Electrochemical performance and reaction mechanism investigation of V 2 O 5 positive electrode material for aqueous rechargeable zinc batteries †

The electrochemical performance and reaction mechanism of orthorhombic V 2 O 5 in 1 M ZnSO 4 aqueous electrolyte are investigated. V 2 O 5 nanowires exhibit an initial discharge and charge capacity of 277 and 432 mA h g (cid:1) 1 , respectively, at a current density of 50 mA g (cid:1) 1 . The material undergoes quick capacity fading during cycling under both low (50 mA g (cid:1) 1 ) and high (200 mA g (cid:1) 1 ) currents. V 2 O 5 can deliver a higher discharge capacity at 200 mA g (cid:1) 1 than that at 50 mA g (cid:1) 1 after 10 cycles, which could be attributed to a di ﬀ erent type of activation process under both current densities and distinct degrees of side reactions (parasitic reactions). Cyclic voltammetry shows several successive redox peaks during Zn ion insertion and deinsertion. In operando synchrotron di ﬀ raction reveals that V 2 O 5 undergoes a solid solution and two-phase reaction during the 1st cycle, accompanied by the formation/decomposition of byproducts Zn 3 (OH) 2 V 2 O 7 $ 2(H 2 O) and ZnSO 4 Zn 3 (OH) 6 $ 5H 2 O. In the 2nd insertion process, V 2 O 5 goes through the same two-phase reaction as that in the 1st cycle, with the formation of the byproduct ZnSO 4 Zn 3 (OH) 6 $ 5H 2 O. The reduction/oxidation of vanadium is con ﬁ rmed by in operando X-ray absorption spectroscopy. Furthermore, Raman, TEM, and X-ray photoelectron spectroscopy (XPS) con ﬁ rm the byproduct formation and the reversible Zn ion insertion/deinsertion in


Introduction
Aqueous rechargeable zinc batteries (ARZBs) have received much attention for application in large-scale energy storage because of their advantages, such as high safety, low cost, and sufficient abundance. 1,2 Aqueous electrolytes have higher ionic conductivity (up to 1 S cm À1 ) than that of non-aqueous electrolytes (about 10 mS cm À1 ), resulting in a higher rate capability. 1 Meanwhile, the utilization of aqueous electrolytes can lower the activation energy for charge transfer at the electrode/ electrolyte interface. The manufacturing costs of ARZBs are expected to be low since water-based electrolytes are non-toxic, inammable, and strict humidity control is not required during cell assembling. 3,4 Moreover, metallic Zn has a high specic capacity (820 mA h g À1 ) and high volumetric capacity (5854 mA h cm À3 ), making it promising as a negative electrode. It also owns sufficiently high overpotentials with respect to the hydrogen evolution, overcompensating the negative value of À0.76 V vs. SHE that makes it usable in water 5 and, therefore, can be directly used as the negative electrode in aqueous-based electrolytes. However, the lack of high-performance positive electrode materials, the heavy mass, and the large polarization of divalent Zn 2+ hinder the practical applications of ARZBs.
Many efforts have been made on the exploration of highperformance positive electrode materials for ARZBs, including the polymorphs of manganese oxide (a-, b-, g-, d-MnO 2 , and spinel-MnO 2 ), 4,6-17 Prussian blue analogues, 18 39,40 and other compounds such as Na 3 V 2 (PO 4 ) 3 , 41 Na 3 V 2 (PO 4 ) 2 F 3 , 42 VS 2 , 43 and Mo 6 S 8 . 44 Among them, vanadiumbased oxides are very promising due to their open framework and relatively high capacity as well as operation voltage of around 0.6-1.0 V. For example, Kundu et al. 5 reported Zn 0.25 -V 2 O 5 $nH 2 O positive electrode, which delivers a high initial discharge/charge capacity of 282/278 mA h g À1 at C/6 rate, excellent rate capability, and high cycling stability with a capacity retention of 80% aer 1000 cycles at 15C rate. Xia et al. 29 demonstrated that Na 0.33 V 2 O 5 nanowires have a high capacity of 367.1 mA h g À1 at 100 mA g À1 when used as the positive electrode in ARZBs. This material also exhibits good rate capability and high capacity retention of 93% aer 1000 cycles. Pang et al. 34 investigated H 2 V 3 O 8 /graphene composite, which shows a high capacity of 394 mA h g À1 at C/3, high rate capability, and excellent cycling stability with a capacity retention of 87% aer 2000 cycles. Despite this signicant progress, it is still urgently required to deeply understand the electrochemical reaction mechanism of positive electrode materials. Because it is more meaningful to the development of ARZBs by understanding the reaction mechanism behind good electrochemical performance than trying to improve performance empirically.
Because of its typical layered structure and relatively high capacity, V 2 O 5 is considered a promising positive electrode for ARZBs. [45][46][47][48] Despite its good electrochemical performance in ARZBs, the structural changes of V 2 O 5 during electrochemical processes are still under debate. Three different viewpoints regarding the structural changes of V 2 O 5 in ARZBs have been proposed so far: (i) Zhou et al. 46 compared the performance of V 2 O 5 in different electrolytes (i.e. Zn(NO 3 ) 2 , Zn(CH 3 COO) 2 , ZnCl 2 , and ZnSO 4 aqueous-based electrolytes) and with different concentrations of ZnSO 4 -based electrolytes. In 3 M ZnSO 4 , V 2 O 5 delivers the best performance with a high capacity of 224 mA h g À1 at 100 mA g À1 and good cycling stability at the high current densities of 1 and 2 A g À1 , respectively. Ex situ X-ray diffraction (XRD) demonstrated the formation of a new phase of Zn x V 2 O 5 upon Zn insertion into V 2 O 5 .
(ii) Zhang et al. 45 also reported a V 2 O 5 cathode material with a capacity of 470 mA h g À1 at 0.2 A g À1 and high capacity retention of 91.1% aer 4000 cycles at 5 A g À1 in 3 M Zn(CF 3 -SO 3 ) 2 electrolyte. They proved that this material can work in extreme conditions at both high (50 C) and low (À10 C) temperatures. The co-insertion of hydrated Zn ions into the V 2 O 5 crystal structure was proposed based on ex situ XRD, X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM).
(iii) Chen et al. 47 proposed a phase transition reaction mechanism via ex situ XRD, Raman, and XPS. Orthorhombic V 2 O 5 underwent a phase transition to zinc pyrovanadate during the rst discharge, where the formed zinc pyrovanadate showed reversible Zn 2+ (de)insertion during subsequent cycles.
However, the electrochemical reaction mechanism of V 2 O 5 in the above-reported works was investigated through ex situ techniques. The ndings are in controversy with each other and a detailed investigation of the structural changes of V 2 O 5 upon Zn-ion insertion/deinsertion is still missing. Moreover, nonequilibrium or intermediate species or states cannot be detected using ex situ studies, while in operando studies can provide a more reliable understanding of the structural evolution of a battery material in "real use".
In addition, it has been frequently reported that a mass of complex byproducts was produced/decomposed during the discharge/charge cycling of vanadium oxides such as  49 Interestingly, byproducts were not observed in the three works related to the study of V 2 O 5 reported above. To better understand the electrochemical mechanism of V 2 O 5 in ARZBs in operando techniques like diffraction using synchrotron radiation are required. Due to their extremely bright, high ux, and tunable also high energy, synchrotron radiation-based characterization provides deep penetration into the sample, highquality data, and real-time diffraction. Particularly, in operando synchrotron study can effectively avoid the unpredictable contamination and irreversible changes of highly reactive samples during material preparation, handling, and transportation, which can provide more reliable and precise data for analysis. Hence, in this work, orthorhombic V 2 O 5 were prepared by a facile hydrothermal method and the detailed structure evolution and reaction mechanism of orthorhombic V 2 O 5 in ARZBs are studied via in operando synchrotron diffraction and X-ray absorption spectroscopy (XAS) together with ex situ Raman, TEM, and X-ray photoelectron spectroscopy (XPS).

Synthesis of V 2 O 5 nanowires
V 2 O 5 nanowires were prepared via a modied hydrothermal method followed by heat treatment. 50,51 Briey, 0.18 g of commercial V 2 O 5 powder (Alfa Aesar, 99.99%) was added to 30 ml of deionized water under vigorous stirring for a few minutes to form a light orange suspension. Then, 2.5 ml 30% hydrogen peroxide (H 2 O 2 ) was dropwise added to the above suspension and kept stirring for 30 min to get a transparent orange solution. The obtained solution was transferred to a 50 ml Teon-lined stainless-steel autoclave and kept at 190 C for 4 days. The precipitate was collected and washed with deionized H 2 O several times and dried at 80 C for 12 h. Finally, the product was annealed at 400 C for 2 h with a heating rate of 5 C min À1 in air atmosphere

Preparation of the electrolyte
The 1 M ZnSO 4 electrolyte was prepared by dissolving a corresponding amount of ZnSO 4 $7H 2 O powder in distilled H 2 O with vigorous stirring at room temperature.

Morphological and structural study
The morphology was studied with a Zeiss Supra 55 Scanning Electron Microscope (SEM) with primary energy of 15 keV. The structural characterization was performed using synchrotron radiation (l ¼ 0.4132 A, 30 keV) at the Material Science and Powder Diffraction beamline (MSPD) of ALBA synchrotron (Barcelona, Spain). 52 The powder was lled in 0.5 mm Ø borosilicate capillary, and the diffraction pattern was collected in capillary geometry. A LabRam HR Evolution Raman microscope from Horiba Scientic equipped with HeNe laser (633 nm, 17 mW) and a CCD detector (Horiba) was used to collect the Raman scattering of the samples. Meanwhile, a 600 grating was used to split the measurement signal with a Â100 objective (NA 0.95) for all the pristine and cycled samples. The data were collected for 30 seconds with 4.25 mW of the laser and ve measurements were added to reduce signal noise. Transmission Electron Microscopy (TEM) imaging and high angle annular dark eld-scanning TEM (HAADF-STEM) electron dispersive X-ray (EDX) mapping were acquired by Themis 300 under 300 kV with Ceta camera and Super-X EDX detector, respectively. The last measured screen current for highresolution TEM (HRTEM) imaging was 998 pA and for EDS mapping was 93.4 pA. X-ray photoelectron spectroscopy measurements were performed using a K-Alpha XPS spectrometer (ThermoFisher Scientic, East Grinstead, UK), applying a micro-focused, monochromated Al K a X-ray source with a spot size of 400 mm. To prevent any localized charge buildup, the K-Alpha + charge compensation system was employed during analysis, using electrons of 8 eV energy and low-energy argon ions. The Thermo Avantage soware was used for data acquisition and processing. 53 The spectra were tted with one or more Voigt proles (binding energy uncertainty: AE0.2 eV). All spectra were referenced to the O 1s peak of vanadium oxide at 530.0 eV binding energy. In addition, the discharged and charged V 2 O 5 and discharged and charged Zn electrodes were sealed in an in situ Raman cell with a quartz window inside a glovebox. Note the "discharged Zn" refers to the counter electrode of V 2 O 5 at the discharged state of 0.3 V from the same cell, while "charged Zn" refers to the counter electrode of V 2 O 5 at the charged state of 1.6 V from the same cell.

Electrochemical characterization
The electrode was prepared by mixing active material V 2 O 5 nanowires with Super C65C (Timcal) and polyvinylidene diuoride (PVDF) binder in a weight ratio of 70 : 20 : 10 with Nmethyl-2-pyrrolidone solvent. The dried electrode mixture with mass loading of $1.4 mg cm À2 was pressed (4 tons) on a stainless steel mesh of 12 mm and dried at 120 C overnight under vacuum. CR2032-type coin cells for electrochemical measurements were assembled in air at room temperature. The cells were built with V 2 O 5 positive electrode, Zn foil as the negative electrode, 1 M ZnSO 4 as the electrolyte, and a piece of glass microber (Whatman) as the separator. The galvanostatic cycling with the potential limitation (GCPL) and cyclic voltammetry (CV) measurements were performed between 0.3 and 1.6 V (vs. Zn 2+ /Zn) with a VMP3 potentiostat (BioLogic) at 25 C. GCPL was performed at different current densities ranging from 50 to 1600 mA g À1 to determine the rate capability of the battery.

In operando synchrotron diffraction and in operando Xray absorption spectroscopy (XAS)
In operando synchrotron diffraction was performed at the Material Science and Powder Diffraction beamline (MSPD) at the ALBA synchrotron. The electrochemical cell consists of 2025-type coin cell with glass windows of 5 mm diameter for beam entrance. The positive electrode was prepared by pressing the dried electrode mixture (as described above) on a stainless steel mesh within a 5 mm hole in the center, a Zn foil with a 5 mm hole in the center was used as the negative electrode. In operando synchrotron diffraction was conducted with radiation l ¼ 0.4132 A wavelength (30 keV) and the position-sensitive detector MYTHEN. Data in steps of 0.006 over an angular range of 1.8-42 in 2theta were gathered with an effective exposure time of 60 s during the 1.5 cycles with a current density of 50 mA g À1 . The coin cell was continuously oscillated AE5 around the incoming beam direction to improve the powder averaging (i.e. increasing the number of crystallites fullling Bragg condition and contributing to the observed reections). Diffraction data were analyzed by the Rietveld method using the Fullprof soware package. 54 In operando XAS measurements were performed at beamline P65 at the synchrotron source PETRA III (DESY, Hamburg). XAS was carried out during the rst charge/discharge process at the current of 50 mA g À1 in the same coin-cell conguration as above but with a Kapton window. X-ray absorption spectra of vanadium were recorded in quick-XAS (6 min per spectrum) mode in uorescence geometry using a PIPS (passivated implanted planar silicon) diode. The V K-edge for V 2 O 5 was investigated, and the energy was calibrated utilizing vanadium foil as commonly applied in XAS experiments. V 2 O 3, VO 2 , and V 2 O 5 were used as standard materials. All data were collected at room temperature with a Si(111) double crystal monochromator, and all spectra were processed using the DEMETER soware package. 55

Structural and morphological characterization
The crystal structure of the prepared V 2 O 5 nanowires was investigated by synchrotron diffraction and HRTEM imaging, as displayed in Fig. 1a. All reections can be indexed to the orthorhombic V 2 O 5 with space group Pmn2 1 , and the lattice parameters are a ¼ 11.515(1) A, b ¼ 4.374(1) A, c ¼ 3.566(1) A, in good agreement with the previous work. 56 The strong intensities of the reections conrm the high crystallinity of the obtained V 2 O 5 nanowires material. The SEM image (Fig. 1b) demonstrates that V 2 O 5 material is composed of nanowire-like nanostructure with lengths up to several micrometers. TEM imaging further reveals the nanowire-like morphology of V 2 O 5 while HRTEM image (Fig. 1c) displays the highly crystalline sample on the [010] zone axis with lattice fringe of 0.58 nm and 0.34 nm corresponding to the (200) and (110) plane of V 2 O 5 , respectively. This ts well with the result obtained from synchrotron diffraction. nanowires electrode displays a at plateau at around 1.00 V, followed by a sloping-like plateau at about 0.50 V. During the 1st charge process (Zn ions de-insertion), a slope and a at plateau at 1.20 V together with a slope up to 1.6 V can be observed. Compared with the 1st discharge, the 2nd discharge prole shows a shorter plateau at around 1.00 V and similar sloping-like plateau at 0.50 V. In the 2nd charge process, a slope similar to that for the 1st charge is observed, but no clear plateau at 1.20 V can be seen. During cycling, the voltage proles signicantly change: the plateau at 1.20 V completely disappears, and only one slope and a sloping-like plateau at 0.50 V can be observed (see the 5th discharge prole). On the 5th charge prole, two sloping-like plateaus at 0.70 V and 1.0 V are observed. The V 2 O 5 nanowires electrode delivers an initial discharge and charge capacity of 277 and 432 mA h g À1 , respectively, at a current density of 50 mA g À1 . The discharge capacity of 277 mA h g À1 almost reaches the theoretical one based on the insertion of 1 mol of Zn 2+ in V 2 O 5 (294 mA h g À1 ).

Electrochemical properties
The huge extra charge capacity might be attributed to side reactions such as O 2 evolution from the aqueous-based electrolyte. The electrode delivers a discharge capacity of 302 mA h g À1 at the 2nd cycle, a value higher than that for the rst cycle, which might be due to an activation of the active material (theoretical capacity of 442 mA h g À1 by considering insertion of 1.5 mol of Zn 2+ into V 2 O 5 ), accompanied by Zn insertion, as also reported by other previous studies. 6,57 The  V 2 O 5 positive electrode exhibits a dramatic decrease of capacity during the following 20 cycles at 50 mA g À1 (94 mA h g À1 at the 22nd cycle) and delivers a discharge capacity of only 21 mA h g À1 aer 100 cycles. Moreover, the cycling stability of V 2 O 5 positive electrode was studied at a higher specic current of 200 mA g À1 , as presented in Fig. 2b. An initial discharge capacity of 278 mA h g À1 is obtained, followed by a sudden capacity decline down to 180 mA h g À1 at the 2nd cycle. In the following four cycles, the behavior differs from the one of the electrode cycled at 50 mA g À1 : it displays a capacity increase, indicating a long activation process of the active material, which is oen observed in positive electrode materials for ARZBs. 6,57 However, this activation phenomenon takes more cycles under high current density, demonstrating its possible correlation with the involved chemical ion kinetics during cycling. 22 Aer 100 cycles, the capacity at 200 mA g À1 is higher (45 mA h g À1 ) than that delivered at 50 mA g À1 . This could be attributed to a different type of activation process under different current densities. The side reactions (parasitic reactions) under low current have more relevance than at high current and, consequently, the degradation is faster. Moreover, a signicant difference is observed between the two rst charge proles recorded at both current densities: the charge prole recorded at 200 mA g À1 displays a slope-like plateau (Fig. S1 †); a clear at plateau in the voltage prole is observed at the low current of 50 mA g À1 . The difference of voltage proles suggests a different mechanism reaction and therefore results in a type of activation process under high and low current.

Electrochemical mechanism
In order to clarify the Zn-ion storage mechanism in the V 2 O 5 material, CV was performed at a scan rate of 0.1 mV s À1 in the voltage range of 0.30-1.60 V (vs. Zn 2+ /Zn). Fig. 3 displays two peaks centered at 0.92 and 0.50 V during the 1st reduction process and a broad peak at 1.20 V with a shoulder at 1.05 V during the 1st oxidation process. In the following scans, three features are observed for both reduction and oxidation processes, respectively. The reduction peak at 0.92 V gradually shis to higher potential, becomes weaker, and nally disappears. However, a new reduction peak at 0.88 V emerges and grows up and the reduction peak at 0.50 V gradually shis to 0.57 V, indicating polarization (and resistance) decrease of this process. At the same time, a new oxidation peak appears at 0.74 V. The oxidation peak at 1.05 V shis to 1.00 V aer the rst cycle and increases upon cycling. The oxidation peak at 1.20 V shis to 1.12 V in the 3rd cycle; this shi is accompanied by decreasing in current intensity, indicating the decrease of polarization.
To investigate the structural evolution of V 2 O 5 upon Zn-ion insertion/deinsertion, in operando synchrotron diffraction was performed during the initial 1.5 cycles. The contour maps of selected diffraction patterns and corresponding voltage-time proles are provided in Fig. 4. Note that an electrochemical activation is observed at the high current of 200 mA g À1 during the rst 5 cycles, with very interesting changes. Here we go deeper on understanding the very initial changes at low current and in the future, it would be interesting and necessary to study what is happening during the further cycles at higher current density.
Before the discharge (at 1.   58 Therefore, the value of x ¼ 0.44 is an estimate for the lower limit of the inserted Zn content. Upon further Zn ions insertion (Region III), some new reections begin to appear at 2. 16 , 4.34 , 9.26 , 9.35 , which are assigned to the byproduct of ZnSO 4 Zn 3 (OH) 6 $5H 2 O with space group P 1. Once the byproduct is formed, all its Bragg reections remain at the same 2q-positions throughout the discharge process, but their intensities increase. Meanwhile, most of the reections maintain their positions along with the increase of their intensities (Region III), while only two reections of Zn 0.44 V 2 O 5 at 3.81 and 7.01 shi to high and low angles, respectively. It indicates that a solid solution process happens in Region III to form a nal phase Zn 0.94 V 2 O 5 together with the abovementioned two byproducts with a total discharge capacity of 277 mA h g À1 (see Fig. S2c † for the Rietveld renement of 36th pattern at rst fully discharged V 2 O 5 ). During the 1st charge, the reections do not evolve in a symmetric reverse backway, as indicated by the appearance and disappearance of a small reection at 7.17 , where two 2-phase regions are observed (Region II). The reections return back to their initial 2q-positions for the pristine V 2 O 5 state, but with much lower intensities (see Fig. S2d † for the Rietveld renement, based on the 69th pattern at fully charged V 2 O 5 ). This is possibly caused by the amorphization of the crystalline active material or by the vanadium dissolution in the electrolyte 32,36 (see Fig. S4 †). Interestingly, both byproducts disappear again along with the 1st charging process. The evolution of the reections in the 2nd discharge is analogous to that during the 1st cycle. In Region I, the reections show the same behavior as those for the rst discharge process, suggesting a solid solution reaction. In Region II and III, the electrode undergoes the same process: a shorter two-phase transition and solid solution than in the rst discharge, but without the appearance of the reections related to the byproduct Zn 3+d (OH) 2 V 2 O 7 $2H 2 O (see Fig. S2e † for the Rietveld renement, based on the 97th pattern at the 2nd fully discharged V 2 O 5 ). The formation of the other byproduct ZnSO 4 Zn 3 (OH) 6 $5H 2 O is also detected again during the second discharge process. The Zn 3+d (OH) 2 V 2 O 7 $2H 2 O has an open layered structure and has been reported by Alshareef et al. 26 as positive electrode material in ARZBs. The material has a capacity of 213 mA h g À1 at 50 mA g À1 and shows a shi of the reection 001 during Zn-ion insertion and deinsertion. 26 This means that the Zn 3+d (OH) 2 V 2 O 7 $2H 2 O byproduct, generated during cycling, still contributes to the overall capacity of the electrode. Recently, Chen et al. 47 reported that Zn 3 (OH) 2 V 2 O 7 -$2H 2 O is also formed via the 2-phase coexistence transition in an aqueous Zn-V 2 O 5 battery, and then it functions as a host structure in the following cycles, where it shows a shi of reection 001 during Zn ion insertion. However, in our case, the shi of the 001 reection is not observed, which might suggest that Zn 3+d (OH) 2 V 2 O 7 $2H 2 O is not active or just the c-axis parameter does not change. Furthermore, our results directly show for the rst time the formation and decomposition of both byproducts Zn 3+d (OH) 2 V 2 O 7 $2H 2 O and ZnSO 4 Zn 3 (OH) 6 $5H 2 O during cycling.
In order to investigate the variation of the oxidation state and the local electron environment of vanadium during the discharge/charge (i.e. Zn insertion/deinsertion) process, in operando XAS was performed on the V 2 O 5 nanowires as positive electrode within an in operando coin cell. Fig. 5 shows the normalized V K-edge spectra collected during the initial discharging and charging processes and reference spectra of standard vanadium oxides, where V 2 O 5 , VO 2 , and V 2 O 3 have the oxidation state of +5, +4, and +3, respectively. The edge position of the V K-edge in the pristine V 2 O 5 positive electrode overlaps with that of the standard V 2 O 5 reference, indicating that the oxidation state of V in V 2 O 5 is +5. Moreover, an intense pre-edge peak for the V K-edge of pristine V 2 O 5 is observed, which is ascribed to the transitions between the 1s and bound phybridized d-states. 59,60 Along with progressive discharging, the main absorption edge shis towards lower binding energies, conrming the reduction of the oxidation state of vanadium upon the Zn-ion insertion. Meanwhile, the pre-peak (A in Fig. 5a) also shis gradually to lower binding energy with the simultaneous decrease of intensity, conrming the reduction of V and the deformation of the local V environments during Znion insertion. This is due to the co-existence of a distorted tetragonal pyramid and VO 6 octahedra. The edge resonance (B in Fig. 5a) displays distinct changes in both intensity and shape, which is caused by the absorption of photons accompanied by core-electron excitations. 60,61 During the discharge process, two broad peaks centered at 5494 eV and 5507 eV shi to lower energy and with a decrease of their intensities, before they converge into one very broad peak centered at 5500 eV from the initial stage to 0.99 V (peak B in Fig. 5a). Aer that, the formed broad peak centered at 5500 eV and the peak at 5486.5 eV continuously shi to lower energy accompanied by an increases of both intensities (from 0.99 V to 0.3 V). Two distinct isosbestic points 62 at $5474 eV and $5502 eV (red arrows in Fig. 5a and b) are observed during both discharge (from 0.99 V to 0.3 V) and charge (from 0.3 V to 1.32 V) processes. This conrms the 2phase coexistence transitions upon Zn-ion insertion/ deinsertion into/from V 2 O 5 structure, as already proposed from in operando synchrotron diffraction. At the fully discharged state at 0.30 V, the observed energy of the V K-edge lies almost in the middle of those of the V 2 O 5 and V 2 O 3 reference spectra. This means that the oxidation state of V is very close to V 4+ , in good agreement with the electrochemical data (see Fig. S5 †). However, the energy of the edge does not completely overlap with the spectrum of VO 2 , reecting structural differences between the discharged state with x ¼ 1.00 in Zn x V 2 O 5 and the reference material VO 2 . During the charging process (Zn-ion deinsertion), a completely reversible behavior can be observed. Pre-peak and edge resonance (A 0 and B 0 in Fig. 5a and b) prove a reversible process. At the fully charged state of $1.6 V, the V spectrum returns back to its initial state, indicating that the V ions are fully oxidized to the oxidation state of +5. The evolution of V K-edge spectra reveals that the V ions are reduced and reversibly oxidized during the Zn-ion insertion and deinsertion, respectively, accompanied by the local structural changes around the V ions and in full agreement with in operando synchrotron diffraction.
Raman spectra were collected to investigate vibration modes as ngerprints of the short-range structure of the samples during Zn-ion insertion/deinsertion, as displayed in Fig. 6 and Table S1. † In the pristine V 2 O 5, oxygen atoms lie in four distinct sites in a [VO 5 ] pyramid unit, denoted as O(1)-O(4). The stretching mode of V-O(1) bond is located at 994 cm À1 and its bending vibrations are located at 405 and 284 cm À1 . Raman peaks at 482 cm À1 and 701 cm À1 are ascribed to the bending    TEM and elemental mapping were carried out to further study the structural and morphology evolution of V 2 O 5 during cycling. It is unable to observe the nanowire-like feature of Zn x V 2 O 5 on the fully discharged electrode, whereas a sheet-like morphology can be clearly seen (Fig. 7a). Fig. S6 † shows that the O, S, V, and Zn elements are uniformly distributed in the sheetlike material. The sheet-like material is probably attributed to the byproducts of ZnSO 4 Zn 3 (OH) 6 $5H 2 O, as demonstrated by synchrotron diffraction. Fig. 7b conrms the disappearance of sheet-like morphology and the recovery of nanowire-like feature of V 2 O 5 aer the charge process, demonstrating the decomposition of byproducts.
XPS was applied to investigate the surface chemistry and surface elemental composition of pristine V 2 O 5 , rst discharged V 2 O 5 , and rst recharged V 2 O 5 . As displayed in Fig. 8, the V 2p spectrum of pristine V 2 O 5 can be tted with two doublets: a main one with V 2p 3/2 at 517.6 eV and a second one with weak intensity at 516.3 eV, which demonstrates that V exists mainly in the oxidation state +5 with a minor contribution of vanadium  +4. It can be seen that the O 1s spectrum of pristine V 2 O 5 can be tted with one peak at 530.3 eV, corresponding to V-O group. The V 2p spectrum in the discharged state is not visible anymore due to the formation of the byproduct ZnSO 4 Zn 3 (-OH) 6

Conclusion
In summary, orthorhombic V 2 O 5 nanowires were prepared via a facile hydrothermal approach. In the 1 M ZnSO 4 electrolyte, V 2 O 5 nanowires deliver an initial discharge/charge capacity up to 277 and 432 mA h g À1 , respectively, at a current density of 50 mA g À1 , which almost reaches the theoretical capacity based on 1 Zn 2+ insertion with two electrons per formula unit V 2 O 5 (294 mA h g À1 ). The V 2 O 5 positive electrode exhibits a dramatic decrease of capacity during the following 20 cycles at 50 mA g À1 (94 mA h g À1 for 22nd) and delivers a very low discharge capacity of 21 mA h g À1 aer 100 cycles. Moreover, it delivers an initial discharge capacity of 278 mA h g À1 at 200 mA g À1 , followed by an activation process of the material. The capacity is higher for 200 mA g À1 than for 50 mA g À1 , which could be attributed to a different type of activation process under both current densities and different resulting degrees of side reactions (parasitic reactions). CV displays in the rst scan two reduction peaks, centered at 0.92 and 0.50 V, and a broad oxidation peak at 1.20 V with a shoulder at 1.05 V. In the following four scans, the CV curves indicate signicant changes in both reduction and oxidation peaks. In operando synchrotron diffraction focused on the rst 1.5 cycles, reveals that V 2 O 5 rst undergoes a solid solution and 2-phase coexistence transitions upon Zn-ion insertion. It also conrms the formation of two byproducts, Zn 3+d (OH) 2 V 2 O 7 $2(H 2 O) and ZnSO 4 Zn 3 (OH) 6 $5H 2 O during the Zn-ion insertion. The electrode undergoes a reversible process upon Zn-ion deinsertion with the decomposition of both byproducts. The V 2 O 5 electrode goes in the 2nd discharge process through the same two-phase reaction as that in the 1st discharge without the formation of the byproduct Zn 3+d (OH) 2 -V 2 O 7 $2(H 2 O). In operando XAS conrms the reduction/oxidation of vanadium during the Zn insertion/deinsertion. Moreover, ex situ Raman and XPS also prove the reversibility of the reactions during cycling. The electrochemical performance of V 2 O 5 can be improved by electrode engineering such as surface coating with carbon/graphene oxide and electrolyte optimization with more concentrated salt and water-organic solvent.

Author contributions
Q. F. conceived the idea and discussed with J. W., A. S., L. Z., A. M., E. W., X. L., Z. D., M. K., H. E., and S. D.; Q. F. and J. W. performed material synthesis, sample preparation, characterizations, electrochemical measurements, and analyzed the data. L. Z. performed Raman and analyzed the data. Q. F., J. W., A. S., A. M., E. W., M. K., and H. E. conducted in operando measurements and analyzed the data. X. L. performed XPS measurements and analyzed the XPS data. Z. D. carried out TEM and analyzed the data. Q. F. wrote the preliminary dra with input from J. W.; Q. F., J. W., M. K., H. E., and S. D., discussed the results and revised the manuscript. All authors contributed to interpreting the ndings, reviewing, and commenting on the manuscript.

Conflicts of interest
The authors declare no competing nancial interests.