery electrode applications.

In most papers it was noted that strong electrostatic interactions between divalent Mg²⁺-ions and lattice oxygen cause slow diffusion of Mg²⁺ in α-V₂O₅, and this is one of the main reasons for the unsatisfactory magnesium storage properties. Nanostructuring can increase the active surface area and reduce the diffusion distance to improve the electrochemical performance of α-V₂O₅. Various nanostructured vanadium oxide cathode materials with superior electrochemical performance were synthesized.

Hydrothermally synthesized V₂O₅ nanowires (Figure 2a) were used as a cathode in a magnesium ion system with 1 M Mg(ClO₄)₂/AN as an electrolyte and Mg_xMo₆S₈ (x ≈ 2) as an anode [75]. V₂O₅ nanowires provided an initial discharge/charge capacity of 103/110 mAh·g⁻¹ and a maximum discharge capacity of 130 mAh·g⁻¹ in the sixth cycle at a C/20 rate in a full cell.

The synchrotron diffraction pattern of the pristine V₂O₅ nanowires (Figure 2b) showed a high degree of crystallinity in the material. All reflections were indexed in the orthorhombic α-V₂O₅ structure model with space group Pmn21 and lattice parameters a = 11.511 Å, b = 4.373 Å and c = 3.565 Å. The mechanism of Mg²⁺ insertion and extraction in V₂O₅ during the first two cycles was studied by in operando synchrotron diffraction (Figure 2c). It was shown that during the first discharge, the V₂O₅ phase accommodates Mg²⁺-ions via a solid solution mechanism up to the stoichiometry of Mg_0.14V₂O₅, while the lattice parameters a and c increased and b decreased. During further magnesium uptake, a decrease in the amount of Mg_0.14V₂O₅ phase and an increase in the Mg_0.6V₂O₅ phase with constant cell parameters for both phases were observed; at the end of the first discharge, the phase ratio of Mg_0.14V₂O₅:Mg_0.6V₂O₅ was 13:87. In the second cycle, the material showed almost the same behavior during discharge, and at the end of the second charge V₂O₅ returned to its original structure.

Flower-like V₂O₅ microspheres composed of 25 nm thick nanosheets (Figure 2d) were synthesized via a surfactant-assisted hydrothermal procedure in [76]. The cathode produced from V₂O₅ microspheres with an average diameter of 3 μm delivered the initial discharge capacity of 126.2 mAh·g⁻¹ at 50 mA·g⁻¹ in 0.25 M Mg(AlCl₂EtBu)₂/THF, demonstrated good cycling stability (90.7 mAh g⁻¹ after 80 cycles) and enhanced rate capability (60 mAh·g⁻¹ at 200 mA·g⁻¹). The improved electrochemical performance of the V₂O₅ microflowers was explained by the increased specific surface area (13.7 m²·g⁻¹) and flexibility. The microspheres with larger size (7 and 15 μm) or structural irregularities demonstrated lower capacities and cycling stability (Figure 2e). The charge–discharge mechanism was investigated by X-ray diffraction and X-ray photoelectron spectroscopy.

Monodisperse spherical V₂O₅ particles with a diameter of 230–250 nm were obtained in [77] (Figure 2f). The V₂O₅ spheres exhibited an initial discharge capacity of 225 mAh·g⁻¹ (Figure 2g) and a stabile discharge capacity of 190 mAh·g⁻¹ at 10 mA·g⁻¹ in a dry Mg(ClO₄)₂/AN electrolyte. The rate performance of V₂O₅ spheres (Figure 2h) and long-term cycling stability at different current rates were good, and the specific discharge capacity of 55 mAh·g⁻¹ was achieved at 50 mA·g⁻¹ with 5% and 13% capacity fading after 50 and 100 cycles, respectively. The retention of 95% coulombic efficiency after 100 cycles and the stability of the phase and morphology confirmed the stability of the material during Mg²⁺-ion intercalation/deintercalation.

V₅O₁₂·6H₂O nanoflowers formed by self-assembly of nanosheets were obtained by a one-step solvothermal method (Figure 3a) in [78]. The synthesized V₅O₁₂·6H₂O had uniform morphology with an average size of a nanoflower of 650 nm while the individual nanosheets were about 25 nm thick (Figure 3b). For comparison, V₅O₁₂·6H₂O nanoparticles were also synthesized. The galvanostatic tests revealed that both V₅O₁₂·6H₂O nanoflowers and nanoparticles showed a similar activation process, the nanoparticles having a much lower capacity (Figure 3c). It was shown that the structural water in the interlayer of V₅O₁₂·6H₂O improved the stability of the crystal structure and created more active sites for Mg²⁺ storage. The nanoflower morphology increased the active surface area, and enhanced the contact between the electrode material and the electrolyte, improving the diffusion kinetics of Mg²⁺. The V₅O₁₂·6H₂O nanoflowers had a high initial specific capacity of 234.3 mAh·g⁻¹ and reversible capacity after activation of 153.2 mAh·g⁻¹ at 10 mA·g⁻¹ (in a 1:1 mixture of 0.8M PhMgCl and 0.4M AlCl₃ in anhydrous THF electrolyte) (Figure 3d), a good rate capability in the current range 50–500 mA·g⁻¹ (Figure 3e), and long cycling stability (73.8% capacity retention after 1500 cycles at 100 mA·g⁻¹, i.e., 0.017% capacity loss per cycle). It was proposed that the layered structure of the V₅O₁₂·6H₂O nanoflowers can accommodate the volume changes during the intercalation/deintercalation of Mg²⁺, and the interlayer water molecules can serve as pillars enhancing the structural stability.

Ultrathin 2D V₆O₁₃ nanosheets were synthesized via a microwave-assisted method [79]. Two pairs of redox peaks observed in CVs at 0.9/0.68 V and 1.15/1.33 V (vs. Mg/Mg²⁺) correspond to the magnesiation/demagnesiation process, suggesting that it is a two-step process. The V₆O₁₃ nanosheet cathode delivered a maximum discharge specific capacity of 324 mAh·g⁻¹ at 20 mA·g⁻¹ in 1 M Mg(ClO₄)₂/AN, corresponding to 3.22 mol of Mg²⁺ ions’ insertion per unit formula of V₆O₁₃. At current densities of 40, 60, and 80 mA·g⁻¹, the reversible specific capacities were 278, 244, and 214 mAh·g⁻¹, respectively, indicating the good rate capability of the material. A high reversible capacity of 200 mAh·g⁻¹ was retained after 30 cycles at 40 mA·g⁻¹ with a ~100% coulombic efficiency.

Vanadium oxide nanotubes are advantageous for Mg²⁺-intercalation due to their large interlayer distance, open ends, and large inner and outer diameters. The shape of the tubes is favorable for an ion insertion process because the diffusion path in the solid is shorter and the heterogeneous kinetics are faster at higher surface-to-bulk ratios. In addition, the tubes can provide electrolyte-filled channels or voids for fast Mg²⁺-ion transportation to the insertion sites.

VO_x nanotubes (with the nominal composition of VO_2.37[C₁₈H₄₀N]_0.26) were obtained from V₂O₅ and octadecylamine through a hydrothermal reaction [80]. The nanotubes had open ends and were 1–3 μm long, and their outer/inner diameter was 60–100 nm/15–40 nm, respectively. In an 0.25 M Mg(AlBu₂Cl₂)₂/THF electrolyte, the material exhibited good cycle performance, and the initial specific discharge capacity was 81 mAh·g⁻¹ at 5 mA·g⁻¹. The Mg²⁺-insertion into the VO_x nanotubes was confirmed by the XPS and XPD results.

VO_x nanotubes were obtained using a microwave-assisted hydrothermal method and low and high concentrations of the amine template, denoted as HT and LT, respectively [81]. The amount of the amine template did not affect the morphology of the VO_x nanotubes; both types of nanotube had an average outer diameter of 80–100 nm and were 1–5 μm long. The performance of VO_x nanotubes with various oxidation states (V³⁺/V⁴⁺/V⁵⁺) as a cathode for RMB was studied in 0.5 M Mg(ClO₄)₂/AN solution. The kinetics of magnesium insertion and extraction depended highly on the oxidation state and bonding structure of the nanotubes’ surface. The formation of V³⁺-ions in highly reduced VO_x nanotubes resulted in high initial discharge capacity (218 mAh·g⁻¹ for HT-VO_x and 230 mAh·g⁻¹ for LT-VO_x at 0.2 C (60 mA·g⁻¹)), excellent cycling performance, and lower charge transfer resistance at the electrode/electrolyte interface compared to VO_x nanotubes containing vanadium ions of higher oxidation states. Although the initial capacities of HT-VO_x were lower than those of LT-VO_x, HT-VO_x demonstrated better cycling performance, with capacity retention of 70.8% after 20 cycles. The coulombic efficiency of both materials increased while cycling.

The Mg²⁺-ion insertion/extraction performance of VO_x-NTs was investigated in Mg(ClO₄)₂ tetramethylsilane-ethyl acetate (TMS-EA) electrolyte in comparison with an AN electrolyte [82]. The VO_x-NTs were prepared using a microwave-assisted hydrothermal method starting with V₂O₅ and octadecylamine. As seen in the SEM and TEM images (Figure 3f,g), the VO_x-NTs were open-ended, multilayered tubular structures, with alternating arrangements of VO_x and amine layers. VO_x-NTs showed an initial capacity of more than 200 mAh·g⁻¹ in de-aerated 0.5 M Mg(ClO₄)₂/AN electrolyte solution, but had poor capacity retention. Superior cycling performance and coulombic efficiency were observed for the VO_x-NTs in TMS and TMS-EA, although the initial discharge capacities in TMS-based electrolytes were lower (Figure 3h). It was shown that the TMS-EA solvent improved the cell performance due to the higher stability of TMS against oxidation and the strong Mg²⁺ coordination ability of EA. The initial discharge capacity of VO_x-NT in TMS-EA solution was 124 mAh·g⁻¹ at 0.2 C (60 mA·g⁻¹), and the capacity retention was 70% after 80 cycles.

Figure 3. (a) Scheme of the preparation of V₅O₁₂·6H₂O nanoflowers, (b) SEM of a V₅O₁₂·6H₂O nanoflower, (c) cycling performance of V₅O₁₂·6H₂O nanoflowers and nanoparticles at 50 mA·g⁻¹, (d) discharge/charge profiles of V₅O₁₂·6H₂O nanoflowers, (e) rate capability of V₅O₁₂·6H₂O nanoflowers and nanoparticles. (Reprinted with permission from [78]. Copyright 2023, Elsevier). (f,g) High-resolution and (inset) low-resolution SEM and TEM images of VO_x-NTs, (h) discharge capacity and coulombic efficiency of VO_x-NT electrodes in AN, TMS, and TMS-EA electrolytes. (Reprinted with permission from [82]. Copyright 2016, American Chemical Society).

Figure 3. (a) Scheme of the preparation of V₅O₁₂·6H₂O nanoflowers, (b) SEM of a V₅O₁₂·6H₂O nanoflower, (c) cycling performance of V₅O₁₂·6H₂O nanoflowers and nanoparticles at 50 mA·g⁻¹, (d) discharge/charge profiles of V₅O₁₂·6H₂O nanoflowers, (e) rate capability of V₅O₁₂·6H₂O nanoflowers and nanoparticles. (Reprinted with permission from [78]. Copyright 2023, Elsevier). (f,g) High-resolution and (inset) low-resolution SEM and TEM images of VO_x-NTs, (h) discharge capacity and coulombic efficiency of VO_x-NT electrodes in AN, TMS, and TMS-EA electrolytes. (Reprinted with permission from [82]. Copyright 2016, American Chemical Society).

The electrochemical Mg²⁺-ion intercalation and extraction in multi-walled vanadium oxide nanotubes VO_x-NTs was investigated in [83]. The VO_x-NTs had an open-ended structure of multilayer scrolls with an outer diameter of 50 nm and a length of a few µm, and the VO_x interlayer spacing was 2.7 nm. The initial specific discharge capacity for VO_x-NTs in 1 M Mg(ClO₄)₂/AN and in 1 M Mg(TFSI)₂/G2 was 146 ± 36 and 33 ± 12 mAh·g⁻¹ at 5 mA·g⁻¹, respectively, since, as a rule, the cells with the Mg(TFSI)₂/G2 electrolyte provide lower initial capacities compared to Mg(ClO₄)₂/AN. In both electrolytes, the VO_x-NTs demonstrated poor capacity retention. The intercalation of Mg²⁺-ions was confirmed by ICP-AES. In operando scattering studies have shown that Mg²⁺-ion intercalation into VOx-NTs is accompanied by an increase in the interlayer spacing and a loss of the long-range stacking order.

2.3. Bilayer V₂O₅

V₂O₅ gels can be synthesized by a sol–gel process from a variety of precursors. Xerogels and aerogels differ in the drying process that removes the solvent from the nanoscale pores and capillaries after the material is synthesized. Xerogels are dried at ambient conditions, resulting in a more compact and dense material than aerogels. Aerogels are dried by filling the pores and capillaries with a supercritical fluid and then cooling so that the pores are filled with a gas. This process allows the pores to remain intact and increases the surface area of the material. V₂O₅ aerogel has a surface area ~30 times larger than that of V₂O₅ xerogel [27].

V₂O₅ gels have an enhanced capacity for insertion/extraction of magnesium ions due to their water content. Furthermore, the increased surface area supports more efficient diffusion and provides more sites for insertion/extraction of magnesium ions, thus increasing the ability of the material to enable high-rate charge and discharge. In addition, the capacity increases due to the large surface area because there is less isolated bulk material that the intercalating ions cannot reach.

Amorphous V₂O₅ aerogels with high surface area, consisting of interpenetrating networks of water and V₂O₅ ribbons, are capable of accommodating 0.6–2 moles of Mg²⁺ per mole of V₂O₅ as a result of chemical insertion [84], or two moles formally corresponding to the reduction of V⁵⁺ to the V³⁺ state [85].

Bilayer V₂O₅·nH₂O is generally considered to be amorphous due to the lack of long-range structural order, but at the nanoscale, a well-organized repetition of bilayers is observed. The synthesis route affects the morphology of V₂O₅·nH₂O. The reduced structural order, wide interlayer space, and short diffusion length in V₂O₅·nH₂O allow reversible cation accommodation [86]. The interlayer structural water can increase the interlayer spacing, which is favorable for the intercalation of Mg²⁺. The structural water molecules can partially shield the charge of Mg²⁺, which leads to the improvement in the diffusion kinetics of Mg²⁺ in bilayer V₂O₅ [86].

The insertion of Mg²⁺ into V₂O₅ xerogel and V₂O₅ xerogel/C composite in aqueous Mg(NO₃)₂ electrolytes was studied [87]. The intercalation/deintercalation reaction of Mg²⁺- ions was much faster in the case of the V₂O₅ xerogel/C composite as compared to the pure V₂O₅ xerogel. The addition of 10% carbon black during the synthesis of V₂O₅ xerogel/C composite allowed a significant improvement in the electrochemical behavior of this electrode material; the specific capacity of V₂O₅ xerogel/C composite in Mg(NO₃)₂ was 107 mAh·g⁻¹, while pure V₂O₅ xerogel yielded 50 mAh·g⁻¹.

V₂O₅ xerogel was prepared under microwave (MW) irradiation [88]. The structure and electrochemical properties of the V₂O₅ xerogel were compared with those of V₂O₅ prepared by conventional heat treatment. XRD revealed that the V₂O₅ xerogel prepared by MW irradiation was low-crystalline. The first discharge capacity of V₂O₅ prepared by MW irradiation was 175 mAh·g⁻¹ at a 0.1 C rate in 0.3 M Mg(ClO₄)₂/PC solution, and the second discharge capacity increased to 463 mAh·g⁻¹, indicating that Mg²⁺-insertion increased after the first cycle. The capacity of V₂O₅ prepared by heat treatment at 200 °C was 138 and 190 mAh·g⁻¹ for the first and the second cycles, respectively, and the capacity of V₂O₅ prepared by heat treatment at 300 °C was 77 mAh·g⁻¹.

An amorphous V₂O₅ xerogel/graphite composite containing 10 wt.% synthetic graphite showed an initial deintercalation capacity of 77 mAh·g⁻¹ at 10 mV·s⁻¹ (~18 C rate) in Mg(NO₃)₂ aqueous electrolyte [89]. After ten cycles, the value decreased to 63.5 mAh·g⁻¹. In GCD tests, a relatively high magnesium storage capacity of 62, 53, 47, and 44 mAh·g⁻¹ was obtained at 2, 3, 4, and 5 A·g⁻¹, respectively.

The scenario of Mg²⁺ and H₂O co-intercalation in nanocrystalline V₂O₅ xerogel was analyzed by first-principles calculations [90]. The models of fully relaxed structures of the fully magnesiated and demagnesiated V₂O₅ xerogel, where two individual V₂O₅ layers are bound by long interlayer V–O bonds and the intercalated Mg atoms and H₂O molecules are located in the space between two bilayers, are shown in Figure 4a,b.

Analysis of the stable phases of Mg-intercalated V₂O₅ xerogel at different voltages and in electrolytes with different water contents revealed a range of concentrations of intercalated Mg-ions in the V₂O₅ xerogel and H₂O in the electrolyte where there is no thermodynamic driving force for the water molecules to shuttle with Mg²⁺-ions during electrochemical cycling (Figure 4c,d). It was also demonstrated that shuttling of water molecules with the Mg²⁺-ions in wet electrolytes yields higher voltages than in dry electrolytes.

Highly hydrated nanoribbons of bilayer V₂O₅ with a large interlayer distance (13.5 Å) were electrodeposited on porous highly conductive carbon nanofoam [91]. This method also allowed effective incorporation of defects, water, and hydroxyl groups, which in turn promoted electron transfer and (de)intercalation of highly charged Mg²⁺ ions. The “as-prepared” individual V₂O₅ nanoribbon is shown in an HRTEM image in Figure 4e. Mg²⁺ ions were incorporated into the as-prepared bilayer V₂O₅ cathodes by galvanostatic discharge at 20 μA to the potential 0.2 V (vs. Mg/Mg²⁺). The specific capacity of 240 mAh·g⁻¹ obtained after this preconditioning procedure suggests reduction of vanadium V⁵⁺ to V⁴⁺. The electrode demonstrated reversible Mg²⁺ (de)intercalation in 1 M Mg(ClO₄)₂/AN electrolyte with a capacity of 150 mAh·g⁻¹ at C/15 rate. The O 1s XPS spectra before cycling and after 10 cycles of charging showed that strongly bound structural hydroxyl groups remain in the structure of bilayer V₂O₅ during cycling (Figure 4f,g), and act as a lubricant for reversible (de)intercalation of solvated Mg²⁺ ions between the layers of V₂O₅. Molecular dynamics simulations of a three-bilayer V₂O₅ model, shown in Figure 4h,i, suggested that upon insertion of Mg²⁺-ions into the V₂O₅ structures, the spacing between bilayers decreases from ~12 to ~11 Å due to the interaction of Mg²⁺-ions with bilayer apical oxygen and structural hydroxyl groups, but a significant amount of H₂O molecules remains in the structure, solvating the Mg²⁺ ions. The HRTEM image (Figure 4j) showed that after Mg²⁺-ions were intercalated, the layered V₂O₅ structure featured a spacing of 6.0 to 6.4 Å, and the bilayer distance was not uniform across the whole imaged area.

The reversible insertion/extraction of Mg²⁺ into the V₂O₅·nH₂O electrode accompanied by the co-intercalation of solvent molecules was demonstrated [92]. The V₂O₅·nH₂O achieved a discharge capacity of 50 mAh·g⁻¹ (0.25 Mg per V₂O₅·nH₂O) in a Mg(TFSI)₂/G2 (diglyme) electrolyte at a current density of 20 mA·cm⁻². It was shown that the process of Mg²⁺ intercalation into V₂O₅·nH₂O involves the formation of new phases, while the bilayer spacing expands/contracts. The large interlayer spacing of bilayer V₂O₅ allowed the reversible co-intercalation of Mg²⁺ and solvent molecules.

Bilayered nanostructured V₂O₅·nH₂O 2D nanopapers with oxygen defects (BL-HVOd NPS) with improved Mg²⁺ insertion/extraction kinetics were obtained [93]. The interlayer water molecules effectively stabilized the expanded interlayer spacing (10.6 Å) in BL-HVOd NPS and shielded the electrostatic interaction between Mg²⁺-ions and BL-HVOd NPS lattice, improving diffusion kinetics during repeated cycling (the log(D_Mg²⁺) value was in the range from −8.3 to −9.5 during charge and discharge). The nanopaper structure of BL-HVOd NPS enhanced electrolyte/electrode contact and reduced the diffusion path of Mg²⁺-ions, improving the rate performance. Cells with a BL-HVOd NPS cathode, AC anode, and 2 M Mg(CF₃SO₃)₂ aqueous electrolyte demonstrated a reversible capacity of 162.8 mAh·g⁻¹ at 0.2 A·g⁻¹ and high cyclic stability with 88% capacity retention after 2000 cycles at 10 A·g⁻¹. This was attributed to the contribution of oxygen defects.

Bilayer V₂O₅ is a promising cathode material for RMBs, but the low electronic conductivity of bilayer V₂O₅ is still a challenge that hinders its further application in RMBs. The main strategy to solve this problem is the modification of V₂O₅ with highly conductive materials or the pre-intercalation of metal ions to improve the conductivity of bilayer V₂O₅ and promote the ionic and electronic transport.

2.4. Nanocomposites of V₂O₅ with Conductive Carbons

The use of conductive carbons as additives for the improvement in electrical conductivity of V₂O₅ has also been studied in a large number of works.

Mg²⁺ ion intercalation into a V₂O₅/carbon composite, consisting of aggregated carbon particles (30–100 nm) covered with a thin layer (<100 nm) of V₂O₅ xerogel, in 1 M Mg(ClO₄)₂/AN electrolyte, was reported in [94,95]. Mg²⁺ insertion into V₂O₅ xerogel/C composites and accompanying structural changes were studied [94]. In the cyclic voltammogram for the V₂O₅ xerogel/C composite in 1 M Mg(ClO₄)₂/AN, two broad cathodic peaks associated with the Mg²⁺-insertion process were observed, pointing at the existence of two different Mg intercalation sites in the V₂O₅ xerogel. Ex situ FT-IR spectra suggested that Mg²⁺ is inserted into inner layer sites of V₂O₅ at −0.15 V and into the interlayer sites at −0.65 V vs. Ag/Ag⁺. The reversibility of Mg²⁺ intercalation/deintercalation was confirmed by FT-IR and XRD. The V₂O₅/carbon composite showed higher capacity than a conventional V₂O₅ electrode. It was assumed that the Mg²⁺ diffusion into the interlayer is slow, but since the V₂O₅/C composite has a large interlayer distance and short diffusion length compared to the normal xerogel of V₂O₅, the interlayer site could be utilized more effectively for Mg²⁺-insertion.

With a V₂O₅ gel/carbon composite, insertion of 1.84 mol Mg per 1 mol V₂O₅ in the first cycle at a scan rate of 0.1 mV·s⁻¹ was found, resulting in a specific capacity of 540 mAh·g⁻¹ (per V₂O₅ mass) [95]. In GCD tests, at a current density of 1.0 A·g⁻¹ the composite electrode showed a capacity of about 600 mAh·g⁻¹, and at a high current density of 20 A·g⁻¹ the capacity of about 300 mAh·g⁻¹ was maintained. The capacity of the carbon-free V₂O₅ xerogel electrode at 1.0 A·g⁻¹ was only 150 mAh·g⁻¹, and at 20 A·g⁻¹ it was less than 10% of the capacity of the composite electrode. Thus, homogeneous mixing of a V₂O₅ sol with carbon particles, resulting in the formation of a thin layer of V₂O₅ gel around the carbon particles, was effective in improving the Mg²⁺ intercalation behavior. The reduction in the Mg²⁺ diffusion length in V₂O₅ and the improvement in the electronic conductivity by the carbon particles resulted in high capacities, especially at high current densities. The composites of vanadium oxide with graphene are expected to advance the RMB performance by providing rapid electron transport and easy Mg²⁺-diffusion due to the large surface area and excellent structural stability.

The synthesis of V₂O₅/graphene (GNP) nanoparticles by a ball milling technique was attempted in [96] to enhance the electrochemical performance of V₂O₅. The full cell with a Mg anode, (V₂O₅)_1-x/GNP_x cathode, and MgNO₃∙6H₂O/tetraethylene glycol dimethyl ether electrolyte yielded an initial capacity of ~90 mAh·g⁻¹ at a low current density, while Mg/V₂O₅ cells exhibited the initial discharge capacity of ~100 mAh·g⁻¹ under the same testing conditions. The capacities dropped to half the initial value on the second cycle and then the cells short-circuited, probably due to the incompatible electrolyte.

A composite of V₂O₅ with graphene oxide (GO/V₂O₅) was prepared by a solvothermal method [97]. SEM revealed that the V₂O₅ microparticles were wrapped around GO sheets, and TEM micrographs demonstrated the tight contact between GO and V₂O₅ microparticles. The good contact and dispersion of the composite components provided improved electrochemical performance of the material. The GO/V₂O₅ cathode was tested in coin cells with a Mg foil anode and 0.25 M Mg(AlCl₂EtBu)₂/THF as an electrolyte. The GCD measurements showed an initial discharge capacity of GO/V₂O₅ of 178 mAh·g⁻¹ at 0.2 C, which faded to 160 and 150 mAh·g⁻¹ in the second and third cycle, respectively, due to formation of a Mg passivation layer, but still remained at 140 mAh·g⁻¹ after 20 cycles.

rGO-decorated hydrated V₂O₅ nanowire composites with different contents of crystal water have been synthesized by a sol–gel method with subsequent freeze-drying [98]. The V₂O₅·nH₂O/rGO aerogel with a water content of 12.3% and large lattice spacing of 1.13 nm had a highly porous, interconnected 3D structure with V₂O₅ nanowires anchored on rGO (Figure 5a). The V₂O₅·nH₂O/rGO electrode showed a capacity of 280 mAh·g⁻¹ at 100 mA·g⁻¹ in a 0.5 M Mg(TFSI)₂/AN electrolyte, a capacity retention of 81% after 200 cycles, and a coulombic efficiency of 99%. At 50 mA·g⁻¹ the capacity was 320 mAh·g⁻¹. The nanocomposite displayed a wide working temperature range (~ 30–55 °C). As shown in Figure 5b, at 55 °C a high discharge capacity of ~200 mAh·g⁻¹ was achieved, much higher than the capacity of 120 mAh·g⁻¹ at room temperature. A reversible capacity of 40 mAh·g⁻¹ was delivered at −30 °C. The improved electrochemical performance of the material was attributed to the shielding effect of crystal water as well as the synergistic effects of rGO, which provided Mg diffusion pathways, high surface area, and structural stability.

A three-dimensional structure of V₂O₃ nanoparticles with reduced graphene oxide (rGO) with improved conductivity and structural stability was prepared by spray-drying of a dispersed solution [99]. In the 0.3 M Mg(TFSI)₂/AN electrolyte solution, V₂O₃@rGO microspheres demonstrated a high specific capacity (291.3 mAh·g⁻¹ at 50 mA·g⁻¹) and improved rate performance (185.3 mAh·g⁻¹ at a high current density of 2 A·g⁻¹) compared to pure V₂O₃ (~80 mAh·g⁻¹). Excellent cycling stability (88.5% capacity retention after 1000 cycles at 0.5 A·g⁻¹) with coulombic efficiency close to 100% was achieved. V₂O₃@rGO delivered a capacity of 472.1 mAh·g⁻¹ in the GITT study, and the calculated diffusion coefficient of Mg²⁺ in the V₂O₃@rGO structure was 3.4 × 10⁻¹¹ cm²·s⁻¹.

A hierarchical V₂O₃@C structure with an accordion-like vanadium oxide/carbon heterointerface was synthesized from vanadium-based metal organic frameworks (Figure 5c) [100]. The mesoporous V₂O₃@C nanorods had a length of 250–300 nm, elongated pores of 30–50 nm (Figure 5d), and a BET surface area of 114.6 m²·g⁻¹. The V₂O₃@C demonstrated a capacity of 354.8 mAh·g⁻¹ at a current density of 500 mA·g⁻¹ and an ultra-high capacity of 130.4 mAh·g⁻¹ at 50 A·g⁻¹ in 0.3 M Mg(TFSI)₂/AN(H₂O) with a coulombic efficiency of 99.6% and a capacity retention of 60.0% (1000 cycles, 500 mA·g⁻¹). To verify the advantages of hierarchical accordion-like heterointerfaces, the rate performance of the V₂O₃@C electrodes was compared with that of bulk-V₂O₃/AC and bulk-V₂O₅/AC electrodes prepared with activated carbon in the same ratio as in V₂O₃@C. As shown in Figure 5e, the specific capacity of bulk-V₂O₃/AC at 200 mA·g⁻¹ (203.2 mAh·g⁻¹) is four times higher than of bulk-V₂O₅/AC (54.9 mAh·g⁻¹) due to the anodic hydration of the V₂O₃ phase, occurring in bulk-V₂O₃/AC upon the first charging. The crystalline structure of V₂O₃ was reconstructed into a V₃O₇∙H₂O@C through an anodic hydration reaction upon the first cycle. The cell was disassembled under an inert atmosphere and the V₃O₇∙H₂O@C electrode was retested in a 0.3 M Mg(TFSI)₂/AN electrolyte, where the specific capacity of the V₃O₇∙H₂O@C electrode was 259.3 mAh·g⁻¹ at 500 mA·g⁻¹. The capacity difference of 33.9 mAh·g⁻¹ between the two electrolytes indicated that the capacity contributed by protons in the water-containing electrolyte is about 13% of the total capacity. The higher overpotential of the V₃O₇∙H₂O@C electrode in water-free electrolyte solution resulted from the slower diffusion of Mg²⁺-ions. In a full cell with a Mg anode and a 0.3 M Mg(TFSI)₂  +  0.25 M MgCl₂/DME electrolyte within the voltage range of 2.25 V vs. Mg/Mg²⁺, the V₂O₃@C exhibited a capacity of 245.1 mAh·g⁻¹.

Figure 5. (a) Scheme of V₂O₅·nH₂O@rGO nanocomposite with electron/ion transport pathways, (b) cycling performance of V₂O₅·nH₂O@rGO at 1.0 A·g⁻¹, (c) discharge capacity V₂O₅·nH₂O@rGO at various temperatures at 1.0 A·g⁻¹. (Reprinted with permission from [98]. Copyright 2015, Elsevier). (c) Scheme of the synthesis of V₂O₃@C from MOFs, (d) SEM image of V₂O₃@C, (e) rate capabilities of V₂O₃@C, bulk-V₂O₃/AC and bulk-V₂O₅/AC. (Reprinted with permission from [100]. Copyright 2024, John Wiley and Sons).

Figure 5. (a) Scheme of V₂O₅·nH₂O@rGO nanocomposite with electron/ion transport pathways, (b) cycling performance of V₂O₅·nH₂O@rGO at 1.0 A·g⁻¹, (c) discharge capacity V₂O₅·nH₂O@rGO at various temperatures at 1.0 A·g⁻¹. (Reprinted with permission from [98]. Copyright 2015, Elsevier). (c) Scheme of the synthesis of V₂O₃@C from MOFs, (d) SEM image of V₂O₃@C, (e) rate capabilities of V₂O₃@C, bulk-V₂O₃/AC and bulk-V₂O₅/AC. (Reprinted with permission from [100]. Copyright 2024, John Wiley and Sons).

2.5. Foreign Cation Pre-Intercalated Vanadium Oxides

The rich valence state variation of vanadium and the easy distortion of V–O polyhedra lead to the ability of V–O structure adaptation to incorporate different cations. The resulting derivatives with an A–V–O structures (A—metal ion or ${\mathrm{NH}}_4^{+}$) are called vanadates or foreign cation pre-intercalated vanadium oxides.

The pre-intercalation strategy has been widely recognized as a way to improve the properties of vanadium oxide cathodes for RMBs by increasing the interlayer space due to the pillaring effect combined often with improved stability of the obtained structures, allowing reversible insertion/extraction of Mg²⁺-ions. The analysis of the results of different works shows that, in most cases of pre-intercalated ions, a dual effect of structure expansion and void hydration takes place.

It should be noted that the pre-intercalation of cations, with its positive effects on increasing the interlayer spacing of V₂O₅ bilayers, also has some drawbacks. The most important one is that the introduction of electrochemically inactive cations leads to a decrease in specific capacitance. Therefore, the necessary optimization of cathode materials requires a balance between structural and stability factors on the one hand, and the improvement of specific capacities on the other hand.

The differences in the active material composition (especially water content), electrode processing and morphology, and electrolyte system make it difficult to compare results from different groups.

Among the cations used for pre-intercalation of vanadium oxides for RMBs are mono-, bi-, and tri-valent ions: Na⁺ [101,102,103,104], ${\mathrm{NH}}_4^{+}$ [105,106], Mg²⁺ [107,108,109,110], Mn²⁺ [111], Ca²⁺ [112], and Al³⁺ [113].

NaV₃O₈·1.69H₂O nanobelts were synthesized from commercial V₂O₅ powder by a solvothermal procedure under ambient conditions (Figure 6a) [102]. The cells with a Mg anode and all-phenyl complex electrolyte demonstrated an initial discharge capacity of 150 mAh·g⁻¹ and 110 mAh·g⁻¹ after five cycles at a current density of 10 mA·g⁻¹. The GCD curves of NaV₃O₈·1.69H₂O at different current densities (Figure 6b) showed several plateaus during the charging and discharge process, due to the multistep redox reactions responsible for the insertion/extraction of Mg²⁺ ions. The limited specific capacity of the cells was attributed to the trapping of Mg²⁺ ions in the lattice of NaV₃O₈. The cathodes demonstrated high cyclic stability with 80% capacity retention after 100 cycles at 50 mA·g⁻¹.

Pre-intercalating of alkali ions (Li⁺, Na⁺, K⁺) into layered vanadium oxide was studied in [101]. It was shown that the interlayer spacing of Li-, Na-, K-intercalated V₃O₈ increased regularly with the size of the pre-intercalated cations. The cycling performance of cathodes for Mg²⁺ storage was enhanced with the bigger radius of pre-intercalated ion (Figure 6c); after 30 cycles the capacity retention of KV₃O₈ was 88.6%, compared to 85.78% of NaV₃O₈ and 42.2% of LiV₃O₈, although the specific capacity of KV₃O₈ was the lowest (37.56 mAh·g⁻¹). The initial specific capacity of LiV₃O₈ and NaV₃O₈ at 100 mA·g⁻¹ was 252.2 and 204.16  mAh·g⁻¹, respectively (Figure 6d). Based on structural analysis and electrochemical tests, it was concluded that pre-intercalation with Na⁺ resulted in a more stable interlayer structure, allowing free Mg²⁺ diffusion and preventing destructive collapse of the layers during the charge–discharge.

The interlayer spacing of V₂O₅ was expanded in [103] by introducing Na⁺-ions in the crystal lattice. NaV₆O₁₅ (NVO) free of crystal water was synthesized and studied as a cathode material for RMBs with anhydrous 0.5 M Mg(ClO₄)₂/AN electrolyte solution. It was shown that the introduction of Na⁺-ions enhanced the diffusion kinetics of Mg²⁺ ions (D_Mg²⁺ during the discharge process was in the range of 7.55 × 10⁻¹³ to 2.41 × 10⁻¹¹ cm²·s⁻¹ according to GITT), and improved the stability of the layered structure. NVO exhibited high initial discharge capacity of 213.4 mAh·g⁻¹ at the current density of 10 mA·g⁻¹, good capacity retention (87% after 100 cycles at 20 mA·g⁻¹), and good rate capability. The ex situ XRD showed that the mechanism of Mg²⁺ storage in NVO is reversible intercalation/de-intercalation. The DFT calculation results indicated that during the intercalation process, Mg²⁺-ions tend to occupy the semi-occupied sites of Na⁺ in the NVO.

Water-pillared Na₂V₆O₁₆·1.63H₂O nanowires obtained in [104] displayed high performance in magnesium storage. In a cell with a Mg(TFSI)₂/DME electrolyte and an AC anode, Na₂V₆O₁₆·1.63H₂O exhibited a high specific capacity of 175 mAh·g⁻¹ at 0.05 A·g⁻¹, long-term cycling performance, and ≈100% coulombic efficiency. Under the same testing conditions, annealed Na₂V₆O₁₆ delivered a low specific capacity of only 40 mAh·g⁻¹ (Figure 6e). Na₂V₆O₁₆·1.63H₂O cathodes demonstrated a discharge voltage plateau at around 2.0 V (vs. Mg/Mg²⁺), ascribed to the Mg²⁺ ion intercalation into the bilayer structure. According to ex situ XRD analysis, Na₂V₆O₁₆·1.63H₂O possesses a very stable structure for reversible Mg²⁺ ion intercalation/deintercalation. During the discharge, Mg²⁺-intercalation is accompanied by shrinkage of the interlayer space (Figure 6f). During subsequent charge–discharge processes, the Mg²⁺-ions can reversibly intercalate/deintercalate into the Na₂V₆O₁₆·1.63H₂O layers. The water molecules acting as “pillars” stabilize the layered structure and effectively shield the high charge density of Mg²⁺-ions.

Magnesiated V₂O₅ xerogel with the formula Mg_0.1V₂O₅·1.8H₂O was synthesized in [109] by a low-temperature procedure. The material had an initial discharge capacity of 300 mAh·g⁻¹ in 0.5 M Mg(TFSI)₂/AN. A capacity of ~250 mAh·g⁻¹ was maintained over eight cycles at a C/10 rate, consistent with an insertion of 1 equivalent of Mg²⁺ per formula unit. No change in the interlayer distance (12.3 Å) was observed upon cycling of Mg_0.1V₂O₅·1.8H₂O electrodes.

Mg-inserted V₂O₅·xerogel was prepared via an ion removal sol–gel method in [114]. The Mg_0.1V₂O₅·2.35H₂O cathode displayed a high electrode potential of 3.0 V (vs. Mg/Mg²⁺), high energy density of 420 mWh·g⁻¹, and a discharge capacity of 140 mAh·g⁻¹ at a 0.1 C rate.

Mg²⁺ pre-intercalated hydrated vanadium oxide nanowires with bilayer structure, Mg_0.3V₂O₅·1.1H₂O, were synthesized in [110] from α-V₂O₅ using a hydrothermal approach. The reaction of α-V₂O₅ with H₂O₂ lowers the valence state of V, resulting in structural transformation of VO₅ square pyramids to VO₆ octahedra, and a large number of H₂O molecules are embedded into the interlayer. This results in a bilayer structure with a larger interlayer space, providing ample channels for the subsequent insertion of Mg²⁺ ions. As a result of the coordination between Mg²⁺ and lattice oxygen, the insertion of Mg²⁺ ions into the layers leads to layer slippage and shrinkage (Figure 7a). In order to investigate the role of Mg²⁺ ions and lattice water molecules, V₂O₅·nH₂O and Mg_0.3V₂O₅ nanowires were also prepared. In Mg_0.3V₂O₅·1.1H₂O, the intercalated Mg²⁺ ions work together with crystal water to provide wide channels for electrolyte ion transport during the charge–discharge process. Mg_0.3V₂O₅·1.1H₂O possessed high electronic conductivity, fast Mg²⁺ reaction kinetics, and good structural stability. The charge and discharge curves of Mg_0.3V₂O₅·1.1H₂O, V₂O₅·nH₂O and Mg_0.3V₂O₅ at 0.1 A·g⁻¹ in a cell with 0.3 M Mg(TFSI)₂/AN electrolyte and an activated carbon anode are shown in Figure 7b. Mg_0.3V₂O₅·1.1H₂O exhibited three discharge voltage plateaus at 3.02, 2.25, and 1.40 V (vs. Mg/Mg²⁺), which can be attributed to the multi-step Mg²⁺ intercalation in the bilayer structure. Mg_0.3V₂O₅·1.1H₂O delivered high capacity (164 mAh·g⁻¹ at 0.1 A·g⁻¹), corresponding to the insertion of 0.5 Mg²⁺ ions per formula unit. V₂O₅·nH₂O had a similar charge–discharge curve, with a lower discharge capacity of 114 mAh·g⁻¹. For Mg_0.3V₂O₅, no obvious plateaus were observed on the charge–discharge curve and it had the lowest discharge capacity (91 mAh·g⁻¹). As can be seen in Figure 7c, Mg_0.3V₂O₅·1.1H₂O demonstrated good rate performance (50 mAh·g⁻¹ at 4.0 A·g⁻¹) and high cycling stability (capacity retention of 80.0% after 10,000 cycles at 2.0 A·g⁻¹). In a full RMB cell with a Na₂Ti₃O₇ anode and the Mg_0.3V₂O₅·1.1H₂O cathode, the specific capacity of about 62 mAh·g⁻¹ at the current density of 100 mA·g⁻¹ was achieved, and the battery demonstrated a stable voltage plateau of 1.5 V and a cycle life of 100 cycles.

Mg²⁺-pillared hydrated vanadium oxide Mg_xV₅O₁₂·nH₂O nanofibers (Figure 7d), with expanded interlayer spacing were obtained in [108]. Mg²⁺ pillars and structural H₂O molecules stabilized the material framework and promoted Mg²⁺ diffusion during cycling. At a current density of 0.05 A·g⁻¹ the Mg_xV₅O₁₂·nH₂O electrode provided a high capacity of 160 mAh·g⁻¹, corresponding to 1.3 Mg²⁺ inserted per V₅O₁₂ structural unit. The Mg_xV₅O₁₂·nH₂O electrode showed excellent long-term stability (81% capacity retention after 10,000 cycles at 2 A·g⁻¹), the coulombic efficiency was >99%. In situ XRD measurements (Figure 7e) had shown that the Mg_xV₅O₁₂·nH₂O electrode undergoes dynamic structural changes upon Mg²⁺-intercalation/deintercalation. The (001) diffraction peak shifted to a higher angle region during the discharge process (reduction), indicating the gradual contraction of the interlayer spacing. During charging (oxidation), the (001) peak returned to the initial position by the end of a deep charge. The Mg²⁺-intercalation caused a decrease in interlayer spacing to 1.10 nm, and it recovered to 1.19 nm upon Mg²⁺ deintercalation (Figure 7f,g).

Expanding the interlayer spacing of layered materials is an effective way to improve Mg²⁺-ion storage performance. On the one hand, as the interlayer spacing increases, the interaction between intercalated guest ions and host lattices weakens, which greatly improves the mobility of cations in cathode materials. On the other hand, the number of available sites in cathode materials also increases due to the enlarged lattice spaces, leading to improved reversibility of Mg²⁺-ion storage capacity.

2.6. Vanadium Oxides Pre-Intercalated with Organic Molecules and Conducting Polymers

The intercalation of conducting polymers into vanadium oxides has been shown to be advantageous for improving the functional properties of electrode materials. Conducting polymers intercalated into layered vanadium oxide can increase the interlayer spacing and stabilize the layered structures by reducing the coulombic interactions between the guest cations and the host framework, as has been shown on numerous examples for aqueous Zn-ion batteries [115].

Rechargeable magnesium batteries with a V₂O₅ xerogel cathode with polyethylene oxide (PEO) incorporated between the oxide layers showed a significant improvement in the reversible capacity of magnesium ions [116]. PEO was introduced in the V₂O₅ interlayer during the V₂O₅ sol–gel synthesis (Figure 8a,b); this approach allowed the expansion of the interlayer spacing in V₂O₅ and the reduction in the interaction of intercalated divalent Mg²⁺ with the host lattice. X-ray diffraction showed that the interlayer spacing in V₂O₅ xerogel was increased to 12.6–13.6 Å by the inclusion of PEO. Galvanostatic charge–discharge profiles of V₂O₅ xerogel and V₂O₅-PEO nanocomposites with different amounts of PEO are shown in Figure 8c. In a cell with a Mg anode and a dry Mg(ClO₄)₂/AN electrolyte, the V₂O₅-PEO-1 nanocomposite exhibited a discharge capacity of 100.3 mAh·g⁻¹ at 10 mA·g⁻¹, ~5 times higher than of V₂O₅ xerogel (and ~2 times higher than of V₂O₅-PEO-2), improved stability, and enhanced rate capability. The Mg²⁺-ion capacity for the V₂O₅ xerogel, V₂O₅-PEO-1, and V₂O₅-PEO-2 cathode materials was determined as 0.06, 0.34, and 0.18 per V₂O₅ formula unit (i.e., x in Mg_xV₂O₅), respectively. In cyclic voltammograms, the V₂O₅-PEO-1 nanocomposite showed higher currents compared to V₂O₅ xerogel and V₂O₅-PEO-2, indicating that the introduction of an optimal PEO ratio within V₂O₅ improved the Mg ion charge storage performance. The Mg²⁺ diffusion coefficient in V₂O₅-PEO-1 nanocomposite was 4.7 × 10⁻¹¹ cm²·s⁻¹, that is, 2 times higher than the Mg²⁺ diffusion coefficient in the V₂O₅ xerogel. The Mg²⁺ diffusion coefficient in V₂O₅-PEO-2 was 2.5 × 10⁻¹² cm²·s⁻¹, lower than that for the V₂O₅ xerogel. Therefore, Mg²⁺-ion diffusion was not only influenced by the interlayer spacing (which is the largest for V₂O₅-PEO-2), but also by the interlayer composition. It was suggested that a higher PEO ratio in the composite could result in lower charge storage for reasons such as a lower diffusion rate of Mg²⁺-ions in PEO-rich regions, lower electronic conductivity due to the insulating nature of PEO, and coordination of PEO molecules with the V₂O₅ lattice, blocking sites for Mg²⁺ intercalation.

PANI-V₂O₅ 2D organic–inorganic superlattices were synthesized and tested as cathode materials for RMBs in [117]. PANI-V₂O₅ (PVO) was synthesized via the intercalation of aniline monomers and subsequent interlayer polymerization (Figure 8e). In the acidic environment, the anilinium cations diffused into the interlayer space of V₂O₅ and underwent oxidative polymerization, V₂O₅ being a mild oxidizing agent. Compared to the XRD pattern of V₂O₅ (VO), a new diffraction peak of hydrated V₂O₅ (HVO) was located at 7.08°, confirming that the interlayer spacing of HVO is greater than that of VO, and for PVO, the new diffraction peak was shifted to a lower angle 6.47°, indicating larger interlayer spacing in PVO than in HVO and the insertion of PANI (Figure 8f). The formation of protonated PANI in PVO was confirmed by the Raman spectra (Figure 8g). PANI not only expanded the interlayer spacing of V₂O₅, but also provided additional charge storage sites. Benefitting from the above features, the PVO demonstrated high capacity (280 mAh·g⁻¹ at a current density of 0.1 A·g⁻¹) and excellent rate performance (135 mAh·g⁻¹ at a current density of 4 A·g⁻¹) in 0.3 M Mg(CF₃SO₃)₂/AN electrolyte. The PVO cathode delivered capacities of 275, 250, 220, 175, 155, and 130 mAh·g⁻¹ at current densities of 0.1, 0.2, 0.5, 1.0, 2.0, and 4.0 A·g⁻¹, respectively. The PVO had an excellent cycling stability, and the retained capacity after 500 cycles at 4 A·g⁻¹ was 80 mAh·g⁻¹.

Polyaniline was in situ intercalated into V₆O₁₃ [118]. With the increase in the amount of aniline monomer, PANI-V₆O₁₃ composites with different enlarged d-spacings (11.8, 13.4, and 14.7 Å) were obtained. The contents of V₆O₁₃ and PA in the PA100-V₆O₁₃ with 13.4 Å d-spacing were 88.18% and 9.60%, respectively. In the cell with an activated carbon anode and a Mg(TFSI)₂/DME electrolyte, the PA100-V₆O₁₃ sample demonstrated the highest reversible capacity (195 mAh·g⁻¹ at 0.1 A·g⁻¹), with a retention of 173 mAh·g⁻¹ after 100 cycles. PANI-V₆O₁₃ exhibited high rate capability (46 mAh·g⁻¹ at 10 A·g⁻¹) and outstanding cycling stability (2500 cycles at 5 A·g⁻¹). The higher reversible capacity of Mg²⁺ storage of PA100-V₆O₁₃ was mainly ascribed to the expanded interlayer spacing. The sample PA150-V₆O₁₃ with a higher polyaniline content and a larger interlayer spacing had higher initial discharge capacity but less satisfactory cycling performance, probably due to poor structural stability which reduced the capacity retention. The Mg²⁺ diffusion coefficients in PANI-V₆O₁₃ calculated from GITT data varied over the entire insertion/extraction processes and were in the range from 2.1 × 10⁻⁹ to 7.5 × 10⁻¹² and from 2.7 × 10⁻¹⁰ to 3.2 × 10⁻¹¹ for the insertion and extraction processes, respectively. Thus, in polyaniline-intercalated V₆O₁₃, the Mg²⁺ intercalation kinetics were enhanced due to the π-conjugated polyaniline molecules weakening strong coulombic interactions between Mg²⁺-ions and anions in the host material.

The layered V₂O₅-PEDOT composite with an enlarged interlayer spacing of 9.86 Å was synthesized in [119] via in situ polymerization of 3,4-ethylenedioxythiophene and sequential intercalation of PEDOT and cetyltrimethylammonium bromide (CTAB) into V₂O₅. In the XRD pattern of the as-prepared V₂O₅-PEDOT (VOP), a new peak appears at 8.96°, corresponding to an increased interlayer spacing of 9.86 Å, much larger than the d-spacing of V₂O₅ (4.38 Å). The intercalation of PEDOT pillars results in dark coloring of the VOP powder; these pillars change the structure of bulk V₂O₅ by drastic volume expansion, resulting in delamination of the VOP into a nest-like structure of interlaced nanowires about 50 nm in diameter, as seen in the SEM image (Figure 9a–c). The magnesium storage performance of V₂O₅-PEDOT cathode was evaluated in a cell with a Mg anode and all-phenyl complex APC-CTAB/THF electrolyte. The CTAB was added for further improvement in the Mg²⁺ diffusion kinetics. Two pairs of broad redox peaks located at around 1.25/0.80 and 1.78/1.57 V (vs. Mg/Mg²⁺) were observed in the cyclic voltammograms of V₂O₅-PEDOT. The reversible capacity of V₂O₅-PEDOT in the full cell was 288.7 mAh·g⁻¹ at 0.1 A·g⁻¹, and the electrode demonstrated high cyclability (over 500 cycles at 0.5 A·g⁻¹ with capacity retention 68%).

A V₂O₅/PEDOT (VOP) composite was synthesized by intercalating poly-3,4-ethylenedioxythiophene into V₂O₅ under stirring at room temperature for 1 week (Figure 9d) [120]. After 4 days of stirring, a new diffraction peak appeared in the XRD pattern of VOP at 7.6°, corresponding to an interlayer spacing of 11.98 Å. After 7 days of stirring, two new peaks appeared at 4.65° and 9.32°, corresponding to an interlayer spacing of 19.02 Å of the (001) plane and 9.49 Å of the (002) plane (Figure 9e). It was observed that the bilayer structure of orthorhombic V₂O₅ was retained even after the insertion of PEDOT. The porous structure of the as-prepared VOP resulting from the aggregation of the bilayer structures is shown in the SEM image (Figure 9f). The electronic coupling of PEDOT with V₂O₅ interlayers led to a change in PEDOT structure from the quinoid to the benzoid, and the corresponding blue shift of the principal peaks of PEDOT was observed in UV-vis spectra (Figure 9g). A reversible and fast Mg²⁺ ion insertion/extraction in/out of an enlarged interlayer spacing of 19.02 Å (in the charged state at +1.0 V) and 20.16 Å (in the discharge state at −1.0 V) was achieved. The VOP electrodes in 0.3 M Mg(TFSI)₂/AN delivered a high specific capacity of 339.7 mAh·g⁻¹ at 0.1 A·g⁻¹, high rate capacity of 256.3 mAh·g⁻¹ at 0.5 A·g⁻¹, and long-term cyclic stability with a 0.065% decay rate and high capacity of 172.5 mAh·g⁻¹ after 500 cycles. VOP delivered much higher specific capacities (320, 271, 220, 172, and 117 mAh·g⁻¹) than pristine V₂O₅ (72.0, 67, 58, 50, and 43 mAh·g⁻¹) at the current densities of 0.1, 0.2, 0.5, 1.0, and 2.0 A·g⁻¹, respectively. The enhancing effect of water activation on the kinetics was confirmed. When 3 M H₂O was added to 0.3 M Mg(TFSI)₂/AN, the discharge capacities of VOP at 0.1 A·g⁻¹ increased from 165 to 348 mAh·g⁻¹, and the capacity of pristine V₂O₅ increased from 82 to 113 mAh·g⁻¹. The improvement in the electrochemical performance of VOP and pristine V₂O₅ was attributed to the lowered desolvation energy of the solvated Mg²⁺-ions.

V₂O₅·H₂O was synthesized and intercalated with polyacrylonitrile (PAN) by hydrothermal synthesis procedure [121]. The exact chemical composition of the PAN-intercalated oxide was V₂O₅·0.34H₂O-PAN, and that of vanadium oxide used for comparison was V₂O₅·0.33H₂O. According to XRD data, all peaks of the V₂O₅-PAN and V₂O₅·H₂O were consistent with double-layer V₂O₅ (JCPDS No. 40-1296). The layer spacing increased from 9.91 Å in the pure V₂O₅·H₂O to 10.63 Å in V₂O₅-PAN. It was shown that intercalation of PAN also induced cation reduction and generated abundant anion vacancies, thus improving ion diffusion and electron transfer kinetics. V₂O₅-PAN electrodes had perfect structural stability and electrochemical reversibility during charge–discharge. After stabilization, V₂O₅-PAN achieved a high specific discharge capacity of ~180 mAh·g⁻¹ at 50 mA·g⁻¹ in 0.5 M Mg(TFSI)₂/AN electrolyte, and its life-span at 2 A·g⁻¹ was 18,000 cycles.

The general tendency of positive effects observed with the introduction of conducting polymer molecules is explained by the formation of “pillars” expanding the interlayers, and also by the improvement in electrical conductivity and the effect of electrostatic shielding of the host from the high charge density of Mg²⁺. However, to date, the role of the composition/nature and the optimal fraction of conducting polymers introduced into the layered structure need further investigation.

3. Mechanisms of Intercalation and Role of Structural Water

The general operating principle of RMBs in dry aprotic electrolytes is thought to be similar to that of other metal ion batteries. It is based on the reversible intercalation/deintercalation of Mg²⁺-ions into the active material of the cathode. However, there is some controversy in the studies of the mechanism of Mg²⁺ intercalation in dry electrolytes; the reaction mechanism of α-V₂O₅ in RMBs with a dry electrolyte is not well understood.

A systematic study of Mg insertion into orthorhombic V₂O₅ was performed by combining electrochemical and structural methods [122]. The results for an electrochemically cycled V₂O₅ cathode in a full cell with Mg metal anode obtained by atomic-resolution transmission electron microscopy showed the local formation of the theoretically predicted ε-Mg_0.5V₂O₅ phase; however, the intercalation level of Mg was low.

The investigations of the electrochemical behavior of orthorhombic α-V₂O₅ in dry and “wet” alkyl carbonate-based electrolytes and the insertion-driven structural changes of the active phase studied by ex situ X-ray diffraction showed that proton intercalation dominates the reaction of α-V₂O₅, even in a dry electrolyte, and the intercalation of Mg²⁺-ions is negligible [123].

In operando studies of the reaction mechanism of α-V₂O₅ cathodes in RMBs by synchrotron diffraction and in situ X-ray absorption near-edge spectroscopy (XANES) together with ex situ Raman and X-ray photoelectron spectroscopy [75] have shown the reversibility of magnesium ion intercalation and provided information on the evolution of the crystal structure and the change in oxidation degrees during charge–discharge cycling. It was shown that α-V₂O₅ transforms to a Mg-poor phase (Mg_0.14V₂O₅) during discharging, and then undergoes a two-phase transition.

There are three related issues, the role of which in the case of vanadium oxide cathodes for RMBs should be considered in more detail:

(i)

Crystalline V₂O₅ is capable of reversible intercalation of Mg²⁺ in “wet” organic electrolytes, containing enough water molecules to coordinate to Mg²⁺-ions and shield their high charge during intercalation, thus allowing faster diffusion of Mg²⁺-ions within the host material.

(ii)

The incompatibility of metallic Mg anodes with wet electrolytes leads to the focus of the research on the bilayer V₂O₅, hydrated, and water-pillared vanadium oxides, where water molecules are naturally present in the lattice [104,108,111]. The interlayer water molecules effectively shield the interaction between Mg²⁺ and the host lattice, improving the reversibility of Mg²⁺-ion intercalation/deintercalation and structural stability of the cathode material during cycling.

Following the pioneering work [61], many researchers noted that V₂O₅ exhibits improved capacity in aprotic electrolytes containing small amounts of water. In particular, the study of the electrochemical performance of V₂O₅ cathodes in Mg(ClO₄)₂/PC and Mg(TFSI)₂)/G2 electrolytes [124,125] showed that controllable amounts of water lead to increased specific capacities.

Water molecule intercalation into V₂O₅ structures improves Mg²⁺-ion diffusion kinetics due to enlarged interlayer spacing. In addition, the water molecules in the cathode increase electrostatic shielding, which is a very important factor for improving the migration rate of Mg²⁺-ions within the host material [47,124].

The experimental observations of the positive influence of water molecules in non-aqueous electrolytes on the charge storage properties of V₂O₅ (higher currents in CVs, higher capacities, lower polarization) were later extended by the experimental results pointing to the mechanism in which both protons and magnesium ions participate as charge carriers in the redox process.

Some works demonstrate co-intercalation of protons together with magnesium ions, providing higher capacity of vanadium oxide cathodes. A number of works demonstrate the dominant role of reversible proton insertion in the capacity of RMB in aqueous electrolytes and non-aqueous electrolytes containing small amounts of water.

The study of the influence of water additive in the organic electrolyte had shown that the presence of water enables higher capacities for an α-V₂O₅ cathode [125]. The capacity achieved for the wet electrolyte was ~260 mAh·g⁻¹, which is close to the theoretical value. In dry electrolyte, a discharge capacity of only ~60 mAh·g⁻¹ was achieved. Solid state NMR revealed that the major contribution towards the higher capacity in high water-containing electrolyte solution (1 M Mg(TFSI)₂/G2 with 2600 ppm H₂O) originates from the reversible proton insertion. On lowering the water level of the electrolyte (1 M Mg(TFSI)₂/G2 with 15 ppm H₂O), the evidence of reversible Mg intercalation was obtained by EDX of the discharged V₂O₅, which demonstrated the presence of a higher concentration of Mg in the discharged state compared to the charged and pristine V₂O₅ electrodes.

The intercalation of Mg²⁺ in α-V₂O₅ was again investigated in dry and water-containing organic electrolytes [126]. An improved electrochemical performance was observed after adding water to a dry organic electrolyte; this effect was explained by the co-intercalation of protons and Mg²⁺. The discharged product of α-V₂O₅ in a 0.5 M Mg(ClO₄)₂ + 2.0 M H₂O/AN electrolyte was Mg_0.17H_xV₂O₅ (x = 0.66–1.16), indicating that the capacity is mainly due to proton intercalation.

The magnesium ion storage capability of a VO₂(B) cathode in dry and wet 0.5 M Mg(ClO₄)₂/AN electrolytes was studied in [56]. The water content in dry and wet electrolytes was 40 and 650 ppm, respectively. In the dry electrolyte, VO₂(B) deliveredan initial discharge capacity of 50.6 mAh·g⁻¹ at a current density of 20 mA·g⁻¹, with a coulombic efficiency of 94.4%. Under the same electrochemical test conditions, the initial discharge capacity of VO₂(B) in the wet electrolyte was 251.0 mA·g⁻¹, with a coulombic efficiency of 99.4 %. The high discharge capacity observed in the wet electrolyte was attributed to predominant proton intercalation instead of magnesium intercalation. It was also shown that VO₂(B) may suit as a high-performance cathode material only if its particles are relatively small, to exclude long distances for Mg²⁺-ion diffusion along the diffusion channels within the material.

The strongly polarizing nature of Mg²⁺-ions leads to slow intercalation kinetics. The structural water in host electrode materials facilitates migration of Mg²⁺-ions due to the “lubricating” effect; in addition, water molecules act as charge screening media to reduce the coulombic repulsion between the ions and the host lattice.

The enhanced insertion reaction of Mg²⁺-ions into the V₂O₅(H₂O)_y (2 < y < 3) xerogel, containing single and/or double layers of H₂O molecules in the interlayer space of the oxide, was reported as early as 1992 in [127]. Bonded H₂O molecules allowed smooth insertion and de-insertion of Mg²⁺-ions during the electrochemical cycling of V₂O₅ xerogel electrodes in a dry 1 M Mg(ClO₄)₂/AN electrolyte. The specific charge density 170 Ah·kg⁻¹ was measured in the first cycle for V₂O₅ xerogel, while less than 50 Ah·kg⁻¹ was observed in the first cycle for pure V₂O₅.

Layered H₂V₃O₈ nanowires demonstrated high Mg²⁺-ion storage with a capacity of 304.2 mAh·g⁻¹ at 50 mA·g⁻¹ in a three-electrode cell with 0.5 M Mg(ClO₄)₂/AN electrolyte and activated carbon cloth as counter and reference electrodes [128]. In the cyclic voltammograms at a scan rate of 0.2 mV·s⁻¹, one cathodic peak at ~2.0 V and two anodic peaks at 2.3 and 2.5 V (vs. Mg/Mg²⁺) were observed, which were ascribed to the insertion and extraction of magnesium ions. GCD curves showed that H₂V₃O₈ nanowires possess a high working voltage platform of about 2.0 V. H₂V₃O₈ nanowires demonstrated 85.9% capacity retention after 20 cycles at 50 mA·g⁻¹.

Hydrated vanadium oxide H₂V₃O₈ nanowires were synthesized via a one-step hydrothermal method [129]. The H₂V₃O₈ nanowires had a diameter of 100–200 nm (Figure 10a,b) and an interlayer spacing of 0.34 nm (Figure 10c), corresponding to the d(011) spacing of H₂V₃O₈. The material exhibited reversible magnesiation/demagnesiation with an initial discharge capacity of 80 mAh·g⁻¹ and 231 mAh·g⁻¹ at 10 mA·g⁻¹ at 25 °C and 60 °C, respectively (Figure 10d), and an average discharge voltage of ~1.9 V (vs. Mg/Mg²⁺) in a dry (48 ppm H₂O) 0.5 M Mg(ClO₄)₂/AN electrolyte. The discharge capacities of H₂V₃O₈ at 60 °C were 231, 201, 170, and 97 mAh·g⁻¹ at the current densities of 10, 20, 40, and 80 mA·g⁻¹, respectively (Figure 10e). The capacity observed at 25 °C in dry electrolyte was much lower than the value of 304.2 mAh·g⁻¹ reported for H₂V₃O₈ in [128] due to the difference in the water content in the organic electrolyte. In the wet electrolyte (5790 ppm H₂O), it increased to ~260 mAh·g⁻¹. The structural water of H₂V₃O₈ (V₃O₇·H₂O) remained stable during cycling. The chemical formula for the discharged electrode at 60 °C was calculated as Mg_1.22H₂V₃O₈, considering that one electron charge transfer corresponds to 94.75 mAh·g⁻¹ per formula unit of H₂V₃O₈. The analysis of the discharged electrode suggested the chemical formula Mg_0.97H₂V₃O₈ (Figure 10f). The difference of ~20% was explained by the thorough washing of the electrodes during the sample preparation for elemental analysis, when the bulk-intercalated Mg²⁺-ions remained and the surface ions were washed out. A homogeneous distribution of Mg²⁺-ions was observed in the vanadium oxide particles of the discharged electrode, while they were absent in the initial or charged samples (Figure 10 g–i).

It was shown that the kinetic and the electrochemical properties of H₂V₃O₈ (HVO) are improved by adding water to a dry organic electrolyte [130]. HVO nanofibers were synthesized by exfoliation of V₂O₅ followed by a hydrothermal process. H₂V₃O₈ delivered an initial specific capacity of 303 mAh·g⁻¹ at a current density of 50 mA·g⁻¹ in a wet Mg(ClO₄)₂·3.1H₂O/AN electrolyte, four times higher than in a dry electrolyte (67 mAh·g⁻¹). Although the presence of water in the aprotic electrolyte was beneficial for the specific capacity, the capacity retention in the wet electrolyte was 80% after 30 cycles at 50 mA·g⁻¹ and 61% after 200 cycles at 100 mA·g⁻¹, while in the dry electrolyte the capacity retention was 99% after 30 cycles at 50 mA·g⁻¹ and 109% after 200 cycles at 100 mA·g⁻¹, demonstrating the long-term stability and the reversibility of the charge–discharge process in the dry electrolyte. In an “ambient” Mg(ClO₄)₂·1.4H₂O/AN electrolyte, the HVO cathode delivered an initial specific capacity of 167 mAh·g⁻¹ at a current density of 50 mA·g⁻¹ and showed a capacity retention of 73% after 30 cycles, suggesting that the water content in the electrolyte affects the reversibility of the charge–discharge processes. The structural changes in HVO as a function of the water content in the electrolyte were studied by in operando XRD during galvanostatic cycling. It was observed that the HVO structure almost fully recovered after two discharge/charge cycles in dry electrolyte, while in the presence of water, irreversible changes in the lattice parameters occurred in the initial cycle followed by a reversible discharge/charge cycle. In the presence of water, the changes in lattice parameters and unit cell volume during discharge/charge were 8 times larger, supporting the hypothesis that H₂O molecules from a hydration shell co-intercalate with Mg²⁺ ions, thus leading to a decrease in the cycling stability of HVO in electrolytes with water content. In the wet and “ambient” electrolytes, the insertion of ~0.60 equivalent Mg²⁺ into HVO took place. In the wet electrolyte, STEM/EDX (Figure 10 j–m) revealed that the inserted Mg²⁺-ions were mainly accumulated close to the surface and at the edges of the HVO fibers, but did not diffuse into the bulk. This explained the high reversibility of the insertion process and the small lattice parameter changes.

Introducing water molecules into the lattice material together with cation pre-intercalation is also an effective strategy for improving the electrochemical performance of V₂O₅ in RMBs. Mn²⁺-ions were introduced into the hydrated vanadium oxide by a hydrothermal method in [111]. The experimental results revealed the synergetic effect of both Mn²⁺ ions and crystal water in the Mn_0.04V₂O₅·1.17H₂O nanobelts, improving the structural and cycle stability of the cathode in the Mg(TFSI)₂/AN electrolyte. The Mn_0.04V₂O₅·1.17H₂O material exhibited a high specific capacity (145 mAh·g⁻¹ at 50 mA·g⁻¹), good rate capability (50 mAh·g⁻¹ at 4 A·g⁻¹), and long cycle stability (82% capacity retention after 10,000 cycles at 2 A·g⁻¹), better than that of water-free Mn_0.04V₂O_5. The in situ and ex situ material characterizations revealed that the structure of Mn_0.04V₂O₅·1.17H₂O was stable during the charge–discharge process.

Water-lubricated and aluminum ion-pillared vanadate H₁₁Al₂V₆O_23.2 (HAlVO) was reported as a high-performance cathode material for Mg²⁺ ion storage [113]. The capacity fade mechanism of water-free aluminum vanadate AlV₃O₉ (AlVO) was also investigated. The charge transfer process in water-lubricated and water-free aluminum vanadates was analyzed by DFT calculations, and the different charge transfer processes in the two materials and the charge shielding effect of the water molecule in HAlVO were revealed. The comparison of the cyclic voltammograms of HAlVO and AlVO showed that the shape of the CVs and the position of the redox peaks of the two materials are different: HAlVO had more intense reduction and oxidation peaks than AlVO (Figure 11a). HAlVO displayed higher initial specific capacity and a more stable cycle performance (165 mAh·g⁻¹ in the 1st and 50th cycles), while AlVO exhibited a capacity of only 125 mAh·g⁻¹ in the 1st cycle, which decreased to 105 mAh·g⁻¹ in the 50th cycle (Figure 11b,c). Water-lubricated HAlVO had better cycle performance at all current densities. The diffusion coefficients were obtained for HAlVO and AlVO from GITT curves at different stages of Mg²⁺ insertion. At the initial stage of Mg²⁺ insertion, D_HAlVO was 4.45 × 10⁻¹¹ cm²·s⁻¹ and D_AlVO was 2.77 × 10⁻¹¹ cm²·s⁻¹. When the amount of inserted Mg²⁺ reached 100%, D_HAlVO decreased to 1.01 × 10⁻¹¹ cm²·s⁻¹ and D_AlVO to 2.12 × 10⁻¹² cm²·s⁻¹. The faster decrease in the diffusion rate of Mg²⁺ in AlVO than in HAlVO was explained by larger interlayer spacing of HAlVO and the shielding effect of interlayer water molecules, providing more active storage sites for Mg²⁺, enough diffusion space, and less charge repulsion when a large amount of Mg²⁺-ions is intercalated into HAlVO. The scheme of reversible insertion/extraction of Mg²⁺ in HAlVO is shown in Figure 11d. Pre-intercalated Al³⁺ ions act as “pillars” which stabilize the V-O layered structure both in discharged and charge states, which improves the cyclic stability of the electrode. When Mg²⁺ ions are inserted, the water molecules in the interlayer space facilitate achieving local electroneutrality and a lower Mg diffusion barrier; this improves the rate performance of HAlVO. The DFT calculation results suggested that the water molecules also reduce the insertion energy barrier of Mg²⁺, thus improving the specific capacity of HAlVO. The charge transfer occurs mainly from the Mg²⁺ to O of the water molecule, and less from Mg²⁺ to O in the V-O layer, thus probably preventing the structural decay of the cathode during cycling.

A Mg²⁺ pre-intercalated hydrated V₁₀O₂₄·nH₂O layered material was obtained [107]. Mg_0.75V₁₀O₂₄·4H₂O (MVOH) had an interlayer spacing of 13.9 Å, much larger than that of the hydrated Mg_xV₅O₁₂ (11.9 Å) reported in [108]. The nanoflower morphology greatly increased the specific surface area of the MVOH (33.07 m²·g⁻¹) and provided short diffusion paths for Mg²⁺ from bulk solution to active sites, accelerating the insertion/extraction of Mg²⁺. The diffusion coefficient of Mg²⁺ in the MVOH calculated from the EIS data was as high as 1.08 × 10⁻⁹ cm²·s⁻¹, indicating fast migration kinetics inside the MVOH crystal lattice. In a three-electrode cell with a platinum counter electrode, a saturated calomel reference electrode (SCE), and 2 M Mg(CF₃SO₃)₂ aqueous electrolyte, the MVOH cathode exhibited a high discharge capacity of 350 mAh·g⁻¹ (at 0.05 A·g⁻¹) and 70 mAh·g⁻¹ at a higher current density of 4 A·g⁻¹. The full cell with a MVOH cathode, perylene-3,4,9,10-tetracarboxylic dianhydride (PTCDA) as an anode, and 2 M Mg(CF₃SO₃)₂ in PEG/H₂O (1:1) electrolyte, delivered a discharge capacity of 133 mAh·g⁻¹ at a current density of 0.02 A·g⁻¹ and 42 mAh·g⁻¹ at a high current density of 4 A·g⁻¹, demonstrating high rate capability and satisfactory cycling stability with a capacity retention of 62% after 5000 cycles at 4 A·g⁻¹. DFT calculations performed on the basis of structural characterization results demonstrated that pre-intercalated Mg²⁺-cations and structural H₂O molecules stabilize the layered structure of MVOH. Based on the determined pre-intercalated structure of MVOH, DFT calculations were performed to identify the energetically favorable sites for Mg²⁺ to occupy and the nearest neighboring site for Mg²⁺ to hop to. Based on the relaxed structures, the waltz-like like shuttle mechanism of H₂O molecules coordinated with Mg²⁺ transport was demonstrated during the charge–discharge of the MVOH cathode. It was shown that the Mg²⁺-ion migrates together with the coordinated water molecules, which adjust their orientation by rotating or flipping to facilitate the zig-zag migration of Mg²⁺-ions (Figure 11e). After the extraction of Mg²⁺ from MVOH, the Mg²⁺-H₂O bonds break, and the lattice H₂O molecules rearrange and form H-bonds in the interlayer of MVOH to stabilize the lamellar structure. To verify the Mg²⁺ shuttling of lattice H₂O molecules, the MVOH cathode was investigated by in situ Raman spectroscopy during discharging/charging, and a variation in the peak at 871 cm⁻¹, corresponding to the V-OH₂ bonds, at different discharged/charged states, was observed (Figure 11f). To further verify the Mg²⁺ shuttle function of lattice H₂O molecules, the fully discharged and charged MVOH cathode was analyzed by time of flight–secondary ion mass spectrometry (TOF-SIMS). The much higher intensity of the MgVO⁻ signal in the discharged MVOH than in the charged MVOH indicates successful insertion of Mg²⁺ during discharge, and the higher MgOH⁺ signal in the discharged MVOH than in the charged MVOH indicates strong coordination of lattice H₂O molecules with Mg²⁺-ions inserted during discharge, confirming the H₂O–Mg²⁺ shuttle mechanism. The same intensity and distribution of the H⁺ signal in discharged and charged MVOH indicate that the H₂O-Mg²⁺ waltz-like shuttle mechanism eliminates the co-insertion or co-extraction of H⁺ together with Mg²⁺ (Figure 11g–i).

To sum up, the investigations of vanadium oxide-based cathodes in RMBs show that the water content, even in trace amounts, has an important influence on the mechanism of intercalation processes in organic electrolytes. Although the exact mechanism of the effect of the water molecules present in organic electrolytes on the increase in the specific capacity of vanadium oxide-based cathodes is not fully understood, it has been emphasized by many researchers. The following explanations have been proposed:

(i)

Water molecules can solvate the magnesium ions and shield their divalent charge to facilitate insertion into the interlayer spaces of VO-based cathodes.

(ii)

Water molecules inserted together with magnesium ions solvate the interstitial space and inner -V=O groups, and can lubricate Mg²⁺ diffusion.

(iii)

Insertion or co-insertion of protons (from water molecules) takes place.

(iv)

Water molecules can react with the surface -V=O groups to form stable structural hydroxyl groups to reduce the migration barrier of magnesium insertion.

A better understanding of these effects is also necessary for the proper design of compatible cathodes and electrolytes.

For more illustrative comparison, the main electrochemical characteristics of selected vanadium oxide-based cathode materials are listed in

Table 2. Summary of electrochemical performance of selected vanadium oxide-based cathodes in RMBs.

Cathode Material	Synthesis Method	Morphology/ Interlayer Spacing, Å	Electrolyte	Specific Capacity, mAh·g−1 (Current Density, A·g−1)	Capacity Retention (Number of Cycles and Current, A·g−1)	Ref.
V2O5 thin film	vacuum deposition	200 nm film	0.5 M Mg(ClO4)2/AN	180 (0.15 mV·s−1)	83% (36, 0.15 mV·s−1)	[71]
α-V2O5	commercial product	powder	1 M Mg(TFSI)2/G2, 15 ppm H2O	80 (20 μA·cm−2), initial 60 (20 μA·cm−2)	94% (2–10, 0.05)	[125]
α-V2O5	commercial product	powder	1 M Mg(TFSI)2/G2, 2600 ppm H2O	250 (20 μA·cm−2)	-	[125]
α-V2O5	commercial product	powder	0.5 M Mg(TFSI)2/PY14TFSI, at 110 °C	295 (C/5)	83.8% (50, C/5)	[72]
α-V2O5	precipitation, calcination	spheres/11.52	0.2 M Mg(ClO4)2/AN	225.0 (0.01), initial 190 (0.01) 100 (0.03) 55 (0.05) 40 (0.08)	95% (50, 0.05) 87% (100, 0.05)	[77]
V2O5 film	AACVD	thin film	0.075 M MgCl2	427(5.9) 170 (2.4)	82% (2000, 5.9)	[73]
V2O5·nH2O	ion-exchange	bilayered	1 M Mg(TFSI)2/G2	50 (20 µA·cm−2)	100% (10, 20 µA·cm−2)	[92]
V2O5·nH2O	hydrothermal method	bilayered, 2D nanopapers/10.6	2 M Mg(CF3SO3)2/H2O	162.8 (0.2) 128 (0.3) 116.7 (0.5) 104.4 (1.0) 100.4 (2.0) 91.9 (4.0) 86 (6.0) 81.3 (10.0)	70.1% (70, 0.2) 88% (1000, 10.0)	[93]
VOx-NT	hydrothermal	nanotubes	0.25 M Mg(AlBu2Cl2)2/THF	75 (0.005)	-	[131]
VOx-NT	microwave- assisted hydrothermal process	nanotubes	0.5 M Mg(ClO4)2/AN	218 (0.06)	70.8% (20, 0.06)	[81]
VOx-NT	microwave- assisted hydrothermal process	open-ended multilayered nanotubes	0.5 M Mg(ClO4)2/TMS-EA	124.0 (0.06)	50% (80, 0.06)	[82]
V2O5	hydrothermal synthesis	3 μm microspheres/25 nm thick nanosheets	0.25 M Mg(AlCl2EtBu)2/THF	126.2 (0.05), initial 92 (0.1) 80 (0.15) 62 (0.2)	71.9% (80, 0.05)	[76]
V5O12·6H2O	hydrothermal synthesis	nanoflowers	0.8 M PhMgCl + 0.4 M AlCl3/THF	234.3 (0.01) 53.2 (0.5)	82.1% (400, 0.05) 74.5% (1500, 0.1)	[78]
V6O13	Microwave- assisted	2D nanosheets/monoclinic	1.0 M Mg(ClO4)2/AN	324.0 (0.02) 278.0 (0.04) 244.0 (0.06) 214.0 (0.08)	71.9% (30, 0.04)	[79]
VO2 (B)	hydrothermal process	nanorods/12.093	1.0 M Mg(ClO4)2/AN, wet	391.0 (0.025) 370 (0.05) 341 (0.1)	94.7% (10, 0.025)	[55]
V2O5 nanoclusters on porous carbon frameworks	ambient hydrolysis deposition	amorphous (XRD)/orthorhombic (51V NMR)	0.2 M [Mg2(μ-Cl)2(DME)4][AlCl4]2/DME	350 (0.04), initial 225 (0.04) 180 (0.08) 120 (0.320) 100 (0.640)	75% (50, 0.04) 67% (100, 0.320)	[74]
V2O5/GO	Solvothermal process	microparticles, /11.512	0.25 M Mg(AlCl2EtBu)2/THF	178 (0.2C), initial	78.6% (20, 0.2C)	[97]
V2O5·1.42 H2O@rGO	melt quenching process and freeze drying	nanowires /~11.3	0.5 M Mg(TFSI)2/AN	320 (0.05) 280 (0.1) ~100 (2.0)	81% (200, 1.0)	[98]
V2O5·0.4H2O/C	evaporation of solution	xerogel, bilayer	4.7 M Mg(NO3)2	88 (1.0), initial 102 (1.0), reversible 62 (2.0) 53 (3.0) 47 (4.0) 44 (5.0)	82.7% (10, 1.0)	[89]
V2O3@GO	hydrothermal/spray-drying/annealing in Ar	3D graphene- supported nanoparticles	0.3 M Mg(TFSI)2/AN	291.3 (0.05) 280.1 (0.1) 259.1 (0.2) 232.5 (0.5) 210.4 (1.0) 185.3 (2.0)	88.5% (1000, 0.5)	[99]
V2O3@C	hydrothermal/annealing in air/carbonization	mesoporous nanorods	0.3 M  Mg(TFSI)2/AN (H2O)	381.0  (0.2) 354.8  (0.5) 292.7 (5.0) 208.0 (20.0) 130.4  (50.0)	60.0% (1000, 0.5)	[100]
V2O5-PEDOT	chemical intercalation/polymerization	/9.86	0.4 M APC/THF + 0.2 M CTAB	288.7 (0.1) 153.8 (0.3) 110.7 (0.5) in full cell	58% (100, 0.1) ~68% (500, 0.5)	[119]
V2O5/PEDOT	chemical intercalation/polymerization	porous, bilayered/ 19.02	0.3 M Mg(TFSI)2/AN + 3M H2O	339.7 (0.1), initial 256.3 (0.5), initial 320.4 (0.1), reversible 271.2 (0.2) 219.6 (0.5) 172.3 (1.0) 117.5 (2.0)	60.85% (200, 0.1) 67.3% (500, 0.5)	[120]
V6O13 -PANI	solvothermal, in situ intercalation	nanobelts /13.4	0.5 M Mg(TFSI)2/DME	195 (0.1), initial 168 (0.1), 147 (0.2) 127 (0.5) 99 (1.0) 65 (5.0) 46 (10.0)	89% (100, 0.1) 100% (2500, 5.0)	[118]
PANI-V2O5 (V2O5 ⋅   0.35C6H6N   ⋅   0.92H2O)	intercalation of monomers and subsequent interlayer polymerization	2D organic- inorganic superlattices /15.3	0.3 M Mg(TFSI)2/AN, 0.2 M Mg(CF3SO3)2-MgCl2-AlCl3/DME (full cell)	280 (0.1), initial 275 (0.1), reversible 250 (0.2) 220 (0.5) 175 (1.0) 155 (2.0) 130 (4.0) 115 (0.1), full cell	70% (500, 4.0) 69.5% (50, 0.1) (full cell)	[117]
V2O5· 0.34H2O·PAN	hydrothermal synthesis	/10.63	0.5 M Mg(TFSI)2/AN	~180 (0.05), initial 143.2 (0.05) ~75 (0.5) 120.6 (0.1) (full cell, initial)	91% (18,000, 2.0) 41.5% (80, 0.1), full cell	[121]
PEO-V2O5	sol-gel synthesis	nanosheets /12.6	0.5 M Mg(ClO4)2/AN	125 (0.01), initial 100.3 (0.01)	77% (35, 0.01)	[116]
H2V3O8	hydrothermal process	nanowires /0.467 nm	0.5 M Mg(ClO4)2/AN	304.2 (0.05)	89.5% (20, 0.05)	[128]
H2V3O8	hydrothermal process	nanowires /0.34 nm	0.5 M Mg(ClO4)2/AN 60 °C	231 (0.01) 201 (0.02) 170 (0.04) 97 (0.08)	95.2% (30, 0.01)	[129]
H2V3O8	hydrothermal process	nanowires	Mg(ClO4)2· 3.1 H2O/AN	303 (0.05)	80% (30, 0.05) 61% (200, 0.1)	[130]
H2V3O8	hydrothermal process	nanowires	0.25 M Mg(ClO4)2/AN	67 (0.05)	99% (30, 0.05) 109% (200, 0.1)	[130]
NaV6O15	hydrothermal process, annealed in air	rod-like, monoclinic /10.08	0.5 M Mg(ClO4)2/AN	213.4 (0.01), initial 210.1 (0.01) 137.0 (0.02) 111.7 (0.05) 80.2 (0.1) 52.3 (0.2) 27.2 (0.5)	87% (100, 0.02)	[103]
Na2V6O16· 1.63 H2O	hydrothermal process	nanowires /7.9	0.5 M Mg(TFSI)2/DME	228 (0.02) 193 (0.03) 175 (0.05) 105 (0.08) 91 (0.1) 65 (0.2)	71% (450, 0.2)	[104]
NaV3O8	hydrothermal method, annealed in air	nanowires /6.85	0.5 M Mg(ClO4)2/AN	260.0 (0.05) 184.0 (0.1) 123.9 (0.2) 86.6 (0.5) 62.4 (1.0)	85.8% (30, 0.1) 88.3% (100, 0.5)	[101]
NaV3O8· 1.69H2O	solvothermal at RT	nanobelts	0.4 M (PhMgCl)2-AlCl3/THF	150.0 (0.01), initial 110.0 (0.01), 5 cycles 55.0 (0.02)	80% (100, 0.08)	[102]
NH4V4O10	hydrothermal method	nanosheets/11.57	0.5 M Mg(ClO4)2/AN	175 (0.2 C) 128 (0.5 C) 110 (1.0 C) 69 (2.0 C)	36.5% (100, 1C)	[106].
Zr-NH4V4O10	hydrothermal method	nanorods/11.572	0.5 M Mg(ClO4)2/AN	328 (0.04) initial 300 (0.04) 240 (0.1) 207 (0.2) 175 (0.4) 147 (0.8) 100 (0.01), full cell	86% (150, 0.04)	[105]
Mg0.3V2O5· 1.1H2O	solvothermal	nanowires (150 nm)/11.9	0.3 M Mg(TFSI)2/AN	162 (0.1) 145 (0.2) 134 (0.5) 120 (1.0) 85 (2.0) 50 (4.0)	~100% (10,000, 1.0) 80% (10,000, 2.0)	[110]
Mg0.1V2O5· 1.8H2O	sol-gel method	agglomerates of sub-micron sized particles/ 12.3	0.5 M Mg(TFSI)2/AN	300 (C/10), initial 250–280 (C/10)	93% (7, C/10)	[109]
MgxV5O12· nH2O	solvothermal at near ambient conditions	nanofibers/ 11.9	0.3 M Mg(TFSI)2/AN	~160 (0.05)	89% (300, 0.05) 81% (10,000, 2.0)	[108]
Mg0.75V10O24· nH2O	one-step precipitation method	nanoflowers/13.9	2 M Mg(CF3SO3)2/H2O	350 (0.05) 149 (1.0) 70 (4.0)	64% (100, 3.0)	[107]
Mn0.04V2O5· 1.17H2O	hydrothermal method	nanobelts/ 10.9	0.3 M Mg(TFSI)2/AN	145 (0.05) 131 (0.1) 110 (0.2) 108 (0.5) 97 (1.0) 80 (2.0) 50 (4.0)	82% (10,000, 2.0)	[111]
H11Al2V6O23.2	Hydrothermal method	Urchin-like microspheres/ 13.3	0.3 M Mg(TFSI)2/AN	165 (0.1) 127 (0.2) 98 (0.5) 75 (1.0) 60 (2.0) 47 (4.0)	98% (50, 0.1) 87% (3000, 1.0)	[113]

.

4. Conclusions

RMBs are probably one of the most promising next-generation energy storage technologies due to their high theoretical energy density, high safety, abundant resources, and low cost. The design of novel cathode materials for RMBs is the focus of investigations for the future development of RMBs.

Vanadium oxide-based cathode materials present many advantages, including low cost, high safety, and great opportunity to tune the ion transport channels, which resulted in their improved electrochemical performance in RMBs.

Although great progress has been made in the development of vanadium oxide-based cathodes for RMBs, a number of challenges remain to be overcome in the future. The main challenge in the development of cathodes for RMBs is the sluggish kinetics of Mg²⁺-ion intercalation and diffusion in electrode materials due to the strong interaction between doubly charged Mg²⁺-ions and vanadium oxide host materials.

For the development of new cathodes with improved electrochemical performance, different synthetic procedures that allow new morphological modifications of vanadium oxide-based materials, pre-intercalation of metal ions and organic (mono- and polymeric) molecules into the interlayer space can be proposed. Pre-intercalated species act as pillars between the layers, stabilize the crystal structure of the material, and prevent amorphization and dissolution of oxides during repeated insertion/extraction of Mg²⁺-ions. A pre-intercalation strategy allows the acquisition of materials with increased reversibility of Mg²⁺ (de)intercalation processes and increased interlayer/channel distances, which significantly improve the Mg²⁺-ion diffusion kinetics and possibly reduce the desolvation barriers. Deeper understanding of the effects of different pre-intercalation strategies, their optimization, and their combination with other approaches will facilitate the achievement of high-performance RMB cathodes. Further searching for possible modifiers that act as pillars represents a perspective means of cathode improvement.

Modification of vanadium oxides with highly conductive materials, such as nanocarbons, and especially conducting polymers, is also a promising means of enhancing the intrinsic conductivity of vanadium-based electrode materials and facilitating Mg²⁺ migration.

Excessive removal of interlayer water results in the loss of capacity of highly crystalline vanadium oxide structures. Mild annealing of vanadium oxides during synthesis should be considered to preserve an adequate amount of crystal water necessary to lubricate the interlayer space and weaken an interaction between oxygen sub-lattices and Mg²⁺ ions, providing electrostatic shielding. In combination with pre-intercalated metal ions or organic molecules, interlayer water also plays an important role in maintaining sufficient interlayer space. The role that traces of water in the organic electrolytes play in facilitating Mg²⁺ ion insertion also needs to be elucidated for further design of high-performance RMBs.

In addition to designing cathode materials with improved electrochemical performance, electrode/electrolyte interphase control is an important issue for stable electrochemical performance in RMBs. More attention should also be paid to the development of electrolytes compatible with cathodes and anodes.