In Situ and Operando Analyses of Reaction Mechanisms in Vanadium Oxides for Li‐, Na‐, Zn‐, and Mg‐Ions Batteries

Due to their diversity in the composition, lattice structures and physical/chemical properties, and various oxidation states (2+, 3+, 4+, and 5+), (VxOy) nanomaterials have attached much attention for developing new rechargeable batteries, including lithium‐ion batteries (LIBs), sodium‐ion batteries (NIBs), zinc‐ion batteries (ZIBs), and magnesium‐ion batteries (MIBs) as well as new energy storage concepts such as light‐rechargeable batteries. However, to further improve the electrochemical performance of VxOy‐based batteries, it is crucial to understand the various electrochemical mechanisms taking place in these materials for LIBs, NIBs, ZIBs, and MIBs. This review covers a systematical discussion of in situ and operando analysis methods carried out on V2O5, VO2, LixVyOz, NaxVyOz, ZnxVyOz, and MgxVyOz for LIBs, NIBs, ZIBs, and MIBs and the fundamental insights they have provided in the energy storage mechanisms in these batteries.


Introduction
To cope with our society's continued high energy consumption and push toward renewable energy, there is a demand for electrochemical energy storage devices with enhanced capacity. [1][2][3] Lithium-ion batteries (LIBs) are at the forefront of high energy and power density applications needed in electric vehicles and sodium-ion (Na + ), zinc-ion (Zn 2+ ), and magnesium-ion (Mg 2+ ) chemistries are promising for large-scale storage of energy harnessed from renewable energy sources. [4][5][6][7][8][9] The active electrode materials used in battery electrodes have a direct impact on the electrochemical performance including the capacity, energy and power density, cycling stability, safety, and cost. Vanadium oxide (V x O y ) nano materials have attracted increasing attention for rechargeable batteries because of its crystal structure Table 1. The properties of vanadium oxides such oxidation states and lattice system. [10][11][12][13] Figure 1. Schematic of Li-, Na-, Zn-, and Mg-ion batteries using vanadium oxides and in situ/operando analysis such as XRD, XAS, Raman, UV-vis, TEM, and AFM. [35][36][37][38][39]

In Situ/Operando XRD Analysis
XRD is based on the scattering of X-rays by periodically spaced lattices in a crystal. [40][41][42][43] XRD is used routinely to characterize and monitor the synthesis of battery materials, but it is also a powerful technique for operando monitoring of changes in the electrode material. [40][41][42][43] In situ/operando XRD analysis avoids the challenges with post-mortem XRD analysis of electrodes, such as changes of the electrode while retrieving them from the battery (e.g., short circuits) and changes to the material when exposed to air. [40][41][42][43] In situ/operando XRD analysis, on the other hand, can trace real-time phase changes and lattice parameter changes of electrode materials, which reveal additional insight into the energy storage and/or degradation mechanisms. [40][41][42][43] To measure in situ/operando XRD analysis, the cell designs should allow interaction with the X-ray beam while maintaining a perfect seal to prevent air and moisture contamination in the cell. [44][45][46] A further discussion of the design guideline for in situ/operando XRD cells as well as the operation of XRD can be found in the previous reports. [44][45][46][47][48][49] This section reviews papers conducting in situ/operando XRD analysis of batteries using different vanadium oxide formulations in LIBs, NIBs, ZIBs, and MIBs.

V 2 O 5
V 2 O 5 has been widely used as a cathode material for LIBs due to its relatively high theoretical capacity of 294 mAh g −1 compared to other cathodes such as LiCoO 2 (≈140 mAh g −1 ), LiMnO 4 (≈148 mAh g −1 ), LiFePO 4 (≈170 mAh g −1 ), LiNi 0.8 Mn 0.1 Co 0.1 O 2 (≈279 mAh g −1 ), and LiNi 0.8 Co 0.15 Al 0.05 O 2 (≈274 mAh g −1 ). [12,[50][51][52][53][54][55] It is known that V 2 O 5 has a typical layered structure, which is constructed by a periodic sequence of distorted VO 5 square pyramids (Figure 2a). [12] These layers of pyramids are stacked in the planes of (00l) with weak van der Waals interaction. [12] The intrinsic interlayer spacing is estimated at about 4.37 Å, possessing open ion diffusion channels between layers. This offers opportunities for intercalation of a range of different metal ions. [12] The Li + intercalation and de-intercalation process of V 2 O 5 can be described as follows (Equation (1)) 2 5 2 5 x x x (1) Figure 3a shows the potential profiles for the Li + intercalation and de-intercalation process of a V 2 O 5 electrode at a current density of 50 mA g −1 . [19] These measurements show a 1st cycle discharge and charge capacity of 286 and 285 mAh g −1 , respectively, with a Coulombic efficiency of 99.7%. In addition, multiple plateaus in the potential profile during the discharging process correspond with Li + intercalation into the V 2 O 5 layered structure, as explained below. To explore the structural evolution of V 2 O 5 during discharge and charge, in situ/operando XRD analyses were performed at low rate of 0.05 C (Figure 3b). [19] Figure 3b shows the (001) peak shifts toward a lower diffraction angle from 3.5 and 3.0 V (vs Li/Li + ) during discharge (Stage 1). This shift was associated with the first transformation from pristine α-V 2 O 5 to εand δ-Li x V 2 O 5 in the composition range of 0 < x < 1. [50,[56][57][58][59] When V 2 O 5 was discharged to stoichiometries of 1 ≤ x < 1.3 in the voltage window from 3.0 to 2.5 V (Stage 2), a sudden split of the (001) peak is observed. At this stage, the phase transformation from δ-Li x V 2 O 5 to γ-Li x V 2 O 5 proceeds while excess Li ions are inserted, where the VO 5 pyramids undergo an orientation change from "up-up" (denoted by the symbols "++") to up-down ("+−") ( Figure 3c). [50,[56][57][58][59] At stage 3, the intercalation of more than one Li ion (1.4 < x) leads to the transformation into ω-Li x V 2 O 5 , where the plateau was observed from 2.5 to 2.0 V. [9,19] The peaks split from the (001) plane, merging to a peak toward the same position of the OCV state, indicating ω-Li x V 2 O 5 , was going back to its pristine phase such as γ-, δ-, ε-, and α-V 2 O 5 . [35,50,[56][57][58][59] This work investigated the changes of V 2 O 5 such as lattice parameters, unit cell volume, and phases during a Li+ intercalation and de-intercalation. The intercalation of more than one Li + into V 2 O 5 is a key consideration to further increasing the capacity of LIBs.

VO 2 (B)
Monoclinic VO 2 (B) is a promising cathode material for LIBs because of its unique layered structure, by edge-sharing VO 6 octahedra with cavities, which provides a direct pathway for Li ion transport into VO 2 (B) (Figure 2b). Various nanostructured monoclinic VO 2 (B) have shown improved electrochemical results. [15,[60][61][62][63][64] To understand the enhanced electrochemical performance, investigating the structural evolution during the discharge and charge processes is important. However, the electrochemical energy storage mechanism of VO 2 (B) has not been studied extensively. Liu et al. reported Li ion insertion and de-insertion mechanism in monoclinic VO 2 (B) nanorods via in situ/operando high-energy synchrotron XRD (HEXRD) (Figure 3d-f). [15] The voltage profile of VO 2 (B) nanorods in the range of 2.0-3.6 V at a current density of 30 mA g −1 shows a pair of well-resolved cathodic and anodic voltage plateaus (Figure 3d). Figure 3e shows the 2D contour plot of the HEXRD patterns recorded at different discharge/charge states and the corresponding voltage profile. First, HEXRD patterns returned nearly to their OCV position after the first discharge and charge cycle, which indicates structural evolution is reversible. The discharge process can be divided into three different stages: I) During the discharge process, Li ions tend to be inserted into the largest sized tunnels first (i.e., the 8.2 Å 2 tunnel along the b direction) followed by the second largest tunnels (i.e., the 5.0 Å 2 tunnels along the c direction) (Figure 3f). In the initial discharge stage, the (110), (310), and (020) plane peaks move slightly to lower angles. These changes are minor, indicating that the insertion of Li ions into VO 2 (B) nanorods does not significantly change the crystalline lattices of VO 2 (B) without a phase transition, which corresponds to the sloping potential in discharge and charge curves as shown in Figure 3d. On the other hand, the (001) and (200) plane peaks are barely changed, indicating that the insertion Li ions in the largest tunnels along the b direction is favorable, which causes no significant lattice expansion. II) Next, at an intermediate state of discharge around 2.4 V, a wide-range plateau appears in the voltage profile and in the diffraction patterns the (110), (310), and (020) reflection splits into two peaks with the appearance of a peak at a lower angle, which b) Contour plot of the in situ/operando XRD patterns of V 2 O 5 during the discharge and charge processes, divided into three stages. c) Illustration of α, ε, δ, γ, and ω phases of orthorhombic V 2 O 5 . Red pyramids and green circles are V 2 O 5 and Li ions, respectively. Reproduced with permission. [19] Copyright 2021, Elsevier. d) Discharge and charge curves, e) contour plot of diffraction patterns as a function of time and f) crystal structure of monoclinic VO 2 (B) nanorods. Reproduced with permission. [15] Copyright 2018, Elsevier. g) Discharge and charge profiles of the LiV 3 O 8 material cycled at C/10 between 3.8 and 2.0 V, h) expanded view of selected peak regions, i) crystal structures of LiV 3 O 8 , Li 2.5 V 3 O 8 , and Li 4 V 3 O 8 . Reproduced with permission. [27] Copyright 2017, American Chemical Society.
corresponds to a new Li y VO 2 phase. III) Finally, at deep discharge, only Li y VO 2 remains and the (110), (310), and (020) peak position shifts continuously to lower angles as the discharge progresses, suggesting that the Li y VO 2 lattice expands further as increasing amounts of Li ions are inserted in the material.

LiV 3 O 8
LiV 3 O 8 (LVO) is a monoclinic layered material that is a promising cathode material for LIBs because of its highrate performance and relatively high theoretical capacity of 362 mAh g −1 (Figure 2c). [65][66][67][68][69] LVO has an excess vacant site in the interlayer of the layered structure, supporting the occupation of more ions. [66] Second, the weak LiO bond, which arises from the lower electron density of the O 2ions in the lattice, has been suggested to permit the hopping of ions and then a fast ion diffusion in LVO. [66][67][68][69] Zhang et al. reported the voltage profiles of the 1st, 50th, and 100th cycles to investigate the electrochemical stability of the LiV 3 O 8 electrode, which shows capacity fade occurring predominantly around the 2.6 V plateau (Figure 3g-i). [27] This capacity fade is believed to be associated with a two-phase transformation process. To monitor these phase transformations, in situ/operando XRD was conducted during a full discharge and charge cycle in the range 3.6-2.0 V (Figure 3h). Figure (111) of β-phase, while α-phase (100) reflections decreased in intensity, indicating a two-phase transformation. At the fully discharged state, the α-phase entirely transformed to β-phase, and the β (100) and β(111) peaks shifted to a higher and lower angle, respectively. However, in situ data showed the V 3 O 8 layer shrinks by 0.3 nm when transforming from Li 2.5 to Li 3.5 to form β-phase. The degree of contraction increased during cycling, [27] which could be linked to the capacity fade observed in the cycling tests.

NaVO 2
Among various electrode materials for NIBs, layered vanadium oxides have gained significant attention owing to their high theoretical capacity and natural abundance. In addition, the vanadium oxides have sufficiently large interlayer spacing to accommodate Na ions, which are larger than Li ions into their lattice structure. However, vanadium oxides still suffer from poor ion diffusion coefficient, poor electronic conductivity, and large volume fluctuation. [21,65,[70][71][72] To improve electrochemical performance of vanadium oxides for NIBs, recent studies have doped Na into the host material. [28,29,65,[73][74][75] For instance, Claude's group synthesized the layered Na x VO 2 systems and studied the structural transitions during intercalating and de-intercalation in NIBs with in situ/ operando XRD experiments. [30,76] To understand the structural changes and the correlation with the electrochemical reaction of this material, shifts in XRD peaks are tracked in situ and are shown alongside electrochemical galvanostatic curves in Figure 4a. In the 0.67 > x > 0.50 composition range, the (002) peak shifts toward a lower diffraction angles as Na ions are de-intercalated. This corresponds to a gradual increase of the d-spacing (2·d (002) = c·sinβ) due to increasing oxygen repulsions as the amount of Na ions decreases between the layers. In the 0.67 > x > 0.64 composition range, the (002) peak position shifts continuously to lower angles as the charge progresses. The (002) peak broadens between 2.10 and 2.23 V, which might be due to the existence of a bi-phase domain for the two phases. [76] A similar observation is made for the plateau between 0.62 > x > 0.55. The plateau shown in the 0.55 > x > 0.50 composition range from 2.27 V is the characteristic of the change from the two phases to single phase by the gradual change of the superstructure peaks intensities (Figure 4a,b). The presence of a superstructure is proof of a structural change from NaVO 2 to Na 1/2 VO 2 .

NaV 6 O 15
NaV 6 O 15 can be formed by inserting one-third moles of Na + ions into V 2 O 5 units (i.e., Na 1/3 V 2 O 5 ), where the preintercalated Na + ions between 2D V 4 O 12 layer act as pillars to stabilize the structure upon insertion/extraction of Na + ions. [28,29,[77][78][79] In addition, these pillars enhance not only the interlayer space of materials but also the ion diffusion rate. [29,80,81] However, monitoring structural evolution of sodium vanadium oxides and corresponding understanding electrochemical performance in real time condition has not been explored much. Dong et al. reported structure evolution of NaV 6 O 15 electrode during Na + intercalation via in situ/operando XRD. [28] Figure 4c shows the voltage plateaus of the NaV 6 O 15 electrode between 2.7 and 3 .0 V (vs Na/Na + ) during discharging and charging. In addition, even at current densities of 10 A g −1 , capacities of 96.5 mAh g −1 were retained (131.5 mAh g −1 at 0.1 A g −1 ). To understand the reaction mechanism of NaV 6 O 15, in situ/ operando XRD analysis was carried out (Figure 4d). The (200) plane gradually shifted to the lower angle during discharging while after charging, the (200) plane completely returned to the position of the open circuit voltage (OCV), where no new phase was observed. The high reversibility of the electrodes is attributed from single-phase charge-discharge mechanism during cycling, which causes good cyclic retention of 87% after 2000 cycles. In situ/operando XRD analysis of vanadium oxides for NIBs has still rarely reported. Therefore, more analysis needs to be done on other vanadium oxides as well to understand the structural and lattice evolution. In addition, organic electrolytes for NIBs tend to result in limited capacities, cycle stabilities, and rate capabilities. This is because of the low conductivity of the organic electrolyte and a lack of understanding of the Na-ion storage. For more details on the current challenges and developments on NIBs, we refer the readers to review papers such as. [6,13,75]

Zn 0.25 V 2 O 5 •nH 2 O
Manganese-based oxides and Prussian blue materials have shown reversible Zn 2+ intercalation and therefore have been investigated as cathode materials for ZIBs. However, these have some limitations in rate performance and specific capacity. [82][83][84][85] Kundu et al. reported that the Zn 0.25 V 2 O 5 ⋅nH 2 O nanobelt possesses two-electron redox centers (vanadium) and a layered structure containing intercalated divalent cations and water, showing structural stability during cycling with high specific capacity. [82] Ref. [86] shows the galvanostatic discharge and charge profile of a Zn 0.25 V 2 O 5 •nH 2 O nanobelt at a current density of 300 mA g −1 in 1 m ZnSO 4 , using a Zn anode. A capacity of 282 mAh g −1 is measured in the 1st discharge, corresponding to a 2.2 electron redox process, of which 278 mAh g −1 is recovered upon charging. In the subsequent 50 cycles, the voltage profile shows minor changes such as a small decrease in capacity, suggesting only slight structural changes. Figure 5a shows in situ/ operando XRD measurement during the 2nd cycle to investigate the insertion mechanism in Zn 0.25 V 2 O 5 . The discharge process can be separated into three different stages (0.25 < x < 0.55, x ∼ 0.55, and 0.55 < x < 1.4). Up to x ∼ 0.55, there is only a slight contraction of the interlayer distance, consistent with an increased screening of the interlayer electrostatic repulsion as the inserted Zn 2+ increases. From x = 0.55, the interlayer distance sharply contracts from 12.3 to 11.0 Å, which is similar to the 10.8 Å interlayer spacing observed in the pristine material before the H 2 O intercalation from the electrolyte, which is attributed to the expulsion of water from the interlayer. In the 0.55 < x < 1.4, the intensity and position evolution of the peaks at 26°, 32°, and 40°, clearly show that the structure continues to evolve during Zn 2+ insertion.

VO 2 (B)
VO 2 (B) has been used as a cathode material for ZIBs because of its layered and tunnel-type structure, which provides the benefit of fast intercalation and de-intercalation of zinc ions (0.74 Å). [14,86] The intercalation behaviors of VO 2 (B) in organic and aqueous electrolyte including Li and Na ions, have been widely investigated, while those in Zn 2+ aqueous electrolytes are reported less. [14,63,[87][88][89] Figure 5b shows a high and stable reversible capacity of 357 mAh g −1 was achieved at 0.25 C. To study the Zn 2+ intercalation behavior into VO 2 (B) electrodes, Junwei et al. conducted in situ/operando XRD measurements, monitoring the structural change during the discharge and charge process (Figure 5c). The 2D contour map of the diffraction patterns collected during discharging exhibits a symmetric pattern to that during charging, suggesting a reversible . a) In situ/operando XRD obtained during charging process of NaVO 2 . The evolution of (002) peak is aligned to the galvanostatic charge curve. b) Illustration of the structural change from NaVO 2 to Na 1/2 VO 2 . Reproduced with permission. [76] Copyright 2012, American Chemical Society. c) Charge and discharge curves of NaV 6 O 15 , d) in situ/operando XRD patterns during charging and discharging process of NaV 6 O 15 . e) Illustration of the crystal structure of typical NaV 6 O 15 , projection of NaV 6 O 15 along the b-axis, the connection way of [VO 6 ] octahedral and [VO 5 ] square pyramids, the sites occupied by sodium ions. Reproduced with permission. [28] Copyright 2021, Elsevier.

LiV 3 O 8
The layer structured LiV 3 O 8 (LVO) is an interesting electrode material for ZIBs because of the fast ion diffusivity and structural stability. [66,69] As discussed above, LiV 3 O 8 is a monoclinic layered material that is a promising cathode material for LIBs because of its high-rate performance and relatively high theoretical capacity of 362 mAh g −1 . In fact, the ionic radii of Li + and Zn 2+ are similar at 0.76 and 0.74 Å, respectively, which indicates that the Zn 2+ intercalation system of LVO is also possible. [66] Alfaruqi et al. synthesized a layered-type LVO as an intercalation cathode for ZIBs with high capacities. Figure 5e shows the discharge and charge curves at a current density of 16 mA g −1 in the potential range of 0.6-1.3 V. Two plateau regions were observed in the discharge curves, indicating Zn 2+ intercalation. During the 1st discharge, the specific capacity for the Zn 2+ intercalation is slightly higher than that during the de-intercalation of the 1st charge. However, from the 2nd cycle, the amount of intercalated Zn 2+ is almost the same with that of de-intercalated Zn 2+ . The phase and structural change of LVO was confirmed by in situ/operando XRD (Figure 5f). During discharge to 0.6 V, the evolution in (110) facet is significant because the lattice  [82] Copyright 2016, Springer Nature. b) Cycling performance of VO 2 (B) nanofibers at 0.25 C (inset: typical galvanostatic charge-discharge curve). c) 2D contour map of in-situ/operando XRD peaks intensities versus galvanostatic charge and discharge times. d) Schematic view of Zn 2+ intercalation/de-intercalation VO 2 (B) nanofibers projected along the direction of the b-and c-axis. Reproduced with permission. [14] Copyright 2018, Wiley-VCH. e) Initial five voltage profiles of the LiV 3 O 8 electrode at a current density of 16 mA g −1 . f) Electrochemical discharge profile of the fabricated spectro-electrochemical cell cycled within 1.2-0.6 V at a current rate of 16 mA g −1 and in situ/operando XRD patterns obtained within selected scanning angle (2θ) domains of 13.2-18°. g) Schematic of the Zn-intercalation mechanism in the present LiV 3 O 8 cathode. Reproduced with permission. [66] Copyright 2017, American Chemical Society.
parameter "a" is largely influenced by the amount of Zn 2+ . The discharge process can be divided into three regions. In the 1st region from 1.28 to 0.82 V, the (100) peak shifted to a higher angle, which is indicative of a single-phase domain displaying a solid-solution behavior, with a corresponding gradually sloping potential curve (Figure 5e-g). In the 2nd discharge region (0.81-0.7 V), the (100) peak splits into two peaks, which is attributed to the formation of a new Zn-rich phase like ZnLiV 3 O 8 during Zn 2+ intercalation. In the 3rd region for final discharge to 0.6 V, the (100) peak returns to a single peak, indicating a change from two-phase to a single-phase domain. During subsequent charging, the (100) peak does not entirely return to the same position as the pristine electrode, which relates to the difference in capacity the number and location of plateaus between the 1st charge and discharge process in the voltage profile (Figure 5e). This work suggests the formation of a stoichiometric ZnLiV 3 O 8 phase with a maximum discharge capacity of 256 mAh g −1 at a current density of 16 mA g −1 .

V 2 O 5
MIBs have been considered as an alternative to replace LIBs because of various advantages of Mg metal such as abundance, low cost, and safety. [90][91][92][93] However, active electrode materials such as the Chevrel phase is suffering from slow ion diffusion. [90] V 2 O 5 is used as a cathode material for MIBs because of its interstitial vacancies being large enough to insert Mg 2+ ions into their layered structure. [90,94,95] However, the V 2 O 5 capacity is limited to ≈180 mAh g −1 as electrodes for MIBs. [29,[95][96][97] To understand this limited capacity, investigation of the structural changes of V 2 O 5 in MIBs is crucial, [95] yet only few publications have been published on this topic to date. A recent paper provides information on the reversible crystal structure evolution and understanding the capacity of V 2 O 5 nanowires for MIBs. Figure 6a shows the voltage profile of the V 2 O 5 electrodes in a Mg(ClO 4 ) 2 /AN electrolyte at C/20 during cycling. The 1st discharge curve exhibited a short plateau at around 0.9 V and a long plateau at around 0.4 V while only one flat plateau is observed at around 1.1 V for the 1st charge curve. In addition, the capacities at 1st discharge and charge were 103 and 110 mAh g −1 , respectively. The 2nd cycle also shows a similar behavior with a lower plateau compared with that of the 1st discharge. The 5th cycle shows a different voltage profile and an enhanced capacity compared with that of 1st and 2nd cycles, which might be attributed to an irreversible phase change during cycling. To understand the electrochemical mechanism during the 1st and 2nd cycles, in situ/operando XRD was performed. Figure 6b shows contour maps in the selected ranges during the first two cycles. In the initial part of the first discharge, Mg 0.14 V 2 O 5 is formed (region I). As discharging progresses, most peaks such as (200), (101), (400), (001), and (301) shifted to lower angles, indicating an expansion of the unit cell from gradual Mg 2+ insertion (Figure 6b,c). At the end of discharge, new reflections appear while those from the Mg 0.14 V 2 O 5 phase decreased, indicating that a new phase like Mg 0.6 V 2 O 5 with Mg-rich Mg 0.49 V 2 O 5 is present (region II). During charge process, the intensities of the Mg-rich phase gradually decreased while intensities of Mg-poor phase increased. The positions of these reflections remained at a slightly lower angle than that for the OCV state, which is in agreement with the capacity irreversibility discussed above (Figure 6a). This paper reported the structural change for only the first and second cycle, and it would be interesting to analyze the structure of the material after the fifth cycle.   [24,98,99] However, their rate capability and cycle life are limited due to poor electrical conductivity and poor structural stability, resulting from larger interlamellar spacing due to excess water between layers. [24,82,100,101] To overcome these limitations, Xu et al. developed a Mg 2+ preintercalated hydrated vanadium oxide (Mg 0.3 V 2 O 5 ⋅1.1H 2 O). [24] In addition, this work studies the structure and reaction mechanism via in situ/operando XRD. The charge and discharge curves of the electrodes at a current density of 100 mA g −1 in the potential range of 1.4-3.4 V (vs Mg/ Mg 2+ ) are shown in Figure 6d. For the discharge curve, three plateaus were observed at around 3.02, 2.25, and 1.40 V, which is indicative of a multistep Mg 2+ intercalation behavior in the layer structure. The discharge capacity of Mg 0.3 V 2 O 5 ⋅1.1H 2 O is 164 mAh g −1 , corresponding to 0.5 Mg 2+ ions per formula unit, which is higher than the 114 and 91 mAh g −1 reported for V 2 (Figure 6e). During discharging, the main (001) diffraction peak at 8.7° 2θ shifted slightly to a higher angle as Mg 2+ is intercalated into electrode material. Meanwhile, the (003) reflection also moved toward a higher angle from 25.7° to 26.5° 2θ. At a full discharge state, the interlayer spacing of layer Mg 0.3 V 2 O 5 ⋅1.1H 2 O decreased from 11.9 to 11.4 Å, which represents a small lattice change of 4.2%. It is derived from a benefit of the pillaring effect of the preintercalated Mg 2+ to maintain structural stability. During the following charge process, the interlayer spacing and peak position showed a full recovery from Mg 2+ extraction. In addition, in situ/operando XRD patterns for the 2nd discharge and charge demonstrated the high reaction reversibility and stable crystal structure.

In Situ/Operando X-Ray Absorption Spectroscopy (XAS)
XAS is used to determine the local geometric and/or electronic structure of materials. The technique can be divided into XANES and EXAFS. XANES provides information on the chemical state of the absorbing atoms, local symmetry, and chemical bonding, while EXAFS is sensitive to the spatial arrangements of nearest neighbors around the absorbing atoms. Since these experiment usually requires synchrotron X-rays, there are fewer examples of in situ/operando XAS studies of vanadium oxide battery materials than for XRD, which can be performed both on lab-based instruments and at synchrotrons. Compared to ex situ measurements, in situ/operando XAS offers similar advantages as for XRD, namely, measurement without sample exposure and the possibility to measure several points during electrochemical cycling using a single sample thus avoiding subtle differences in sample preparation. Thus, in situ/operando XAS can provide more reliable and precise data for analysis. In situ/operando XAS cells are very similar to those for XRD analysis. However, XAS analysis is mostly measured in the transmission mode. [44,102,103] Therefore, the cells should be designed with two X-ray transparent windows at the top and bottom of the cell and thick electrode. Further details on these cells as well as the principles of XAS are provided in other papers. [44,[46][47][48][49]102,103] This section reviews papers conducting in situ/operando XAS analysis of vanadium oxide-based cathodes in LIBs, ZIBs, and MIBs.

V 2 O 5
During lithium insertion into V 2 O 5 , the position of the XANES main edge continuously shifts to a lower binding energy which is indicative of a decrease in the average valence state of vanadium (Figure 7a). [104][105][106][107] Comparison of the edge position with standards (crystalline V 2 O 5 , VO 2 , and V 2 O 3 for V 5+ , V 4+ , and V 3+ , respectively) shows that pentavalent vanadium is initially reduced to tetravalent vanadium during the intercalation of 0-2 Li + per V 2 O 5 , and then tetravalent vanadium is reduced to trivalent vanadium upon additional insertion of Li + . [106] In the fully discharged state, a shoulder emergesred arrow in Figure 7a-which is consistent with the reduction of tetravalent vanadium to trivalent vanadium. [104][105][106] The pre-edge peak also shifts to lower energy upon lithiation in response to the decreasing oxidation state of vanadium. [106,107] Simultaneously, the pre-edge peak decreases in intensity. Within a specific coordination geometry, the area under the pre-edge peak shows a strong dependence on the oxidation state and local structure of vanadium, such as the degree of local distortion of the first coordination sphere. Therefore, in the reduced state the local structure of vanadium has a higher degree of octahedral symmetry. [106,107] Mansour et al. gave a detailed analysis of the EXAFS during the first discharge of V 2 O 5 . [106] The first peak in the Fourier transform in the 1-2 Å range is related to the first shell V-O coordination, and the second peak in the 2-3 Å range to the second shell V-V coordination. [106,107] While the V-O distance remains relatively unchanged during lithiation, the increasing amplitude of the first shell coordination indicates an increasing degree of local symmetry and hence a narrower range of V-O distances. There are significant and complex changes to the second shell V-V coordination, but overall the V-V distance decreased with increasing lithium content in line with the changes seen between standards V 2 O 5 , VO 2 , and V 2 O 3 , i.e., transitioning from pentavalent to tetravalent to trivalent vanadium oxidation state. [106] The significant changes in the local geometry upon lithium intercalation are proposed to be a key contributor to the poor cycle life of V 2 O 5 . Unfortunately, Liu et al. do not include EXAFS analysis for their reported graphene-modified nanostructured vanadium pentoxide hybrids. [104,105] Finally, on the first cycle the changes in the XANES for V 2 O 5 are reported to be highly reversible. [104,105,107] Upon charge, the energy shifts of the pre-edge peak and main edge, and the intensity of the pre-edge peak, return very closely to their original value indicating good electronic reversibility in the first cycle.

VO 2 (B)
XANES spectra collected during lithium insertion into VO 2 (B), with a starting V 4+ oxidation state, also reveal a continuous shift of the main edge toward a lower binding energy (Figure 7b). [15,111] As was also seen for V 2 O 5 , a shoulder forms in the main edge and pre-edge nearing the discharged state, confirming the presence of both tetravalent and trivalent vanadium oxidation states concurrently. [15,36] The pre-edge also decreases in intensity, as was seen for V 2 O 5 , indicating a higher degree of local symmetry. [15,36] The shift in the main edge is found to be morphology dependent, with VO 2 with a smaller particle size (made by microwave-assisted synthesis) yielding a greater shift in edge position-i.e., greater reduction of V 4+ -and higher discharge capacity compared to VO 2 from solvothermal synthesis. [36] Linear combination analysis (LCA) using VO 2 and V 2 O 3 standards shows the valance state changed from V 4+ to V 3.4+ , equivalent to a lithiation capacity of 194 mAh g −1 and consistent with the reported electrochemistry. [15] Trends in the XANES spectra collected on the first and second cycles, e.g., the edge position and the intensity of the pre-edge and resonance peak as shown in Figure 7c, show good early reversibility of the VO 2 electronic structure. [15] Unfortunately, to our knowledge, there has been no thorough in situ investigations of the EXAFS for VO 2 during Li intercalation and de-intercalation.

V 2 O 5
Zhao et al. used in situ/operando XAS to investigate NH 4+ preintercalated V 2 O 5 ·nH 2 O as a cathode for aqueous rechargeable Zn-ion batteries. [108] Despite the different electrolyte system, the trends in the XANES during intercalation of Zn 2+ into V 2 O 5 (1 m ZnSO 4 electrolyte) are similar to that described above for Li + intercalation in organic carbonate-based electrolyte (e.g., 1.2 m LiPF 6 in ethylene carbonate (EC) and ethyl methyl carbonate (EMC) 3:7 by weight). In brief, during Zn 2+ intercalation the main edge and pre-edge shift to lower energies and the preedge decreases in intensity; these changes are reversed on Zn 2+ extraction during charge (Figure 8a,b). As expected, the changes in the EXAFS reveal a strong perturbation of the local structure around vanadium during Zn 2+ intercalation (Figure 8c,d). An increase in the amplitude of the first shell V-O again highlights the increasing symmetry during vanadium reduction. In contrast to Li + insertion into V 2 O 5 , which showed the greatest change in the second shell V-V feature, Zn 2+ insertion mainly affects the first V-O shell structure, inducing a shift to lower radial distances while the V-V shell structure only shows changes in intensity.

VO 2 (B)
The same groups represented in ref. [108] also reported in situ/operando XAS of Zn 2+ insertion/de-insertion with a VO 2 (B) cathode in aqueous 1 m ZnSO 4 electrode. [109] Again, the trends in the XANES (Figure 8e) closely reflect those for Li + insertion, namely, the pre-edge and main edge shift toward lower energy as tetravalent vanadium is reduced to trivalent vanadium and the pre-edge decreased in intensity with the increase in local symmetry. LCA with VO 2 and V 2 O 3 standards suggests reduction of V 4+ to V 3.26+ and a fully discharged composition of Zn 0.37 VO 2 , which is consistent with the 255 mAh g −1 of capacity measured in the electrochemistry (i.e., V 3.21+ and Zn 0.395 VO 2 ). Zn 2+ insertion into VO 2 strongly modifies the local structure, as shown in the phase uncorrected Fourier transform EXAFS spectra in Figure 8f. The V-O feature increases in intensity and shifts to shorter radial distances indicating a higher degree of symmetry and shorter VO bond length, while the V-V coordination sphere shows changes in shape and position reflecting structural changes due to Zn 2+ intercalation.

V 2 O 5 and Mg 0.3 V 2 O 5
In situ/operando XAS of vanadium oxides in MIBs has been reported recently for V 2 O 5 . [90] The pre-edge and main edge progressively shift to lower energy on discharge (Figure 9a) and the reverse occurs on charge (Figure 9b), indicating the reversible reduction of pentavalent vanadium and the Figure 7. a) In situ/operando XANES spectra at the V K-edge of a V 2 O 5 -graphene cathode during the first discharge process.Reproduced with permission. [104] Copyright 2015, Nature Communications. b) In-situ/operando XANES spectra at the V K-edge characterization of VO 2 (B) during the first discharge process. [15] c) Dependence of intensity of the XANES pre-edge and edge resonance (main peak, inset) on composition in the first charge/discharge cycle. [15] insertion/de-insertion of Mg 2+ with V 2 O 5 . In the discharged state, the edge position is stated to be "in the middle between that of standards V 2 O 3 and V 2 O 5 " and closely matches VO 2 , although the reported discharge capacity suggests a composition of Mg 0.51 V 2 O 5 and vanadium oxidation state closer to 4.5+. In LIBs and ZIBs, the pre-edge feature for both V 2 O 5 and VO 2 showed a significant decrease in intensity in the discharged state-see Figures 7a,c and 8b,e, respectively. [104][105][106][107][108] By comparison, the pre-edge intensity changes very little during Mg 2+ insertion into V 2 O 5 (Figure 9a,b), which is possibly related to the lower degree of vanadium reduction for Mg 2+ insertion (from electrochemistry: V 4.5+ ) compared to Li + (V 3.5+ ) and Zn 2+ (between V 3+ and V 4+ ). [90,104,105,108] The in situ/ operando EXAFS spectra are not reported in ref. [90], although ex situ data for a Mg 0.3 V 2 O 5 ·1.1H 2 O cathode (pristine composition) discharged to 164 mAh g −1 or the equivalent of ≈0.5 Mg 2+ per formula unit has been recently reported by Xu et al. [24] The Fourier transform EXAFS spectra in Figure 9d (with corresponding XANES in Figure 9c) show an increase in the length of the VO bond while the VV distance marginally shortens, which again shows much milder changes in the local structure compared to that seen in LIBs and ZIBs.

LIBs
Raman spectroscopy is a powerful technique for detecting variations in local disorder, bond lengths, and bond angles. [41,110] The cells for in situ/operando Raman need to be designed with a window for laser light to reach the electrode. [44,73,74,111] A more detailed description of cell designs and the use of Raman for batteries can be found in refs. [44,73,74,111]. Baddour et al. reported the effect of Li insertion/de-insertion into orthorhombic crystalline V 2 O 5 thin films using in situ/operando Raman spectroscopy. [109] Figure 10a shows the discharge-charge curve of V 2 O 5 from 3.8 to 2.8 V at a C/3 rate, which exhibits two well-defined insertion processes at about 3.4 and 2.3 V. In addition, this insertion and de-insertion reaction is reversible with a high Coulombic efficiency around 99%. To confirm the change of the local structure for the insertion/de-insertion of Li into V 2 O 5 , in situ/operando Raman analysis was studied as shown in Figure 10b,c. An assignment of the Raman spectrum has been previously reported on the local structure of V 2 O 5 characterized by a coordination of vanadium with oxygen. [110,111] A   Figure 8. V K-edge XANES of the NH 4+ preintercalated V 2 O 5 ⋅nH 2 O electrode during a) discharge and b) charge processes. Fourier transforms of V K-edge EXAFS (k 3 -weighted) during c) discharge and d) charge processes. Reproduced with permission. [36] Copyright 2020, American Chemical Society. e) In situ/operando V K-edge XANES spectra of the VO 2 (B) electrode during discharge, and f) Fourier transforms (k 3 -weighted) of the V K-edge EXAFS as a function of state of discharge. Reproduced with permission. [108] Copyright 2020, Royal Society of Chemistry.
complete recovery of the Raman spectrum was shown after the insertion and de-insertion of Li ions (Figure 10b 1) A decrease in the intensity at 145 cm -1 and a peak shift from 145 to 153 cm -1 are observed, which corresponds to an increase of disorder within the V 2 O 5 layer as Li ions are inserted. 2) Several modes at 404 and 527 cm -1 shift toward higher wavenumber, which is consistent with a stiffening of the lattice along the a direction, where the displacement of oxygen takes place.
3) The vanadyl stretching mode of pristine V 2 O 5 is observed at 997 cm -1 . During Li insertion, this VO stretching mode shifts to 980 cm -1 . It suggests the spacing in the layers in the c direction has increased. In addition, a new peak is observed at 958 cm -1 at higher Li insertion levels (x ≥ 0.75), which is related to the formation of a new Li-rich phase. [110] The above bond changes such as intensity and location return to the original state after Li insertion and de-insertion. Baddour et al. reported in situ/operando XRD analysis only for the first cycle. However, investigations into multiple cycles are also very important to confirm reversible structural change of V 2 O 5 . Zhang et al. studied in situ/operando Raman during insertion and de-insertion of Li ions into V 2 O 5 nanowire for four cycles [43] Figure 10d,e shows the in situ Raman test for four cycles (voltage window 2.0-4.0 V, current density 100 mA g −1 ) and it is observed that the peaks that represent б(O3-V-O2), ν(d3) and ν(d1), which are at around 200, 700, and 1000 cm -1 , decrease rapidly with cycle number. This may suggest the distortion of crystallinity of the local structure of V 2 O 5 . Further, the Raman band at around 980 cm -1 shift toward lower wavenumbers and widen, which means that Li insertion into V 2 O 5 expands the shortest VO bond along the b axis.

ZIBs
Mai and co-workers measured in situ Raman to evaluate the structural transformation of V 2 O 5 cathode during intercalation and de-intercalation of Zn ions and Figure 11a shows the schematic illustration of the measurement set up. [37] As shown in Figure 11b (third charge/discharge curve in voltage range of 0.3-1.4 V and respective Raman spectra at different state of charge), particularly the intensities of the Raman peaks associated with vanadyl oxygen stretching mode (≈1034 cm -1 ) and doubly coordinated oxygen stretching mode (≈766 cm -1 ) increase during charging process (de-intercalation of Zn ions), and then gradually decreases when discharged to 0.3 V. The insertion of Zn ions into layered V 2 O 5 weakens the intensity of the Raman signals and the Raman signal is reversible when de-intercalating Zn ions from V 2 O 5. This suggests that the electron transfers in the redox couple of V 5+ and V 4+ are reversible during the discharge and Figure 9. a,b) In situ/operando XANES for the V K-edge during the first discharge and charge process; the isosbestic points are indicated by red arrows. Reproduced with permission. [90] Copyright 2019, American Chemical Society. c,d) V K-edge XANES spectra of Mg 0.3 V 2 O 5 ·1.1H 2 O and the corresponding Fourier transform extended X-ray absorption spectroscopy (FT-EXAFS) spectra in R space. Reproduced with permission. [24] Copyright 2019, Elsevier.
charge. Further, Li et al. studied in situ Raman of the K 2 V 3 O 8 cathode in ZIB system at different states of charge. [112] It is observed that the intensity of the characteristic band located around 891, 926, 464, and 526 cm -1 assigned to the K 2 V 3 O 8 cathode change noticeably in charge and discharge processes. For instance, the peaks intensities first weaken and then reach a maximum at the end of the charge stage as shown in Figure 11d,e. This change of peak intensities is related to the deformation of the [VO 4 ] tetrahedron and the [VO 5 ] pyramids. [112,113] It is attributed from the extraction of K + in the first charge.

In Situ/Operando Combined Optical Characterization and UV-Vis, TEM, and AFM of V 2 O 5
The optical characterization and UV-vis measurements of V 2 O 5 have been carried out to understand changes in the bandgap of these materials as a function of their state of charge. This is important for applications in light rechargeable batteries which have been demonstrated for V 2 O 5 in both LIB [38] and ZIB [114] systems. As illustrated in Figure 12a,b, the color of V 2 O 5 changes from orange to light gray when discharged to 2 V which implies changes in the optical properties of V 2 O 5 . The UV-vis analysis (Figure 12c) suggests that insertion of Li ions into V 2 O 5 increases the bandgap followed by shifting the optical absorption peak toward higher energies. The design of in situ/operando cell was shown in refs. [38,44]. It was mentioned earlier that the phase transformation from the α-phase into ε-phase undergo when insertion of Li [115,116] This change in the optical properties of V 2 O 5 under Li ions insertion is reversible, where the Li ions inserted light gray colored V 2 O 5 changes to its original orange color when extracting Li ions (Figure 12b). In comparison, VO 2 (B) ZIB electrodes did not show significant changes in optical properties as a function of the state of charge based on ex situ measurements. [117] Gu et al. probed the insertion/de-insertion behavior of Li ions in V 2 O 5 nanoribbons by in situ/operando TEM. For in situ/operando TEM analysis, it is the important to keep the  [109] Copyright 2009, Elsevier. d,e) In situ/ operando Raman spectra of V 2 O 5 for four cycles at a current density of 100 mAh g −1 . Reproduced with permission. [43] Copyright 2019, Springer Nature. cells sealed and thin with a high vacuum system. [44] After Li insertion for 180 s with a constant voltage of −2 V, the width of the V 2 O 5 nanoribbon changes from 28 to 31 nm (Figure 12d,e). The ratio of expansion is estimated to be 10% along the (100) direction. Significant volume expansions are mainly observed along the (100) direction because Li ions are mainly stored between the (200) layers. The inset of Figure 12f shows the diffraction pattern of the V 2 O 5 nanoribbon after 180 s of Li insertion, which matches the ω-Li 3 V 2 O 5 cubic phase. It is attributed to the large lattice spacing of 5.77 Å between the (200) planes in the Li channel (Figure 12d). The inset of Figure 12f shows the V 2 O 5 nanoribbon returns to the origin V 2 O 5 with crystal planes (110) and (200) after de-insertion up to 4 V.
AFM is a scanning method for measuring high-resolution surface topology, and this method can measure sample in a liquid environment relatively easily because this analysis does not use electron beams and vacuum system. [39] Like the other analytical cells, a window of in situ/operando AFM cells need to be properly sealed to prevent evaporation of the electrolyte. [44] More detailed design guidelines for AFM cells were reported in refs. [44,118]. Therefore, it allows the monitoring of nanometric grains during electrochemical reactions [ = ]. Aurbach and Cohen reported unique surface phenomena of V 2 O 5 in LiPF 6 and LiClO 4 for LIBs using in situ/operando AFM. [39] Figure 12g shows topography images of V 2 O 5 in a LiPF 6 /PC solution as electrolyte at three stages: pristine, full insertion, and full de-insertion during the first cycle. After Li full insertion, nanosized round-shaped grains appear on the boundaries of the V 2 O 5 particles, with an average diameter of 20 nm. These particles remained after the full de-intercalation, indicating the irreversible reaction. On the other hand, V 2 O 5 in a LiClO 4 /PC solution has no significant change in the micromorphology during Li insertion/de-insertion (Figure 12h). This study demonstrated the impact of the passivation phenomena using images of surface films formations in LIB systems.

Summary
V x O y is an interesting cathode material for LIBs, NIBs, ZIBs, and MIBs, though it should be cautioned that these electrodes only work with a metal anode, which limits their application. Because of the various compositions, lattice structure, and material morphologies that can be achieved, various V x O y materials can be synthesized including In order to further improve and commercialize these V x O y , an in-depth understanding of the science aspect such as structural degradation mechanisms, charge storage mechanisms, and interfacial electrode/electrolyte reactions relating to LIBs, NIBs, ZIBs, and MIBs is crucial. In situ/ operando analysis can reveal the real-time structural change, which can be linked directly with the electrochemical properties. Figure 11. a) Schematic illustration for the testing set-up of in situ Raman spectroscopy. b) In situ/operando Raman spectroscopy results and the corresponding charge and discharge profile of the third cycle at 100 mAh g −1 , c) schematic illustration of insertion/extraction of Zn 2+ in layer structured V 2 O 5 host. Reproduced with permission. [37] Copyright 2020, Elsevier. d) Electrochemical in situ/operando Raman mapping and e) waterfall plot of the cathode of Zn/K 2 V 3 O 8 ZIB for the initial cycle. Reproduced with permission. [112] Copyright 2020, American Chemical Society.
This review paper provides an overview of in situ/operando studies using XRD, XANES, UV-vis, Raman, TEM, and AFM to study V x O y cathodes in LIBs, NIBs, ZIBs, and MIBs, although most in situ and operando studies focus on LIBs. The main findings from each operando technique are summarized below and in Table 2  c) The UV-vis absorbance spectra of the photocathode ex situ at the respective states of discharge and charge marked in Figure 12a. Reproduced with permission. [38] Copyright 2021, American Chemical Society. d) High-resolution image (HRTEM) showing the V 2 O 5 nanoribbons grown along the [010] direction in situ/operando TEM images of V 2 O 5 during e,f) discharging and f, inset) charging. Reproduced with permission. [35] Copyright 2019, American Chemical Society. g) In situ/operando AFM images of V 2 O 5 cathodes in g) LiPF 6 in propylene carbonate (PC) electrolyte and h) LiClO 4 in PC electrolyte for the first cycle. Reproduced with permission. [39] Copyright 2004, Elsevier.

4)
In situ/operando optical characterization and UV-vis were used to track the change in the bandgap of V 2 O 5 electrodes for LIB as a function of the state of charge. By comparison, ZIBs did not show a significant change in bandgap in the voltage range tested so far. 5) In situ/operando TEM provided morphological and atomic lattice information about V 2 O 5 nanoribbon for LIBs, showing volume expansion (app 10%) and volume recovery. 6) In situ/operando AFM revealed differences in surface morphology after cycling depending on the type of electrolyte used.