Rational design of TiO₂–V₂O₅–C nanostructure grafted by N-doped graphene with enhanced photocatalysis and lithium ion store performances

Abstract

Highly ordered mesoporous crystalline C–TiO₂–V₂O₅ core–shell microspheres encapsulated by porous carbon further embedded in N-doped graphene network (GN–TV–C) nanostructures were fabricated by a simple combination of hydrothermal-calculation method. Such material exhibits a highly efficient photocatalytic activity for water splitting, as well as a high specific capacity and exceptional cycle ability in LIBs. The extrusive features of such material include the assembly of components in a manner that enables an effective integration between the constituents and the ability to modify the electronic properties of GN–TV–C. The positive synergistic incorporation between TiO₂ and V₂O₅, the high electrical conductivity, and the three-dimensional hierarchically mesoporous nanostructure of these composites result into a highly active photocatalytic ability, showing patterns of increased light harvesting ability and promoted exciton dissociation, and excellent electrochemical performance in terms of a high rate capability and stable cycling. Profiting from the dual insurance of the flexible carbon layer derived from glucose and elastic GN walls with superior specific surface areas, a significant enhancement in the electron transfer and electronic diffusion channels, and a highly enhanced structural stability of the TiO₂-based electrode material were simultaneously obtained. In addition, the synergistic function among TiO₂, V₂O₅ and GN involving the optimized energy gap, the compromised particle assembly and surface defects, as well as their distinctive core–shell nanostructures were extensively studied. The carbon shell serves as a blocking layer that retards the interfacial recombination during photocatalysis, thereby protecting the active materials from pulverization during the superior cycles in the energy store. The current study may provide us an alternative approach for improving the performances of TiO₂ nanocrystals used in energy storage and photocatalysis applications.

---

Introduction

At present, titanium dioxide (TiO₂) has garnered considerable attention, endow unique nature and broad range of applications including environmental and energy areas from photocatalysis, solar cells, and electrochemical fields, as well as in lithium ion batteries and hydrogen generation.^1–6 Despite its nontoxic nature, abundance, and stability, it is equally true that the wide band gap of rutile TiO₂ (3.0 eV) can be excited only with UV light, which is considerably restricted when the sun light reaches the earth's surface. In addition, it also suffers from poor lithium storage properties and exhibits limited capacity at high current rates owing to poor rate performances, thereby significantly hindering its large scale applications. In order to overcome the disadvantages of TiO₂, doping composites need to be urgently used to narrow its band gap and to enhance its capacity and cycle ability. Among the versatile candidates, V₂O₅ possesses several attractive properties such as low cost, high energy density, easy synthesis, abundance, and high theoretical capacity (∼400 mA h g⁻¹), which enables it to be used as a pseudocapacitive anode for LIBs.⁷ It shows an enhanced charge storage capacity because of its ability to intercalate ions in a wide range of sites, and owing to the relatively low band gap energy (about 2.3 eV) of V₂O₅, which allows it to absorb radiations from the entire solar spectrum.⁸ The heterojunction of TiO₂–V₂O₅ (vanadium oxide loaded on a titanium oxide surface) can serves as an effective, practical catalyst for photocatalytic applications in water splitting. It can also form high quality anodic material.

However, numerous unfavorable factors, such as the short lifetime of the charge carriers⁹ and poor ionic and electronic conductivity, restrict its capacity and rate capability.^10–12 To facilitate the separation and transfer of photogenerated charge carriers so as to improve the kinetics of water-splitting processes and enhance the conductivity of the electrode material in order to build blocks for preferable channels for electron and ion separation and utilization.^13–16 Quite recently, several studies have shown that TiO₂-based materials exposed to highly-reactive facets exhibit superior photocatalytic activities, as well as enhanced lithium storage capabilities.^17–21 N-doped graphene (GN) introduced into the GN-semiconductor photocatalysts has been certified to extend the visible light harvest ability. It also prolongs the lifetime of the exciton and facilitates the transfer of the photogenerated charge carriers; thus, improving the adsorption capacity of the reactant, for which GN is supposed to behave as an electron reservoir in order to capture/shuttle the electrons photogenerated from the semiconductor.^13–17 On the other hand, the GN are also observed to show a stable cycle ability and excellent rate capability as anodic materials in lithium ion batteries. Therefore, the heteroatom doping further generates a narrow band gap of photocatalysts and extrinsic defects in the interwall space for Li storage.

However, the practical application of TiO₂–V₂O₅ is restricted by the instability of the interaction between the TiO₂–V₂O₅ nanoparticles and the GN surface both for the photocatalyst and anodic material. Zhang recently reported that the rendering of glucose can act as the facet-controlling agent and as a connection linker owing to its large number of exposed hydroxyl groups.¹⁸ The addition of glucose during the preparation of TiO₂-based materials increases the surface active sites, which serves as an electron trap by capturing electrons and promotes charge separation for the photocatalytic reaction. This subsequently promotes the surface redox reactions and inhibits the photocorrosion problem of V₂O₅ in the photocatalysis case^19,20 and prohibits the direct contact of V₂O₅ within electrolytes. This in turn prevents its dissolution in electrolytes that can cause unfavorable side reactions; thus, leading to rapid capacity fading during cycling in the energy store problem, which is a bottleneck problem for V₂O₅ as an anodic material.

Herein, we report a simple template-free self-assembly synthesis of uniform continuous carbon layer-encapsulated TiO₂–V₂O₅ mesoporous nanostructures that closely anchor on the GN sheets (GN–TV–C). The close incorporation of TiO₂–V₂O₅ within GN is via the bulk intercalative powder between the carbon layer from the glucose and from the GN sheets. This design possesses nanostructure engineering and combination project with which have been demonstrated two effective approaches for improving the TiO₂ based materials both for photocatalysts and LIBs. The visible spectrum of the as-prepared material shows extended light absorption from 400 to 600 nm. It shows remarkable photocatalytic activity for the water splitting in aqueous solution under visible light irradiation. Moreover, the lithium storage performance of such material was significantly enhanced in terms of excellent rate capability and superior cycle endurance. A reversible lithium storage capacity of 288.52 mA h g⁻¹ for up to 1000 cycles without any notable decrease was obtained. The excellent photocatalytic and electrochemical behaviors of such complicated material are attributed to the core–shell nature and well-connected nanoscale heterostructure (the strong positive synergistic effect between GN and high active TiO₂–V₂O₅ nanoparticles) with the improvement of both photocatalysis and electrochemical performance. The configuration of the carbon layer retards the electron–hole recombination and prevents the structural deformation upon cycling. It also shortens the electron–ion transfer path, while the embedded GN serves as an electron trap and preserves the electrical integrity between the electrochemical active species and the current collector. This ensures a good contact among the carriers with water molecules, electrodes and electrolyte, respectively, during the progress of the above reaction.

Experimental

Material synthesis

Graphene oxide (GO) nanosheets were first synthesized from natural graphite powder by a modified Hummers method.¹⁰ The pure TiO₂ was prepared by a sol–gel method; 3 mL of Ti(OC₄H₉)₄ (tetrabutyl titanate, TBT) was added to a 20 mL aqueous solution under vigorous mechanical stirring at 70 °C and placed in a water bath for 6 h. The milk white solution was collected by centrifugation and repeatedly washed with deionized water and acetone, and then dried at 80 °C in static air. The observed nanocomposite was further calcinated in a muffle furnace at 800 °C for 5 h. V₂O₅ was coated on the TiO₂ surface by an impregnation method.²¹ The prepared TiO₂ paste (1.8 g) was added in an ammonium metavanadate (NH₄VO₃, 0.0690 g) solution. Subsequently, the mixture was heated at 60 °C in water bath and refluxed for 6 h. In the following procedure, the composite was dried at 120 °C for 8 h, and then annealed at 350 °C with a detailed heating rate of 2 °C min⁻¹ and placed in air for 4 h. Subsequently, 0.5 mg of the as-synthesized white powers,150 mg glucose 133 mg urea (CO(NH₂)₂) were added into 67 mL GO aqueous solution. Then, the mixture was sealed in a Teflon-lined stainless steel autoclave and maintained at 180 °C for 12 h; to prepare pure TiO₂–V₂O₅ (TV) and N doped TiO₂–V₂O₅ (GN–TV) as comparison, the synthesis of GN–TV was carried out in the absence of glucose and the preparation of TV was conducted without the addition of glucose, urea and GO under the same conditions as those used in the synthesis of the composite. Scheme 1 presents the preparation process.

Scheme 1 Schematic illustration of the synthesis procedure for the GN–TV–C nanostructure.

Characterization and optics

X-ray diffraction (XRD) was performed on a Philips X'Pert Pro Super diffractometer with Cu Kα radiation (λ = 1.54178 Å). The optical properties of the samples were analyzed using UV-visible diffuse reflectance spectroscopy (DRS; V-560, JASCO). The field emission scanning electron microscopy (FE-SEM) images were taken on a JEOL JSM-6700F SEM. The transmission electron microscopy (TEM) was carried out on a JEM-2100F field emission electron microscope at an acceleration voltage of 200 kV. The high-resolution TEM (HRTEM), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and corresponding energy dispersive spectroscopy (EDS) mapping analyses were performed on a JEOL JEM-ARF200F TEM/STEM with a spherical aberration corrector. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo scientific ESCA Lab250 spectrometer comprising a monochromatic Al Kα as the X-ray source, a hemispherical analyzer and a sample stage with multi-axial adjustment to obtain the composition on the surface of samples. The Raman spectra were measured using a Renishaw inVia Raman System 1000 with a 532 nm Nd: YAG excitation source at room temperature. The Brunauer–Emmett–Teller (BET) surface area was measured using a BELSORP-mini (BET Japan) at liquid nitrogen temperature (77 K). The photoluminescence (PL) spectra for the solid samples was investigated on an Edinburgh FL/FS900 spectrophotometer with an excitation wavelength of 360 nm. The photocurrent measurements were performed in homemade three electrode quartz cells with a PAR VMP3 Multi Potentiostat apparatus. The electrolyte was 0.2 M aqueous Na₂SO₄ solution (pH 6.8) without other additives. The electrochemical impedance spectroscopy (EIS) measurements were performed in the presence of 10 mM K₃[Fe(CN)₆]/K₄[Fe(CN)₆] and 0.5 M KCl by applying an AC voltage with 5 mV amplitude in a frequency range from 1 Hz to 100 kHz, under open circuit potential conditions.

Photoelectrochemical measurements were carried out in a three-electrode system with a potentiostat (HSV-110, Hokuto Denko) and an electrochemical cell at room temperature. All measurements were carried out under an inert helium atmosphere in 1 M K₂SO₄. Typically, 4 mg of the catalyst and 30 μL of the Nafion solution (Sigma-Aldrich, 5 wt%) were dispersed in 1 mL of the water–isopropanol solution with a volume ratio of 3 : 1 by sonicating for 1 h to form a homogeneous ink. Then, 5 μL of the dispersion (containing 20 μg of catalyst) was loaded onto a FTO electrode with 3 mm diameter (loading 0.285 mg cm⁻²). Hg/Hg₂SO₄ in 1 M K₂SO₄ saturated solution was employed as the reference electrode, a Pt rod (AlfaAesar, 99.9995%) was designed as the counter electrode, and the GN–TV–C/FTO functioned as the working electrode. All the voltages were calculated relative to the reversible hydrogen electrode according to the Nernst equation:

V vs. RHE(volt) = V vs. Hg/Hg₂SO₄(volt) + 0.236(volt) (1)

The Nyquist plots were measured with frequencies ranging from 1000 kHz to 1 Hz at an overpotential of 200 mV. The impedance data was fitted to a simplified Randles circuit to extract the series and charge-transfer resistances. The irradiation wavelength was controlled by a 300 W Xe arc lamp (PLS-SXE300, Beijing Perfect light Co., Ltd.) equipped with a UV-cut filter to cut off light at certain wavelengths (λ > 420 nm). Light intensity, as measured by a visible-irradiance meter, was about 4.0 mW cm⁻² at the position of the working electrode. The dark currents were orders of magnitude lower than photocurrents in all cases within the measured voltage ranges.

The water splitting experiment was conducted in H-type reactor with 50 mL of 0.1 M K₂SO₄ aqueous solution in each solution. A small overpotential as high as 200 mV was needed to support the stable H₂ and O₂ evolution rates. A bubbler purged the electrolyte in the reactor to push the evolved gases into a gas chromatograph for analysis. After that, the gas was fed back into the reactor via a recirculation pump. The gas chromatograph (SRI Instrument, Inc.) was equipped with a molecular sieve 13× packed column and a helium ionization detector. Before each reaction, the system was calibrated with H₂ and O₂ gases of accurate concentration.

Electrochemical tests

The investigation of electrochemical performance was performed using coin-type LIB cells (2025) assembled in an argon-filled glove box. For anode preparation, a mixture of active material, carbon black, and polyvinylidene fluoride (PVDF) binder with a weight ratio of 80 : 10 : 10 was dispersed in N-methylpyrrolidone (NMP) solution, and then the slurry was uniformly pasted on a Cu foil current collector. A typical electrode was dried at 120 °C for 24 h under vacuum before being assembled into coin cells in an argon-filled glove box. A Celgard 2400 microporous polypropylene membrane was used as the separator and Li foil was used as the counter electrode. The nonaqueous electrolyte used was 1 m LiPF₆ dissolved in a mixture (1 : 1 : 1, in wt%) of ethylene carbonate (EC)/dimethyl carbonate (DMC)/diethyl carbonate (DEC). Galvanostatic cycling experiments of the cells were performed on a LAND CT2001A battery test system in the voltage range of 1.00–3.00 V versus Li⁺/Li at room temperature. The coin cells were activated at a current density of 0.5 C (167 mA g⁻¹) for the first cycle, and then cycled under different current densities within the voltage range of 1.00–3.00 V using a LAND-CT2001A battery test system (Jinnuo Wuhan Corp., China). Subsequently, the cells were cycled under different current densities within the voltage range 1.00–3.00 V. Electrochemical impedance spectroscopy (EIS) measurements were conducted on an electrochemical workstation (CHI 660 D, CHI Company) under a frequency range from 0.1 Hz to 100 kHz.

Results and discussions

Structure characterization

The crystal phase properties of the as-prepared TiO₂ composite were also confirmed by powder X-ray diffraction (XRD) (Fig. 1a). All the samples presented a single-phase diffraction pattern identical to that of the rutile TiO₂ (PDF#99-0090). The XRD peak position inclines to shift to lower 2q angles as the incorporation of V₂O₅, although this shift was very small as evidenced by the amplified spectroscopy result (Fig. S1†). The crystallite size of TV could be obtained according to the Scherrer's formula D = 0.9λ/β cos θ, where λ is the average wavelength of the X-ray radiation, β is the line-width at half-maximum peak position, and θ is the XRD peak. The average crystallite sizes of the bulk TV and GN–TV–C samples calculated using the Scherrer equation were 24 nm and 8.3 nm, respectively, indicating that the carbon shell and GN could induce smaller TV crystals.

Fig. 1 XRD patterns of the TV samples (a) Raman spectra of the TV composites (b) TGA curves of the GN–TV–C,GN–TV and GN composites under Oxygen flow (c) and XPS survey spectrum of GN–TV–C at 700 °C in NH₃ atmosphere for 1 h (d).

It appeared that the V₂O₅ component in these samples was distributed mainly near the surface, but the degree of segregation was too small to be detected by XRD, owing to the relatively low concentration of the component (V/Ti = 0.01). Graphene and nitrogen were also not distinctively detected in the pattern of all the Ti samples because of their less content or amorphous feature. However, the Raman shift highlighted the presence of graphene in the GN–TV–C composite (Fig. 1b), whereas the XPS spectroscopy confirmed the existence of nitrogen. The two distinct peaks correspond to the D band at ca. 1350 cm⁻¹ and the G band at ca. 1595 cm⁻¹, respectively.²² The D band is commonly attributed to a series of structural defects, whereas the G band is observed for all graphitic structures. It is noted that the intensity ratio of the D and G bands, I_D/I_G, represents the relative concentration of the sp³ hybridized defects relative to the sp² hybridized graphene domains. The average I_D/I_G values for GN–TV–C and GN–TV are 1.25 and 1.12, respectively, demonstrating that the presence of GN introduces some defects and the load of the carbon layer from glucose further increase the disorder of GN–TV–C to some extent.

The carbon and nitrogen contents in the GN–TV–C composites were determined by TG analysis (Fig. 1c). In the case of GN–TV, 5.25% of weight is achieved at a high temperature (800 °C) because of the complete combustion of GN under air atmosphere. The TGA of GN is nearly overlapped with that of GN–TV, indicating the complete reduction of GO to graphene. The oxidative decomposition of the carbon shell is believed to be attributed to the different weight loss of GN–TV–C and GN–TV. The content of the carbon shell in the GN–TV–C composites can be calculated to be 4.98%. It proves that the TV composites were effectively restricted in the pores of GN–TV–C, which is consistent with the results of BET measurements as shown in Fig. S2.†

The N₂ adsorption/desorption isotherms were performed to study the porous feature of GN–TV–C microspheres. As shown in Fig. S2a,† a distinct hysteresis loop between 0.8 and 1.0 was observed, which is a characteristic of mesoporous materials. The pore-size distribution curve of GN–TV–C microspheres monitored by Barrett–Joyner–Halenda (t-plot) method have micropores relatively with a size distribution of 0.65–1.9 nm and a sharp peak at 30 nm (Fig. S2b†), demonstrating the mesoporous structure GN–TV–C nanoparticles. The Brunauer–Emmett–Teller (BET) surface area of GN–TV–C nanostructures is calculated to be 142 m² g⁻¹, a pore volume of 0.587 cm³ g⁻¹, which is a significant improvement as compared with that of bulk TV nanoparticle (35 m² g⁻¹ and 0.139 cm³ g⁻¹) and the GN–TV nanocrystals (102 m² g⁻¹ and 0.413 cm³ g⁻¹), indicating the efficiency of the carbon shell and GN sheets in increasing the surface area of TiO₂-based nanohybrids. The substantial increase of surface area and pore volume is mainly because of the contribution of GN nanosheets with a large outer surface and the porous carbon shells derived from the carbonization of glucose with inner specific area. It is believed that the larger specific surface area of mesoporous nanostructure promotes an increase in the photocatalytic and electrochemical active sites, which are beneficial for improving the photocatalytic and electrochemical performances of GN–TV–C.

Furthermore, the X-ray photoelectron spectroscopy (XPS) spectra with the total survey spectrum (Fig. 1d) also verified the existence of GN–TV–C. The C1s XPS (Fig. 2a) indicates the abundance of various oxygen-containing functional groups on the graphene surface. The strong binding energy peak at 284.7 eV is assigned to the C–C bonding, 287.8 eV is assigned to the C=O bonding, and the weak binding energy peak at 289.3 eV is assigned to the O–C=O bonding, indicting the strong chemical reactivity between TiO₂–V₂O₅ and graphene, in which the TiO₂–V₂O₅ particles and GN in excellent interface binding were uniformly interconnected, which is further confirmed by SEM and TEM images of these materials (Fig. 3–5).

Fig. 2 XPS spectrum of GN–TV–C at 700 °C in NH₃ atmosphere for 1 h: C1s narrow scan (a) N1s narrow scan (b) Ti_2p narrow scan (c) O1s narrow scan (d).

Fig. 3 Images of pure TV samples at different magnifications (a) and (b) GN–TV–C composite at different magnifications (c) and (d).

Fig. 4 EDS map of the selected rectangular area of FESEM image of GN–TV–C nanocrystals and results the of EDS analysis.

Fig. 5 TEM image (inset is the scheme of TV particles) (a) HRTEM image (b) selected area electron diffraction (SAED) pattern for pristine TV (c) TEM image at different magnifications (inset is the scheme of GN-TV-C) (d) and (e) SAED for GN–TV–C (f).

As shown in Fig. 2b, the low intensity of N1s signal indicate that N is more incorporated with graphene than with TiO₂–V₂O₅. Two characteristic peaks of the Ti_2p spectrum at 458.2 and 463.9 eV are attributed to the Ti_2p3/2 and Ti_2p1/2 (Fig. 2c), respectively. According to the relationship between the oxidation state of Ti and the separation of peak energy (Ti_2p), such energy gap of 5.7 eV infers that the average oxidation state of elemental Ti in the GN–TV–C composite is around −3.9, showing the dominance of Ti⁴⁺ in the rutile TiO₂. The two peaks located at ∼532.5 and ∼533.3 eV attributed to the O1s (Fig. 2d). In addition, for GN–TV–C core–shell–shell nanostructure, the TiO₂ core has been fully coated with the V₂O₅ shell format and continuous carbon layer deriving from glucose, the binding energy of V mainly located at 517 eV and 525 eV are covered by the large amounts of O signal; thus, reducing the signal intensity of the inside element and the V element is obviously not distributed in the XPS spectrum. However, V is certified in the EDS file (Fig. 4). The heterojunction between the core–shell–shell TiO₂–V₂O₅–C nanostructures embedded in continuous GN network can facilitate the quick transfer of the photogenerated electrons from TiO₂–V₂O₅–C to GN. The shorten transfer time for long lived-exciton impedes the electron–hole recombination and their ability to participate in electron transfer is integral for these excitons to serve as light harvesting antennae. In addition, the flexible carbon layer and elastic GN sheets significantly improve the conductivity of the electrode material, accommodate the large volume change of GN–TV–C anode upon lithiation–delithiation, leading to an excellent cycling stability and high coulombic efficiency even at high current densities, protecting from V₂O₅ from dissolution in aqueous and electrolyte, prevent its photocorrosion under visible light irradiation and decay in long life cycle in LIBs, and thus, improve the structural stability. This investigation might facilitate the high-efficient and stable hybrid photocatalysts applications in water splitting using solar energy and energy store in the development of sustainable resources. In particular, the adjustable space in the core–shell accommodates well a dispersion of active nanocrystals and the retention of intact protective graphene layer finally contributes to the significantly improved cycle ability. The continuous carbon network can not only provide higher accessible surface area for the GN–TV–C nanostructures in the composite, but also allow the enhanced light absorption capability and the electrolyte ions to rapidly diffuse inside. Therefore, it is highly desirable to fabricate pore-rich materials for their potential applications in both catalysis and energy storage.

The particle morphology and size of GN–TV–C samples are confirmed by SEM and TEM, as shown in Fig. 3. As shown in Fig. 3a and b, the TV SEM images show microsized clusters with obvious agglomerated particle-like nanoparticles. The grain sizes are around 30 nm for pure TV after annealing at 700 °C. In the case of GN–TV–C (Fig. 3C), particle-like TiO₂–V₂O₅–C nanoparticles with a diameter of 20 nm anchored on the GN surface with significantly decreased aggregation tendency. The uniform carbon encapsulation and the flexible GN sheets helped in preventing the agglomeration of the primary crystallites of TV during calcination.

To unravel the elemental composition and the amount of distribution of elements present in TV–C core–shell heterostructures, energy dispersive X-ray spectroscopy (EDS) elemental mapping was conducted. The elemental maps of this nanocomposite also suggest a uniform distribution of nitrogen, carbon, titanium, vanadium, oxygen, which is consistent with the XRD and XPS results.

The mapping results (Fig. 4) of the selected rectangular area of FESEM image of GN–TV–C clearly depicted the coexistence of T, V, O, N and C elements in the heterostructures. It is noted that the distribution of all the elements is homogeneous and uniform, but the distribution of Ti and V exhibit a considerably larger area than the other elements, indicating that TiO₂ cores are densely and uniformly encased by V₂O₅ layer, thereby resulting into core–shell–shell heterostructures.

Fig. 5 shows the typical TEM images of the TiO₂ samples, as shown in Fig. 5a, and pure TiO₂ nanoparticles consist of densely packed irregular nanoparticles, which is consistent with the SEM results with a high degree of crystallinity. As shown in Fig. 5b, The observed HRTEM image clearly reveals the lattice spacing of 0.32 nm, which corresponds to the (110) crystal planes of the rutile TiO₂ and the {101} orientation of the V₂O₅ nanoparticles, respectively. The TV nanocrystals exhibit a polycrystal of the TiO₂ (Fig. 5c).

As for GN–TV–C, the aggregation phenomenon occurred in the case of the TV particles, which is in agreement with the SEM result. Moreover, an excessive amount of carbon led to a thick carbon-coating layer (10–20 nm). The TEM images in Fig. 5d show the primary crystallites of the TV material with sizes ranged from 5–10 nm. TV had smaller primary crystallites (5–10 nm) than pure TV, which agrees with the XRD result. It is clear that many obvious bright regions appear in the sphere, revealing an underlying hierarchical porous structure in Fig. 5d, which complies well with the result of BET. The image also shows the GN wrapping around the core–shell–shell TiO₂–V₂O₅–C nanostructure.

The particles become indiscernible and completely coated by a thick layer of carbon irrespective of the high crystalline of the TV core–shell particle (read by the XRD result) (Fig. 5d). There are many voids inner the carbon layer to alleviate the volume change during the lithium–delithium progress (inset Fig. 5d). This intermixing allows the TV and carbon matrix to provide especially effective mechanical reinforcement, accommodating the volume change of the active particles, while also trying to separate them and inhibit their agglomeration during extended cycling. Therefore, the reinforcing GN matrix and the elastic carbon layer guarantee the stability structure of GN–TV–C as intended. The brightness difference between the core and the shell in the magnified TEM image portrayed in the inset of Fig. 5d substantiates the existence of a core–shell configuration of the heterostructures.

By analyzing the interface of the core and shell in HRTEM taken from the selected rectangular region of GN–TV–C heterostructures (as marked in Fig. 5e), it is observed that the lattice fringes are continuous from core to the shell (Fig. 5e). Thus, TEM and HRTEM analysis, explicitly validate the core–shell morphology of TV-C heterostructures. As clearly seen from the TEM images (Fig. 5f), the polymeric morphologies were affected by thick carbon formation and the SAED in Fig. 5f showing a slightly clear six-fold pattern confirms that the crystal structure of the original graphite is retained in the exfoliated sheets and would improve by the hydrothermal reaction. This suggests that they are dispersed on the GN sheets with good contact, the intimate contact makes the interfacial charge transfer available at the interface of the composites. Thus, on the basis of the FESEM, FESEM, EDS elemental mapping, TEM, HRTEM and point-EDS analysis, it is safe to conclude the core–shell–shell morphology of TiO₂–V₂O₅–C.

The GN–TV–C crystals were concluded to be formed during the dissolution and recrystallization of TiO₂–V₂O₅ progress, which involved the Ostwald ripening process.²³ As confirmed by the TEM images, the rutile TiO₂ mesocrystals were formed through the oriented self-assembly process of rutile nanocrystal subunits elongated along the {110} direction under equilibrium conditions, while the V₂O₅ finally preserves its most highly active {101}. It is well-known that {110} and {101} are the most thermodynamically stable facets for rutile TiO₂ and V₂O₅. A large amounts of oxygenated groups, hydroxyl, carbonyl and carboxyl groups can be induced on the glucose surface during the hydrothermal reation.²⁴ Glucose can be expected to act as a role of surfactant,¹⁸ in order to verify the speculation, the surface tension of the aqueous solution was investigated. Consequently, the tension reduced from 68.96 mN m⁻¹ without glucose to 98.73 mN m⁻¹ after the addition of glucose. Therefore, the glucose indeed plays a predominant role as a surfactant during the configuration of the core–shell–shell of TiO₂–V₂O₅–C, as a tension to impose on the pristine TV particles, and remains in a meta-stable phase in the aqueous form because of a higher surface energy of the inner core. The TV nanoparticles were dissolved from inside to outside intensely enclosed by the glucose macular. The rapid dissolution and slow expansion glucose–TV led to the formation of the porous carbon shell. Therefore, the TV particles were gradually encapsulated by the carbon microspheres because the surface tension of the sphere is the lowest, and the ripening process for the core–shell–shell TiO₂–V₂O₅–C micronanostructures was completed. The glucose adsorbed encapsulating the active TV by interweaving the sides of GN at the same time could be competing with the face-growth inhibitors, slowing the growth of the involved face of TV and the production of mesoporous channels in the TV nanocrystals, which introduces favorable charge carrier and ion diffusion. The surface energy, which is responsible for the crystallization was involved during the whole progress of the crystals growth. The TV nanostructures appear in a uniformly distributed TiO₂–V₂O₅–C morphology in the aqueous medium, whereas the nanocrystals incline to a homogeneously arranged porous core–shell–shell feature. Moreover, the nanocrystal subunits were partly recrystallized during the oriented self-assembly process. The well crystallized TiO₂–V₂O₅–C densely grown on GN matrix with mesoporous nanostructure can facilitate the quick transfer of the photogenerated electrons from TiO₂–V₂O₅–C to GN to cause the effective charge separation in the photocatalysis case, as well as provide the accommodation of volume change during the charge and discharge processes in LIBs application; thus, indicating its possible application as a visible-light driven photocatalyst and as an attractive anode material candidate for LIBs. Scheme 2 illustrates the entire progress of the reaction.

Scheme 2 Schematic illustration of the formation GN–TV–C nanostructures.

Optical characterization

The UV-vis absorption spectra of TV, GN–TV, and GN–TV–C are shown in Fig. 6a. Pure TV particles have a characteristic absorption sharp edge around 438 nm with a band-gap of 2.85 eV (Fig. 6b). As for GN–TV sample, it exhibits a wider light absorption range of beyond of 489 nm because of the incorporation of GN, which significantly enhance the light absorption capability because of the great surface area and its band gap calculated to be 2.75 eV. The entire absorption peak in the visible region of GN–TV–C nanocomposite shows a larger red shift (with a small band gap of 2.68 eV) compared with that of TV and GN–TV due to rational electron level, core–shell–shell nanostructure, favorable surface modification of GN, its graft with V₂O₅ with narrower band gap enables its extended light harvest ability, while the core–shell nanostructure is verified to show great promise for the fast separation of charge carriers and low charge recombination rate,²⁵ the close contact with GN is also beneficial to suppress the exciton recombination. On the one hand, the highly conductive GN with great specific surface area facilitates the rapid electron separation with hole and serves as a reservoir as electrons, which is convenient for the hydrogen evolution, on the other hand, the presence of proper residual oxygen in carbon matrix (the XRD results) enhances light absorption and inhibits the recombination of charge carriers.²⁶ The synergistic of unique morphology and composite technique endow the GN–TV–C the dramatically enhanced photocatalysis capability, which can be expected to significantly decrease the charge-carrier recombination.

Fig. 6 The UV-vis diffuse reflectance spectra of the prepared TV samples (inset are the color variations of TV materials) (a) Plots of the band gap energy in direct transition recorded from the TV composite (b) current density (mA cm⁻²) of externally short-circuited GN–TV–C photoelectrodes under the same conditions using simulated sunlight exposure (c) and PL spectra of the TiO₂ nanoparticles (d).

The band gap energy of semiconductors can be determined on the basis of the following formula:

α(hν) = C(hν − E_g)^n/2 (2)

where α, h, ν, E_g, and C are the absorption coefficient, Planck constant, light frequency, band gap energy, and a constant, respectively; n is 1 for direct transition and 4 for indirect transition. The nature of the transition is possible to be determined by plotting a graph of (αhν)² versus hν, thereby the band gap energies can be deduced by extrapolating the straight-linear portions of the plot to intersect with the photon energy axis. Compared with TV and GN–TV, the absorption edge for GN–TV–C present an obvious blue shift, but the extend absorbance to the visible light region is mainly attributed to the decoration of GN, this dense incorporation of GN will not only reduce the recombination efficiency of the photogenerated electron–holes and improve the photocatalytic performance, but also has a good structural stability and prolongs the service life of the composites, i.e. the large 2D at structure of the GN sheets can significantly facilitate the optical transparency and provide more efficient contact between light and H₂O. The obvious narrowed band gap suggests the electronic interaction between individual components in TiO₂ and V₂O₅, which together with the enhanced visible light absorption could contribute to the formation and transfer of photogenerated charge transfer carriers across the interface, which is consistent with the photocurrent result of the following result.

To further investigate the characteristics of GN–TV–C arrays, photoelectrochemical measurements were performed in a three electrode. Fig. 6c presents the time courses for the photocurrent of TV based materials under visible light irradiation at +0.2 V (vs. Hg/Hg₂SO₄). Obviously, the current abruptly increases and decreases as modulating the light source on and off. The incorporation of GN in the TV supply enhanced the photocurrent under visible-light irradiation, the core–shell–shell configuration of TiO₂–V₂O₅–C due to the addition of glucose was most effective for improving the visible photoresponse. It is to be noted that the co-doping GN and carbon shell substantially improves the photocurrent of the final GN–TV–C. Moreover, the photocurrent of GN–TV–C (3% V₂O₅) is higher than that of GN–TV–C (5% V₂O₅), thicker TiO₂–V₂O₅ contacting surface may become a new carrier and holes recombination center. This indicates that the amount and ratio of Ti and V are associated with the enhancement in photoactivity under visible light.

The result agreed well with the following PL measurement (Fig. 6d): PL is a predominant parameter to weigh the charge separation property of the photocatalysts. The higher PL intensity means a high rate of electron–hole recombination,²⁷ in other words, the weaker intense the PL emission peak is the more it suppress the photocatalytic activity. In comparison with pristine TV and GN–TV, obviously, GN–TV–C (3% V₂O₅) exhibits the weakest intense PL peak than others from Fig. 6d, demonstrating that a minimum recombination of electron–holes in the TiO₂ based sample, leading to the improvement of the photocatalytic activity, further the addition of V₂O₅ up to 5% leads to increased electron–hole recombination. It can be shown that hybrid TiO₂ with GN can cause a clear diminution of the PL emission because the high conductive GN can trap photogenerated electrons through the interface to facilitate the charge separation to achieve a higher separation efficiency. Moreover, the accumulated holes on the surface of carbon layer participate in the oxygen evolution reaction, which guarantee the high oxygen production efficiency. This result was because of the fact that the electrons were excited from the valence band of TV to the conduction band, and then transferred to high conductive GN, thereby preventing the direct recombination of electrons and holes. The mitigated exciton recombination leads to an increased photocurrent.

Photocatalysis performance

Solar-driven water splitting with 250 mV is achieved under visible light (>420 nm) using the as-synthesized TV samples. For electrode preparation, the prepared TV was deposited on a fluorine-doped tin oxide (FTO) substrate, introducing a Pt electrode as the counterpart electrode. Fig. 7a displays the evolution of H₂ and O₂ gases from the TV samples dispersed in electrolyte. The 2 : 1 stoichiometry between H₂ and O₂ with the linear increase in gas concentrations confirms the water-splitting photoactivity. Among the three TV samples, the GN–TV–C (3% V₂O₅) exhibit the highest efficiency (3.987 and 1.993 μmol h⁻¹ for H₂ and O₂ from 2.0 mg of nanotree heterostructures per hour),while those of the other two samples just reached 2.96 and 1.43, 1.18 and 0.58 μmol h⁻¹, respectively, for GN–TV and TV, which agree with all the aforementioned results. GN sheet based on a rational percentage could serve as a rapid electron transfer path way for high carrier separation and the electron collection at the conductive substrate. The core–shell–shell TiO₂–V₂O₅–C nanocrystals could prevent the direct contact between the exposed GN surface (rich in negatively charged electrons) and the electrolyte (rich in positively charged holes), which could substantially reduce the recombination rate of the photogenerated carriers. Such photocatalyst can endure continuous 12 h working time test without an obviously decreased working efficiency (Fig. 7b), further certifying the stability of the strong nanostructure.

Fig. 7 The evolution of H₂ and O₂ gases of TV samples (a) stability of photocatalytic water for GN–TV–C nanocomposite under visible light irradiation (b).

The photocatalysis mechanism is shown in Fig. 8. The photo-excited electrons would transfer from TV to FTO, and then to Pt, and the separated electrons on the surface of Pt would combine with adsorbed H⁺ to produce H₂. Simultaneously, the accumulated holes on the surface of TV would firmly retained within the surface of carbon layer because of its positive band layer,²⁶ which would oxidize the H₂O to release O₂. Compared with bare TV, the GN–TV–C heterostructures had prolonged the life time of separated exciton pairs, resulting in the improvement of the H₂ and O₂ production.

Fig. 8 The photocatalysis mechanism for GN–TV–C.

Electrochemical performances

Cyclic voltammetry was conducted in order to investigate the effect of the electrochemical properties of the TV based samples by using a scanning rate of 0.5 mV s⁻¹, voltage range: 1.00–3.00 V. A single pair of sharp oxidation and reduction peaks, a plateau at 2.0 V and a sloped region of 1.7–1.0 V, corresponding to the typical lithiation and de-lithiation process of TiO₂, respectively, can be clearly discerned. Interestingly, V₂O₅ presents the electrochemically inactive sample among the three samples, but is observed to contribute to the capacities. Fig. 9a presents the comparison of the CV profiles of pristine TV, GN–TV and GN–TV–C in the first cycle. For pristine TV, the oxidation and reduction peaks are observed at 2.21 and 1.72 V, respectively, with the polarization being 0.49 V. As for GN–TV, the oxidation and reduction peaks appears at 2.16 and 1.78 V, respectively, and the polarization is 0.38 V. As for GN–TV–C, the oxidation and reduction peaks occurs at 1.96 and 1.72 V, respectively, with the polarization just decreases to 0.24 V. The broadened peaks of pristine TV present the most poor kinetics, lithium intercalation and deintercalation is sluggish. The lower polarization and higher peak of GN–TV–C sample compared to that of TV correlate with the unique core–shell–shell nanostructure coating. Because the active TV is confined to the amorphous porous carbon shell on the nanoscale, it maintains a close contact with the interwoven GN channels, the electron transfer and lithium ion diffusion paths are dramatically decreased; hence, resulting in significantly enhanced reversibility. In addition, the inertness of carbon with more exposed active cites leads to better Li kinetics, poor adsorption leads to Li clustering, significantly affecting the performance of the battery.²⁸

Fig. 9 Cyclic voltammetry curves of all TV samples at 0.5 mV s⁻¹ (a) charge–discharge curves CV curves of the composite GN–TV–Cat a scan rate of 0.5 mV s⁻¹ (b) charge–discharge curves of TV sample (c) GN–TV composite (d) GN–TV–C (e) and The rate performance curves of TV samples (f).

Cyclic voltammograms (CV) of GN–TV–C from the 1^st to 3^rd cycle at a scan rate of 0.5 mV s⁻¹ in the voltage range of 1.0–3.0 V is illustrated in Fig. 9b. It can be observed that the intensity of the redox peaks increased during the second and third scan, implying a possible activating process in the electrode material. The second and third cycles with higher cathodic peaks and lower anodic peaks compared to the initial cycle are nearly overlapped, supporting the excellent electrochemical reactivity of modified TiO₂ composites. The plots replication certifies the good reversibility of the lithium extraction–insertion reactions in terms of small polarization, which well complies with the aforementioned discussions. Fig. 9c–e describes the galvanostatic charge–discharge voltage profiles of cells for all TV samples at progressively increased C rates from 0.1 C to 30 C between 1.00 and 3.00 V vs. Li⁺/Li. It should be noted that the voltage plateau is lengthened for the GN and C modified TV materials, which should be ascribed to the higher electrochemical reactivity of modified TV and excellent kinetics. This result is consistent with the CV profiles. The modified TV materials exhibit higher capacities than TV at a low rate C/10, and a maximum capacity of 345.59 mA h g⁻¹ is observed in GN–TV–C, while that of the GN–TV and TV reach 300.50, 298.58 mA h g⁻¹, respectively. In addition, it is found that a serious polarization tendency at high C rates, while the modified samples have a slight polarization tendency. Fig. S3† compared the charge–discharge files of 3% V₂O₅ and 5% V₂O₅, it is observed that that the capacity of later decreased obviously, too much V₂O₅ can introduce more structure instability due to the irreversible phase transitions outweigh its contribution to the capacity of TV during charge–discharge progress.²⁹

Fig. 9f shows the cycling performances at various current density ranging from 100 mA g⁻¹ to 30 A g⁻¹. It can be clearly shown that, as compared to the GN–TV and pure TV samples, the specific capacity of the GN–TV–C sample is substantially higher at all the investigated charge–discharge rates. For example, the GN–TV–C electrode displays a considerably superior rate performance of 303.69 and 293.81 mA h g⁻¹,which is 1.3 times that of the GN–TV electrode (243.79 and 238.95 mA h g⁻¹) and is almost 2 times of that for pure TV (201.41, 179.26 mA h g⁻¹) at current densities of 20 and 30 A g ⁻¹, respectively. Recovering the current density back to 50 mA g⁻¹ and 100 mA g⁻¹, reversible capacities of up to their initial capacities, verifying the good reversibility and structural stability of the GN–TV–C anode composite and the capacity keeps ascending during subsequent cycling, approaching its initial capacity. The excellent rate capability arises from the 3D interconnected porous core–shell–shell carbon structure and the GN interweaved favorable channels, which not only offer multidimensional channels for the electrolyte access and sufficient void space to buffer the volume change of TV during cycling, but also effectively shorten the electronic–ionic transport path. In particular, the excellent mechanical stability of the carbon layer greatly confines the active TV particles in the core–shell nanostructure impeding the corrosion from electrolytes, and moreover, the good conductivity of the 3D interconnected GN matrix results in the long cycle ability of GN–TV–C. The reversible formation of flexible carbon buffering layer and elastic graphene sheets restrain the volume change of TV during the charge and discharge processes.

The heterostructure GN–TV–C exhibits an excellent rate capability compared to the GN–TV and TV samples; however, the difference in conductivity between them is be taken into account. Electrochemical impedance spectroscopy (EIS) measurements were performed to investigate the conductivity of the cathodes composed of TV based electrodes. As shown in Fig. 10a, the GN–TV–C electrode possesses a considerably lower resistance than that of the GN–C and TV electrodes (ca. 73.2 vs. 145.9 and 330.4 Ω, respectively), indicating that the encapsulation of core–shell–shell nanostructure and GN matrix can significantly enhance the electrochemical kinetics. The synergistic effect of the successful integration of the TV structure and interweaved GN matrix is beneficial to the structure stability during the photocatalysis and charge and discharge progresses. Ideal electrode materials are both good electronic and ionic conductors. The electrons and Li ions must reach or leave the reaction point simultaneously. However, for TV, it not only suffers from extremely low intrinsic electronic conductivity because of the lacking of effective conductive carbon layers on its surface. Thus, the larger particle size for pure TV may decrease the electronic conductivity in the bulks and result in the higher charge-transfer impedance R_ct.³⁰

Fig. 10 Electrochemical impedance spectroscopy (EIS) result of the TV samples (a) and long-term stability for GN–TV–C material 30 Ag⁻¹ (b).

In particular, the lower electron transfer resistance can effectively accelerate the transfer mechanism of photogenerated charges in the hetero-structured samples, positively suppressing the combination of the photogenerated electron–hole pairs and further enhances the reactive activities of these samples, which is beneficial for the enhancement of photocatalytic performance and extending the sunlight harvest ability. Therefore, the synergistic morphology and composite engineering together with the particular mesoporous nanostructure introduce a promising candidate for both photocatalysis, as well as energy storage.

The robust nanostructure of GN–TV–C was further investigated (Fig. 10b), the morphology of GN–TV–C were examined after the 2000 times cycling test. Fig. S4† show that the TEM image of the core–shell structure of TiO₂–V₂O₅–C is remained escaping from pulverization after 1000 times cycling, which is attractively pursued when designing anode materials. With the dual protections from GN and carbon layer, the severe volume changes of TV can be effectively accommodated. The carbon layer can also protect the TV electrode from electrolyte corrosion and maintain the structural stability of the TV; thus, resulting into a considerably improved rate and low temperature capability and cycle stability of the GN–TV–C, in good agreement with the aforementioned results.

Conclusions

In summary, coupling morphology technology and composite engineering to GN–TV–C lead to the enhancement of the photocatalytic activity and the energy storage. In this project, a rational combination of TiO₂ with V₂O₅ (3% V₂O₅) encapsulated by uniform carbon nanocages which is embedded in interconnected GN channels was fabricated for water splitting and lithium ion batteries. Owing to its optimized band gap and shorter distance between the core and its surface, the photogenerated electrons and holes could be easy transferred to the surface for reaction, effectively extending the light harvest ability, improving the utilize efficiency of exciton and photoelectron stability. Besides, the carbon layer films that possess more amorphous regions induce more exciton production in the photocatalysis procedure, and allow more space to effectively alleviate the volume change during lithium insertion–extraction progress, and provide more active sites in both the cases. For the latter, owing to the shortened Li⁺ diffusion distance, high contact surface area, good electronic conductivity, and superior structure stability of carbon-encapsulated core–shell nanostructure, high reversible rate capability and cycle ability are achieved. This simple and efficient synthetic protocol incorporation of morphology and surface modification can be further extended to build up other composited nanomaterials based on earth-abundant elements with promising applications in catalysis, sensing, supercapacitors, and lithium batteries and sodium batteries.

Acknowledgements

The authors thankfully acknowledge the support of National Science Foundation of China (no. 51172160) for financial supports.

References

1. T. Butburee, Y. Bai, J. Pan, X. Zong, C. H. Sun, G. Liu and L. Z. Wang, J. Mater. Chem. A, 2014, 2, 12776.
2. M. Q. Yang, Z. Nan and Y. J. Xu, ACS Appl. Mater. Interfaces, 2013, 5, 1156.
3. G. H. Qin, X. Y. Zhang and C. Y. Wang, J. Mater. Chem. A, 2014, 2, 12449.
4. W. Li, F. Wang, S. Feng, J. Wang, Z. Sun, B. Li, Y. Li, J. Yang, A. A. Elzatahry, Y. Xia and D. Zhao, J. Am. Chem. Soc., 2013, 135, 18300.
5. Q. P. Wu, J. Zhao, G. H. Qin, C. Y. Wang, X. L. Tong and S. Xue, Appl. Catal., B, 2013, 142, 142.
6. G. H. Qin, H. J. Zhang and C. Y. Wang, J. Power Sources, 2014, 272, 491.
7. J. Liu, H. Xia, D. Xue and L. Lu, J. Am. Chem. Soc., 2009, 131, 12086.
8. S. L. Xie, T. Zhai, W. Li, M. H. Yu, C. L. Liang, J. Y. Gan, X. H. Lu and Y. X. Tong, Green Chem., 2013, 15, 2434.
9. S. H. Choi and Y. C. Kang, Chem.–Eur. J., 2014, 20, 1.
10. Y. Qiao, X. L. Hu, Y. Liu, C. J. Chen, H. H. Xu, D. F. Hou, P. Hu and Y. H. Huang, J. Mater. Chem. A, 2013, 1, 10375.
11. A. R. Armstrong, G. Armstrong, J. Canales, R. García and P. G. Bruce, Adv. Mater., 2005, 17, 862.
12. J. Wang, Y. Zhou, Y. Hu, R. O'Hayre and Z. J. Shao, J. Phys. Chem. C, 2011, 115, 2529.
13. B. Weng, J. Wu, N. Zhang and Y. J. Xu, Langmuir DOI:10.1021/la4048566.
14. Q. Li, B. Guo, J. Yu, J. Ran, B. Zhang, H. J. Yan and J. R. Gong, J. Am. Chem. Soc., 2011, 133, 10878.
15. Y. T. Liang, B. K. Vijayan, O. Lyandres, K. A. Gray and M. C. Hersam, J. Phys. Chem. Lett., 2012, 3, 1760.
16. T. Xu, L. Zhang, H. Cheng and Y. Zhu, Appl. Catal., B, 2011, 101, 382.
17. N. Zhang, Y. Zhang, M. Q. Yang, Z. R. Tang and Y. J. Xu, J. Catal., 2013, 299, 210.
18. B. C. Qiu, M. Y. Xing and J. L. Zhang, J. Am. Chem. Soc., 2014, 136, 5852.
19. S. Y. Guo, S. Han, B. Chi, J. Pu and J. Li, ACS Appl. Mater. Interfaces, 2014, 6, 4743.
20. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev. 10.1039/c3cs60378d.
21. W. Zhao and Q. Zhong, RSC Adv., 2014, 4, 5653.
22. D. Wei and Y. Liu, Adv. Mater., 2010, 22, 3225–3241.
23. B. Li, G. Rong, Y. Xie, L. Huang and C. Feng, Inorg. Chem., 2006, 45, 6404.
24. G. H. Qin, Q. P. Wu, J. Zhao, Q. Q. Ma and C. Y. Wang, J. Power Sources, 2014, 248, 588.
25. H. Guo, W. Wang, L. Liu, Y. He, C. Li and Y. Wang, Green Chem., 2013, 15, 2810.
26. Q. Li, N. Zhang, Y. Yang, G. Z. Wang and D. H. L. Ng, Langmuir, 2014, 30, 8965.
27. S. Usai, S. Obregón, A. I. Becerro and G. Colón, J. Phys. Chem. C, 2013, 117, 24479.
28. M. Q. Yang and Y. J. Xu, J. Phys. Chem. C, 2013, 117, 21724.
29. X. F. Zhang, K. X. Wang, X. Wei and J. S. Chen, Chem. Mater., 2011, 23, 5290.
30. M. E. Stournara, Y. Qi and V. B. Shenoy, Nano Lett., 2014, 14, 2140.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra08931f

---