Oxygen vacancies and N‐doping in organic–inorganic pre‐intercalated vanadium oxide for high‐performance aqueous zinc‐ion batteries

Abstract

Pre‐intercalation of metal ions into vanadium oxide is an effective strategy for optimizing the performance of rechargeable zinc‐ion battery (ZIB) cathodes. However, the battery long‐lifespan achievement and high‐capacity retention remain a challenge. Increasing the electronic conductivity while simultaneously prompting the cathode diffusion kinetics can improve ZIB electrochemical performance. Herein, N‐doped vanadium oxide (N‐(Zn,en)VO) via defect engineering is reported as cathode for aqueous ZIBs. Positron annihilation and electron paramagnetic resonance clearly indicate oxygen vacancies in the material. Density functional theory (DFT) calculations show that N‐doping and oxygen vacancies concurrently increase the electronic conductivity and accelerate the diffusion kinetics of zinc ions. Moreover, the presence of oxygen vacancies substantially increases the storage sites of zinc ions. Therefore, N‐(Zn,en)VO exhibits excellent electrochemical performance, including a peak capacity of 420.5 mA h g⁻¹ at 0.05 A g⁻¹, a high power density of more than 10 000 W kg⁻¹ at 65.3 Wh kg⁻¹, and a long cycle life at 5 A g⁻¹ (4500 cycles without capacity decay). The methodology adopted in our study can be applied to other cathodic materials to improve their performance and extend their practical applications.

INTRODUCTION

Owing to the widespread development of renewable energy infrastructures, such as those exploiting solar and wind-power, the need for safe and environmentally friendly large-scale energy storage devices is urgent. Lithium-ion batteries (LIBs) are promising energy storage devices, compared to other secondary batteries, owing to their higher energy density.^1–3 However, due to the lithium resource limitations and safety hazards of combustible organic electrolytes, the development of aqueous batteries is accelerating. Among the aqueous metal ion batteries, aqueous zinc-ion batteries (ZIBs) have received extensive attention because of the distinctive merits of metallic zinc, including high theoretical capacity (820 mA h g⁻¹ and 5851 mA h cm⁻³), small ionic radius (0.74 Å), low redox potential (−0.76 V vs. standard hydrogen electrode), and high compatibility in aqueous electrolytes.^4–6 However, owing to slow diffusion kinetics of Zn²⁺ in the host framework, the application of aqueous ZIBs is limited by their energy density and cycle life.^7,8

The development of advanced cathode materials is crucial to achieve high-energy-density ZIBs. To date, several materials have been exploited as Zn²⁺ intercalation hosts, including manganese oxides, vanadium oxide materials, Prussian blue analogs, and organic compounds.^9–12 However, these materials have either limited capacity or poor cyclic stability. Conversely, layered vanadium-based materials are considered promising candidate cathodes for ZIBs owing to their open skeleton structure, and large theoretical capacity.¹³ Vanadium is characterized by different oxidation states (V⁵⁺, V⁴⁺, V³⁺, and V²⁺); the distortion of the VO polyhedron leads to various vanadium-based compounds with different compositions and structures.¹⁴ However, vanadium oxide is susceptible to capacity decline owing to its low electronic conductivity and sluggish diffusion kinetics. Therefore, exploring layered vanadium oxide cathodes with high capacities, long lifetimes, and outstanding rate performances is challenging.

To optimize the electrochemical performance of vanadium oxide materials, several strategies have been explored, including embedding structured water, metal cations, and organic components into the vanadium oxide interlayer. Structured water functions as a charge shield for Zn²⁺ ions, reducing their effective charge and interaction with host frameworks to improve the zinc-ion diffusion kinetics.¹⁵ Interlayer metal cations (such as Na⁺, Ca²⁺, and Mg²⁺) work as pillars to maintain the vanadium-oxide structure stability.¹⁶ Organic molecules increase the interlayer spacing of the vanadium-based compounds and improve the zinc ion mobility.^17,18 The combination of the long organic amine chain (en) with inorganic metal ions pre-intercalating into the vanadium oxide can improve the cathode performance. Moreover, creating defects drastically improves the material performance.¹⁹ As an example, phosphate groups improve the surface reactivity and reaction kinetics of the electrodes.^19,20 To date, the literature on nitrided vanadium oxide as cathode for aqueous ZIBs is scarce.

Herein, we report a nitrogen-doped vanadium oxide cathode with high capacity, long cycle stability, and superior rate performance. N-doping and oxygen vacancies were concurrently introduced in vanadium oxide through low-temperature NH₃ treatment. Positron annihilation lifetime spectroscopy (PALS) and electron paramagnetic resonance (EPR) demonstrated the presence of oxygen vacancies. X-ray photoelectron spectroscopy (XPS) revealed that oxygen vacancies provide abundant active sites for the storage of zinc ions. DFT analysis confirmed that oxygen defects and N-doping reduced the bandgap of N-(Zn,en)VO. The narrowed bandgap enhances the charge carrier excitation to the conduction band, which is beneficial for the transfer of electrons during the oxidation–reduction reaction.²¹ The nudged elastic band (NEB) revealed that oxygen vacancies and nitrogen doping reduced the diffusion energy barrier of Zn²⁺ and accelerated the migration of Zn²⁺. Therefore, N-(Zn,en)VO delivered a peak capacity of 420.5 mA h g⁻¹ at 0.05 A g⁻¹ and a long cycle life of 4500 cycles with no capacity decay at 5 A g⁻¹. This design effectively improves the cyclic stability and specific discharge capacity of the cathode and opens a new avenue for the design of advanced aqueous ZIB cathode materials.

RESULTS AND DISCUSSION

The one-step hydrothermal method in Figure S1 was adopted to synthesize the (C₁₂H₂₈N)_xV₇O₁₆·nH₂O nanotubes (denoted as (en)VO).²² (Zn,en)VO was prepared by ion exchange in a mixed solution of zinc chloride in water and ethanol. N-(Zn,en)VO with N-doping and oxygen vacancies was obtained at 200°C in NH₃ atmosphere. The crystal structures of the synthesized materials were evaluated by X-ray diffraction (XRD). The diffraction peaks at 3.2°, 6.5°, 9.8°, 29.1°, and 32.6°, corresponding to the (001), (002), (003), (200), and (210) planes of (en)VO crystals, respectively, are consistent with previous.^22,23 The crystal structure of (en)VO in the (001) direction (2θ = 3.185°) indicates an interlayer spacing of 27.7 Å (Figure S2A) beneficial for the storage of Zn ions. Fourier transform infrared (FTIR) spectroscopy further verified the successful synthesis of (en)VO (Figure S2B). The organic amine chains were located in the interlayer of the VO skeleton and acted as pillars (Figure S2B). Figure 1 A reports the XRD patterns of (Zn,en)VO and N-(Zn,en)VO. The characteristic peak of the (001) plane of (Zn,en)VO shifted to a higher angle (Figure S2C), and the corresponding interlayer distance was reduced from 27.7 to 10.66 Å, because the insertion of zinc ions replaced part of the inserted amine molecules, in line with the FTIR spectra analysis. The absorption related to v_s (CH₂) and v_as (CH₂) at 2850.4 and 2922.6 cm⁻¹, respectively, was weakened upon the zinc ion insertion, indicating that the amine chain content in (Zn,en)VO decreased (Figure 1B). The FTIR spectra of N-(Zn,en)VO showed a new absorption peak at 1260 cm⁻¹, attributed to the interaction between the doping nitrogen and the lattice. Concurrently, the intensity of the absorption at 1600 cm⁻¹ increases, indicating the stronger surface activity of the nitrided materials, which easily combine with the water molecules in the air.²⁴ The absorptions at 3195 and 3322 cm⁻¹ were assigned to the hydrogen bonds formed between water and N, O. Raman spectroscopy was performed to further analyze the structural changes of the materials during the preparation. Four peaks are present in the spectrum of (en)VO in Figure 1C. The peak at 141 cm⁻¹ originates from the bending vibration of the OVO chain.²⁵ The signals related to the VO bending vibrations appear at 280 and 406 cm⁻¹.²⁵ The peak at 991 cm⁻¹ confirms the presence of terminal VO bonds.²⁶ After the pre-intercalation of zinc ions, the layer spacing of vanadium oxide shrinks (Figure 1A). A new peak at 850 cm⁻¹ appears in (Zn,en)VO, corresponding to the symmetric contraction vibration of the [VO₄] tetrahedra.²⁷ According to the elemental analysis performed by EPMA and ICP-MS (Tables S1 and S2), the ratios of V to Zn and V to C were 7:1.45 and 7:0.54, respectively. Therefore, the determined molecular formula of (Zn,en)VO was [Zn_1.45,(C₁₂H₂₈N)_0.54]V₇O₁₆. The thermogravimetric (TG) results showed a mass loss of 1.72% in the range 125–252°C attributed to the loss of structural water (Figure S2D). Therefore, each N-(Zn,en)VO molecule contained 0.79 structural water. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images show a uniform morphology for the N-(Zn,en)VO nanotubes with a length of 1 μm and a width of approximately 100 nm (Figure S2E,F). As shown in Figure S3A,B, ion doping and nitriding treatment have no effect on the morphology of the samples. According to the Brunauer–Emmett–Teller (BET) model, the specific surface area of different samples was characterized. Specifically, the (en)VO and (Zn,en)VO nanotubes have a specific surface area of 49.3 and 31.8 m² g⁻¹, respectively. N-(Zn,en)VO exhibited a specific surface area of 32.4 m² g⁻¹. The nitriding treatment has not obviously changed the specific surface area of the materials (Figure 1D). The large specific surface area of N-(Zn,en)VO facilitated the contact between the material and the electrolyte. Figure 1E shows that the nanotubes consisted of 5–7 parallel vanadium oxide layers, with the thinner thickness beneficial to zinc ion transport. High-resolution transmission electron microscopy (HRTEM) indicated the (001) plane lattice spacing of N-(Zn,en)VO equal to 10.7 Å (Figure 1D). Finally, transmission electron microscopy–energy dispersive spectroscopy (TEM–EDS) element mapping images revealed a uniform distribution of V, O, N, and Zn in N-(Zn,en)VO and (Zn,en)VO nanotubes (Figures 1F and S3C). Table S3 shows the element content in the EDS spectra of N-(Zn,en)VO and (Zn,en)VO (Figure S3D,E), corroborating the doping of nitrogen and the generation of oxygen vacancies.

FIGURE 1. (A) XRD patterns, (B) FTIR spectra, (C) Raman spectrum, and (D) Nitrogen adsorption–desorption isotherm of (en)VO, (Zn,en)VO and N-(Zn,en)VO. (E) TEM and HRTEM images of N-(Zn,en)VO. (F) TEM-EDS elemental mapping of N-(Zn,en)VO

XPS was performed to analyze the chemical states of the elements. The spectrum of V 2p in N-(Zn,en)VO consists of two characteristic peaks at 517.1 and 515.8 eV, corresponding to V⁵⁺ and V⁴⁺, respectively (Figure S4A). According to the peak area ratio, the average valence state of vanadium in N-(Zn,en)VO was 4.49. As a comparison, the valence state of vanadium in (Zn,en)VO was 4.55, indicating that the vanadium is partially reduced due to N-doping. It should be pointed out that N possesses a smaller electronegativity (3.04) than O (3.44), thus there is less electron donation from N to V compared with the donation from O to V.^28a The O 1s XPS spectra of N-(Zn,en)VO was fitted into three peaks (Figure S4B), which were related to the lattice oxygen (O_L, 530.0 eV), chemisorbed oxygen (O_C, 531.2 eV), and oxygen vacancies (O_V, 532.4 eV).^28a From the XPS peak area ratio, the proportion of O_V increase from 6.6% to 9.0% after nitridation, indicating the generation of oxygen vacancies during the N-doping. The generation of oxygen vacancies can be explained as follows, N in ammonia partially replaces lattice oxygen in the (Zn,en)VO under heating conditions to form nitrides (Figure S4C).^28b Due to the substitution of N³⁻ for O²⁻, the electroneutrality of the [VO] material is destroyed, thus every two N-doping is accompanied by an oxygen vacancy to balance the charge. The extracted O combines with H to form H₂O. The corresponding reaction as following equation:[Image Omitted. See PDF] [Image Omitted. See PDF]

At g = 2.001, N-(Zn,en)VO showed a stronger EPR signal than (Zn,en)VO, which indicates a higher oxygen vacancy content in N-(Zn,en)VO (Figure S4D). The spectrum of N 1s in N-(Zn,en)VO in Figure S4E is composed of two characteristic peaks at 399.7 and 401.5 eV, attributed to N-V and NH₂ groups, respectively. Compared with that in (Zn,en)VO, the ratio of N-V in N-(Zn,en)VO was relatively enhanced after NH₃ treatment. The combination of Ar ion sputtering and XPS analyses of different nitridation times corroborated the nitrogen doping into N-(Zn,en)VO (Figure S5).

Furthermore, PALS is informative of the defect type, size, and relative concentration, based on the measurement of positron lifetime.^29,30 Typically, positrons were emitted into the cathodic material, rapidly thermalized by inelastic collision, and finally diffused throughout the lattice until annihilation. Figure S6 shows the typical positron lifetime spectra of (Zn,en)VO and N-(Zn,en)VO. Table 1 lists the three components of the positron lifetime and their corresponding intensities. The shortest lifetime (τ₁) usually reflects the annihilation of positrons at small size defects or in defect-free crystals.²⁹ The defects and vacancies reduce the surrounding electron density, which increases the positron lifetime.³¹ Therefore, the value of τ₂ is much larger than τ₁. For (Zn,en)VO, τ₂ was 360 ps and the intensity (I₂) was 65.4%. For N-(Zn,en)VO, τ₂ increased to 374 ps because of nitriding, which is a result of the interaction of vanadium-oxygen vacancy associates (Table 1).³² The τ₂ intensity of N-(Zn,en)VO (I₂ = 73.1%), is larger than that of (Zn,en)VO, indicating that N-(Zn,en)VO contains oxygen vacancy to a larger extent. The longest lifetime component (τ₃) was attributed to the positron annihilation at the material interfaces and large voids. The τ₃ intensity was relatively low in both materials, indicating the limited formation of nanoscale voids of oxygen vacancy.³² Combined with the EPR signal intensity, these results clearly highlighted the high content of oxygen vacancies in N-(Zn,en)VO.

TABLE 1 Positron lifetimes and relative intensities for N-(Zn,en)VO and (Zn,en)VO

Sample	τ1 (ps)	τ2 (ps)	τ3 (ns)	I1 (%)	I2 (%)	I3 (%)
N-(Zn,en)VO	199	374	2.332	25.0	73.1	1.9
(Zn,en)VO	201	360	2.189	32.2	65.4	2.4

The electrochemical performance of the N-(Zn,en)VO electrode was tested with a coin-type cell at 25°C, using a zinc foil anode, glass fiber separator, and 3 M Zn(CF₃SO₃)₂ electrolyte. The battery exhibited an outstanding stability over 4500 cycles at 5 A g⁻¹ (Figure 2A). The initial capacity (189.5 mA h g⁻¹) reached the highest value of 260.1 mA h g⁻¹ after 1500 cycles. The increased capacity in the initial several cycles is attributed to the activation of N-(Zn,en)VO. Usually, it was associated with the gradual penetration of electrolyte and utilization of internal active materials during cycling.³³ At the same time, the kinetic reaction of Zn/N-(Zn,en)VO batteries was improved during cycling. After 4500 cycles, the capacity was 198.1 mA h g⁻¹, and did not undergo degradation, compared to the initial capacity. Contrarily, (Zn,en)VO exhibited a lower specific capacity (134.4 mA h g⁻¹), with a capacity retention of only 62% at 5 A g⁻¹ of 3000 cycles. Meanwhile, the self-discharge curves after 100 h show that N-(Zn,en)VO has higher voltage retention than (Zn,en)VO (Figure S7A). This indicates that the nitriding treatment optimized the battery durability. The N-(Zn,en)VO electrode also showed excellent cycling performance at 15, 0.05, and 0.5 A g⁻¹ (Figures S7B,C and 2B). Specifically, the initial capacity of N-(Zn,en)VO at 0.5 A g⁻¹ is 294.4 mA h g⁻¹ and maintained a capacity of 304.3 mA h g⁻¹ after 500 cycles (Figure 2B). The illustration in Figure 2B and the Video S1 show that a coin battery powered a digital temperature-hygrometer. As indicated by the comparison of the capacity and cycling performance of cathodic materials reported in the literature (Table S4), N-(Zn,en)VO exhibited a superior capacity and excellent cycling performance. Moreover, the battery exhibited outstanding rate performance (Figure 2C). Specifically, capacities of 342.7, 318.4, 275.0, 230.9, 184.5, 144.6, and 122.8 mA h g⁻¹ at current densities of 0.2, 0.5, 2, 5, 10, 15, and 20 A g⁻¹, respectively, were measured for the N-(Zn,en)VO electrode. At the current restored to 0.2 A g⁻¹, the average capacity was 342.2 mA h g⁻¹, indicating its excellent rate capability, with the corresponding discharge and charge profiles shown in Figure S8A. Under the same current density conditions, (Zn,en)VO exhibited the lower capacities of 267.5, 261.4, 229.9, 176.7, 101.3, 55.4, and 33.3 mA h g⁻¹ (Figure 2C). Compared with 0.2 A g⁻¹, the capacity retention at different current densities is shown in Figure S8B. N-(Zn,en)VO shows better capacity retention at high current densities, reasonably related to the intrinsic conductivity increase and fast ion transport ability of the material. Furthermore, the N-(Zn,en)VO electrode exhibited excellent energy and power densities (based on the mass of the active material). N-(Zn,en)VO displayed an energy density of 250.2 Wh kg⁻¹ with 142.8 W kg⁻¹ at the current density of 0.2 A g⁻¹ (Figure 2D). Notably, the cathode showed a distinguished power density of 10 393 W kg⁻¹ and an energy density of 65.3 Wh kg⁻¹ at 20 A g⁻¹, which is superior or comparable to the previously reported cathodes H₂V₃O₈, Mg_xV₂O₅, ZnHCF, V₂O₅@PANI, ZnMn_1.86O₄, MoS₂, δ-MnO₂ Na₃V₂(PO₄)₂F₃, Zn₃V₂O₇(OH)₂·2H₂O, and VS₂ (Table S5).^7,16c,34

FIGURE 2. Cycling performances of the N-(Zn,en)VO at (A) 5 A g−1 and (B) 0.5 A g−1. The inset shows that a coin battery can power the digital temperature-hygrometer. (C) Rate performance of the N-(Zn,en)VO and (Zn,en)VO at selected current densities. (D) Ragone plots of N-(Zn,en)VO cathode and other reported cathode materials

The quantitative capacitance effect of the electrode was analyzed using cyclic voltammetry (CV), to reveal the electrochemical kinetics of N-(Zn,en)VO. Figure 3 A exhibited CV profiles of the N-(Zn,en)VO electrode at different scan rates from 0.1 to 1.5 mV s⁻¹. The two pairs of oxidation/reduction peaks in the CV curve corresponded to the change in the vanadium valence state. At higher scan rates, the CV profiles maintained a similar shape, indicating the N-(Zn,en)VO electrode tolerance to ion intercalation/extraction.³⁵ Moreover, the differential capacity dQ/dV curve of the N-(Zn,en)VO electrode exhibited two pairs of peak at 0.48 V/0.75 V and 0.92 V/1.01 V (200th cycle at 0.5 A g⁻¹), which are consistent with the CV results (Figure S9A). The results indicate that the capacity of N-(Zn,en)VO is mainly ascribed to the valence state change of vanadium. The capacitance effect was analyzed by the linear relationship in Equation (1) between the peak current (i) and scan rate (v):^7,16c [Image Omitted. See PDF]where a and b are adjustable parameters. Generally, the value of b is between 0.5 and 1, where b = 0.5 represents the ion diffusion process and b = 1 the capacitive control process.^36,37 Figure 3C shows the b value obtained from the slope of log (i) versus log (v). The b values of peaks 1–4 were 0.67, 0.98, 0.98, and 0.66, respectively, indicating that the charge storage process was affected by both diffusion-controlled and capacitive contributions. In the high-voltage range, the chemical reaction is dominated by the capacitive behavior (peaks 2 and 3), whereas ion diffusion affects the reaction process in the low-voltage range (peaks 1 and 4). Furthermore, the capacitance contribution can be calculated using Equation (2):^36,37 [Image Omitted. See PDF]where k₁·υ and k₂·υ^1/2 represent the contributions of the capacitance effect and ion diffusion, respectively. As displayed in Figure S9B, 71.1% of the total charge was allocated to the capacitance response (green area) at 1 mV s⁻¹. As the scan speed changed from 0.1 to 1.5 mV s⁻¹, the capacitance contribution increased from 40.3% to 75.3% (Figure 3D). For comparison, the CV curves of (Zn,en)VO at the corresponding scanning speeds are shown in Figure S10, and discussed in Supporting Information. At the same scan rate, the capacitance contribution of N-(Zn,en)VO was higher than that of (Zn,en)VO, consistently with the better rate performance of N-(Zn,en)VO. Figure 3E displays the CV curves of the N-(Zn,en)VO and (Zn,en)VO electrodes at 0.2 mV s⁻¹. For the N-(Zn,en)VO electrode, a lower cathodic peak potential and a larger peak area were measured, with the anodic peak shifted to a higher potential compared to that of (Zn,en)VO. Therefore, N-(Zn,en)VO exhibited a smaller polarization voltage (192 mV) than (Zn,en)VO (307 mV), correlated to a better reaction reversibility.^38,39 The smaller polarization voltage of N-(Zn,en)VO than (Zn,en)VO is attributed to the improved reaction kinetics of the electrode and the reduced diffusion energy barrier of Zn²⁺.^20,39 To quantify the electrode reactivity, the activation energy (E_a) of Zn²⁺ diffusion was measured according to R_ct in the temperature range of 25–85°C (Figure S11). E_a was calculated by the Arrhenius Equation (3):⁴⁰ [Image Omitted. See PDF]where A is the frequency factor, R is the gas constant and T is the temperature. Figure S11 shows a higher E_a for (Zn,en)VO (15.13 kJ mol⁻¹) than for N-(Zn,en)VO (8.73 kJ·mol⁻¹), revealing that the nitridation treatment effectively improved the electrochemical activity of the cathodic material. The improvement in performance after N-doping is further analyzed by the Mott–Schottky plot, as shown in Figure S12A. N-(Zn,en)VO and (Zn,en)VO show negative slopes, revealing that they are p-type semiconductor.²⁰ Besides, the slope of N-(Zn,en)VO is smaller than that of (Zn,en)VO, which means that N-(Zn,en)VO processes higher charge carrier density due to the presence of oxygen vacancies and N-doping.²⁰ Therefore, nitridation-treated N-(Zn,en)VO exhibited higher capacity and longer lifetime.

FIGURE 3. (A) CV curves of N-(Zn,en)VO electrodes at various scan rates. (B) dQ/dV curve of the N-(Zn,en)VO of the 200th cycle at 0.5 A g−1. (C) log (peak current) versus log (scan rate) plots according to the CV data at selected oxidation/reduction peaks. (D) The capacitive contributions of N-(Zn,en)VO at different scan rates. (E) CV curves of N-(Zn,en)VO and (Zn,en)VO electrodes at 0.2 mV s−1. (F) Calculation of Zn2+ diffusion coefficient of N-(Zn,en)VO based on GITT. (G) Alternating-current impedance plots of N-(Zn,en)VO after different cycles. (H) Summary of Rct fitting results after different cycles (inset, equivalent circuit). (I) Zn2+ diffusion coefficient of N-(Zn,en)VO calculated based on EIS results

The diffusion coefficient (D_Zn) of the Zn²⁺ ions in the N-(Zn,en)VO cathode was estimated using the galvanostatic intermittent titration technique (GITT). Figure S12B shows a typical GITT curve for the second cycle. The battery was discharged or charged at a galvanostatic current of 50 mA g⁻¹, with a pulse time of 20 min, and then relaxed for 120 min to achieve the voltage equilibrium. The detailed calculation process is presented in the Supporting Information. The calculated D_Zn of N-(Zn,en)VO was in the range 10⁻⁹–10⁻¹⁰ cm² s⁻¹ (Figure 3F), which is higher or comparable to that of other vanadium-based materials, such as (Zn,en)VO (10⁻⁹–10⁻¹¹ cm² s⁻¹, Figure S12D), V₂O₅ (10⁻¹⁰–10⁻¹¹ cm² s⁻¹), V₅O₁₂·6H₂O (10⁻¹⁰–10⁻¹¹ cm² s⁻¹), and Zn_0.25V₂O₅·nH₂O (10⁻⁹–10⁻¹⁰ cm² s⁻¹).^35,36,41 The appropriate kinetics of the N-(Zn,en)VO cathode achieved rapid Zn²⁺ migration, leading to a good rate performance. To gain insights into N-(Zn,en)VO electrochemical behavior, electrochemical impedance spectroscopy (EIS) measurements were conducted at open circuit voltage in the frequency range 0.01–100 kHz. The charge transfer resistance (R_ct) was evaluated by EIS, whose equivalent circuit is shown in Figure 3H. The initial R_ct of Zn/N-(Zn,en)VO was 131.3 Ω (Figure 3G,H). After five cycles, the resistance dropped sharply to 5.8 Ω, and then slightly increased to 24.4 Ω after 500 cycles. The decrease in R_ct after five cycles is reasonably explained by the activation of the cathodic materials and the capture of a small amount of zinc ions, which are connected to the [VO] layer and contribute to more charge transfer channels during the cycle.^42,43 Moreover, D_Zn of zinc ions was calculated from the linear correlation between the low frequency and the real part of the impedance in the EIS curve (Figure S13). The calculated D_Zn of N-(Zn,en)VO in the pristine state (8.6 × 10⁻¹⁶ cm² s⁻¹, Figure 3I) gradually increased to 1.09 × 10⁻¹⁴ cm² s⁻¹ after 50 cycles, confirming the initial activation process of the electrode.

To explore the effect of different nitriding temperatures on the electrochemical performance of the cathode, samples were prepared at 200, 300, and 400°C. Figure S14 shows the XRD patterns, EIS spectra, CV curves, XPS spectra, and electrochemical performance of the samples treated at different temperatures. The contents of nitrogen doping and oxygen vacancies were enhanced with increasing nitridation temperature based on XPS results. Combined with the XPS analysis of samples with different nitridation times (Figure S5D), the oxygen vacancies in the materials can be effectively regulated by adjusting the nitridation temperature and reaction time. The results of XRD, EIS, CV, and galvanostatic charge/discharge tests indicated nitriding at 200°C as the optimized condition leading to the best electrochemical performance (Figure S14). The excellent performance was attributed to the stable existence of interlayer organic molecules under low-temperature nitridation treatment, confirming the crucial effect of the organic amine on the cathode performance. The organic amine long chain expands the interlayer distance of vanadium oxide responsible for the storage of Zn²⁺ ions. Furthermore, the pre-intercalation of metal ions can stabilize the vanadium oxide layered structure while improving the conductivity of the material and promoting the zinc ion migration (Figure S15). Therefore, additionally to the role of oxygen vacancies and nitrogen doping introduced by the nitridation treatment, the organic–inorganic pre-intercalation contributed to the vanadium-oxide electrochemical performance improvement.

DFT calculations were conducted to further explore the effects of nitrogen doping and oxygen defects on the N-(Zn,en)VO electrode performance. The calculation models of (Zn,en)V₇O₁₆, (Zn,en)V₇O_16−x (with O vacancies) and N-(Zn,en)VO (with O vacancies and N-doping) are shown in Figure S16. Figure 4A–C display the partial density of states and total density of states (PDOS and TDOS) of the three models. The valence and conduction bands are mainly composed of V 3d and O 2p orbitals. The negative and positive values of DOS represent the down-spin and up-spin electrons, respectively.³³ (Zn,en)V₇O_16−x (0.48 eV) exhibited a smaller bandgap than (Zn,en)V₇O₁₆ (0.83 eV) due to the introduction of oxygen vacancies. After nitrogen doping, the bandgap of N-(Zn,en)VO was further reduced to 0.29 eV (Figure 4C), revealing that both oxygen vacancies and nitrogen doping decreased the bandgap. The energy band structures of the three calculation models further revealed the narrowing of the bandgap (Figure S17), which enhanced the excitation of the charge carrier to the conduction band, favoring the transfer of electrons in the redox reaction, consistently with the increase in conductivity measured in the experiment (Figure S15B).³² Figures 4D and S18 indicate the possible binding sites of zinc ions after structural optimization. The calculated binding energies of the corresponding sites (Figure 4E) highlight the highest binding energy (−0.96 eV) for the (Zn,en)V₇O₁₆ VO site, followed by the binding energy of the NH₂ site (−0.51 eV), indicating the intercalated-zinc-ion tendency to be stored in the VO site. Furthermore, with the introduction of oxygen vacancies and nitrogen doping, the binding energy of VO further increased to −1.32 and −1.60 eV. This stronger binding energy enhances the zinc-ion storage capacity of the material. The NEB method was used to calculate the diffusion energy barrier of the zinc ions in these models. Based on the binding energies of different insertion sites, Figures 4F and S19 show the possible Zn²⁺ migration pathways, whose corresponding energy profiles are reported in Figure 4G. Specifically, the diffusion energy barrier of Zn²⁺ in (Zn,en)V₇O₁₆ was 0.513 eV. The presence of oxygen vacancies and N-doping substantially reduced the migration barrier. The energy barriers of (Zn,en)V₇O_16−x (0.406 eV) and N-(Zn,en)VO (0.325 eV) (Figure 4G) emphasize the fast zinc-ion diffusion kinetics of N-doped defective vanadium oxide (N-(Zn,en)VO).

FIGURE 4. The PDOS and TDOS of (A) (Zn,en)V7O16, (B) (Zn,en)V7O16−x, and (C) N-(Zn,en)VO. (D) The possible binding sites of zinc ions in the N-(Zn,en)VO. (E) The binding energy of specific binding site in (Zn,en)V7O16, (Zn,en)V7O16−x, and N-(Zn,en)VO. (F) The Zn2+ migration pathways for N-(Zn,en)VO. (G) The energy barriers of Zn2+ in the three calculation models

To elucidate the electrochemical mechanism of N-(Zn,en)VO, ex-situ XRD and XPS were conducted to investigate the phase evolution during the charging/discharging processes. Figure 5A shows the ex-situ XRD pattern of the N-(Zn,en)VO cathode in the first cycle. In the pristine state, the main peak of the (001) plane is located at 9.2°, and the corresponding interlayer spacing is 9.6 Å (Figure S20A). When the electrode was immersed in the electrolyte (open circuit), the (001) reflection changed from 9.2° to 8.2°, and the corresponding interlayer spacing increased from 9.6 to 10.80 Å. This result was explained with the intercalation of solvent water. At the initial stage of discharge, the peak corresponding to the (001) plane shifted to a lower angle with the insertion of zinc ions, corresponding to the VO layer spacing expansion (Dis-0.8). However, when further discharged to 0.5 and 0.2 V, the interlayer spacing of the cathode maintained at 11.0 Å. This phenomenon was reasonably caused by the increased screening of the interlayer electrostatic repulsion with the higher Zn²⁺ content.⁴¹ Compared with the open circuit state, the peak position in the fully charged state shifted slightly to a lower angle, indicating that some zinc ions remained between the layers (Cha-1.4). The TEM images of N-(Zn,en)VO in the fully discharged/charged state (Figure S20B,C) show a nanotube-like structure. The lattice spacing corresponding to the (001) plane was compatible to the XRD results. When the soaked electrode was completely dried (120°C, 12 h) the XRD diffraction pattern returned to a pristine position (Figure 5B), confirming that the solvent water accessed the interlayer when the electrode was immersed in the electrolyte. At the same temperature, the TG curve of the electrode soaked in electrolyte exhibited more mass loss, corroborating this conclusion (Figure S20D). The intercalated solvent water and the inherent crystal water in N-(Zn,en)VO act as lubricant and reduce the effective charge of Zn²⁺ ions, facilitating the Zn²⁺ ion migration.¹⁵

FIGURE 5. (A) XRD patterns at pristine state and the selected discharge/charge states (points A–G), and corresponding galvanostatic charge and discharge curves at 0.1 A g−1. (B) The XRD patterns of N-(Zn,en)VO in different states: the pristine state, the electrode immersion in the electrolyte for 6 hours, and the completely dry electrode. (C–E) XPS spectra of V 2p, O 1s, and C 1s regions at the pristine, first discharged (0.2 V), and first charged (1.4 V) states

XPS was conducted to analyze the chemical state changes of the elements in different charge and discharge states. Figure S21A shows the survey spectra of the N-(Zn,en)VO cathode in different cycling states, confirming the presence of V, O, N, C, Zn, and F in the samples, where F mainly derived from the PVDF binder. The Zn 2p spectra (Figure S21B) revealed a certain amount of Zn in the pristine N-(Zn,en)VO originating from the ion exchange. The increase in Zn intensity in the discharged samples confirmed the intercalation of Zn ions. The fully charged Zn 2p showed some remaining zinc ions, namely trapped Zn²⁺ connecting the [VO] layers to provide more charge transfer paths, consistently with the EIS analysis. Figure 5C shows the V 2p XPS spectrum of N-(Zn,en)VO in the first cycle. For the pristine state, the peaks of 517.5 and 516.8 eV (V 2p_3/2) were attributed to V⁵⁺ and V⁴⁺, respectively. With the intercalation of Zn²⁺ and the reduction of vanadium during the discharge, a new peak, attributed to V³⁺, appeared at 515.6 eV. When charged to 1.4 V, the spectra returned to their initial state, and the V³⁺ signal disappeared. To further verify the reversible change of the material, the V 2p of the fifth charged state was tested. The spectrum of the fifth charged state is consistent with the pristine and first charged states, confirming the high reversibility of N-(Zn,en)VO during the cycle (Figure S22A). In the O 1s spectra of the pristine state (Figure 5D), the peaks at 530.0, 531.2, and 532.4 eV were assigned to lattice oxygen (OV), physically adsorbed water (OH), and oxygen defects, respectively.⁴⁴ The signal of oxygen defects disappeared in the discharged state, indicating that certain zinc ions occupied the oxygen vacancies upon zinc ion intercalation, and confirming that the oxygen vacancies enabled more storage sites for the zinc ions and increased the specific capacity of the N-(Zn,en)VO. Upon charging, the signal of oxygen defect reappeared as zinc ions were released. The C 1s signal of N-(Zn,en)VO in the initial state was composed of three peaks, corresponding to CH/CC, CN, and CO (Figure 5E). A new peak, ascribed to CF₃SO₃⁻, appeared at higher binding energies in the charged and discharged states, corresponding to adsorbed zinc trifluoromethanesulfonate.¹⁵ Figure S22B shows the N 1s spectra in different states, where the peaks at 399.7 and 401.5 eV, corresponding to NV and NH₂, respectively, highlighted the amine chain stability during the cycle.

To further reveal the evolution of the electrode structure and morphology during the long cycle, XRD and SEM images at different cycle numbers were examined. Figure 6 A shows the XRD patterns of the fully charged and discharged samples with different cycles at 5 A g⁻¹. In the first cycle, the peaks corresponding to the other phases were absent. In the first discharged state, the main peak corresponding to the (001) plane broadened with the insertion of zinc ions, consistently with other organic–inorganic pre-intercalated vanadium oxides.¹⁷ With the de-intercalation of the zinc ions, the peak intensity of N-(Zn,en)VO increased in the charged state. Moreover, new peaks corresponding to Zn₃(OH)₂V₂O₇·2H₂O appeared at the 500th and 1000th cycle. The intensity of Zn₃(OH)₂V₂O₇·2H₂O increased at the 1000th cycle compared with that at the 500th cycle, indicating that the new phase gradually accumulated upon cycling. Figure 6B shows the surface morphology of the electrode in the pristine state. The N-(Zn,en)VO nanotubes were mixed with conductive additives and binders distributed on the Ti current collector. As shown in Figure 6C and S23A, two nanosheet structures with different specifications appeared on the electrode surface in the 500th discharge state. In the subsequent charged state (Figures 6D and S23B), larger nanosheets remained, whereas the smaller nanosheets disappeared. The XRD analysis assigned the large nanosheet structure to Zn₃(OH)₂V₂O₇·2H₂O. This reversible disappearance of the small nanosheets reasonably corresponded to Zn_x(OH)_y(CF₃SO₃)·nH₂O,^17,45 which was not detected by XRD because of its low abundance. The change in morphology at the 1000th cycle is similar to that at the 500th cycle (Figure S23C–F). Additionally, the thickness of the nanosheets was measured in the fully charged state at the 500th and 1000th cycles. As shown in Figures 6E,F and S24, the average thickness of the nanosheets after 500 cycles was 122 nm, and the average thickness after 1000 cycles was 193 nm, indicating a gradual increase of the Zn₃(OH)₂V₂O₇·2H₂O content upon cycling, consistently with the previous XRD analysis.²⁰ Moreover, the XPS spectra after 1000 cycles are consistent with the initial spectra, further confirming the N-(Zn,en)VO stability (Figure S25). According to the XPS area ratio, O_V accounted for 12.9% in the first charge state (Figure 5D), and increased to 18.2% at the 1000th charge state (Figure S25B). The oxygen vacancies generated during long cycling may be related to the microstrain during the migration of zinc ions.⁴⁶ The abundant oxygen vacancies provide more storage sites for Zn ions, enhancing the specific discharge capacity of the cathode, which further reveals the capacity increase in the initial several cycles. The Zn₃(OH)₂V₂O₇·2H₂O formation primarily originated from the interaction between OH⁻ and N-(Zn,en)VO, which promoted the insertion of H⁺ ions into the cathode. To corroborate this mechanism, the battery performance was tested in 0.3 M Zn(CF₃SO₃)₂ organic electrolyte and organic electrolyte with different water contents (Figure S26A). In the 0.3 M Zn(CF₃SO₃)₂ pure acetonitrile electrolyte, N-(Zn,en)VO showed a lower open circuit voltage and a limited discharge capacity (0.77 V, 55.7 mA h g⁻¹). Contrarily, N-(Zn,en)VO exhibited an open circuit voltage of 0.91 V and a specific discharge capacity of 223.0 mA h g⁻¹ in a 0.3 M Zn(CF₃SO₃)₂ aqueous electrolyte. Therefore, the hydrogen content increase was beneficial to the material electrochemical performance. The EIS spectra further confirmed this conclusion (Figure S26B). As the water content in the electrolyte increased, the initial R_ct of N-(Zn,en)VO gradually decreased. Combined with the electrode morphology analysis, apart from the Zn²⁺ storage mechanism, H⁺ participated in the electrode reaction and contributed to the discharge capacity.

FIGURE 6. (A) The ex situ XRD patterns of N-(Zn,en)VO at different cycle. SEM image of N-(Zn,en)VO at (B) pristine state. SEM images of N-(Zn,en)VO at (C) fully discharged state and (D) fully charged state after 500 cycle. (E,F) SEM images of nanosheets in the fully charged state at 500th and 1000th cycles

CONCLUSIONS

In summary, we successfully synthesized an N-doped (Zn,en)VO nanotube cathode with abundant oxygen vacancies. The presence of oxygen vacancies enhanced the storage sites for zinc ions and increased the specific discharge capacity of the cathodic electrode. The introduction of N-doping and oxygen vacancies reduces the material bandgap, enhances the charge carrier density, and promotes the occurrence of redox reactions for N-(Zn,en)VO. Moreover, nitrogen doping and oxygen vacancies reduced the diffusion barriers and accelerated the migration of Zn²⁺ ions. Benefiting from the defect treatment and organic–inorganic pre-intercalation, Zn/N-(Zn,en)VO batteries exhibited fast reaction kinetics, enhanced zinc-ion storage performance, excellent rate performance, and long cycle span. The design strategy developed in our study can be further extended to other materials to improve the cathode performance and promote the practical applications of aqueous ZIBs.

AUTHOR CONTRIBUTIONS

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

ACKNOWLEDGMENTS

This work is supported by the National Natural Science Foundation of China (Grant no. 52072224, 51902187, and 51732007), the Natural Science Foundation of Shandong Province (ZR2020YQ35 and ZR2018BEM010), the Qilu Young Scholar Funding of Shandong University, the Young Elite Scientists Sponsorship Program by CAST (YESS, 2019QNRC001) and the support from Collaborative Innovation Center of Technology and Equipment for Biological Diagnosis and Therapy in Universities of Shandong. The authors would like to thank Shiyanjia Lab.

CONFLICT OF INTEREST

The authors declare no conflict of interest.