The structure architecture of the interpenetrating carbon/vanadium nitride networks, oxygen functional group-containing surfaces, and hierarchical porous structure is schematically shown in Scheme 1. The polymer casting solution including PAN, PEG, and BCP (PAN-b-PMMA-b-PAN) was spin coated and immersed into a non-solvent water bath to induce phase separation for membrane formation. In the process, vanadyl acetylacetonate was added to the casting solution, as vanadyl acetylacetonate and PAN show good blending compatibility in DMF at a suitable temperature. We sought to determine the optimal conditions for uniform dispersion of V into the polymer system in the polymer precursor/V source hybrid system. Through thermotreatment under a NH3/N2 atmosphere, interpenetrating carbon/VN networks were obtained. In the networks, VN was dispersed uniformly and was wrapped by the carbon scaffold. During the phase process, PEG moved to the surface of the membrane, which generated oxygen-containing functional groups on the surface. The functional groups through sintering improved the infiltration of the electrode in electrolyte solution. Solvent exchange, PEG immigration, and self-assembly of BCP created an asymmetric 3D polymer membrane with hierarchical porous structure. This unique structure was preserved in the final carbon/VN material after heat treatment. Solvent exchange and PEG migration leave micropores with a variety of pore sizes. On the other hand, self-assembled PMMA blocks from BCP were used as a “sacrificed chain” and generated abundant mesopores during high-temperature carbonation. In addition, under a NH3 and N2 mixed atmosphere at 800 °C, VN/C (I) was doped with nitrogen and underwent further modification.
SEM, TEM, and N2 adsorption–desorption isotherms of VN/C (I) were used to study the morphology as well as pore structure and distribution, as shown in Fig. 1. Typical SEM cross-sectional views of the VN/C (I) sample showed a classic membrane structure formed by the liquid–liquid phase-separation method (Fig. 1a–c). The representative morphology of the asymmetric membrane with a thickness of approximately 5.0 μm was observed with a regular and uniform shape and porous structure (Fig. S1). By comparing the SEM images of top and bottom sections of the VN/C (I) (Fig. 1b–c) and VN/C (II) (Fig. S1b–c), VN/C (I) exhibited more mesoporous features, as shown in the high-magnification SEM images. According to the SEM images, as shown in Fig. 1d and S2, a uniform concave size appeared in the VN/C (I) sample that was caused by the aggregation of BCP and loss of PMMA blocks during phase inversion and thermotreatment, respectively. However, the surface of the VN/C (II) sample was smooth, fine, and compact with few pores, which significantly decreased ion accumulation at its surface. In addition, from the TEM images, as shown in Fig. 1e–f and S3, abundant homogeneous mesopores approximately 40–50 nm in size were observed throughout the VN/C (I) material, but none were observed in VN/C (II). The porous structures of the membranes were also analyzed by nitrogen adsorption. As shown in Fig. 1g–h and S4, and Table S1, the VN/C (I) sample exhibited a typical type-IV curve with an average pore diameter of 1.9 nm and total pore volume of 0.24 cm3 g−1. A wide pore size distribution range from 1.7 to 250 nm was observed, which manifested as a loose network with micro-, meso-, and macropores. In comparison, the VN/C (I) sample exhibited a larger specific surface area (SSA, 523.5 m2 g−1) than that of VN/C (II) (504.2 m2 g−1), which was also much higher than previously reported VN and carbon composite materials [38, 41,42,43,44]. Notably, the pore distribution range with larger pores, formed by adding PEG and BCP to the casting solution, can act as ion buffer pools by storing electrolytes and facilitating ion transmission to improve the efficiency of the material ratio surface area.
Through the phase-separation and thermotreatment process, an interpenetrating polymer/V-sources network was formed followed by fabrication of interpenetrating carbon/VN networks. Figure 2 shows the correlation representational data that highlight the homogeneity of the VN/C (I) sample. The high-resolution TEM image (Fig. 2a) exhibited a large number of small VN quantum dots with 5–8 nm in size, which was evenly and densely embedded in the carbon substrates. The low-crystalline nature of VN/C (I) was indicated by surface area electron diffraction (SAED) measurements (Fig. 2c). The XRD pattern of the VN/C (I) sample (Fig. 2b) exhibits a broad peak at approximately 22°, indicating the presence of amorphous carbon derived from the PAN precursor in the homopolymer or BCP. This carbon scaffold improves the electrical conductivity of the material and increases the utilization and stability of VN quantum dots. In addition, a slightly stronger peak at 43.6° and two weak peaks at approximately 37.4° and 63.4° were, respectively, ascribed to the (2 0 0), (1 1 1), and (2 2 0) diffractions of the VN (ICDD PDF 35-768) [27, 31]. The XRD results and SAED pattern indicate that the prepared VN/C (I) contained amorphous carbon and VN with low crystallinity. Figure 2d shows the TEM elemental mapping images of VN/C (I), which revealed uniform distributions of C, N, O, and V throughout the material. In conclusion, these results indicated that the VN nanoparticles were evenly distributed in the substrate material. The carbon provided active sites for electrolyte ions by preventing VN grain growth and aggregation [22, 23, 27, 32,33,34,35,36].
To further explore the surface modification, XPS of the VN/C (I) (Fig. 3) and VN/C (II) (Fig. S5) samples was performed. The overall XPS spectra show that the surface of samples consisted of C, N, V, and O, and the corresponding analytical results are summarized in Table S2. The respective proportions of C, N, V, and O in VN/C (I) and VN/C (II) were 88.0, 3.6, 2.3, 6.1 at%, and 87.8, 4.2, 2.7, 5.3 at%, respectively. Three main peaks at 284.7, 285.6, and 286.2 eV were observed in the C 1 s spectrum in Fig. 3b and were attributed to C–C, C–N, and C–O bonds, respectively [9, 13, 45, 46]. As illustrated in Fig. 3c, the N 1 s signal could be partitioned into four characteristic peaks at 398.3, 400.0, 401.1, and 403.2 eV corresponding to pyridinic N (N-6), pyrrolic N (N-5), graphitic N (N-Q), and oxygenated N (N–O), respectively [19, 46,47,48]. As shown in Fig. 3d, the peaks centered at 531.1, 532.6, and 534.5 eV were assigned to V–N–O, C = O/N–O, and C–OH, respectively [19, 31, 39, 40]. In addition, the peaks at approximately 514.1 and 521.6 eV typical of vanadium in the VN structure, and the peaks at 517.1 and 524.4 eV arose from the V–O bonds on the surface of VN/C (I) (Fig. 3e) [22,23,24, 30, 31, 49]. Comparing the C 1 s, N 1 s, and O 1 s spectra of the two samples, it is clear that more oxygen-containing groups such as C–O (20.0 at%), N–O (20.9 at%), and C–OH (29.7 at%) were present in VN/C (I) compared to the VN/C (II) sample (C–O 17.1 at%, N–O 12.3 at%, and C–OH 16.9 at%). These results indicated that PEG formed oxygen functional groups through sintering at the surface, which improved the infiltration of the VN/C (I) electrode.
To investigate the electrochemical capacitive performance of the prepared samples, CV, GCD, and EIS were measured using a three-electrode system in 6 M KOH aqueous electrolyte with a SCE and Pt as the reference and counter electrodes, respectively. The CV curves of the VN/C (I) and VN/C (II) samples were obtained at a scan rate of 10 mV s−1 in the potential range of − 1.2–0 V, as shown in Fig. 4a. Both curves were quasi-rectangular in shape and featured broad redox humps, indicating typical double-layer capacitive behavior with Faradaic reactions [48,49,, 50, 51]. However, the CV curves of the VN/C (I) showed much bigger curve areas compared to those of VN/C (II), indicating a higher capacitance. As shown in Fig. 4b, the GCD curves of the samples were measured at a current density of 0.5 A g−1. VN/C (I) showed a linear and slightly asymmetric triangle shape resembling the characteristics of a normal double-layer capacitor and indicating satisfactory electrochemical reversibility. The discharging time required for the VN/C (I) sample was longer than that of VN/C (II), indicating the better capacitance of VN/C (I). Moreover, from the relevant calculations, the mass specific capacitances of VN/C (I) and VN/C (II) were 392.0 and 245.1 F g−1, respectively, at a current density of 0.5 A g−1. EIS tests were also performed over a frequency range of 0.01 Hz to 100.0 kHz (Fig. 4c). The impedance curves contained one semicircle at a high frequency and a linear feature at low frequency. In addition, the internal resistance of the VN/C (I) (0.55 Ω) electrodes, acquired from the intercept of the plots on the real axis, was much smaller than that of VN/C (II) (0.64 Ω). This indicated good infiltration of the electrolyte caused by the introduction of oxygen-containing functional groups. Because of the weakened conductivity, the diameters of the semicircles of the VN/C (I) samples were larger than that of VN/C (II). Moreover, the conductivity was measured using a 4-point probe resistivity measurement system (RTS-9), and values of 8.3 and 9.7 S cm−1 for VN/C (I) and VN/C (II), respectively, were obtained. These results agreed well with the smaller charge transfer resistance of VN/C (II) indicated by the EIS analysis. In addition, the plots of the VN/C (I) Warburg angle were higher than those of the VN/C (II), indicating that the abundant pore structure was beneficial for the diffusion of electrolyte ions and resulted in a small diffusion impedance. Figure 4d shows the specific capacitances of the samples at different current densities. When the current density increased from 0.5 to 30 A g−1, the capacitance retention values for VN/C (II) and VN/C (I) were 46.1 and 50.5%, respectively. Thus, in terms of comprehensive capability, the VN/C (II) electrode was shown to be more suitable for use in supercapacitors.
The electrochemical behavior of VN/C (I) at various current densities was also investigated. As shown in Fig. 5a, all CV loops were nearly quasi-rectangular in shape and almost no deformation was observed at high scan rates, indicating a small internal resistance. The low internal resistance was likely due to good wettability of the VN/C (I) electrode with electrolyte and its hierarchical porous structure, which was important for electron transport. The GCD curves (Fig. 5b) obtained at various current densities from 0.5 to 5 A g−1 exhibited a nearly linear and typical triangular symmetrical trend, demonstrating good electrode-reaction reversibility of the VN/C (I). The calculated specific capacitances were 392, 322, 295, 280, 276, 267 F g−1 at different current densities of 0.5, 1, 2, 3, 4, 5 A g−1, respectively. In addition, an excellent cycling stability of 83.5% was obtained at a current density of 2 A g−1 after 5000 cycles (Fig. 5d). Figure S6 shows the low- and high-resolution TEM images of VN/C (I) after 5000 cycles, which maintained their original morphology, indicating high stability. The TEM images show numerous small VN quantum dots homogeneously embedded in the porous carbon substrate. This explicitly indicates that the carbon matrix can prevent VN aggregation and simultaneously act as an active material for charge storage during the charging/discharging process. In conclusion, all the electrochemical results show convincingly that the prepared VN/C (I) is a promising electrode material.
To accurately assess the performance of the developed material for practical application, an asymmetric supercapacitor featuring a two-electrode system was assembled using VN/C (I) in 6 M KOH as the negative electrode and Ni(OH)2 as the positive electrode. Figure 6a shows the CV curves of the hybrid device over the voltage range of 0–1.6 V at various scanning rates between 10 and 50 mV s−1. In addition, a couple of wide oxidation reduction peaks at 1.0 V were observed, which were likely caused by the pseudocapacitive reactions related to the positive Ni(OH)2 and negative VN/C (I) electrodes. Figure 6b shows the linear potential–time relationship of the GCD curves of the hybrid device at different current densities from 1 to 5 A g−1 at working potential window of 1.6 V. The specific capacitance measured at the current density of 1 A g−1 was calculated to be 122 F g−1, and the retained capacitance was 91 F g−1 when the current density increased to 5 A g−1. As shown in Fig. 6c, the EIS of the hybrid device was tested in from 0.01 Hz to 100 kHz at room temperature. Notably, a small intercept at the real axis at approximately 0.87 Ω was observed, indicating a lower intrinsic resistance of the supercapacitors (SCs). Figure 6d shows a high rate performance where approximately 74.6% of specific capacitance was retained as the current density was raised from 1 to 5 A g−1. Moreover, the Ragone plots, as shown in Fig. 6e, showed the relationship between energy and power densities. The Ni(OH)2//VN/C (I) device showed an excellent energy density of 43 Wh kg−1 and high-power density of 800 W kg−1. From the comparison of power and energy densities in Fig. 6e and more detailed information in Table S3, the performance of the SCs was superior to related materials in the recently published papers. The cyclic stability of the SCs was determined by repetitive operation of the galvanostatic charging/discharging process (Fig. 6f). The Ni(OH)2//VN/C (I)-based SCs demonstrated excellent life cycle stability with 82.9% initial capacitance retention after 8000 cycles at a current density of 1.0 A g−1. Overall, the application potential of the prepared hybrid device was demonstrated by thorough characterization of all relevant electrochemical characteristics.