Double Perovskite LaFe1−xNixO3 Coated with Sea Urchin-like Gold Nanoparticles Using Electrophoresis as the Photoelectrochemical Electrode to Enhance H2 Production via Surface Plasmon Resonance Effect

The surface plasmon resonance (SPR) effect and the hetero-junction structure play crucial roles in enhancing the photocatalytic performances of catalysts for the water-splitting reaction. In this study, a series of double perovskites LaFe1−xNixO3 was synthesized. LaFe1−xNixO3 particles were then decorated with sea urchin-like Au nanoparticles (NPs) with the average size of approximately 109.83 ± 8.48 nm via electrophoresis. The d-spacing became narrow and the absorption spectra occurred the redshift phenomenon more when doping increasing Ni mole concentrations for the raw LaFe1−xNixO3 samples. From XPS analysis, the Ni atoms were inserted into the lattice of the matrix, resulting in the defect of the oxygen vacancy, and NiO and Fe2O3 were formed. This hybrid structure was the ideal electrode for photoelectrochemical hydrogen production. The photonic extinction of the Au-coated LaFe1−xNixO3 was less than 2.1 eV (narrow band gap), and the particles absorbed more light in the visible region. According to the Mott–Schottky plots, all the LaFe1−xNixO3 samples were the n-type semiconductors. Moreover, all the band gaps of the Au-coated LaFe1−xNixO3 samples were higher than 1.23 eV (H+/H2). Then, the hot electrons from the Au NPs were injected via the SPR effect, the coupling effect between LaFe1−xNixO3 and Au NPs, and the more active sites from Au NPs into the conduction band of the semiconductor, improving the hydrogen efficiency. The H2 efficiency of the Au-coated LaFe1−xNixO3 measured in ethanol was approximately ten times larger than the that of Au-coated LaFe1−xNixO3 measured in 1-butanol at any testing temperature because ohmic and kinetic losses occurred in the latter solvent. Thus, the activation energies of ethanol at any testing temperature were smaller. The maximum real H2 production was up to 43,800 μmol g−1 h−1 in ethanol. The redox reactions among metal ions, OH*, and oxides were consecutively proceeded under visible light illumination.


Introduction
The development of alternative catalysts for the oxygen evolution reaction (OER) and hydrogen evolution reaction (HER) by tuning the structure and composition of catalysts based on transition metals, for example, metal oxides (or hydroxides), phosphates, chalcogenides, and metal-carbon materials [1], has attracted significant research attention. In particular, materials with the ABO 3 perovskite structure have been found to serve as catalysts with high activities, selectivities, and stabilities [2] owing to their optimum electronic and chemical structures. Especially, double perovskites have received attention in recent years, because double perovskites (AA'BB'O6), in which the A-and B-site elements are substituted by a transition metal, possess several intriguing chemical and physical properties, such as electronic structures that range from insulating to metallic, half-metallic spin-polarized electrical conductivity, and superconductivity [3]. Guan

Spin-Coating LFNO Powder on ITO (Indium-Tin-Oxide) Glass
First, ITO glass (Ruilong, Taiwan, 7Ω) with the dimensions 20 × 20 × 0.7 mm 3 was treated using ozone for 10 min. A gel solution with the following components was prepared: 0.1 g of LFNO, 500 µL of Nafion (Aldrich, Milwaukee, WI, USA, 5 wt %), and 3 mL of DI water. The gel solution was stirred for 2 h at 25 • C, after which 45 µL of the gel solution was dropped with a pipe in a spin-coater (Apisc, Taiwan, SP-M1-S) running at 6500 rpm for 15 s at 25 • C. The coated glass was then dried in the oven at 150 • C for 6 h.

Decoration of Sea Urchin-like Au NPs
Electrophoresis was conducted using an electrochemical analyzer (CHI, Model 6000E Series, Austin, TX, USA). Coated glass was used as the working electrode, and a Pt wire was connected with the counter and reference electrodes. An Au colloid was prepared by mixing 50 mL of 5 × 10 −4 M HAuCl 4 •3H 2 O (Alfa Aesar, Tewsbury, MA, USA, 99.99%) and 5 mL of 4.5 × 10 −2 M sodium citrate dihydrate (J.T Baker, Suite 100 Bridgewater, NJ, USA, 99.0%). Thereafter, an Au solution consisting of 90 µL of Au colloid, 45 mL of 10 −4 M HAuCl 4 •3H 2 O, 7.5 mL of 3.0 × 10 −3 hydroquinone (J.T Baker, Suite 100 Bridgewater, NJ, USA, 99.0%), and 60 µL of 2 wt % sodium citrate dihydrate was prepared. The solution was mixed thoroughly, and the coated glass was introduced into the Au solution. A voltage of 2 V was applied for 2 h for electrophoresis. There were some reasons to use electrophoresis method: (1) electrophoresis is direct and not complicated to load the Au NPs on the thin films, (2) electrophoresis can save time, (3) electrophoresis only needs a small volume of samples, and (4) electrophoresis has a reasonable cost. In this study, the larger surface area, which induces more active sites, and the star-like morphology of Au NPs can enhance the SPR effect.

A Analysis of Properties
The morphology was observed via scanning electron microscopy (SEM, ZEISS-AURIGA AFE-SEM, Jena, Germany); energy dispersive X-ray spectroscopy (EDS) was conducted in conjunction with SEM. The crystal planes were observed via high-resolution transmission electron microscopy (HR-TEM, JEOL JEM-3010, Tokyo, Japan). X-ray photoelectron spectroscopy (XPS, Thermo K-Alpha, X-ray source: Al K-α, Rochester, NY, USA) was conducted to obtain the binding state of each element. The valence band energies of the compounds were obtained from ultraviolet photoelectron spectroscopy (UPS) plots (Thermo VG-Scientific, Sigma Probe, UPS source: He(I), Thermo VG-Scientific). The photon energy of the incident light was 21.1 eV.
Ultraviolet-visible (UV-vis) spectroscopy (PerkinElmer, Lambda 950, Waltham, MASS, USA) was conducted to obtain the absorbance wavelengths and band gaps from 400 to 800 nm. The crystal phases were analyzed using an X-ray analytical system (Bruker AXS Gmbh, Karlsruhe, Germany). Electrochemical analysis was conducted using several methods, such as Mott-Schottky plots, cyclic voltammetry (C-V) curves, and linear sweep voltammetry (LSV) curves. The C-V and LSV curves were measured in 0.1 M KOH solution. LSV was conducted under illumination air mass (AM) 1.5G (1000 W/m 2 ). Furthermore, the H 2 evolutions in ethanol (J.T Baker, Suite 100 Bridgewater, NJ, USA, 99.9%) and 1-butanol (Duksan, Seoul, South Korea,99.9%) were determined from the corresponding I-V curves, with the applied voltage ranging from 0 to 1.5 V; testing was also conducted at different temperatures under AM 1.5G. Figure 1 shows the XRD analysis of the raw LFNO. From Figure 1 the diffraction peaks can be matched to the orthorhombic perovskite structure with the Pnma space group (JCPDS no. 37-1493) [22]. The orthorhombic structure peaks are located at 24 was conducted to obtain the binding state of each element. The valence band energies of the compounds were obtained from ultraviolet photoelectron spectroscopy (UPS) plots (Thermo VG-Scientific, Sigma Probe, UPS source: He(Ⅰ), Thermo VG-Scientific). The photon energy of the incident light was 21.1 eV.

Phase and Morphology Analyses
Ultraviolet-visible (UV-vis) spectroscopy (PerkinElmer, Lambda 950, Waltham, MASS, USA) was conducted to obtain the absorbance wavelengths and band gaps from 400 to 800 nm. The crystal phases were analyzed using an X-ray analytical system (Bruker AXS Gmbh, Karlsruhe, Germany). Electrochemical analysis was conducted using several methods, such asMott-Schottky plots, cyclic voltammetry (C-V) curves, and linear sweep voltammetry (LSV) curves.. The C-V and LSV curves were measured in 0.1 M KOH solution. LSV was conducted under illumination air mass (AM) 1.5G (1000 W/m 2 ). Furthermore, the H2 evolutions in ethanol (J.T Baker, Suite 100 Bridgewater, NJ, USA, 99.9%) and 1-butanol (Duksan, Seoul, South Korea,99.9%) were determined from the corresponding I-V curves, with the applied voltage ranging from 0 to 1.5 V; testing was also conducted at different temperatures under AM 1.5G. Figure 1 shows the XRD analysis of the raw LFNO. From Figure 1 the diffraction peaks can be matched to the orthorhombic perovskite structure with the Pnma space group (JCPDS no. 37-1493) [22]. The orthorhombic structure peaks are located at 24 [23] phases are observed, and La2O3 and La(OH)3 exhibit hexagonal structures [24]. The peaks corresponding to La(OH)3 are 39.36° and 43.46°, while the peaks of La2O3 are 62.44° and 64.02°. In general, a higher intensity of XRD peaks indicates better crystallinity; thus, the lattice arrangement was highly ordered for the samples regardless of the Ni concentration. The particle size increases as the Ni concentration increases due to the difference in the ionic radius of Fe 3+ and Ni 2+ [25]; the particle sizes shown in Figure 2 are listed in Table   Figure 1. X-ray patterns of the raw LFNO samples with different Ni concentrations.

Phase and Morphology Analyses
The particle size increases as the Ni concentration increases due to the difference in the ionic radius of Fe 3+ and Ni 2+ [25]; the particle sizes shown in Figure 2 are listed in Table S1. The specimen where x = 0.07 Ni has the largest particle size, approximately 7.82 ± 0.80 µm, because La 2 O 3 and La(OH) 3 appear in the matrix and enlarge the coverage area. Moreover, the morphologies of the sea urchin-like Au NPs decorated on the LFNO thin film are shown in Figure 3. The flower-like Au NPs are successfully adhered to the films for all LFNO samples with different Ni concentrations. The average size of the Au NPs is approximately 109.83 ± 8.48 nm. Nanomaterials 2022, 12, 622 5 of 22 S1. The specimen where x = 0.07 Ni has the largest particle size, approximately 7.82 ± 0.80 μm, because La2O3 and La(OH)3 appear in the matrix and enlarge the coverage area. Moreover, the morphologies of the sea urchin-like Au NPs decorated on the LFNO thin film are shown in Figure 3. The flower-like Au NPs are successfully adhered to the films for all LFNO samples with different Ni concentrations. The average size of the Au NPs is approximately 109.83 ± 8.48 nm.    Figure 4 and Figure 5 show the selected area diffractions (SAD) results and HR-TEM images of the raw LFNO powders, respectively. The SAD results, shown in Figure 4, are consistent with the XRD patterns. These raw LFNO powders exhibit high degrees of crys-   Figure 4, are consistent with the XRD patterns. These raw LFNO powders exhibit high degrees of crystallinity, and the orthorhombic perovskite structures are dominant. As shown in the HR-TEM images in Figure 5, the d-spacing decreases when the sample is doped with a higher Ni concentration, which occurs because Ni atoms and NiO NPs induce lattice shrinkage [26]. Thus, the hydrogen molecules (H + ) cannot be accommodated on the denser interlayer spacing, and hydrogen gas (H 2 ) can be much more rapidly released. Then, the d-spacing is decreased with high Ni concentrations, and H + or H 2 is not incorporated into the lattices of the raw LFNO samples. Therefore, the H 2 generation efficiency is improved.

Crystalline Analysis
Figures 4 and 5 show the selected area diffractions (SAD) results and HR-TEM images of the raw LFNO powders, respectively. The SAD results, shown in Figure 4, are consistent with the XRD patterns. These raw LFNO powders exhibit high degrees of crystallinity, and the orthorhombic perovskite structures are dominant. As shown in the HR-TEM images in Figure 5, the d-spacing decreases when the sample is doped with a higher Ni concentration, which occurs because Ni atoms and NiO NPs induce lattice shrinkage [26]. Thus, the hydrogen molecules (H + ) cannot be accommodated on the denser interlayer spacing, and hydrogen gas (H2) can be much more rapidly released. Then, the d-spacing is decreased with high Ni concentrations, and H + or H2 is not incorporated into the lattices of the raw LFNO samples. Therefore, the H2 generation efficiency is improved.

Component Analysis
The EDX analysis results of the Au-coated LFNO are shown in Figure 6, wherein the elemental quantitative analyses are presented in sub-figures (c), (f), (i), (l), and (o). These results confirm that sea urchin-like Au NPs are adhered to each LFNO thin film. Figure 7 shows the XPS patterns of all samples. Figure 7a,d,g,j,m are the O 1s state, and there are two peaks to show up for all O 1s state curves. One peak with intensity up to approximately 530.20 eV is corresponding to the oxygen vacancy (Ov). Protons from the moist environment are exchanged with doped holes and oxygen vacancies formed by Bsite cation doping [27], thus generating the more oxygen vacancies. Then, Ni is doped into the lattice of LaFeO3 also to induce the Ov enrichment. Another peak is the metal-oxygen (M-O) bond. The M-O bonds of Figure 7a,d,g,j,m indicate Fe-O bonds [28] and Ni-O bonds [29]. Thus, Fe2O3 and NiO are demonstrated to precipitate into the matrix. No other oxygen states are indicated. Based on Figure 7b,e,h,k,n, La 3d and Ni 2p orbits are overlapped. The peaks corresponding to the normal network structure of La 3+ 3d5/2 and 3d3/2 are at approximately 838.3 and 847.7 eV, respectively. The La state in these results is consistent with the results reported in previous studies [30]. La2O3 is deposited, La-O bonds are decomposed, and oxygen atoms arrive at the interface to induce oxidation. Some peaks of La state disappear due to NiO formation. The La 3d3/2 orbit overlaps with the Ni 2p3/2 orbit. Only one peak of Ni 2+ exists at approximately 870.3 eV [31], and no traces of other Ni states are observed. As seen in Figure 7c,f,i,l,o, one peak of Fe 2p3/2 exists at approximately 711.3 eV, and another peak of Fe 2p1/2 is at approximately 724.9 eV [32]. Fe is maintained at the +3 state, and no other valence sates are observed.

Component Analysis
The EDX analysis results of the Au-coated LFNO are shown in Figure 6, wherein the elemental quantitative analyses are presented in sub-figures (c), (f), (i), (l), and (o). These results confirm that sea urchin-like Au NPs are adhered to each LFNO thin film. Figure 7 shows the XPS patterns of all samples. Figure 7a,d,g,j,m are the O 1s state, and there are two peaks to show up for all O 1s state curves. One peak with intensity up to approximately 530.20 eV is corresponding to the oxygen vacancy (O v ). Protons from the moist environment are exchanged with doped holes and oxygen vacancies formed by B-site cation doping [27], thus generating the more oxygen vacancies. Then, Ni is doped into the lattice of LaFeO 3 also to induce the O v enrichment. Another peak is the metal-oxygen (M-O) bond. The M-O bonds of Figure 7a,d,g,j,m indicate Fe-O bonds [28] and Ni-O bonds [29]. Thus, Fe 2 O 3 and NiO are demonstrated to precipitate into the matrix. No other oxygen states are indicated. Based on Figure 7b,e,h,k,n, La 3d and Ni 2p orbits are overlapped. The peaks corresponding to the normal network structure of La 3+ 3d 5/2 and 3d 3/2 are at approximately 838.3 and 847.7 eV, respectively. The La state in these results is consistent with the results reported in previous studies [30]. La 2 O 3 is deposited, La-O bonds are decomposed, and oxygen atoms arrive at the interface to induce oxidation. Some peaks of La state disappear due to NiO formation. The La 3d 3/2 orbit overlaps with the Ni 2p 3/2 orbit. Only one peak of Ni 2+ exists at approximately 870.3 eV [31], and no traces of other Ni states are observed. As seen in Figure 7c,f,i,l,o, one peak of Fe 2p 3/2 exists at approximately 711.3 eV, and another peak of Fe 2p 1/2 is at approximately 724.9 eV [32]. Fe is maintained at the +3 state, and no other valence sates are observed.

Absorbance Spectra of LFNO
The UV-vis spectra of raw and Au-coated LFNO are shown in Figure 8a,c, respectively. The energy gap (Eg) and photonic extinction (PE) are determined from the UV-vis spectra, and these two energy levels are calculated with Equation (1) [33]. The diagram of (αhυ) n vs. energy is shown in Figure 8b,d. In Equation (1), α is the absorption coefficient, hν is the photo energy, B is a constant, and n is defined the direct semiconductor (n = 2) or the indirect semiconductor (n = 1/2) [34]. Thereafter, a Tauc plot is used to measure Eg and PE. A value of n = 2 indicates that LFNO is a direct band gap material. Table S2 shows the parameters of the energy and absorbance wavelengths, obtained from Equation (1) and Figure 8b,d for different Ni concentrations. The following observations are noted: (1) Eg decreases as the Ni concentration increases because the NiO can be the active center of the infrared emission [35]; (2) PE tends to move to the visible region after the materials are decorated with sea urchin-like Au NPs; and (3) the maximum absorbance wavelength is augmented. In particular, the PE of Au-coated LaFe0.93Ni0.07O3 is only 1.99 eV, which can be attributed to the following reasons: (1) metal nanostructures scatter light strongly on the sharp tips of sea urchin-like Au NPs; therefore, the SPR are enhanced by several orders of magnitude compared with the incident light and tunes into the near-infrared region [36]; (2) a larger particle size induces the rough surface that can disperse the incident light; and (3) the intrinsic property of the LFNO sample contributes to the light absorbance. Moreover, after coating Au NPs, the PEs are also higher than 1.23 eV (H + /H2), meaning the photo-electro-chemical reaction is more toward HER. As seen in Figure 8e, the maximum absorbance wavelength of the prepared Au colloid solution is 535.2 nm. Therefore, the SPR band is at −2.32 eV (vs. vacuum).

Absorbance Spectra of LFNO
The UV-vis spectra of raw and Au-coated LFNO are shown in Figure 8a,c, respectively. The energy gap (E g ) and photonic extinction (PE) are determined from the UV-vis spectra, and these two energy levels are calculated with Equation (1) [33]. The diagram of (αhυ) n vs. energy is shown in Figure 8b,d. In Equation (1), α is the absorption coefficient, hν is the photo energy, B is a constant, and n is defined the direct semiconductor (n = 2) or the indirect semiconductor (n = 1/2) [34]. Thereafter, a Tauc plot is used to measure E g and PE. A value of n = 2 indicates that LFNO is a direct band gap material. Table S2 shows the parameters of the energy and absorbance wavelengths, obtained from Equation (1) and Figure 8b,d for different Ni concentrations. The following observations are noted: (1) E g decreases as the Ni concentration increases because the NiO can be the active center of the infrared emission [35]; (2) PE tends to move to the visible region after the materials are decorated with sea urchin-like Au NPs; and (3) the maximum absorbance wavelength is augmented. In particular, the PE of Au-coated LaFe 0.93 Ni 0.07 O 3 is only 1.99 eV, which can be attributed to the following reasons: (1) metal nanostructures scatter light strongly on the sharp tips of sea urchin-like Au NPs; therefore, the SPR are enhanced by several orders of magnitude compared with the incident light and tunes into the near-infrared region [36]; (2) a larger particle size induces the rough surface that can disperse the incident light; and (3) the intrinsic property of the LFNO sample contributes to the light absorbance. Moreover, after coating Au NPs, the PEs are also higher than 1.23 eV (H + /H 2 ), meaning the photo-electro-chemical reaction is more toward HER. As seen in Figure 8e, the maximum absorbance wavelength of the prepared Au colloid solution is 535.2 nm. Therefore, the SPR band is at −2.32 eV (vs. vacuum).

Photo-Electro-Chemical Characteristics of Raw LFNO Powders
The valence band level, E VB , is calculated using Equation (2) [37], where hν is 21.1 eV, E F B is the electron binding energy, and Φ SA is the work function. Figure 9a-c show the UPS results for all LaFe 1−x Ni x O 3 for each Ni concentration. E F B and Φ SA are measured from the Tauc plot shown in Figure 9b,c, respectively. Finally, the E VB is estimated using Equation (2) and Figure 9b,c. In addition, the conduction band energy, E CB , is evaluated.

Photo-Electro-Chemical Characteristics of Raw LFNO Powders
The valence band level, EVB, is calculated using Equation (2) [37], where hν is 21.1 eV, E F B is the electron binding energy, and ΦSA is the work function. Figure 9a-c show the UPS     Figure 9d shows the Mott-Schottky plots of the samples, where the slope crossover with the potential axis indicates the Fermi level. The carrier density is calculated using Equation (3) [38], where ξ is the dielectric constant of the film, ξ 0 is the vacuum permittivity, e is the electric charge potential, T is the temperature in Kelvin, k is the Boltzmann constant, V is the applied voltage, V FB is the flat-band potential, N D is the donor density, and C SC is the space charge capacitance. The carrier density increases as the Ni concentration increases, and greater carrier density means that electric and thermal conductivity are promoted. All raw LFNO powders are n-type semiconductors based on the positive slopes of Figure 9d. Figure 9e shows the C-V curves, and Figure 9f presents the LSV results. As the Ni concentration increases, the detection range of the redox reaction becomes broader, and LaFe 0.93 Ni 0.07 O 3 demonstrates the broadest detection range. Additionally, the oxidation reaction is suppressed as the Ni concentration increases. For LSV, LaFe 0.93 Ni 0.07 O 3 exhibits the largest HER potential. This result implies that the increased particle size with a high Ni concentration can contribute to a large number of active sites and macro-pores for oxygen transportation [39]. Table S3 shows the values of anodic (E p a ), cathodic (E p c ), and HER derived from the C-V curves and LSV analysis.  [40], where η is the applied bias compensated solar-to-hydrogen efficiency, Ps is the energy of light irradiation, I op is the photo-current density, V op is the applied voltage, and the water electrolysis voltage is 1.23 V. From Figure 10 and Equation (4), the following phenomena are observed. First, the slopes and current densities of the samples tested using ethanol are larger than those of the samples tested using 1-butanol. Second, the H 2 efficiencies of all specimens measured in ethanol are approximately 10 times larger than those measured in 1-butanol at any testing temperature. The reason for this is that the resistance of 1-butanol is higher than the resistance of ethanol, as 1-butanol has four carbon atoms in one molecule. The higher resistance results in a loss of conductivity (ohmic loss) and proton activity (kinetic loss) [41]. Equation (5) [42] is used to calculate the real production, where η is the H 2 generation efficiency, P s is the light power (1000 W/m 2 ), A is the area (10 −4 m 2 for all samples), 237.1 × 10 3 is the energy required for forming one mol of H 2 , and 2.0158 is the molecular weight of one mol of H 2 . The parameters listed in Tables S4 and S5 represent the data obtained from Figure 10 and Equation (4). Tijare et al. found that LaFeO 3 , owing to the high visible-light activity, demonstrated the maximum H 2 generation (3315 µmol g −1 h −1 ) [43]. Moreover, the H 2 evolution rate of CdS-Au/MoS 2 was up to 7010 µmol g −1 h −1 [19]. In this study, the real H 2 production of all samples is a lot larger than 3315 µmol g −1 h −1 , regardless of the temperature or solvent. However, in this study, the greatest amount of real H 2 production is approximately 43,800 µmol g −1 h −1 in ethanol. Some representative perovskites (PVKs) for photocatalytic water splitting are listed in Table 1 [44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59].

H 2 Evolution and Activation Energy of the Au-Coated LFNO
The activation energies, E a , are obtained using the Richardson plot and Equation (6) [60], where I 0 is the reverse saturation current in the I-V curves, T is the temperature in Kelvin, k is the Boltzmann constant, Φ bo is the zero bias effective barrier height, q is the electronic charge, A is the working electrode area, and A* is the Richardson constant. Figure 10i,j show the Richardson plots for samples tested in ethanol and 1-butanol, respectively. From these results, the E a of all Au-coated LFNO specimens measured in ethanol are smaller than the E a measured in 1-butanol. This result implies that electrons are rapidly and easily transferred to H + in ethanol, thereby generating more H 2 . For the definition of the kinetic loss, the sequential proton loss electron transfer mechanism includes two steps: (1) the deprotonation of the antioxidant, and (2) the transfer of an electron to the free radical. The compounds that can react these continuous processes result in the radical adduct formation, hydrogen transfer, and single electron transfer [61]. For 1-butanol, the low activation energy indicates that the pKa is also low, resulting in the slow proton exchange reaction. The values of the activation energies of each Au-coated specimen obtained from Equation (6) and Figure 10i,j are presented in Table S6.
Real production:  Figure 11a shows the energy level diagram of the raw LFNO samples with different Ni concentrations. The values of EVB, Eg, ECB, EFB, and water splitting for the energy level vs. vacuum are shown. Table S7 indicates the parameters of the energy level diagram of raw LFNO (vs. vacuum) from Figure 11a. All the band gaps of the raw LFNO samples are greater than 1.23 eV (H + /H2). Thus, these materials are suitable to act as photocathodes. Figure 11b shows the main reactions of raw LFNO samples under AM 1.5G, and the Fenton-like reaction happens. When the light is luminated on the samples in the solvent, the photogenerated charges, the electrons, and OH • radicals appear. Then, the electrons transfer to Fe 3+ , and OH • radicals move to Fe 2+ . The redox reactions are repeated and recycled between Fe 3+ and Fe 2+ . The reduction reaction of Fe 3+ to Fe 2+ and the oxidization reaction of Fe 2+ to Fe 3+ are described by Equation (7) and Equation (8) [62]. The prime amounts of OH • radicals are produced using Fe 3+ under AM 1.5G, and OH • radicals also simultaneously react with NiO. Equation (9) [63] indicates the oxygen gas generated under the visible light. Ni-Fe compounds are beneficial for the water dissociation into H + and OHions [64].   Figure 11a shows the energy level diagram of the raw LFNO samples with different Ni concentrations. The values of E VB , E g , E CB , E FB , and water splitting for the energy level vs. vacuum are shown. Table S7 indicates the parameters of the energy level diagram of raw LFNO (vs. vacuum) from Figure 11a. All the band gaps of the raw LFNO samples are greater than 1.23 eV (H + /H 2 ). Thus, these materials are suitable to act as photocathodes. Figure 11b shows the main reactions of raw LFNO samples under AM 1.5G, and the Fenton-like reaction happens. When the light is luminated on the samples in the solvent, the photogenerated charges, the electrons, and OH • radicals appear. Then, the electrons transfer to Fe 3+ , and OH • radicals move to Fe 2+ . The redox reactions are repeated and recycled between Fe 3+ and Fe 2+ . The reduction reaction of Fe 3+ to Fe 2+ and the oxidization reaction of Fe 2+ to Fe 3+ are described by Equation (7) and Equation (8) [62]. The prime amounts of OH • radicals are produced using Fe 3+ under AM 1.5G, and OH • radicals also simultaneously react with NiO. Equation (9)

Energy Level Diagram and Scheme of H 2 Generation
O2 + H + + e -+ hυ  1/2 H2O2 + 1/2 O2 Furthermore, Figure 11c shows the H2 evolution scheme of the Au-coated LaFe0.93Ni0.07O3 sample under AM 1.5G. Subsequently, Au supports the more optimized condition, which adds the extra active sites and allows the electrons to transfer rapidly for the hydrogen adsorption and the hydrogen gas formation. The hot electrons overcome the barrier at the interface between Au NPs and LaFe0.93Ni0.97O3 owing to SPR. The SPR energy is −2.32 eV. Therefore, more electrons are tunneled to the conduction band of LaFe0.93Ni0.97O3, leading to H2 generation. For Au-loaded semiconductors, the abilities of harvesting and absorbing light are promoted [65]. In this study, connecting the optical and electrocatalytic reactive interface between the LFNO semiconductor and the Au NP as a PEC cell can drastically enhance the H2 production efficiency in ethanol. This hybrid structure is satisfied with these conditions, the narrow band gaps, the rough surface, the contributions of SPR of Au NPs, and the coupling effect between LFNO-Au. Importantly, the energy in loss is decreaseddue to the plasmon-induced carriers of Au NPs [66], and there are the added active sites which arise from Au NPs. Thus, the photogenerated hot electrons provide high efficiency to generate the H2 reduction reaction.

Conclusions
In this work, double perovskite LaFe1−xNixO3 decorated with urchin-like Au NPs were synthesized. All raw LaFe1−xNixO3 powders exhibited high crystallinities, as indicated by the XRD and SAD patterns. In addition, La2O3 and La(OH)3 were found to coexist. With the increasing Ni concentration, the following advantages were noted: (1) the generation of more active sites, (2) a lower band gap, (3) a higher carrier density, (4) a wider range of redox reaction, (5) HER potential toward a more positive value, and (6) smaller d-spacing. After coating with Au NPs, the PEs of all specimens reduced below 2.10 eV, and the absorbance wavelength occurred in the redshift due to the light scattering using Au NPs, the rough surface, and the intrinsic property of the samples. More light in the visible region was absorbed and harvested due to the hybrid structure. Moreover, larger specimen particle sizes induced more active sites. The average size of the sea urchin-like Au NPs was approximately 109.83 ± 8.48 nm. All the band gaps of the raw LFNO samples were higher than 1.23 eV (H + /H2). The H2 efficiency was 10 times higher in ethanol than that in 1-butanol because ohmic and kinetic losses resulted in the electrons or protons transferring on the surface of the electrode slowly in 1-butanol. That ethanol had lower activation energy resulted in the charges and protons transferring rapidly. The redox reactions were recycled among Fe 3+ , Fe 2+ , NiO, and hydroxyl. More photogenerated carriers were tunneled from Au NPs to the ECB of the raw LFNO samples due to the SPR, the coupling effect, and the additionally active sites from Au NPs. This heterojunction structure was proven to develop the HER process.  Furthermore, Figure 11c shows the H 2 evolution scheme of the Au-coated LaFe 0.93 Ni 0.07 O 3 sample under AM 1.5G. Subsequently, Au supports the more optimized condition, which adds the extra active sites and allows the electrons to transfer rapidly for the hydrogen adsorption and the hydrogen gas formation. The hot electrons overcome the barrier at the interface between Au NPs and LaFe 0.93 Ni 0.97 O 3 owing to SPR. The SPR energy is −2.32 eV. Therefore, more electrons are tunneled to the conduction band of LaFe 0.93 Ni 0.97 O 3 , leading to H 2 generation. For Au-loaded semiconductors, the abilities of harvesting and absorbing light are promoted [65]. In this study, connecting the optical and electrocatalytic reactive interface between the LFNO semiconductor and the Au NP as a PEC cell can drastically enhance the H 2 production efficiency in ethanol. This hybrid structure is satisfied with these conditions, the narrow band gaps, the rough surface, the contributions of SPR of Au NPs, and the coupling effect between LFNO-Au. Importantly, the energy in loss is decreaseddue to the plasmon-induced carriers of Au NPs [66], and there are the added active sites which arise from Au NPs. Thus, the photogenerated hot electrons provide high efficiency to generate the H 2 reduction reaction.

Conclusions
In this work, double perovskite LaFe 1−x Ni x O 3 decorated with urchin-like Au NPs were synthesized. All raw LaFe 1−x Ni x O 3 powders exhibited high crystallinities, as indicated by the XRD and SAD patterns. In addition, La 2 O 3 and La(OH) 3 were found to co-exist. With the increasing Ni concentration, the following advantages were noted: (1) the generation of more active sites, (2) a lower band gap, (3) a higher carrier density, (4) a wider range of redox reaction, (5) HER potential toward a more positive value, and (6) smaller d-spacing. After coating with Au NPs, the PEs of all specimens reduced below 2.10 eV, and the absorbance wavelength occurred in the redshift due to the light scattering using Au NPs, the rough surface, and the intrinsic property of the samples. More light in the visible region was absorbed and harvested due to the hybrid structure. Moreover, larger specimen particle sizes induced more active sites. The average size of the sea urchin-like Au NPs was approximately 109.83 ± 8.48 nm. All the band gaps of the raw LFNO samples were higher than 1.23 eV (H + /H 2 ). The H 2 efficiency was 10 times higher in ethanol than that in 1-butanol because ohmic and kinetic losses resulted in the electrons or protons transferring on the surface of the electrode slowly in 1-butanol. That ethanol had lower activation energy resulted in the charges and protons transferring rapidly. The redox reactions were recycled among Fe 3+ , Fe 2+ , NiO, and hydroxyl. More photogenerated carriers were tunneled from Au NPs to the E CB of the raw LFNO samples due to the SPR, the coupling effect, and the additionally active sites from Au NPs. This heterojunction structure was proven to develop the HER process.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/nano12040622/s1, Table S1: Particle sizes of the raw LFNO samples with different Ni concentrations, Table S2: UV-vis properties of the raw and Au-coated LFNO samples with different Ni concentrations,   Institutional Review Board Statement: Ethical review and approval were waived for this study due to not involving human, or any animal experiment.

Informed Consent Statement: Not applicable.
Data Availability Statement: Image files must not be manipulated or adjusted in any way that could lead to misinterpretation of the information provided by the original image. Irregular manipulation includes (1) introduction, enhancement, moving, or removing features from the original image, (2) grouping of images that should obviously be presented separately (e.g., from different parts of the same gel, or from different gels), or (3) modifying the contrast, brightness or color balance to obscure, eliminate or enhance some information.