Free-Standing Electrospun W-Doped BiVO4 Porous Nanotubes for the Efficient Photoelectrochemical Water Oxidation

While bismuth vanadate (BiVO4) has emerged as a promising photoanode in solar water splitting, it still suffers from poor electron-hole separation and electron transport properties. Therefore, the development of BiVO4 nanomaterials that enable performing high charge transfer rate at the interface and lowering charge recombination is urgent needed. Herein, cobalt borate (Co-B) nanoparticle arrays anchored on electrospun W-doped BiVO4 porous nanotubes (BiV0.97W0.03O4) was prepared for photoelectrochemical (PEC) water oxidation. One-dimensional, free-standing and porousBiV0.97W0.03O4/Co-B nanotubes was synthesized through electrospun and electrodeposition process. BiV0.97W0.03O4/Co-B arrays exhibit a unique self-supporting core-shell structure with rough porous surface, providing abundant conductive cofactor (W) and electrochemically active sites (Co) exposed to the electrolyte. When applied to PEC water oxidation. BiV0.97W0.03O4/Co-B modified FTO electrode displays high incident photon-to-current conversion efficiency (IPCE) of 33% at 405 nm (at 1.23 V vs. RHE) and its photocurrent density is about 4 times to the pristine nanotube. The higher PEC water oxidation properties of BiV0.97W0.03O4/Co-B porous nanotubes may be attributed to the effectively suppress the electron-hole recombination at electrolyte interface due to its self-supporting core-shell structure, the high electrocatalytic activity of Co and the good electrical conductivity of BiV0.97W0.03O4 arrays. This work offers a simple preparation strategy for the integrated Co-B nanoparticle with BiV0.97W0.03O4 nanotube, demonstrating the synergistic effect of co-catalysts for PEC water oxidation.


INTRODUCTION
Renewable, sustainable and environmental-friendly energy sources (such as solar-hydrogen) are extremely required owing to energy exhaustion and the environmental pollution (Chu and Majumdar, 2012;Dominković et al., 2018). Photoelectrochemical (PEC) water splitting serves as an excellent sustainable and environmentally friendly method to cleanly produce H 2 from water via solar light (Li et al., 2013;Modestino and Haussener, 2015;Jiang et al., 2017;Lianos, 2017). As the core components in PEC water splitting cell, semiconductor nanomaterials show a decisive influence on the conversion efficiency of solar-tohydrogen (Alexander et al., 2008;Higashi et al., 2011;Zhang et al., 2013;Li et al., 2014;Shi et al., 2016;Tamirat et al., 2016). Among these candidate semiconductors, the monoclinic BiVO 4 is most potential for PEC water splitting as it offers moderate band gap (2.4-2.5 eV) and appropriate band-edge positions (Cooper et al., 2014;Tan et al., 2017). Theoretically, the BiVO 4 can absorb up to 11% of the solar spectrum and produce upwards of 7.5 mA·cm −2 of photocurrent (Pihosh et al., 2014). Actually, due to its poor electron mobility (0.044 cm 2 /v·s), small hole collection depth [70 nm (Kim and Lee, 2019)], and excessive surface recombination, bare BiVO 4 exhibits low efficiency for PEC water splitting, which hampered its use in energy conversion domains (Zachäus et al., 2017). Therefore, the development of BiVO 4 nanocomposite materials that enable performing high charge transfer rate at the interface and lowering charge recombination is urgent needed for improving the PEC water splitting efficiency.
Currently, the strategy for enlarging charge separation and transport rate mainly involved in the doping foreign elements or co-catalyst (Berglund et al., 2011;Kim et al., 2015;Zhang et al., 2017). For example, Luo et al. demonstrated an effective BiVO 4 photoanode through doping with various metal ions using metal-organic decomposition, and found that only doping W 6+ or Mo 6+ into V 5+ site can supply additional free electrons and enhance the PEC photocurrent (Luo et al., 2011;Yang et al., 2017;Xin et al., 2018). Apart from doping foreign elements, the cocatalysts [such as Ni-Bi, Co-Pi, CoMoO 4 , FeOOH and TiO 2 (Zhou et al., 2012;Cheng B. Y. et al., 2016;Wang et al., 2017;Du et al., 2018)] can improve the water oxidation kinetics by reducing the activation energy of the rate-determining step of the four electron oxidation process. Significantly, PEC water splitting efficiency mainly depended both on charge transfer and on water oxidation kinetic. In this regard, the integration of doping foreign elements and cocatalyst on the BiVO 4 photoanode are more required for the higher PEC water splitting efficiency.
Apart from above factors, nanostructuring is an effective approach to reduce bulk recombination by shortening the diffusion length for charge carriers (Zhang et al., 2014a,b;Han et al., 2017). The BiVO 4 with one-dimensional (1D) morphologies (such as nanowire, nanotube and nanofiber), which has long axial-ratio and high active sites, is a promising photocatalyst (Boettcher et al., 2010;Hernández et al., 2017). These nanostructures, especially nanotube, can provide a large specific surface area, and also short the carrier diffusion length. Recent study demonstrated that 1D nanomaterial has been proved to be outstanding for solar-to-hydrogen conversion (Yao et al., 2019). Significantly, electrospun is a simple, flexible and efficient technology to deal with polymer/inorganic materials into three-dimensional nanofibers with controllable composition, diameter and porosity, and has been concerned in the photovoltaics, chemical sensors, and photocatalysis owing to the one-dimensional open structure, large surface areas, and high porosity (Kumar et al., 2014). While electrospun had been used to fabricate BiVO 4 nanofibers for photocatalysis (Yoon et al., 2015), no studies have been reported on fabricating BiVO 4 nanotubes as a photoanode for water splitting, and the effect of doping and cocatalyst on the nanotubes.
Here, one-dimensional, free-standing and porous BiV 0.97 W 0.03 O 4 /Co-B nanotubes were synthesized through electrospun and electrodeposition process. BiV 0.97 W 0.03 O 4 /Co-B arrays exhibited a unique self-supporting core-shell structure with rough porous surface, providing abundant conductive cofactor (W) and electrochemically active sites (Co) exposed to the electrolyte. When applied to PEC water oxidation, BiV 0.97 W 0.03 O 4 /Co-B modified FTO electrode displayed higher incident photon-to-current conversion efficiency (IPCE) of 33% at 405 nm (1.23 V vs. RHE), and its photocurrent density is about 4 times to the pristine nanotube. The higher PEC water oxidation properties of BiV 0.97 W 0.03 O 4 /Co-B porous nanotubes may be attributed to the effectively suppress the electron-hole recombination at electrolyte interface due to its self-supporting core-shell structure, the high electrocatalytic activity of Co and the good electrical conductivity of BiV 0.97 W 0.03 O 4 arrays. This work offers a simple preparation strategy for the integrated Co-B nanoparticle with BiV 0.97 W 0.03 O 4 nanotube, demonstrating the synergistic effect of co-catalysts for PEC water oxidation.

MATERIALS AND METHODS
The BiV 0.97 W 0.03 O 4 /Co-B nanotube was fabricated through a strategy including three sequential steps: electrospun, hightemperature annealing, and electrodeposition, as illustrated in Figure 1.

Preparation of BiV 0.97 W 0.03 O 4 /Co-B Porous Nanotube
Firstly, Bi(NO 3 ) 3 •5H 2 O and VO(acac) 2 were added and dissolved into a mixture of acetylacetone and DMF. Subsequently, according to the stoichiometric ratio of Bi: V: W = 100: 97: 3, the W(OC 2 H 5 ) 6 was added and stirred for 1 h to form metal ion complex ( Figure S1A). Then, the PVP was added and stirred 12 h, forming the homogeneous and stable precursor solution for electrospinning.
The electrospinning experiment was performed on a selfmade instrument, which consisted of a syringe, a grounded collector, and a high-voltage supply. The homogeneous solution was transformed into the plastic syringe equipped with a needle (inner-diameter 0.4 mm) as the spinneret. The counter plate as collector was covered with aluminum foil, where fluorine doped tinoxide glass (FTO, 3 × 2.5 cm 2 , OPV-FTO-22-07) was also arranged to collect the nanofiber. The electrospinning was carried out at tip-collector spacing of 15 cm, accelerating voltage of 19 kV, flow rate of 0.4 mL/h, and relative humidity of 30%. After electrospinning for 30 min, the nanofiber mats collected on FTO were dried at 110 • C for 10 h, which was shown in Figure S1B. Based on the TG results (see Figure S2), the thermal annealing was carried out at 490 • C for 1.5 h. After naturally cooling down to room temperature, the electrodes of BiV 0.97 W 0.03 O 4 nanotube was successfully obtained, which was shown in Figure S1C. For comparison, some electrodes of pristine BiVO 4 nanotubes were also prepared.
The Co-B was loaded on BiV 0.97 W 0.03 O 4 nanotube by electrodeposition. A three-electrode system was employed with an as-prepared nanotube electrode (working electrode), a Ag/AgCl reference electrode, and a platinum counter electrode. The electrolyte was 0.5 mM Co(NO 3 ) 2 ·6H 2 O in 0.1 M potassium borate buffer (pH = 8.50). The electrodeposition was carried out using an electrochemical workstation (Zahner Zennium) at 0.70 V vs. Ag/AgCl for 1 min.

Characterization
The morphology of nanotube was investigated by scanning electron microscopy (SEM, Zeiss Merlin), energy-dispersive Xray spectroscopy (EDS), and transmission electron microscope (TEM, JEM-2100). The structural properties were characterized by X-ray photoelectron spectroscopy (XPS, ESCALAB Xi + ) and X-ray diffraction (XRD, Bruker Smart-1000CCD diffractometer with Cu Kα radiation, λ = 1.5406 nm). The thickness of photoanode was measured by step profiler (KLA-Tencor D100). Light absorption was measured using a UV-visible spectrophotometer (PE lambda 750) by measuring the reflectance with an integrated sphere attachment. The PEC experiments of as-prepared BiVO 4 and BiV 0.97 W 0.03 O 4 nanotubes were performed in 0.5 M Na 2 SO 4 electrolyte(PH = 7.00) on a electrochemical workstation (Zahner Zennium, Germany) in a three-electrode PEC cell. While the PEC performance of BiV 0.97 W 0.03 O 4 /Co-B nanotube was measured in the same PEC cell with 0.1 M borate electrolyte (pH = 8.50). The PEC cell was composed of Ag/AgCl reference electrode, a platinum counter electrode and working electrode (the asprepared nanotubes on FTO), respectively. The white light LED (average λ = 536 nm, P = 100 mW·cm −2 , Figure S3) and 300 W Xe lamp (CEL-HXF300-T3, P = 100 mW·cm −2 , λ > 420 nm) were used as illumination source. For incident photon-to-current efficiency (IPCE) measurements, a Zahner tunable light source system, model CIMPS TLS03, was employed to exhibit a LED array for monochromatic light excitation. The chopped light voltammetry measurement was carried out with scan speed of 10 mV/s and light period time of 8 s. The photoelectrochemical impedance spectroscopy (PEIS) was performed from 1 to 10 5 Hz frequency with 5 mV amplitude. The Mott-Schottky (M-S) analysis under dark condition was carried out at frequencies 1 kHz with a step width of 50 mV/s.

Morphology and Structure of Nanotube
The morphology of the as-prepared BiV 0.97 W 0.03 O 4 nanotube are inspected by SEM and TEM measurements. Figure S1C shows the photograph of BiV 0.97 W 0.03 O 4 film on FTO glass by elelctrospinning, which shows the good light transmission. Figure 2A shows that the non-woven film is formed by randomly oriented nanotubes, and the nanotubes possess porous hollow tubular structure which has sufficiently large surface active sites for electron and hole separations. The high magnification image of Figure 2B reveals that the porous nanotube exhibits a diameter ranging from 150 to 300 nm, and further verifies a porous structure. Figures 2B,C reveal the wall thickness of nanotubes are about 15 and 40 nm, respectively, which indicate a non-uniform shell structure throughout the nanotube. The elemental mapping is shown in Figure 2D for Bi, V, O, and W, respectively, which further verifies the homogeneous distribution of elements within the nanotube.
The crystal structures of the as-fabricated electrodes are further characterized by XRD (Figure 3). In the pristine BiVO 4 nanotube pattern, the four diffraction peaks in 28.5, 30.5, 34.5, and 35.2 • are related to the crystal planes (121), (040), (200), and (020) of the monoclinic BiVO 4 , respectively (JCPDS 014-0688).  After doping W, the predominant (121) and (020) (200) and (020) can be ascribed to a slight lattice change from monoclinic to tetragonal symmetry. No Co-B diffraction peaks can be found in the XRD patterns because of amorphous Co-B (Kanan and Nocera, 2008). The chemical composition is also verified by XPS, which are presented in Figure 4. In the pristine BiVO 4 nanotube, the two main asymmetric peaks of Bi4f 7/2 and Bi4f 5/2 can be found at 159.0 and 164.7 eV, respectively, corresponding to the Bi 3+ oxidation state; Meanwhile, the two asymmetric peaks of V2p 3/2 and V2p 1/2 can be found at 517.0 and 524.9 eV, respectively, which is corresponding to a V 5+ oxidation state. In the BiV 0.97 W 0.03 O 4 nanotube, the W4f 7/2 and W4f 5/2 peaks located at 35.2 and 37.8 eV, respectively, suggesting the oxidized state of W (W 6+ ) to substitute V 5+ atoms on the surface of the BiVO 4 photoanode, and the positive charge will be compensated by free electrons. The asymmetrical O1s peak ( Figure 4D) in the range of 528-534 eV is fitted into two components centered at 529.9, 530.7 eV which are assigned to the lattice species (O L ) and oxygen vacancies (O V ), respectively. It is worth noting that BiV 0.97 W 0.03 O 4 nanotube possesses more oxygen vacancies than the pristine sample. Moreover, the binding energies of the Bi4f, V2p and O1s slightly shift toward positive direction compared to the pure BiVO 4 , due to the higher electronegativity of the W-dopant . As shown in Figure 4E, the two main asymmetric peaks of Co2p 3/2 and Co2p 1/2 can be found at 780.3 and 795.0 eV, respectively, both corresponding to the Co 2+ . Together, the XRD and XPS results confirm that the W 6+ cation is expectantly doped into the BiVO 4 lattice and also slightly deform the monoclinic structure.

PEC Performance
The band gap was estimated using UV-visible spectroscopy in diffuse reflectance mode. As shown in Figure 5A, there was no obvious enhancement of the UV absorption intensity after doping W element, indicating that the W element was not essential for light absorption. The absorption edges were located at about 505 nm (2.45 eV) and its electrode modification layer thickness was about 800 nm. To study the PEC performance of BiVO 4 nanotube, the linear sweep voltammetry under light illumination was carried out. Figure 5B depicts the    The photocurrents promptly increase when the light is on, and drop when the light is off over a wide potential range, which imply that the photocurrent is generated under light irradiation. It can be also seen that the transient photocurrent peak upon turning the light off/on, which indicates the accumulation of holes at the space charge layer under prolonged irradiation, and the back recombination of electrons with the accumulated holes. Figure 5C shows the current-potential (J-V) curves under AM 1.5G irradiation.  Figure 6B) as before. According to the value of photocurrent density and IPCE, the BiV 0.97 W 0.03 O 4 /Co-B nanotube exhibits the highest PEC performance ever achieved for BiVO 4 nanofiber using electrospinning (Yoon et al., 2015;Antony et al., 2016;Liu et al., 2017;He et al., 2018;Li et al., 2019). Table S1 shows the comparison of photocurrent data reported in the literature with the photocurrent value obtained in the present study. The nanotube-based BiVO 4 photoanode exhibit lower photocurrent density than state-of-the-art photoanode by electrodeposition method about 1.05 mA·cm −2 at 1.23 V vs. RHE. Therefore, an optimization of the film thickness in order to increase amount of active material per area unit is highly desirable, and will be shown and discussed later on. In order to identify the influence of W 6+ and Co-B on the charge transport kinetics of the BiVO 4 nanotube, the analysis of PEIS is carried out at 1.23 V vs. RHE. Figure 7A shows the Nyquist plots of the pristine BiVO 4 , BiV 0.97 W 0.03 O 4 , and BiV 0.97 W 0.03 O 4 /Co-B nanotubes, respectively, and the inset of Figure 7A shows the equivalent electrical circuit. In this equivalent electrical circuit, theR CT denotes the charge transfer resistance from the BiVO 4 photoanode to electrolyte solution, the CPE represents the constant phase element for the electrolyte/photoanode interface, and the Rs is the resistance associated with FTO substrates, the electrolyte, and wire connections in the whole circuit. As shown in Table S2, the total error was below 3%, which presents the Nyquist plots can be fitted well with the equivalent circuit model.  Figure 7B). The flat band potential (E fb ) can be calculated using equation (see supporting information in ESI † ). The E fb of pristine and Wdoped BiVO 4 nanotubes are 0.215, 0.196 V vs. RHE, respectively. Moreover, their carrier densities are 6.4 × 10 19 , 8.9 × 10 19 cm −3 , respectively. This result provides direct evidence that W element is effectively doped into BiVO 4 lattice, and can effectively elevate the donor density, which result in the efficient electron transportation.
Based on the above results, the charge transfer process in the BiV 0.97 W 0.03 O 4 /Co-B nanotube photoanode can be demonstrated in Figure 8. Under light illumination, the BiV 0.97 W 0.03 O 4 nanotube can absorb the photons, and generate electron-hole pairs. Then the photogenerated electrons and holes migrate to the surface of photoanode, and inject into the electrolyte to generate hydrogen and oxygen, respectively. From Figure S4, we can see that there are a lot of recombination of photogenerated electron-hole pairs both in bulk and surface during the overall reaction. The results of PEC test show that the BiV 0.97 W 0.03 O 4 /Co-B nanotube demonstrated better photocatalytic activity compared with other nanofibers that have ever been reported (Yoon et al., 2015;Antony et al., 2016;Liu et al., 2017;He et al., 2018;Li et al., 2019). The reasons can be explained as follows: first of all, the nanotube with large interface area and the porous structure can keep in good contact with the electrolyte, and enrich the active sites. Meanwhile, the W 6+ -doping at V 5+ sites in the BiV 0.97 W 0.03 O 4 nanotube can increase the donor density for the host lattice, and reduce the effective mass of electron, resulting in a higher probability of reaching the active sites of electrode /electrolyte; the nanotube can also minimize the distance of electrons diffusing to the FTO substrate, thereby decreasing the bulk electron/hole recombination; the Co-B cocatalyst can also efficiently extracts photo generated holes from the BiV 0.97 W 0.03 O 4 nanotube, and store long-lived holes as Co 4+ species. Thus, the more holes migrate to the electrode/ electrolyte interface, and the back electron/hole recombination in the space charge layer is significantly retarded. In summary, porous nanotube, doping, and cocatalyst should account mainly for the enhanced PEC performance of BiV 0.97 W 0.03 O 4 /Co-B nanotube photoanode.

CONCLUSIONS
In summary, the cobalt borate (Co-B) nanoparticle arrays anchored on W-doped BiVO 4 porous nanotubes (BiV 0.97 W 0.03 O 4 ) with large specific surface area and short diffusion length have been successfully synthesized by electrospun and electrodeposition process. The as-prepared BiV0 .97 W 0.03 O 4 /Co-B arrays exhibit a unique self-supporting core-shell structure with rough porous surface, providing abundant active sites exposed to the electrolyte. Their diameters range from 150 to 300 nm, and wall thicknesses range from 10 to 40 nm. XPS and XRD reveal that the W-cations are doped into lattice and also slightly deform the monoclinic structure. PEIS and M-S measurements reveal that doping a small amount (3 atom%) of W 6+ can effectively increase carrier density and reduce the charge transfer resistance. The photoanode of BiV 0.97 W 0.03 O 4 /Co-B nanotube possesses the IPCE of 33% at 405 nm at 1.23 V vs. RHE, and its photocurrent density is about 4 times to that of the pristine nanotube. This enhancement is attributed to the suppression of surface electron/hole recombination, and the higher bulk electron transport. These results offer a simple preparation strategy for the integrated Co-B nanoparticle with BiV 0.97 W 0.03 O 4 nanotube, demonstrating the synergistic effect of co-catalysts for PEC water oxidation.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
XY planned the experimental work, wrote the manuscript, and helped in the analysis. XS, HZ, and SZ helped in the analysis, and explained the results. XL and BL writing of original draft manuscript. DL helped in the experimental work and explained the results.