Photocatalytic Hydrogen Production of Nd/Co Co-Doped BiFeO3 Nanoparticles With a Cellular Architecture

Bismuth ferrite (BFO) nanoparticle with general formula Bi1-xNdxFe1-yCoyO3 (x=0, 0.05; y=0, 0.05, 0.10, 0.15, 0.20) were prepared using a two-solvent sol-gel method. Interestingly, most of the samples exhibited a cellular architecture. Bandgap engineering of BFO nanoparticles was achieved by co-doping with Nd and Co. Under illumination with ultraviolet light, the concentration of methylene orange increased. The sample of Bi0.95Nd0.05Fe0.85Co0.15O3 produced a small amount of hydrogen (8.88molg-1 after 1.5;h), but the other samples did not produce detectable levels of hydrogen. In this research, the production of hydrogen occurred under illumination by ultraviolet light, demonstrating the splitting of pure water without the use of a sacrificial reagent. A possible reason for this is that the conduction and valence band edges of BiFeO3 straddle the water redox potential. Consequently, it is possible to realize unassisted water splitting using BFO. The ferromagnetism of all samples increased linearly with the increase of dopant concentration, and the residual magnetization of the Bi0.95Nd0.05Fe0.80Co0.20O3 sample reached to 0.679 emu g−1. Moreover, the magnetic properties of bismuth ferrite and Nd/Co Co-doped bismuth ferrite photocatalyst were also investigated to show the simple separation. These results demonstrate that BFO nanoparticles have potential applications in photocatalytic hydrogen production without the use of a sacrificial reagent.


INTRODUCTION
Multiferroics have attracted the attention of researchers worldwide in recent years is because they exhibit a magnetoelectric coupling effect and have broad application prospects in information storage, electronic devices, and sensors. Bismuth ferrite (BiFeO 3 , or BFO) stands out among multiferroics because of its magnetic and strong ferroelectric properties at room temperature (Dutta et al. 2010). The ferroelectricity of BFO is related to the Bi 6s lone electron pair at the A site, and the magnetic properties are related to the space-modulated spin structure. However, the cycloidal spin structure with a period of 62 nm leads to BFO having weak magnetic properties (Puhan et al. 2019). When other elements are substituted for the A and B positions in BFO nanoparticles, there is a significant effect on the lattice structure and the Bi 6s lone pair, thus affecting ferroelectricity and ferromagnetism (Layek et al. 2013). It has been reported that the incorporation of Nd 3+ at the A site of BFO can alter the ferromagnetism of the material (Yuan et al. 2006). Wang et al. noted that the incorporation of Nd 3+ /Mn 2+ at the A and B positions of BFO affects its structure and properties, and the incorporation of Nd 3+ induces a phase transition from an R3c crystal structure to Pnma (Wang et al. 2019). Co-doping with Co and Eu successfully enhances the magnetic and dielectric properties of BFO nanoparticles. The significant enhancement of the saturation magnetization (M s ) upon codoping with Eu and Co may be due to the destruction of the space-modulated spin structure (Das et al. 2012).
In addition, BFO has a narrow band gap (2.2 eV) and excellent chemical stability, and can be used as a photocatalyst to degrade organic pollutants and decompose water (Luo and Maggard, 2006;Li et al. 2010;Cheng et al. 2016). Bismuth Ferrite nanohybrids display improved photocatalytic activity and degradative ability (Fatima et al. 2020;Fatima et al., 2018;Iqbal et al. 2019a;Iqbal et al. 2019b). Compared with photoelectrochemical (PEC) water splitting, the photocatalytic dissociation of water has many advantages, such as being suited to splitting water of a nearly neutral pH in a onestep process without the need for an applied external bias. However, unassisted overall water splitting under a singleabsorber photocatalytic process must achieve the following two conditions: 1) the valence and conduction gap edges of this photocatalyst must astride across the water oxidation (redox) and proton reduction and potentials; 2) this photocatalyst must possess a adequate narrow bandgap to absorb a majority of the solar spectrum (Fang and Shangguan, 2019). Due to such rigorous requirement, there are very few suitable photocatalyst. Moreover, efficient charge separation is a critical issue. Bismuth ferrite is one kind of special functional material that can be modified to exhibit the required space charge region, with an accumulation layer and a depletion layer, by using n-type and p-type doping. The band bending in this space charge region can be tuned to be favorable for the separation of excited electrons and holes. In addition, for applications in the photocatalyst field, one advantage of BiFeO 3 is its excellent magnetic properties. It is essential to separate photocatalyst after they have been synthesized. Generally, coagulation, flocculation, and sedimentation are widely used as separation processes, but they are complex and expensive (Fang and Shangguan, 2019;Reddy et al. 2019). The separation process is greatly simplified if the photocatalyst is magnetized, as this allows separation by the application of an external magnetic field (Yang et al., 2018).
Chowdhury et al. reported unassisted high-efficiency splitting of pure water using a BiFeO 3 diode (Chowdhury et al. 2018). Cheng et al. reported the production of hydrogen through the use of a ferroelectric polarizationmediated Si/STO/BiFeO 3 tandem photocathode (Cheng et al. 2016). In this work, however, we wanted to prepare a material that could directly produce hydrogen through splitting water solely under illumination, without the use of a sacrificial agent or an external bias field. To achieve this, we simultaneously doped Nd 3+ and Co 2+ into the A and B sites of BFO, respectively. Bi 1-x Nd x Fe 1-y Co y O 3 nanoparticles with different doping concentrations were synthesized using a sol-gel method. The effects of doping on the morphology, structure, magnetic properties, and bandgap of BFO nanoparticles were investigated.

Experiment
The Bi 1-x Nd x Fe 1-y Co y O 3 ( x 0, 0.05; y 0, 0.05, 0. Bismuth nitrate pentahydrate (99.0% purity 0.01 mol) and neodymium nitrate hexahydrate (99.0% purity) are dissolved in a mixture of ethylene glycol (20 ml) and acetic acid (20 ml) in a certain proportion, stirred for 90 min at room-temperature. Since the Bi element is easily volatilized by heat, it is excessively 5%. Ferric nitrate (98.5% purity) and cobaltous nitrate (99.0% purity) are dissolved in a mixture of ethylene glycol (20 ml) and acetic acid (20 ml) in a certain proportion, stirred for 90 min at roomtemperature. After this, both solutions were mixed and put under constant stirring for 3 h at room-temperature. Afterward, the precursor solution was dried at 80°C for 24 h and then grinded into powder. The powder was raised to 600°C at 2°C/min and calcined at 600°C for 3 h to form Bi 1-x Nd x Fe 1-y Co y O 3 nanoparticles. The phase constitutions and structure of the prepared samples were characterized by an X-ray diffraction (XRD) in the range of 2θ (20 0 -60 0 ) and Raman scattering spectra. The morphology of the prepared samples was determined using scanning electron microscopy (SEM). The band-gap of the prepared samples was studied by UV-vis diffused reflectance spectra via UV-vis spectrophotometer. Room temperature magnetic properties of prepared samples were studied by vibrating sample magnetometer (VSM) in the −20 to 20 kOe field range.
The abilities of photocatalytic hydrogen production were conducted in an evacuation system (Beijing Perfect Light Technology Co., Ltd. China) and a Labsolar-III AG closed gas circulation, of which the system maintained the photo-reaction temperature at 5°C with a low-temperature thermostat bath (Poly Science, United States). The nanoparticles were irradiated by the simulated sunlight, which is A 300 W Xennon lamp (PLS-SXE-300UV, Beijing Trusttech Co. Ltd., China) with a UV-cutoff filter (providing visible light λ ≥ 420 nm). a volume of 1.5 ml of gas was hourly sampled and measured by a gas chromatography equipped with a thermal conductivity detector (TCD) and a 5 Å molecular sieve column, so as to identify and quantify the production of gas. At the same time, inert gas argon (Ar) was applied as the carrier gas. The implementation details and schematic diagram of experimental process is shown in Figure 1.

RESULTS AND DISCUSSIONS
The X-ray diffraction (XRD) profiles of different samples are showed in Figure 2. From the result of the XRD patterns, all samples have perovskite structure and no other phases such as Bi 2 Fe 4 O 9 and Bi 25 FeO 40 could be detected. The diffraction peaks of doped BFO nanoparticles had a tendency to shift toward higher diffraction angles when compared with the peaks of pure BFO. This is because the doped ionic radius is small and the unit cell volume is reduced, which shifts the diffraction peaks to higher diffraction angles. Similar phenomena are observed when other rare earth elements are incorporated (Lü et al. 2017) . As the doping ratio increases, the double diffraction peaks of the sample  (104) and (110) diffraction peaks near the diffraction angle 2θ 32°. gradually merge into a single peak in the diffraction angle (2θ) range 31°33°. These results indicate that structural changes have occurred in the after doping. Figure 3 shows the Raman spectra of different samples at 300 K. In our BFO samples, the first four strong peaks at 146.00, 175.25, 213.25, and 435.25 cm −1 manifest the A 1 -1, A 1 -2, A 1 -3, and A 1 -4 modes, and the remaining six small peaks at 283. 25, 333.50, 370.00, 472.50, 558.50, and 611.25 cm −1 are assigned to the E-2, E-3, E-4, E-5, E-6, and E-7 modes, respectively. Table 1 lists the values found in the literature and in this study for the BFO samples. The difference in the peak positions observed by different researchers may be caused by different preparation methods.
As the doping ratio increases, we can see that the peak position of the A 1 -1 mode gradually shifts towards the blue and the relative peak intensity gradually decreases. It is considered that the 472.5 cm −1 mode in the BFO Raman spectrum corresponds to the bending vibration mode of the Fe-O bond, and the 558.5 cm −1 mode corresponds to the stretching vibration mode of the Fe-O bond. The peak intensity of the 472.5 cm −1 mode is gradually increased after the incorporation of Nd 3+ and Co 2+ , indicating that the bond angle of the Fe-O bond changes after the impurity is incorporated.
The A 1 -1 mode of the Bi 0.95 Nd 0.05 Fe 0.80 Co 0.20 O 3 (BNFCO-20) sample disappeared, while a new Raman mode appeared at 677.75 cm −1 . This apparent difference indicates a structural phase change in the sample. The structural phase transition destroys the helical spin structure, which enhances the magnetic properties of the sample. The residual magnetization of the BNFCO-20 sample measured in this experiment is 0.68 emu g −1 higher than that of the undoped BFO sample. Figure 4 shows the surface morphological features of the Bi 1-x Nd x Fe 1-y Co y O 3 nanoparticles by using scanning electron microscopy (SEM). As can be seen, the prepared samples are stacked together to form a regular honeycomb shape. This did not occur in previous experiments. When Nd 3+ and Co 2+ are incorporated, the size of the Bi 1-x Nd x Fe 1-y Co y O 3 nanoparticles gradually becomes smaller, which is in good agreement with the XRD results. The decrease in grain size destroys the cycloidal spin period (62 nm) of the BFO structure, which is consistent with the trend of the magnetic properties of the samples. This agglomeration of the nanoparticles may affect their photocatalytic properties. Figure 5 shows UV-Visible spectra of the Bi 1-x Nd x Fe 1-y Co y O 3 nanoparticles. Compared with pure BFO, the absorption spectra of all doped samples shown a clear red shift. In this paper, the band gap of all samples is calculated by using Eq. 1.
As shown in Table 1, the band gap energy of BFO is greatly reduced when Nd 3+ and Co 3+ are incorporated. The band gap energy of the un-doped BFO sample was 2.063 eV, which is in  The photocatalytic activity of pure and co-doped BFO nanoparticles was studied by the degradation of 30 mg of the powder samples dissolved in 50 ml deionized water along with the addition of 0.01 g/L methylene orange (MO). The degradation of MO with the BFO nanoparticles in the dark condition for 15 h is similar to that of the blank test, which demonstrates that the absorption of RhB on the BFO is limited after the adsorption-desorption equilibrium was reached. From Figure 5B, we can see that the value of methylene orange (MO) peak is not decreased gradually but increased, which is special and interesting for normal photocatalytic degradation . The only possible reason is that the water is splitted and reduced under the illuminating of ultraviolet visible light, which lead to the increase of intensity of methylene orange. However, another possible reason must excluded is that the water might be evaporated under the illuminating of ultraviolet visible light. Consequently, the reaction of photocatalytic degradation without the catalyst under the same condition is conducted, which is prepared to quantitate the evaporative capacity of water during the whole reaction. The results are showed in the Figure 5C. Compared with these two group data as showed in Figures 5B,C we could draw a conclusion that the evaporation of water indeed take place during the illuminating of ultraviolet visible light, but the splitting of water without any external field bias and unassisted overall water splitting indeed occurs. The decrease is 4.5% for the blank group but the increase reaches the 7.1%. In order to further prove the splitting of water, the online detection of production of hydrogen were conducted under the protection of N 2 atmosphere. As showed in Figure 5D, the amount of hydrogen production of Bi 0.95 Nd 0.05 Fe 0.85 Co 0.15 O 3 reached 8.88umol/g after 1.5 h, however the other samples could not be detected the production of hydrogen. In order to easily understand the strength for production of hydrogen gas using doped BFO, a comparative result in form of table is shown in Table 3.
On the basis of above results, a tentative schematic illustration of the energy band diagram of the Nd and Co co-doped BFO photocatalyst is depicted in Figure 6, to explain how the photogenerated electrons in surface contribute to the hydrogen production. As we know the possible reason of unassisted water splitting with BFO is that the valence band edge and the conduction band edge of BFO straddle the water redox potentials. At the same time, the valence band edge and the conduction band edge should bend in favor of the separation of the excited electrons and holes. Firstly the band gap of Bi 0.95 Nd 0.05 Fe 0.85 Co 0.15 O 3 is 1.78 eV and the potential of water is 1.23 eV, which the valence band edge and the conduction band edge of BFO just straddle the water redox potentials. There are report that the optimum band gap of BFO for hydrogen production is 1.77 eV (Vishwakarma et al. 2017). A suitable band gap size is beneficial and necessary. The transfer of mobile charge carriers between the semiconductor and the contact phase produces a space-charge layer. Figure 6 illustrates the space-charge layers produced from the mobility of charge across a semiconductor solution interface (Linsebigler et al. 1995) . The space-charge layer formed is a depletion layer,    and the bands bend upward toward the surface. Because of the positive charges existence on the interface, the accumulation layer is formed and the bands would bend down in the interface. Bands with curved interfaces are stuck at either end of the water band gap like pommel horses. Because BFO is semiconductor material, when semiconductor BFO absorbs ultraviolet-light energy, electrons are excited to the conduction band (CB), while holes are present in the valence band (VB). The band curvature of the interface prevents the recombination of electrons and holes. The electrons present in the CB with protons to generate hydrogen, and the holes combine with reductive reactants to create oxidative products (Shen et al. 2015;Reddy et al. 2019). To sum up, the possible reason for this is that the conduction and valence band edges of BiFeO3 straddle the water redox potential. Consequently, it is possible to realize unassisted water splitting using BFO. Whether the bands are bent in depletion or in accumulation, the width of the space-charge region is proportional to the Debye length Fang and Shangguan, 2019) Where ε 0 is the permittivity of free space, ε r is the dielectric constant of the material, and N D is the donor density. According to Eq. 2, we can see that an increase in the dielectric constant can increase the Debye length and, therefore, the width of the spacecharge region ). The magnetic hysteresis (M-H) loops of the samples at room temperature are shown in Figure 6. It was found that doping Nd 3+ and Co 2+ had a significant effect on the ferromagnetic properties of BFO nanoparticles. It can be seen from Figure 7A that the magnetic hysteresis loop of the un-doped BFO is almost a straight line. This phenomenon proves that the prepared BFO sample does not contain other impurity phases. As shown in Figure 6, as the amount of Co 2+ doping increases, the magnetic hysteresis loop of the sample becomes more and more obvious, and the residual magnetization and saturation magnetization gradually increase. The magnetization can be increased by increasing the doping amount of Co 2+ , which may be mainly caused by the following two aspects. First, as the doping amount increases, the crystal structure gradually changes and the Pnma phase begin to appear in the parent R3c phase. The content of R3c and Pnma is 77.73 and 22.27% respectively in the sample Bi 0.95 Nd 0.05 Fe 0.85 Co 0.15 O 3 . These results came from the Rietveld refinement (not shown). These changes cause the spin coupling of the cycloid to be gradually broken and increase the magnetization. Second, as the incorporation of Co increases, the cycloid spin cycle with a grain size smaller than that of the BFO structure destroys its periodicity, resulting in an increase in magnetization (Reddy et al. 2018). It is found from Figure 7B that an exchange bias effect occurs after the incorporation of 5% of Nd 3+ , that is, the M-H ring undergoes an asymmetric movement along the H axis. The exchange bias effect is due to the coexistence of anisotropic FM and AFM domains (Wu et al. 2018). These results suggested that the magnetic properties of BFO nanoparticles can be enhanced by Co 2+ and Nd 3+ doping, which means that the photocatalyst can be separated easily by applying an external magnetic field.
Cellular architecture of Nd and Co co-doped in BFO nanoparticles, fabricated by two-solvent sol-gel method, are demonstrated as spectral photocatalytic that perform unassisted high efficiency directly produce hydrogen through splitting the water due to the optimum band-gap engineering of BFO nanoparticles was controlled by co-doping of Nd and Co. So the new material that could directly produce hydrogen through splitting the water only under the illuminating without any sacrificial agent without external bias field is urgent. Hence, Cellular architecture of Nd and Co co-doped in BFO nanoparticles, fabricated by two-solvent sol-gel method, are demonstrated as spectral photocatalytic that perform unassisted high efficiency directly produce hydrogen through splitting the water due to the optimum band-gap engineering of BFO nanoparticles was controlled by co-doping of Nd and Co.