Controllable Synthesis of N2-Intercalated WO3 Nanorod Photoanode Harvesting a Wide Range of Visible Light for Photoelectrochemical Water Oxidation

A highly efficient visible-light-driven photoanode, N2-intercalated tungsten trioxide (WO3) nanorod, has been controllably synthesized by using the dual role of hydrazine (N2H4), which functioned simultaneously as a structure directing agent and as a nitrogen source for N2 intercalation. The SEM results indicated that the controllable formation of WO3 nanorod by changing the amount of N2H4. The β values of lattice parameters of the monoclinic phase and the lattice volume changed significantly with the nW: nN2H4 ratio. This is consistent with the addition of N2H4 dependence of the N content, clarifying the intercalation of N2 in the WO3 lattice. The UV-visible diffuse reflectance spectra (DRS) of N2-intercalated exhibited a significant redshift in the absorption edge with new shoulders appearing at 470–600 nm, which became more intense as the nW:nN2H4 ratio increased from 1:1.2 and then decreased up to 1:5 through the maximum at 1:2.5. This addition of N2H4 dependence is consistent with the case of the N contents. This suggests that N2 intercalating into the WO3 lattice is responsible for the considerable red shift in the absorption edge, with a new shoulder appearing at 470−600 nm owing to formation of an intra-bandgap above the VB edges and a dopant energy level below the CB of WO3. The N2 intercalated WO3 photoanode generated a photoanodic current under visible light irradiation below 530 nm due to the photoelectrochemical (PEC) water oxidation, compared with pure WO3 doing so below 470 nm. The high incident photon-to-current conversion efficiency (IPCE) of the WO3-2.5 photoanode is due to efficient electron transport through the WO3 nanorod film.


Introduction
In order to solve both the energy crisis and environmental problems, it is urgent to explore a sustainable, and environmental-friendly energy resources to replace the nonrenewable fossil energy [1][2][3]. Hydrogen, as an energy carrier of solar energy, has the advantages of high energy density, storage, and pollution-free. So, it can be used as an ideal new energy to replace the traditional energy. To date, among the different strategies for producing hydrogen, PEC water splitting is considered to be a promising process in which solar energy can be directly convert into hydrogen and oxygen using a semiconductor [4]. However, the overall efficiency of PEC water splitting is still relatively low due to the high kinetic overpotentials for the water oxidation reaction. Therefore, the key to improving the efficiency of PEC water splitting is to develop a robust semiconductor photoanode for visible-light-driven water oxidation. Since TiO 2 was first reported as photoanode by Fujishima and Honda in 1972 [5], numerous semiconductors (WO 3 , Fe 2 O 3 , Ta 3 N 5 , TaON, etc.) have been employed for photoanode materials [6][7][8][9][10][11][12][13]. WO 3 was investigated as a visible-light-driven photoanode material for PEC water oxidation by Hodes in1976 [14]. Since then, considerable efforts [15][16][17][18][19][20][21] have been devoted to investigating the PEC performances of WO 3 due to the following intrinsic properties: (1) a relatively narrow band gap energy (2.5 eV-2.7 eV), (2) a positive valence band edge position for water oxidation, (3) high stability in acidic condition, (4) a suitable hole diffusion length (~150 nm) for superior electron transport. However, the solar spectrum utilization of WO 3 photoanode is still too low to be used for actual application due to its narrow visible light response range (λ < 460 nm). To enhance the PEC performance, many tactics have been employed to develop efficient WO 3 photoanode for water oxidation, including forming heterojunction, doping, loading co-catalysts, nanostructuring, tuning vacancies and so on [1,4,[22][23][24][25][26][27][28][29][30]. Nanostructured WO 3 photoanodes have attracted more and more research attention to improve their PEC performance, because they offer superior performances than that of unspecified ones because of larger electrode/electrolyte interface area, efficient light absorption, and other structural benefits of nanostructures. So far, various nanostructured WO 3 photoanodes have been synthesized, such as nanosheet [31], nanoplatelets [32], nanoparticles [33,34], nanoflakes [35,36], nanotubes [37], nanobelts [38], nanowire [39,40], and nanorods [41,42], to improve performances in PEC water oxidation. Although the PEC performance of WO 3 photoanode for water oxidation can be enhanced by nanostructure control, the light absorption at long wavelength is difficult to improve. Nowadays, many studies are focusing on extension of light absorption to longer wavelength (λ > 480 nm) by band gap engineering of WO 3 , because efficient light absorption at longer wavelength is significant to improve the efficiency of a photoanode for solar energy conversion. To date, doping WO 3 with transition/other metals [43][44][45][46], nonmetallic elements [25,26,47], or selective molecule [27] to improve its the light absorption at longer wavelengths have been regarded as the most common strategies for band gap engineering.
We previously reported an in situ N 2 -intercalated WO 3 nanorod photoanode, in which N 2 H 4 was used as a dual-functional structure-directing agent for nanorod as well as a nitrogen source for in situ N 2 intercalation [17]. Not only the light absorption at longer wavelengths, but also the PEC performance for water oxidation was dramatically improved. Therefore, a difficult issue of compatibility between nanostructure control and strategy for band gap engineering of WO 3 was resolved. However, we only discussed the calcination temperature dependence on the physiochemical properties. The PEC performance, the effect of addition of N 2 H 4 on the content of N 2 into the WO 3 lattice, as well as the morphology of the N 2 -intercalated WO 3 and the PEC performance for water oxidation have not been investigated yet. Thus, it is necessary to reveal the effect of addition of N 2 H 4 on the morphology and band gap of N 2 -intercalated WO 3 . Herein, we first report the controllable synthesis of N 2 -intercalated WO 3 with different morphologies and band gap using N 2 H 4 as a dual-functional surfactant template. Especially, there is an extremely significant mutually dependent relation between the addition of N 2 H 4 and the PEC performance of N 2 -intercalated WO 3 for water oxidation.

Characterization Structure of N 2 -Intercalated WO 3 Samples
The SEM images showed different morphologies depending on the addition of N 2 H 4 . WO 3 -0 is composed of small irregular nanoblocks with diameters of 10−50 nm (Figure 1a). Figure 1b−f shows the images of N 2 -intercalated WO 3 prepared using different addition of N 2 H 4 . It can be clearly found that the mixed morphologies of nanorods in large scale and small nanoblocks start to be observed from n W :n N2H4 of 1:0.62 (Figure 1b). There are not so many differences in the length and width of the nanorods which are obtained from n W :n N2H4 ratio of 1:1.25 to 1:2.5. However, the length and width increased as the n W :n N2H4 ratio increased from 1:5 to 1:7.5. The specific surface area of WO 3 −2.5 was 2.2 times higher than that of WO 3 -0 (9.6 m 2 g −1 ). The surface area of N 2 -intercalated WO 3 samples decreased to 20.4-17.3 m 2 g −1 with increasing n W :n N2H4 ratio from 1:5 to 1:7.5 due to the agglomeration of nanorod surfaces.
The EDS data of N 2 -intercalated WO 3 indicate the existence of the N element, which derived from N 2 H 4 in precursors, because none of the N element was detected in WO 3 -0; see Figure 2. The W/N atomic ratio was calculated from the EDS data to reveal the dependence of the n W :n N2H4 ratio on it. The results indicate that the W/N atomic ratio increased from 1:0.004 to 1:0.096 with increasing n W :n N2H4 ratio from 1:0.62 to 1:2.5 and thereafter decreased over the 1:2.5 (Table 1).
To instigate the effect of the addition of N 2 H 4 on the crystal structures, the N 2intercalated WO 3 samples are ascertained by XRD ( Figure 3A) and Raman ( Figure 3B). As shown in Figure 2A, it can be clearly observed that all of samples exhibited the monoclinic WO 3 crystals (PDF # 01-083-0950). Additionally, the crystalline structure of N 2 -intercalated WO 3 samples can be significantly affected by the addition of N 2 H 4 . Especially, the intensity of the (002) peak is higher than those of the other neighbor peaks of (020) and (200) for N 2 -intercalated WO 3 samples, which is different from that of WO 3 -0 sample. This could suggest anisotropic progress of crystallization involving the predominant crystallization of (002) from n W :n N2H4 ratio of 1:0.32 to 1:2.5 followed by progressive crystallization of (020) and (200) with the ratio increased to 1:7.5, as depicted by calculation of the crystallite diameter of (002). The larger crystallite diameter of N 2 -intercalated WO 3 samples (27-31 nm) than that of WO 3 -0 (17 nm) can be clearly observed, as shown in Table 1. The PEC performance of WO 3 photoanodes for water splitting was extremely enhanced with highly uniform alignment along the (002) facet [48]. In the Raman spectra, the WO 3 samples exhibited the characteristic peaks of the monoclinic WO 3 at 134.4 cm −1 (lattice vibration), 270.6 cm −1 (δ (O-W-O) deformation vibration), 713.2cm −1 , and 807.1 cm −1 (δ (O-W-O) stretching vibration) in the range of Raman shift from 100~1000 cm −1 . Raman analysis also showed the tendency of Raman peaks to sharpen as the n W :n (NH4)2S ratio increased, which is well consistent with the results of XRD analysis.  As shown in Figure 4, it could be seen that the lattice parameters a, b, c and β were remarkably influenced by the addition of N 2 H 4 for the N 2 -intercalated WO 3 (2)). The values of parameters b, c, and β for all of samples decreased with the addition of N 2 H 4 increase up to 1:2.5, and then increased at higher molar ratio of n W :n N2H4 . However, the value of parameters an increased below 1:2.5 and then decreased with molar ratio of n W :n N2H4 increased. The lattice volumes of N 2 -intercalated WO 3 samples were larger than that of WO 3 −0, and the largest lattice volume of 216.54(1) Å 3 for WO 3 −2.5 was observed due to the insertion of highest contents for N 2 into the lattice. Therefore, it can be confirmed that the N 2 was intercalated into WO 3 lattice by analyzing the change of lattice parameters and volume. Raman spectra at high wavenumber ( Figure 5) were investigated to clarify the configuration and existence of N 2 into the WO 3 lattice, which is responsible for the N content. No peak was observed for WO 3 -0, however, Raman spectra of N 2 -intercalated WO 3 samples exhibited signals at 2327−2342 cm −1 , which can be ascribed to the N≡N vibration of N 2 in the WO 3 lattice on the basis of N≡N vibration of gaseous N 2 (2330 cm −1 ) and the N 2intercalated WO 3 reported previously as well as relevant compounds [49][50][51][52]. Additionally, only one peak at 2330 cm −1 was observed for WO 3 −0.62. It can be attribute to the lower amount of N 2 into the WO 3 −0.62 lattice, because such one peak was observed for N 2intercalted WO 3 prepared by dodecylamine according to ours earlier report [30]. However, it is amazing to note that two peaks at 2328 cm −1 and 2342 cm −1 began to be observed from n W :n N2H4 of 1:1.2 as the content of N 2 increased. This significant difference confirms that the addition of N 2 H 4 affect not only the N content but also the configuration of N 2 into the WO 3 lattice. Further, The N 2 intercalation can be also proved by our previously reported [17,30], the XPS spectrum in an W 4f region was deconvoluted by four bands at 37.0, 34.9, 32.9, 31.2 eV for N 2 intercalated WO 3 The bands at 37.0 and 34.9 eV in higher energy for N 2 intercalated WO 3 (37.1 and 34.9 eV for NH 3 -WO 3 ) are assigned to 4f 5/2 and W 4f 7/2 of the WO 3 lattice similarly to WO 3 −0. The bands at 32.9 and 31.2 eV in lower energy for N 2 intercalated WO 3 can be assigned to the binding energies of 4f 5/2 and W 4f 7/2 interacted with N 2 intercalated. XPS data of WO 3 -N 2 H 4 , in which a peak at 396.9 eV assigned to the intercalated nitrogen bound with the tungsten center. This is obviously different from the band at 396 eV of N-doped metal oxides (WO 3 , TiO 2 , etc.) [53,54] and metal nitrides (TiN) [55]. Additionally, the nitrogen intercalation hardly causes the oxygen defects, which can be beneficial to improving the optical properties of WO 3 .

The Optical Properties of N 2 Intercalated WO 3
The optical properties of WO 3 samples were investigated by UV-vis DRS ( Figure 6A) and the corresponding Tauc plots ( Figure 6B) of the WO 3 −0 (a) and N 2 intercalated WO 3 (bf) samples with changes in the ratio of n H2WO4 :n N2H4 . Compared to the WO 3 −0 (469 nm), only a slight redshift of 11 nm in the absorption edge was observed for the WO 3 −0.62 (480 nm). However, a significant redshift can be seen in the absorption edge with new shoulders appeared in a longer wavelength region than WO 3 −0 with increased the ratio of n H2WO4: n N2H4 . It is obvious that the absorption edges extended to the longer wavelength as the ratio of n H2WO4 :n N2H4 was increased below 1:2.5, and then they decreased when further increased the addition of N 2 H 4 . Furthermore, the formation of the absorption shoulders exhibited the same trend as the formation of the peak at 2342 cm −1 in the Raman spectra, implying that the formation of this peak is beneficial to the generation of the absorption shoulders. According to the Tauc plots, the bandgap of WO 3 −0.62 (2.58 eV) was slightly reduced by 0.06 eV than WO 3 −0 (2.64 eV) due to the formation of a new intermediate N 2p orbital between the CB and the VB after intercalation of N 2 into the WO 3 lattice, as indicated by the earlier report [17,27,30]. As the ratio of n H2WO4 :n N2H4 was increased from 1:1.2, the Tauc plots exhibited two different slopes due to the existence of the new shoulders, with the absorption energies derived from the slopes, as shown in Table 2.   Figure 7 illustrates the relation between the KM values at 500 nm (KM 500 ). It was observed that the KM 500 value is a measure of the increase/decrease of the shoulders at 470-600 nm. Compared with both WO 3 −0 and WO 3 -0.62, the KM 500 increased from 0.11 to 0.31 with an increase in the ratio of n H2WO4 :n N2H4 from 1:1.2 to 1:2.5, and thereafter, decreased from 1:2.5 to 0.13 at 1:7.5. The dependency of KM 500 on the n H2WO4 :n N2H4 ratios applies to the case of the N content, suggesting that the longer wavelength absorption due to the shoulders can be attributed to the intercalation of N 2 into a WO 3 lattice. Mott-Schottky plots from alternating-current impedance measurements were taken to reveal the band structure of the N 2 intercalated WO 3 electrodes. As shown in Figure 8, typical behavior for n-type semiconductors was confirmed for all WO 3 samples, because the reciprocal of the square of capacitance (C −2 [F −2 cm 4 ]) linearly increased with the applied potentials beyond the flat band (E FB ) potentials. The E FB and the donor carrier densities (N D [cm −3 ]) were provided from the x-intercept and the slopes of the straight line, respectively ( Table 2). E FB value of the WO 3 -0 electrode was 0.38 V, which was closed to those of earlier-reported WO 3 electrodes (0.36-0.41 V vs. Ag/AgCl) [32,33]. However, the E FB values of the N 2 intercalated WO 3 electrodes (0.23-0.36 V) were lower than that of WO 3 -0 electrode. The N D values of the N 2 intercalated WO 3 electrodes were higher than that for WO 3 -0 electrode. Especially, the highest N D value for WO 3 −2.5 (4.15 × 10 19 cm −3 ) was calculated, which was 1.12, 1.09, 1.08, 1.03, and 1.06 times higher than those of WO 3 −0, WO 3 −0.62, WO 3 −1.2, WO 3 −5, and WO 3 −7.5 electrodes. The negative shift of the E FB potential and the increase of the N D are usually beneficial to the improvement of the PEC performance for water oxidation. The band structures of the electrodes were estimated, and their energies are summarized in Table 2

Photoelectrochemical Properties
The LSVs for these electrodes were measured with chopped visible light irradiation to study their PEC water oxidation performance. The onset potential for photocurrents was 0.2 V vs. Ag/AgCl on visible-light irradiation for these electrodes due to water oxidation ( Figure 9A). The photocurrent of 1.08 mA cm −2 at 1.0 V for WO 3 −2.5 was the highest.   Figure 10 shows the action spectra of IPCE at 0.5 V vs. Ag/AgCl for different electrodes. The photocurrent could only be observed below 470 nm for WO 3 −0, which is consistent with the bandgap energy of WO 3 [1]. For the WO 3 −0.62 electrode, the onset wavelength for photocurrent generation was at 480 nm (2.58 eV) after N 2 intercalation. The onset wavelengths for WO 3 −1.2, WO 3 −2.5, WO 3 −5, and WO 3 −7.5 are considerably shifted to the wavelengths (530 nm) longer than that of WO 3 −0.62. Moreover, the IPCE results suggest that the photocurrent was generated based on the bandgap excitation, and the bandgap excitation occurs through collateral excitation from intermediate N 2p orbital to CB for the N 2 intercalated WO 3 electrodes. However, for all N 2 intercalated electrodes, the photocurrent at wavelengths longer than 530 nm could not be detected due to the limited current detection level of the employed apparatus.
Photoelectrocatalysis was conducted under the visible light irradiation (λ > 450 nm, 100 mW cm −2 ) at potentiostatic conditions of 0.5 V vs. Ag/AgCl (1.05 V vs. RHE) in a 0.1 M phosphate buffer (pH 6.0) for 1 h using different electrodes (Figure 11). A higher photoanodic current due to water oxidation was observed for the WO 3 −2.5 electrode. The highest charge amount passed and the amount (n O2 ) of O 2 evolved during the 1 h photoelectrocatalysis for WO 3 −2.5 was 2.05 C and 5.19 mmol (97% Faradaic efficiency), respectively, compared with the electrodes prepared at other conditions (Table 3). These results clearly prove that the intercalation of N 2 enhances the PEC performance of WO 3 −2.5 in application to water oxidation. The decay of photoanodic currents was observed for all electrodes. This can be attributed to the formation of inactive tungsten-peroxo adducts [15] and adherence of O 2 bubbles [56].   The Tafel plots and the measurement of electrochemical impedance are useful to evaluate the electron transport and its influence on PEC performance for water oxidation. In the Figure 12A, the Tafel plots of WO 3 −2.5 for water oxidation exhibited the lowest Tafel slope (d, 25 ± 0.3 mVdec −1 ) compared that of WO 3 -0 (a, 35 ± 0.2 mVdec −1 ), WO 3 −0.62 (b, 33 ± 0.4 mVdec −1 ), WO 3 −1.2 (c, 30 ± 0.3 mVdec −1 ), WO 3 −5 (e, 28 ± 0.5 mVdec −1 )and WO 3 −7.5 (f, 29 ± 0.4 mVdec −1 ), suggesting the resistance of electrochemical reaction of WO 3 −2.5 electrodes is less than those of the other WO 3 electrodes. As shown in the Figure 12B, only one semicircle was observed for all WO 3 electrodes in Nyquist plots, indicating that charge transfer process plays an important role in the PEC reaction. WO 3 −2.5 gave smaller semicircles than those of WO 3 electrodes, which implies that the WO 3 −2.5 electrode has lower charge transfer resistance and higher separation efficiency of photogenerated electron-hole pairs than others electrodes. As the ratio of n H2WO4 :n N2H4 increased from 1:0 to 1:2.5, the semicircles decreased and then increased from 1:5 to 1:7.5, corresponding to the amounts of N 2 into the WO 3 lattice.

Synthesis of N 2 -Intercalated WO 3
According to the approach we reported previously [17], N 2 H 4 ·H 2 O (36.5-438 µL, 0.75-9.0 mmol) solution were added to H 2 WO 4 (0.3 g, 1.2 mmol) under vigorous stirring at room temperature to form a yellow suspension with H 2 WO 4 :N 2 H 4 molar ratio (n W: n N2H4 ) of 1:0.62-7.5 in 3 mL water. The white N 2 H 4 -derived precursor was obtained after the suspension was slowly evaporated. The N 2 H 4 -derived precursor powders were heated at 420 • C (1 • C min −1 ) for 1.5 h in flowing O 2 to obtain different WO 3 samples. A control WO 3 sample was prepared in the same manner without addition of N 2 H 4 ·H 2 O.

Fabrication of Electrodes
In a typical procedure, 0.6 mL of water was added to the N 2 H 4 -derived precursor powder (0.8 g), PEG (0.2 mg), and Marpolose (80.0 mg), and it was slowly stirred until a smooth paste without bubbles was formed. A doctor-blading method was employed to coat the resulting paste over a clean FTO glass substrate (1.0 cm −2 area) and dried at 80 • C for 15 min. The N 2 H 4 -derived precursor electrodes were calcined at 420 • C in flowing O 2 flow to give different WO 3 electrodes, which are denoted as WO 3 -0.62, WO 3 -1.2, WO 3 -2.5, WO 3 -5 and WO 3 -7.5, respectively. The control WO 3 electrode was fabricated by the same method using a precursor prepared without addition of N 2 H 4 ·H 2 O, denoted as WO 3 -0.

Measurement
Characterization of the crystalline phase was performed by powder X-ray diffraction (XRD) using a monochromated (Shimadzu International Trade (Shanghai) Co., Ltd., Shanghai, China, XRD-6000, Cu Kα λ = 1.54 Å). The electron microscopy images of surface morphology were observed using a field emission scanning electron microscope (JEOL, TSM-6510LV, Japan). The energy-dispersive X-ray spectroscopic (EDS) data were collected using an electron probe microanalysis (JED-2300, JEOL, Tokyo, Japan) operated at an accelerating voltage of 10 kV. Raman spectra were collected using a Raman microspectroscopic apparatus (Horiba-Jobin-Yvon LabRAM HR, Paris, France) using 532 nm excitation and silicon standard wavenumber (520.7 cm −1 ). A Thermo Fisher Scientific (Thermo Fisher Scientific (China) Co., Ltd., Shanghai, China) ESCALAB Xi+ instrument was employed to collect the XPS spectra, and calibrated by the C 1s peak, appearing at 284.2 eV. A spectrophotometer (Shimadzu UV-2700) in a DR mode with an integrating sphere (ISN-723) was taken to record the UV-visible DRS were recorded. PEC measurements were examined using an electrochemical analyzer (Shanghai Chenhua Instrument Co., Ltd., CHI760E). A two-compartment PEC cell separated by a Nafion membrane. A three-electrode system has been employed using a WO 3 electrode and Ag/AgCl electrode as the working and reference electrodes in one compartment, and a Pt wire in the other compartment as the counter electrode. All the PEC experiments were taken in an aqueous 0.1 M phosphate buffer solution (pH 6.0). The linear sweep voltammograms (LSV) were measured at a scan rate of 5 mV s −1 between −0.2 V and 1.0 V. Light (λ > 450 nm, 100 mW cm −2 ) was irradiated from the backside of the working electrode using a 500W xenon lamp with a UV-cut filter (λ > 450 nm) and liquid filter (0.2 M CuSO 4 , 5.0 cm light pass length) for cutting of heat ray. Electrochemical impedance spectra were measured at an applied potential of 0.68 V vs. Ag/AgCl (1.23 V vs. RHE) in a frequency range from 10 mHz to 20 kHz (amplitude of 50 mV). The output of light intensity was calibrated as 100 mW cm −2 using a spectroradiometer (Ushio Inc., USR-40, Ushio Shanghai Inc., Shanghai, China). Photoelectrocatalysis was conducted under the potentiostatic conditions at 0.5 V at 25 • C with illumination of light (λ > 450 nm, 100 mW cm −2 ) for 1 h. The amounts of H 2 and O 2 evolved were determined from the analysis of the gas phase of counter and working electrode compartments, respectively, using gas chromatography (Shimadzu GC-8A with a TCD detector and molecular sieve 5A column and Ar carrier gas). A monochromic light with 10 nm of bandwidth was employed using a 500 W xenon lamp with a monochromator for IPCE measurements.

Conclusions
N 2 intercalated WO 3 was controllably synthesized using N 2 H 4 with a dual functional role, namely as an N atom source for N 2 intercalation and as a structure-directing agent for the nanorod architecture. The addition of N 2 H 4 dependence on the physiochemical properties and the performance of the PEC water oxidation of the WO 3 -0 and N 2 intercalated WO 3 electrodes were investigated to characterize N 2 intercalation into the WO 3 lattice and reveal the mechanism of the superior performance of PEC water oxidation for the N 2 intercalated WO 3 photoanode. The N 2 intercalated WO 3 exhibited the optimum n W :n (NH4)2S ratio at 1:2.5 for the high concentration of N elements. The N 2 intercalation is responsible for the significant red shift in the absorption edge, with a new shoulder appearing at 470-600 nm compared to that of WO 3 -0. The N 2 intercalated WO 3 photoanode is able to utilize visible light in longer wavelengths below 530 nm for PEC water oxidation, in contrast to utilization below 470 nm for the WO 3 -0 photoanode. The N 2 intercalated WO 3 photoanode is expected to be applied for PEC water splitting cells in artificial photosynthesis to improve the solar energy conversion efficiency.