Porous V2O5/RGO/CNT hierarchical architecture as a cathode material: Emphasis on the contribution of surface lithium storage

A three dimensional vanadium pentoxide/reduced graphene oxide/carbon nanotube (3D V2O5/RGO/CNT) composite is synthesized by microwave-assisted hydrothermal method. The combination of 2D RGO and 1D CNT establishes continuous 3D conductive network, and most notably, the 1D CNT is designed to form hierarchically porous structure by penetrating into V2O5 microsphere assembly constituted of numerous V2O5 nanoparticles. The highly porous V2O5 microsphere enhances electrolyte contact and shortens Li+ diffusion path as a consequence of its developed surface area and mesoporosity. The successive phase transformations of 3D V2O5/RGO/CNT from α-phase to ε-, δ-, γ-, and ω-phase and its structural reversibility upon Li+ intercalation/de-intercalation are investigated by in situ XRD analysis, and the electronic and local structure reversibility around vanadium atom in 3D V2O5/RGO/CNT is observed by in situ XANES analysis. The 3D V2O5/RGO/CNT achieves a high capacity of 220 mAh g−1 at 1 C after 80 cycles and an excellent rate capability of 100 mAh g−1 even at a considerably high rate of 20 C. The porous 3D V2O5/RGO/CNT structure not only provides facile Li+ diffusion into bulk but contributes to surface Li+ storage as well, which enables the design of 3D V2O5/RGO/CNT composite to become a promising cathode architecture for high performance LIBs.


Results and Discussion
Porous V 2 O 5 /RGO/CNT hierarchical architecture. One-step MAH method which is an energy effective dielectric heating system for producing high quality nanomaterials under fast kinetics of crystallization 28 was applied to obtain the 3D V 2 O 5 /RGO/CNT composite. Acid treated o-CNTs were well dispersed in amphiphilic GO solution through π -π interaction, and then reduction of V 2 O 5 precursor had occurred in the mixture of GO and o-CNT with the reducing agent of ascorbic acid by forming V x O y /RGO/CNT ternary composite. Through the oxidation process, a 3D V 2 O 5 /RGO/CNT ternary microsphere was obtained and applied to cathode material. To confirm the individual role of RGO and CNT in the formation of hierarchical microsphere, V 2 O 5 with only RGO and only CNT were prepared separately. The SEM images of Fig. 1a,b show the morphology of only V 2 O 5 /RGO and only V 2 O 5 /CNT, respectively. Instead of the V 2 O 5 nanoparticle assembly, the V 2 O 5 nanoparticles are totally dispersed onto the surface of 2D RGO (Fig. 1a) or agglomerated each other (Fig. 1b) in the similar way when only V 2 O 5 nanoparticles without RGO or CNT undergo the MAH process (Fig. 1c). However, when the RGO and CNT were added together, the V 2 O 5 assembly was obtained as shown in Fig. 1d,e. From the totally different morphologies even through the essentially same synthesis process, the RGO/CNT structure can be regarded as a critical role in forming the V 2 O 5 microsphere. A local interspace confined by RGO and CNT can be a reason for the assembly structure because the V 2 O 5 nanoparticles in the local interspace grow into microsphere via an Ostwald ripening process 29 . Figure 1d,e present the morphology of 3D V 2 O 5 /RGO/CNT composite with the ratio of RGO to CNT at 1:2 and 2:1, respectively. Compared to Fig. 1e, the smaller V 2 O 5 assemblies are observed in Fig. 1d at the same magnification, and the particle size distribution of V 2 O 5 assemblies is not uniform. Moreover, the spheres are less covered with RGO and CNT, as a result, the ratio of RGO:CNT was controlled at 2:1 for subsequent experiments. Based on the local interspace confining the V 2 O 5 nanoparticles and reconstructing them, the amount of RGO was increased twice to confirm the possibility for forming the V 2 O 5 microsphere without the addition of CNT. Only 2D RGOs without CNT can build the confined structure on the assumption that the increased concentration of RGO has more opportunity to obtain a standing RGO between parallel RGOs and thereby makes the local interspace. As shown in Fig. 1f, the V 2 O 5 assembly structure is developed, and irregular V 2 O 5 nanoparticle aggregates over at least 100 nm are randomly anchored on the surface of RGO, in contrast to the only V 2 O 5 /RGO at the low concentration of RGO (Fig. 1a). Although there is an interconnecting RGO between two RGOs as marked with a dotted circle in Fig. 1f, most RGOs seem to be stacked, therefore, it is referred to as 2D V 2 O 5 /RGO and regarded as a control group for investigating the effect of morphological difference between RGO and CNT on V 2 O 5 microsphere structure and its electrochemical properties.
The SEM image of Fig. 2a presents that the V 2 O 5 nanoparticle aggregates below 1 μ m are randomly dispersed in the architecture of 3D V 2 O 5 /RGO/CNT. At higher magnification in Fig. 2b, CNT touching the surface of V 2 O 5 microsphere is observed as indicated by solid arrow, and most notably, certain CNT among the entangled CNTs seems to penetrate the RGO as indicated by dotted arrow. In respect of RGO, the porous RGO interpenetrated by CNTs (dotted arrow) as well as the smooth RGO (solid arrow) are observed, in contrast to the only smooth surface of RGO in 2D V 2 O 5 /RGO, which can indicate the 3D conducting scaffold composed of RGO and CNT in the ternary composite. To investigate the texture properties of composite, N 2 adsorption/desorption measurements of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites were performed as shown in Fig. 2c. The isotherm profile of 3D V 2 O 5 /RGO/CNT composite corresponds to type IV with a hysteresis loop around 0.5-0.7 P/P 0 , suggesting a mesoporous structure, whereas that of 2D V 2 O 5 /RGO is categorized as type II, indicating a negligible development of mesopores (2~50 nm) 30 . The specific surface area and pore volume of 3D V 2 O 5 /RGO/CNT composite are calculated as 70 m 2 g −1 and 0.25 cm 3 g −1 , respectively which are much higher than that of 2D V 2 O 5 /RGO (32 m 2 g −1 and 0.15 cm 3 g −1 , respectively), indicating that the introduction of 1D CNT remarkably enhances the specific surface area and porosity. For comparison, specific surface area and pore volume of V 2 O 5 nanoparticles with only RGO and only CNT are 23 m 2 g −1 and 0.10 cm 3 g −1 and 17 m 2 g −1 and 0.08 cm 3 g −1 , respectively. The pore size distributions of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites are shown in the inset of Fig. 2c. The mesopores ranging from 2 to 10 nm are obviously developed in the ternary composite, whereas the amount of mesopores in 2D V 2 O 5 /RGO is negligible. The mesopores within 2~10 nm in 3D V 2 O 5 /RGO/CNT composite can be resulted from the interspaces of constituent particles, which means the increased gaps between the V 2 O 5 nanoparticles and concurrently emphasizes the role of CNT in the composite for increasing the mesoporosity and consequently specific surface area. According to the XRD patterns in Fig. 2d, all diffraction peaks of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites are indexed to pure orthorhombic crystalline phase of V 2 O 5 (JCPDS-41-1426) 31 . No characteristic RGO and/or CNT peaks in the hybrid composites are observed because the first main peak of RGO and/or CNT overlaps with (110) reflection of V 2 O 5 at around 27° 32 . To determine the carbon content in hybrid composites, TGA was conducted in air. As shown in Fig. 2e, a single step of weight loss between 300-650 °C is observed in both samples, corresponding to the oxidation of the carbon 16 . Excluding physically adsorbed water (below 200 °C), the carbon contents of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites are evaluated to be about 13 and 21 wt%, respectively.
The hierarchical structure of 3D V 2 O 5 /RGO/CNT composite was further examined by using HR-TEM and EDX mapping images. As shown in Fig. 3a, the V 2 O 5 microspheres about 1 μ m are embedded in RGO/CNT matrix. The inset at higher magnification exhibits the CNTs penetrating into the V 2 O 5 assembly, which well describes the porous nature of hierarchical structure. Moreover, the element distributions of V, O, and C are homogeneous in a whole microsphere (Fig. 3b). XPS was carried out to analyze the chemical state of vanadium and the reduction extent of GO in the ternary composite. For the V2p spectrum (Fig. 3c), the binding energy separation between V2p 1/2 (525.1 eV) and V2p 3/2 (517.4 eV) is ~7.7 eV, which corresponds with + 5 oxidation state of V 2 O 5 33 . The peak fitting of C1s spectra in Fig. 3d shows the four main peaks at 284.7, 286.8, 288.3, and 288.3 eV, assigned to C-C/C= C (aromatic ring), C-O (hydroxyl or epoxy group), C= O (carbonyl functional group), and O-C= O (carboxyl group), respectively. Although the C1s spectrum is mixed with RGO and CNT, the highest intensity in the C-C/C= C peak at 284.4 eV suggests the reduction of oxygen functional groups within RGO and the effective restoration of sp 2 carbon network 34 . In situ XRD and in situ XANES analyses of 3D V 2 O 5 /RGO/CNT upon Li + intercalation/ de-intercalation. The structural changes of 3D V 2 O 5 /RGO/CNT composite during discharge and charge  (310)) continuously shift upon Li + intercalation/de-intercalation, and various phase transitions corresponding to the multi-step process that can be divided into four distinct stages in each discharge and charge voltage profile are observed. The position of all reflections after charge process is nearly the same position at pristine state as indicated by black lines, which reveals a structural reversibility upon Li + intercalation/de-intercalation 35,36 . There are numerous metastable phases of Li x V 2 O 5 under chemical or electrochemical Li + intercalation into V 2 O 5 at room temperature 37,38 . When the (001) reflection is focused to examine the phase evolutions in the layered V 2 O 5 structure as a function of Li + composition in the Li x V 2 O 5 (Fig. 4b), V 2 O 5 firstly transforms into α -Li x V 2 O 5 in the composition range of x < 0.1 and then coexists with ε -Li x V 2 O 5 within the composition of x = 0.26. This accords with the phase diagram of the Li x V 2 O 5 system at room temperature, indicating that ε -phase exists solely in the composition range of 0.35 < x < 0.8 37 until the composition of x = 0.93, and then the δ -phase is developed into the main structure as further Li + intercalated. At the composition of x = 1.27, γ -Li x V 2 O 5 appears along with the δ -phase at lower angle and develops until the composition of x = 1.93. After that, the γ -phase changes into ω -Li x V 2 O 5 and maintains the ω -phase until the end of discharge by uptake of x = 2.65. These successive phase transformations from α -phase to ε -, δ -, γ -, and ω -phase are in accord with the general progression of Li + content in the Li x V 2 O 5 phases and the four plateaus at 3.4, 3.19, 2.28, and 2.01 V in the first discharge process [37][38][39] . In addition, the (001) reflection shifts toward lower angles as Li + intercalated, which reflects the increase in the interlayer d spacing of V 2 O 5 layers by uptake of Li + , and the intensity of (001) reflection is decreased at the final ω -phase compared to that of α -phase, which indicates the reduction in structural order upon Li + intercalation 35,36 . More importantly, however, the lower shift returns to nearly the same position at pristine state, and the decreased intensity recovers a respectable amount of pristine state as shown in Fig. 4a, which confirms the structural reversibility of V 2 O 5 in the ternary composite upon Li + intercalation/de-intercalation.
Furthermore, to confirm the changes in oxidation state and chemical environment around the vanadium atom in V 2 O 5 crystals, in situ XANES analysis of the 3D V 2 O 5 /RGO/CNT composite was conducted (Fig. 5) during the first discharge and charge processes at 0.1 C. The vanadium K-edge XANES spectra can be divided into the three features, pre-edge, K-edge, and edge resonance as marked with I, II, and III, respectively in Fig. 5a. During the first discharge, the K-edge continuously shifts to the lower energy value as a function of Li + composition in the Li x V 2 O 5 . This shift represents the decrease in the average oxidation state of vanadium, which corresponds to the increasing amounts of intercalated Li + 40,41 . The pre-edge peak decreases in intensity and shifts toward lower energy value as Li + intercalated. The crystalline V 2 O 5 is composed of VO 5 square pyramids with highly distorted environment 42 . The intensity of pre-edge peak is considerably sensitive to the alteration of local geometric symmetry and directly correlated with the distortion of VO 5 square pyramid 40,41 . An increase in the intensity of pre-edge peak indicates a decrease in vanadyl (V = O) distance 43 , which increases a degree of distortion of local structure and decreases the local symmetry of structure within the VO 5 square pyramid 40,41 . Therefore, the decreased intensity of pre-edge peak during discharge suggests that the local structure of vanadium at the 100% discharged state has higher degree of local geometric symmetry compared to the pristine state. In accord with the K-edge, the pre-edge peak and edge resonance positions are also seen to shift toward lower energy value as Li + intercalated. The intensity of edge resonance is consistent with the symmetry of basal oxygen configuration around the vanadium site 40,41 . As discussed above, the linear decrease in intensity of pre-edge peak represents a reduction of the distortion within VO 5 square pyramid, and the resultant increase of the basal oxygen configuration around the vanadium site is also observed by an increase in the magnitude of edge resonance as Li + intercalated 41 . All the features observed in the XANES spectra during charge process appear on the contrary to discharge process as indicated by blue arrows in Fig. 5b. As shown in Fig. 5c, the oxidation state of vanadium after the first cycle is extremely close to that of the pristine state, which is based on the overlapped K-edge lines between the pristine and 100% charged states. In addition, after the first cycle, the degree of distortion of the VO 5 square pyramid decreases, and the symmetry of basal oxygen configuration around the vanadium site slightly increases compared to the pristine state, which is based on the lowered and the slightly raised intensity of the pre-edge and edge resonance at the 100% charged state, respectively. These two XANES results through comparison of the pristine and 100% charged states reveal the electronic and local structure reversibility in the 3D V 2 O 5 /RGO/CNT composite upon Li + intercalation/de-intercalation, respectively 16 Figure 6a shows the cyclic voltammetry (CV) of 3D V 2 O 5 /RGO/CNT composite during the first five cycles at a scan rate of 0.2 mV s −1 . In the first CV curve, three cathodic peaks at 3.33, 3.11, and 2.13 V correspond to the structural transformation from the α -phase to the ε -, δ -, and γ -phase, respectively (noted as C1, C2, and C3, respectively, for further explanation in the result section), and three corresponding anodic peaks at 2.61, 3.31, and 3.50 V appear in the following anodic scans. As the cycle repeated, the individual redox peaks are decreased slightly. Except for the absence of final phase transition from γ -phase to ω -phase as further lithiated from γ -phase, the structural transformations accord with the in situ XRD results and the previous reports 37, 39 . Figure 6b presents the voltage profiles of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites in the first cycle at 1 C, and C-rate bases on the theoretical capacity of V 2 O 5 (294 mAh g −1 ). The specific capacity is normalized by the mass of only V 2 O 5 in the composites because the capacity of RGO and CNT normalized by the carbon content are negligible in the voltage range of 4.0-2.0 V 44,45 . In agreement with the CV result, the multiple redox plateaus in the voltage profile show the phase transitions from the α -phase to the ε -, δ -, and finally to the γ -phase, respectively. The first discharge and charge capacities of the 3D V 2 O 5 /RGO/CNT composite are 265 and 253 mAh g −1 , respectively, and those of the 2D V 2 O 5 /RGO are 190 and 189 mAh g −1 , respectively, suggesting that more lithium ions with 0.51 Li + are stored in the 3D V 2 O 5 /RGO/CNT than in the 2D V 2 O 5 /RGO composites. The cycle performance (Fig. 6c) presents high capacities and stable capacity retention in  Fig. 6c [46][47][48][49][50][51] . The rate capability of composites was evaluated by increasing the C-rate from 0.1 to 20 C and then returning to 1 C as shown in Fig. 6d. When cycled at 0.1 C, the first charge capacity of 304 mAh g −1 decreases to 290 mAh g −1 for the subsequent 10 cycles, and the capacity become stable as the C-rate increased. Even at a considerably high rate of 20 C, the 3D V 2 O 5 /RGO/CNT composite delivers a high capacity of 100 mAh g −1 with good stability, compared to 68 mAh g −1 of 2D V 2 O 5 /RGO. When the C-rate is reduced back to 1 C after 80 cycles, the 3D V 2 O 5 /RGO/CNT composite recovers a respectable amount of the first 1 C step capacity, demonstrating its superior rate capability.
Moreover, GITT was employed to understand kinetics related with the configuration of 3D V 2 O 5 /RGO/CNT and 2D V 2 O 5 /RGO composites. The chemical diffusivities of Li + in hybrid composites were evaluated by using the GITT curve (Fig. 6e) during the first cycle at 0.1 C for an interval of 10 min followed by a rest of 40 min to obtain the steady state equilibrium voltage (E s ). The Li + diffusion coefficients (DLi + ) of hybrid composites were calculated from the following equation 1 based on Fick's law 52 where m B , V M , M B , and S are mass, molar volume, molecular weight, and active surface area, respectively; L is the characteristic length; Δ E s is the change in steady-state cell voltage for the step in different potential range; and Δ E τ is the total change of cell voltage during the current pulse for time τ . As show in Fig. 6f, "W" shape curves of both samples are indicative of the characteristics of Li + intercalation-host materials 53 . The calculated diffusion coefficients of 3D V 2 O 5 /RGO/CNT are higher than those of 2D V 2 O 5 /RGO on the whole process, implying a favorable Li + diffusivity in the ternary composite. However, the D Li+ of 3D V 2 O 5 /RGO/CNT composite reaches minima values at 3.37, 3.15, and 2.27 V during discharge, corresponding to the plateau voltages of phase transitions in Fig. 6e, and three D Li+ are lower than 2D V 2 O 5 /RGO composite. It can be explained by the more formation of phase transitions (α -phase to ε -, δ -, and γ -phase, respectively) in the 3D V 2 O 5 /RGO/CNT than in the 2D V 2 O 5 / RGO, which is shown by the fact that every plateau of 3D V 2 O 5 /RGO/CNT is longer than that of 2D V 2 O 5 /RGO 54 . In engineered nanoscale materials, a considerable amount of Li + can be stored on the surface of active material besides Li + storage via faradaic process. The 3D V 2 O 5 /RGO/CNT architecture develops the mesoporosity and increases the specific surface area dramatically compared to other structures. To investigate the effect of surface area on Li + storage properties in the ternary composite, the intercalation and capacitive contributions to total capacity were characterized from CV curves at various scan rates by using power law relationship 21,55 as the following equation 2 (Fig. 7), where i is current (A), ν is scan rate (mV s −1 ), and a, b are adjustable parameters. A b-exponent value is determined from the slope of log(| i |) vs. log(ν ) plot, and the b-value of 0.5 indicates the faradaic process such as diffusion-controlled intercalation, conversion, or alloying reactions, while the b-value of 1.0 indicates the surface-limited pseudocapacitance 55 . As shown in Fig. 7a,   and surface-limited reactions, the following equation 3, describing the current response (i) at fixed voltage (V) as the combination of capacitive effect ( ν k 1 ) and diffusion-controlled intercalation effect ( ν k 2 1/2 ), is applied 55 .
i is the current (A) at a given potential, ν is the scan rate (mV s −1 ), and k 1 and k 2 are constants. According to the linear plot of i/ν 1/2 vs. ν 1/2 as shown in Fig. 7b, the electrochemical reactions with phase transitions from the α -phase to the ε -, δ -, and γ -phase (Fig. 6a,b) can be categorized as three regions (C1, C2, and C3, respectively) with the quantitative contribution of intercalation and capacitive reactions as shown in bar graph (Fig. 7c). In both samples, diffusion-controlled intercalation behavior is dominated in the C2 process, and more surface Li + storage in the C1 and C3 than the C2 is observed, which suggests that the faradaic intercalation reaction is developed in the middle of the whole process with a small capacitive effect on total capacity. Moreover, the higher surface Li + storage of 3D V 2 O 5 /RGO/CNT composite is shown in the C1 and C3 processes compared to 2D V 2 O 5 /RGO. It is well known that, for TiO 2 which is regarded as a typical anode material following the intercalation reaction mechanism, the extent of both domains in the initial solid-solution formation characterized by potential drop and in the further Li + adsorption characterized by final sloping curve is directly proportional to the surface area of material 56 . These two characters correspond to the homogeneous Li + introduction followed by the biphasic transition and the surface Li + storage (i.e. interfacial capacity), respectively 56 . Therefore, the initial potential drop (C1) and the final sloping curve (C3) in the V 2 O 5 -based composites can be correlated proportionally with the surface area, consistent with the higher capacitive effect in the 3D V 2 O 5 /RGO/CNT composite having the higher surface area. Especially for the C3 process, the favorable surface Li + storage in 3D V 2 O 5 /RGO/ CNT composite is over diffusion-controlled intercalation, which suggests that the hierarchically porous 3D V 2 O 5 / RGO/CNT architecture can be a rational design to enhance the additional surface Li + storage along with the bulk reaction.
In conclusion, we have successfully prepared a 3D V 2 O 5 /RGO/CNT ternary composite with hierarchical porous structure by using 2D RGO and 1D CNT as conductive network through microwave-assisted hydrothermal method. V 2 O 5 microsphere assembled by nanoparticles is anchored on the surface of RGO and simultaneously penetrated by CNT. This hierarchical porous structure is designed to develop the surface area and mesoporosity of V 2 O 5 microsphere, which enhances electrolyte contact and Li + diffusivity. Through in situ XRD and in situ XANES analyses of the 3D V 2 O 5 /RGO/CNT composite, the phase transformations with structural reversibility upon Li + intercalation/de-intercalation and the electronic and local structure reversibility around vanadium atom during the first cycle are investigated, respectively. Compared to 2D V 2 O 5 /RGO as a control group, the 3D V 2 O 5 /RGO/CNT ternary composite delivers 80 mAh g −1 higher capacity at 1 C after 80 cycles and 32 mAh g −1 higher capacity at a considerably high rate of 20 C, showing an excellent cycle performance (220 mAh g −1 at 1 C) and a superior rate capability (100 mAh g −1 at 20 C), respectively. Moreover, favorable Li + diffusivity and enhanced surface Li + storage over diffusion-controlled intercalation reaction are observed in 3D V 2 O 5 /RGO/ CNT with the advantage of porous structure, which enables the structure of 3D V 2 O 5 /RGO/CNT ternary composite to become a promising cathode design for high performance LIBs.

Methods
Synthesis of graphene oxide nanosheet and functionalized carbon nanotube. GO nanosheets were synthesized from graphite flakes using modified Hummer's method 57 . Commercially available natural graphite powder and NaNO 3 were added into 50 ml of concentrated H 2 SO 4 (sulfuric acid, 98%), and then stirred in ice bath for 30 min. After that, 7 g of KMnO 4 (potassium permanganate) was added by continuous stirring and 100 ml of deionized (DI) water was slowly poured into the reaction mixture. After dilution with DI water (300 ml), H 2 O 2 (hydrogen peroxide, 30%) was added and the mixture was centrifuged at 4000 rpm. The precipitate was washed using 5% HCl (hydrochloric acid) to remove residual ions, washed further with DI water to reach the neutral pH, and then dried at 70 °C overnight. To functionalize CNT (oxidized CNT, o-CNT), 0.5 g of commercial multiwalled carbon nanotubes were treated with 50 ml of concentrated HNO 3 (nitric acid, 65 wt %) under refluxing at 70 °C for 48 h 58 . The mixture was cooled to room temperature, washed and filtered several times using DI water, and then dried in oven at 100 °C to get the functionalized CNT. All chemicals were purchased by Aldrich.
Synthesis of 3D V 2 O 5 /RGO/CNT ternary composite. The 3D V 2 O 5 /RGO/CNT composite was synthesized by the MAH method, followed by oxidation of V x O y /RGO/CNT precursor. A homogeneous dispersion of GO (80 mg) and o-CNT (40 mg) in DI water (120 mL) was formed by ultrasonication, and 1 g of V 2 O 5 powder precursor (purchased from Junsei) was dispersed in the above mixture by magnetic stirring. In addition, the ratio of RGO to CNT was controlled for 2:1 and 1:2 at 80 mg of GO with 40 mg of o-CNT and 40 mg of GO with 80 mg of o-CNT, respectively. An equal molar ratio of ascorbic acid (C 6 H 8 O 6 ) was added to the mixture as a reducing agent, and adjusted to pH 1 using 1 M HCl. The above mixture was transferred into Teflon-vessels and treated in MAH reactor (MARS, CEM Corp.) at 200 °C under 300 torr for 30 min. In the MAH process, reduction of GO to graphene as well as formation of 3D V x O y /RGO/CNT ternary microsphere were achieved simultaneously, and then, the mixture was filtered, washed, and dried at 60 °C overnight in electric oven. The final 3D V 2 O 5 /RGO/ CNT ternary composite was obtained by oxidation of as-prepared V x O y /RGO/CNT composite at 300 °C for 2 h in air.
To Materials characterization. X-ray diffraction (XRD) patterns were recorded in Bruker D2 PHASER diffractometer with Cu Kα 1 radiation (λ = 1.54056 Å) in the range of 10-80°. The morphologies of sample were characterized by using field-emission scanning electron microscopy (FE-SEM, JEOL JSM7000F) and high-resolution transmission electron microscopy (HRTEM, JEOL JEM2100F) coupled with energy dispersive spectrometer (EDS). The carbon content of sample was measured through thermogravimetric analysis (TGA) by a SEICO INST (TG/DTA 7300) at a heating rate of 10 °C min −1 under air flow. X-ray photoelectron spectrometer (XPS) measurement was carried out using VG microtech (ESCA, 2000) with a monochromatic Al Kα source (1486.6 eV), and the results were calibrated by referencing C1s at 284.6 eV. The surface area and pore size distribution were obtained using the N 2 sorption isotherm by using a Brunauer-Emmett-Teller surface area analyzer (BET, Micromeritics ASAP2000) and the Barrett-Joyner-Halenda (BJH) method.
Electrochemical measurements. Working electrode was prepared by mixing the active material of 3D V 2 O 5 /RGO/CNT, super P as a conducting agent, and polyvinylidene fluoride (PVDF) as a binder in N-methyl-2-pyrrolidone (NMP) solvent at a weight ratio of 70:15:15. The mixed slurry was uniformly coated on Al foil current collector, and the electrodes were dried at 120 °C overnight. A CR2032 type coin cell, consisting of the 3D V 2 O 5 /RGO/CNT composite as the working electrode and Li metal foil as a counter and a reference electrode, was assembled in a glove box under Ar atmosphere. A Celgard 2300 membrane was used as a separator, and for an organic electrolyte, 1 M LiPF 6 was dissolved in a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) with a volume ratio of 1:1.
The galvanostatic charge/discharge measurements (WBCS3000, Wonatech) were carried out on the coin cells in the voltage window of 4.0-2.0 V (vs. Li/Li + ) under various current densities. Galvanostatic intermittent titration technique (GITT) was performed during first discharge and charge processes at a constant current flux (0.1 C) for an interval of 10 min followed by open-circuit stand of 40 min to attain the cell to its steady state equilibrium voltage (E s ). Cyclic voltammetry (CV) was measured on Wonatech ZIVE MP2 electrochemical workstation at various scan rates over a voltage range of 4.0-2.0 V (vs. Li/Li + ). All electrochemical tests were conducted at room temperature. The in situ XRD patterns were collected at beamline 6D in Pohang Light Source (PLS-II) using a position sensitive detector (PSD) with the wavelength of 0.61992 Å at 10 s of exposure time. For easy comparison, two theta angles of all the XRD patterns have been recalculated and converted to the corresponding angles for λ = 1.54 Å (Cu Kα 1 radiation). The X-ray absorption spectroscopy (XAS) measurements were performed at beamline 8C in PLS-II. The oxidation states corresponding to V K-edge in each sample were investigated using a Si (111) double-crystal monochromator detuned to 70% of its original intensity to eliminate the high order harmonics. The in situ X-ray absorption near edge spectroscopy (XANES) spectra were collected in transmission mode at 2.5 GeV of electron energy with stored ring current of 400 mA top-up mode. Reference spectra of V metal were collected simultaneously using vanadium foil.