1 Introduction

Photoelectrochemical catalysis technology can be a promising strategy to maintain carbon balance and continuously provide clear energy for human survival, which can convert water, carbon dioxide, and inexhaustible solar energy to the substitution of fossil energy [1]. The photoelectrochemical water splitting (PEC-WS) is considered to be an ideal way to relieve the energy crisis and environmental pollution by the water oxidation to produce hydrogen with photoelectrodes [2]. However, the high photogenerated carrier recombination rate and low oxygen evolution reaction kinetics seriously limit the photoelectric conversion efficiency [3]. At present, various junctions show excellent inhibition of photogenerated carrier recombination [4, 5]. The traditional photocatalyst heterojunctions mainly include the following: type I (cross-bandgap type), type II (staggered bandgap type), and type III (broken bandgap type) [6]. For example, the Schottky junction formed at the interface of semiconductors generally forms a Schottky barrier at the interface to repel most carriers and selectively conduct a few carriers [4]. Among them, oxide semiconductor photocatalysts with heterojunctions have attracted wide attention due to their excellent light trapping ability, efficient charge separation, fast reaction kinetics, and long-term cyclic stability [7]. So far, numerous researches have been devoted to the design and construction of different heterojunctions (type II, Z-Scheme, and S-scheme) to explore the internal mechanism of heterojunction formation and the efficient photogenerated carrier separation [8]. Pandikumaret et al. designed a Co-CuBi2O4/TiO2 p-n heterojunction-modified electrocathode by the hydrothermal method. The Co doping can lead to the distortion of crystal lattice for CuBi2O4, which enhances the electronic conductivity, reduces the charge recombination, and facilitates the formation of the p–n junction [9]. Hematite (α-Fe2O3), as a kind of oxide semiconductors, has a narrow band gap, high absorption efficiency of visible light, suitable valence band level position, and theoretical photocurrent density of 12.6 mA/cm2, which can be an ideal photoanode material for PEC-WS study [10]. However, the photoelectric conversion efficiency is still at a low level due to more serious carrier phase recombination and higher surface state density [11]. The heterojunction is constructed to inhibit the photogenerated carrier recombination and further increases the PEC-WS performance of α-Fe2O3 photoanodes [12].

From the perspective of the PEC reaction process, the PEC conversion efficiency of photoanode materials mainly depends on their light absorption capacity and carrier dynamics including photogenerated carrier separation and interfacial transfer to the active sites for water oxidation reaction [13,14,15]. Porous structures provide a large semiconductor/electrolyte interface area and a short diffusion distance of photogenerated carriers, thereby improving photogenerated carrier separation and transfer efficiency [16, 17]. Metal–organic frameworks (MOFs) have highly controllable pore structures, surface functionalized groups, abundant active sites, and excellent electron transport performance [18, 19]. The introduction of MOFs on the surface of α-Fe2O3 photoanode can effectively enhance the active reaction area and regulate the electronic structures. However, the traditional MOF-based photoelectrode materials exhibit some challenges in terms of stability and PEC-WS efficiency, which are mainly manifested as the precipitation or decomposition of MOF particles, the limitation of light absorption efficiency, and the obstruction of electron transport [20, 21]. Many strategies have focused on modifying MOF materials to improve photoelectric conversion efficiency and stability of α-Fe2O3 photoanode, including element doping, surface plasma effect, and heterojunction [22,23,24]. For example, α-Fe2O3 photoanodes modified with MOF films doped with iron-group elements show excellent PEC-WS performance, which can be primarily attributed to more abundant active sites, the lower photogenerated carrier recombination, and the enhanced charge separation and transfer efficiency [11]. In addition, exciton behavior in conductive MOFs can be regulated through the substitution of organic functional groups and frame-guest interactions, enabling efficient energy transfer from the MOF skeleton to the molecular receptor to improve the photoelectric performance [25, 26]. The introduction of MOFs and element doping have proven their effectiveness in improving the PEC-WS performance of α-Fe2O3 photoanodes. However, it needs to be deeply investigated the effect of MOF and elemental codoping modification on the enhanced PEC-WS performance.

In this work, MOF-5 is first introduced to improve the photogenerated carrier separation and transfer of α-Fe2O3 photoanode along with Ni/Ru codoping. The α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode exhibits the optimal photocurrent density of 2.6 mA/cm2 at 1.23 VRHE, resulting in a 168% increment compared with that of the pure α-Fe2O3 photoanode. The introduction of NH2:MOF-5 and Ni/Ru codoping also increases the PEC stability of the α-Fe2O3 photoanode. This mainly can be attributed to that the more abundant active sites, the more stable heterogeneity between α-Fe2O3 photoanode and MOF, and the higher conductivity of the introduced MOF-5 and Ni/Ru codoping can effectively increase the photogenerated carrier separation and transfer. In addition, the introduction of –NH2, as an electron-donor group, can inhibit charge recombination to further improve the photogenerated carrier transport and the PEC-WS performance.

2 Experimental

2.1 Preparation of α-Fe2O3 photoanode

The FTO conductive glass was sonicated with acetone, ethanol, and deionized water for 15 min. The FTO conductive glass was fixed on the slide with high-temperature tape and placed obliquely into the Teflon reactor, where the FTO conductive side was facing down. Dissolve 4.5 mmol FeCl3·6H2O and 4.5 mmol urea into 50 mL distilled water, and pour into the above Teflon reactor. Then the above fixed solution was heated at 100 °C for 12 h to obtain a yellowish film on the FTO substrate, which was named the FTO/FeOOH photoanode. The FTO/FeOOH photoanode was immersed in 0.1 M SnCl4 ethanol solution for 10 min, then soaked and rinsed with the distilled water, and finally annealed at 500 °C and 680 °C for 2 h and 15 min respectively to obtain the FTO/Sn@α-Fe2O3 photoanode [11, 27]. The FTO/Sn@α-Fe2O3 photoanode is abbreviated as the α-Fe2O3 photoanode for a clear description.

2.2 Preparation of α-Fe2O3 photoanodes with MOF-5 catalyst and elemental doping

A total of 0.33 mol terephthalic acid and 0.84 mol zinc nitrate were dissolved in 50 mL of N,N-dimethylformamide and stirred evenly. The solution was poured into the Teflon reactor filled with the above α-Fe2O3 photoanode, severally heated at 140 °C for 8/12/16 h, washed three times with N,N-dimethylformamide, and finally dried to prepare a α-Fe2O3/MOF-5-8h, α-Fe2O3/MOF-5-12h, and α-Fe2O3/MOF-5-16h photoanodes [28]. The α-Fe2O3/MOF-5-12h photoanode with the optimal PEC-WS performance is renamed as α-Fe2O3/MOF-5 photoanode in the following for further modification. α-Fe2O3/MOF-5(Fe), α-Fe2O3/MOF-5(Ni), and α-Fe2O3/MOF-5(Co) photoanodes were prepared by simultaneously adding 0.84 mol of iron nitrate, nickel nitrate, and cobalt nitrate to the mixed solution of α-Fe2O3/MOF-5 photoanode, respectively. α-Fe2O3/MOF-5@Ru-4000, α-Fe2O3/MOF-5@Ru-5000, and α-Fe2O3/MOF-5@Ru-6000 photoanodes were prepared by adding 4 mL, 5 mL, and 6 mL of 0.24 mmol/L Ru solution to the above mixed solution of α-Fe2O3/MOF-5 photoanode, respectively. The α-Fe2O3/MOF-5@Ru-5000 photoanode with the optimal photocurrent density is renamed as α-Fe2O3/MOF-5@Ru photoanode in the following for a clear description.

The synthetic method of α-Fe2O3/NH2:MOF-5 photoanode was the same as the above, except that terephthalic acid was replaced with 2-aminoterephthalic acid. α-Fe2O3/NH2:MOF-5@Ru was prepared by adding 5 mL of 0.24 mmol/L Ru solution to the mixed solution of α-Fe2O3/NH2:MOF-5 photoanode. The synthetic method of α-Fe2O3/NH2:MOF-5(Ni) photoanode is the same as α-Fe2O3/MOF-5(Ni), except that terephthalic acid was replaced with 2-aminoterephthalic acid. α-Fe2O3/NH2:MOF-5(Ni)@Ru-5000 and α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes were prepared by adding 5 mL of 0.24 mmol/L Ru solution to the mixed solution of α-Fe2O3/NH2:MOF-5(Ni) photoanode.

2.3 Structural characterization

The morphology and microstructure of α-Fe2O3 photoanodes with MOF-5 catalyst and elemental doping were characterized by field emission scanning electron microscopy (SEM, Hitachi S4800) and transmission electron microscopy (TEM, FEI Talos F200X). At the same time, the elemental composition of α-Fe2O3 photoanodes was analyzed using energy spectroscopy X-ray spectroscopy (EDS). X-ray diffraction (XRD) data and UV–vis absorption spectra were recorded with X’Pert Pro MPD with Cu Kα radiation and UV 3600 Shimadzu, respectively. X-ray photoelectron spectroscopy (XPS) was performed on a Thermo ESCALAB 250Xi spectrometer to investigate the surface elemental composition and chemical state of α-Fe2O3 photoanodes.

2.4 Photoelectrochemical measurements

All photoelectrochemical measurements were performed on an electrochemical workstation (CHI 660E) equipped with a three-electrode test system. The electrolyte used was 1.0 M NaOH aqueous solution, Ag/AgCl was the reference electrode, Pt net was the counter electrode, and the prepared α-Fe2O3 photoanode was the working electrode. A 300-W xenon lamp (100 mW/cm2) coupled with an AM 1.5 G filter was provided as the light source. The photocurrent density–voltage curve was measured by linear scanning voltammetry at a scanning speed of 20 mV/s. Mott-Schottky plots were performed with 1 kHz under light irradiation. Electrochemical impedance spectroscopy (EIS) was performed from 100 MHz to 1 kHz. Cyclic voltammetry (CV) was performed at different scan rates (20, 40, 60, 80, and 100 mV/s) in the potential range of 0 ~ 0.1 V (vs. Ag/AgCl).

3 Results and discussion

3.1 Morphology and structural characterizations

Fig. 1 a shows a schematic diagram of the preparing process for α-Fe2O3 photoanodes modified with a MOF and elemental codoping by solvothermal method. SEM images exhibit that α-Fe2O3 film is composed of nanopillars with a diameter of 50 nm and a thickness of ~ 500 nm (Figure S1). With the increasing growth time of MOF, the thickness of MOF film and surface nanoparticles gradually increase. After introducing the MOF-5(Ni) catalyst, a uniformly distributed nanoparticle layer loads on the surface of α-Fe2O3 film, where the thickness still remains at ~ 500 nm (Fig. 1 b and c). After regulating the organic functional group with –NH2 into the α-Fe2O3/NH2:MOF-5(Ni) photoanode, many nanosheets unevenly embed in the middle of nanopillars. The nanopillars are still neatly arranged with a diameter of ~ 50 nm (Fig. 1 d and e) As shown in Fig. 1 f and g, the morphology changes from nanosheets to nanoblocks after introducing the Ru doping into the α-Fe2O3/NH2:MOF-5(Ni) photoanode, where the loaded catalysts are uniformly distributed on the surface of α-Fe2O3 photoanode. And the thickness of α-Fe2O3 photoanodes remains almost unchanged. The α-Fe2O3 photoanodes doped with Fe and Co show a similar morphology (Figure S2).

Fig. 1
figure 1

a Schematic diagram of α-Fe2O3 photoanode modified with MOF and elemental codoping; SEM images of b and c α-Fe2O3/MOF-5(Ni), d and e α-Fe2O3/NH2:MOF-5(Ni), f and g α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes. b, d, and f The top-view images; c, e, and g The cross-sectional images.

Full size image

TEM analyses show that the α-Fe2O3 film is composed of nanoparticles, in which the surface is continuously covered with a thin film of NH2:MOF-5(Ni) catalyst with ~ 16 nm (Fig. 2 a and b). As shown in Fig. 2 c and d, the α-Fe2O3 nanorods are tightly encapsulated by the NH2:MOF-5(Ni)@Ru catalyst with approximately 13 nm. The EDS mapping proves the presence of Fe, O, Zn, Ni, and Ru elements, indicating the successful loading of the MOF-5 and Ni/Ru codoping to the surface of α-Fe2O3 photoanode (Fig. 2e), and the proportion of each element in α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode based on the EDS mapping has been put out in Fig. S3. The XRD spectra prove that the MOF-5 catalyst is successfully introduced and has no significant effect on the phase of α-Fe2O3 films (Figs. 3a, S4, and S5). UV–vis absorbance spectra show that the absorbance cutoff wavelength of α-Fe2O3 photoanode is ~ 600 nm. The α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode exhibits the optimal light absorption performance, implying that the MOF-5 catalyst and Ni/Ru codoping can improve the light absorption capacity (Fig. S6 and S7).

Fig. 2
figure 2

TEM and HR-TEM images of α-Fe2O3/NH2:MOF-5(Ni) (a and b), α-Fe2O3/NH2:MOF-5(Ni)@Ru (c and d) photoanodes, EDS elemental mapping of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode (e)

Full size image
Fig. 3
figure 3

The a XRD patterns of α-Fe2O3 and α-Fe2O3/MOF-5 photoanodes. b XPS survey spectra and c Fe 2p, d Ni 2p, e O 1s, and f Zn 2p in the α-Fe2O3, α-Fe2O3/MOF-5(Ni), α-Fe2O3/NH2:MOF-5(Ni), and α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes, respectively

Full size image

XPS spectra are conducted to study the surface elemental composition and chemical state of α-Fe2O3 photoanodes modified with MOF-5 catalysts and elemental doping. As displayed in Fig. 3b, the characteristic peaks of Fe 2p, Ni 2p, Zn 2p, O 1s, and Ru 3d are observed in the XPS spectra of α-Fe2O3/NH2:MOF-5(Ni)@Ru, which means that the MOF-5 catalyst and Ni/Ru codoping are successfully loaded onto the α-Fe2O3 film. The XPS spectra of Fe 2p demonstrate two characteristic peaks at ~ 711.8 eV (Fe 2p3/2) and ~ 725.4 eV (Fe 2p1/2), respectively, indicating that Fe in α-Fe2O3 exists as Fe3+.[29] The fitted peaks at ~ 716.2 eV and ~ 732.8 eV are satellite peaks of Fe 2p (Fig. 3c). The decreased binding energy of α-Fe2O3 photoanode modified with NH2:MOF-5(Ni) catalyst indicates that the electron density of iron element increases significantly [30]. In particular, the full-width at half-maximum of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode are wider than those of α-Fe2O3, α-Fe2O3/MOF-5(Ni) and α-Fe2O3/NH2:MOF-5(Ni) photoanodes, indicating that there are more abundant valence states in the MOF heterostructure [30]. The peaks of Ni 2p at 874.5 eV and 856.4 eV belong to the Ni 2p1/2 and Ni 2p3/2 corresponding to the spin–orbit feature, indicating the presence of Ni2+ in the α-Fe2O3 photoanodes (Figs. 3d and S8) [31]. The characteristic peaks at ~ 861.9 eV and ~ 881 eV are attributed to satellite structures associated with the formation of γ-NiOOH, improving oxygen evolution reaction (OER) catalytic activity [31]. The fitting results demonstrate that the ratio of Ni2+ over the Ni 2p peaks decreases from 28.93 to 27.60%, and the Ni3+ over the Ni 2p peaks increases from 12.68 to 13.56%. The partial oxidation of Ni2+ to a higher valence (Ni3+) is not only to maintain the electrical neutrality but is more conducive to the oxidation reaction to generate oxygen on the surface of α-Fe2O3 photoanodes. The incorporation of higher valence Ni3+ into α-Fe2O3 photoanodes acts as electron-donating substitution impurities to effectively enhance the electrical conductivity [31]. As depicted in Fig. 3e, the characteristic peaks centered at ~ 530.5 eV and ~ 532 eV in the O 1s spectra are assigned to lattice oxygen (OL) and oxygen vacancy (OV), respectively. Compared with the original α-Fe2O3 photoanode, the oxygen vacancy contents of α-Fe2O3/MOF-5(Ni), α-Fe2O3/NH2:MOF-5(Ni), and α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes severally increase by 19.7%, 15%, and 33.9%, indicating that the introduction of MOF catalysts and elemental codoping can promote the of surface charge transfer [32]. The O 1s spectrum of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode has a positive displacement of binding energy for the two characteristic peaks, which is related to the perturbation of MOF catalyst on the electrons [32]. Zn 2p has two main peaks at 1023.1 eV and 1046.1 eV belonging to Zn 2p3/2 and Zn 2p1/2 (Fig. 3f). Spin orbital splitting observed in Zn 2p demonstrates the successful preparation of MOF-5 catalyst and the presence of Zn2+ in MOF-5 catalyst [33]. Ru 3d and N 1s spectra further demonstrate the presence of Ru and N in the synthesized α-Fe2O3 photoanodes modified with MOF-5 catalyst and Ru doping (Fig. S9 and S10) [34].

3.2 PEC-WS performance

The effect of MOF-5 catalysts with different thicknesses on the PEC-WS performance of α-Fe2O3 photoanodes is first studied by changing the reaction time. The SEM images show that the MOF-5 catalyst is attached to the α-Fe2O3 film in a pentagonal nanoblock, where the nanoblocks on the surface of α-Fe2O3/MOF-5-8h photoanode are small and sparsely distributed (Fig. S1a and 1b). The nanoblocks on the surface of α-Fe2O3 gradually become larger, evenly distributed, and tightly arranged along with the increased reaction time (Fig. S1c-1f). The overgrowth of MOF-5 leads to more charge recombination and a significant decrease in catalytic performance. The α-Fe2O3/MOF-5-12h photoanode possesses the highest photocurrent density at 1.23 VRHE, indicating the optimal PEC-WS performance (Fig. 4a). It is denoted as the α-Fe2O3/MOF-5 photoanode in the following for a clearer description. Similarly, the α-Fe2O3/MOF-5@Ru-5000 and α-Fe2O3/MOF-5(Ni) photoanodes also exhibit the optimal photocurrent density at 1.23 VRHE (Fig. S11). The α-Fe2O3/MOF-5@Ru-5000 photoanode is also expressed below as α-Fe2O3/MOF-5@Ru photoanode. As shown in Fig. 4b, the α-Fe2O3/NH2:MOF-5@Ru photoanode decorated with electron-donating group and noble metal doping exhibits the maximum photocurrent density (1.92 mA/cm2), which is 1.98 times that of the pure α-Fe2O3 photoanode. It also has a starting potential of 0.5 VRHE and a cathode shift of 0.3 V. The changing trend of transient photocurrent density is consistent with that of the J-V curves, where the α-Fe2O3/NH2:MOF-5@Ru photoanode possesses the lowest charge recombination rate (Fig. 4c). This implies that the introduction of noble metal doping and MOF catalyst can effectively enhance the photogenerated carrier separation and transfer [35]. Due to the accumulation of photogenerated holes at the interface between the electrode and electrolyte, a positive current spike is generated at the beginning of the bright state [36]. The attenuation degree of the transient photocurrent density represents the photogenerated carrier recombination rate (ηrec). The ηrec is expressed as (Jtrans − Jsteady)/Jtrans, where Jtrans is the photocurrent density at the spike, and Jsteady represents the photocurrent density at rest. PEIS and Mott-Schottky measurements are conducted to investigate the effects of MOF catalyst, electron-donating group, and elemental doping on the conductivity and photogenerated carrier transport of α-Fe2O3 photoanodes. As shown in Fig. 4d, the α-Fe2O3/NH2:MOF-5@Ru photoanode has the smallest Nyquist arc radius, which is consistent with the results of PEC-WS performance [37]. The equivalent circuit is shown in the inset of Fig. 4d, where Rs is the series resistance, Rtrap is the surface capture resistance, and RCT is the transfer resistance from the surface to the electrolyte. The transport resistance of α-Fe2O3/NH2:MOF-5@Ru photoanode is the smallest (Tab. S1), which can effectively promote the charge separation and transfer in the PEC-WS process [38]. The slopes of Mott-Schottky curves are positive, indicating that these α-Fe2O3 photoanodes are all n-type semiconductors (Fig. 4e). The rank of donor density (ND) calculated by the slope in the Mott-Schottky curves is α-Fe2O3/NH2:MOF-5@Ru > α-Fe2O3/NH2:MOF-5 > α-Fe2O3/MOF-5 > α-Fe2O3 photoanodes. This result proves that the introduction of MOF, electron-donating group, and elemental doping can effectively improve the conductivity of α-Fe2O3 photoanodes and prolong the lifetime of photogenerated holes, thereby enhancing photogenerated carrier separation and transfer [39]. The straight line of Mott-Schottky plots is extrapolated to the x-axis with an intercept value of the flat band potential (Efb). Obviously, compared with the α-Fe2O3 photoanode, the Efb of α-Fe2O3/MOF-5, α-Fe2O3/NH2:MOF-5, and α-Fe2O3/NH2:MOF-5@Ru photoanodes shifts to the lower potential, which is in good agreement with the J-V curve [40].

Fig. 4
figure 4

a and b J-V curves of α-Fe2O3 photoanodes modified with different MOF catalysts and Ru doping. c Transient photocurrent density curves, d Nyquist plots at 1.23 VRHE, and e Mott-Schottky plots of α-Fe2O3, α-Fe2O3/MOF-5@Ru, α-Fe2O3/NH2:MOF-5, α-Fe2O3/NH2:MOF-5@Ru photoanodes. (f and g) J-V curves, h transient photocurrent density curves of α-Fe2O3 photoanodes modified with different MOF catalysts and Fe/Ni/Co/Ru doping, i PEC-WS performance comparison with the partial reported α-Fe2O3 photoanodes.

Full size image

To further improve the conductivity and PEC-WS performance of α-Fe2O3 photoanodes, the elemental doping (Co, Ni, and Fe) is severally introduced into the α-Fe2O3/MOF-5 photoanode. As shown in Fig. 4f, the α-Fe2O3/MOF-5(Ni) photoanode shows the optimal photocurrent density of 2.01 mA/cm2 at 1.23 VRHE. As depicted from the SEM images of α-Fe2O3/MOF-5(Fe), α-Fe2O3/MOF-5(Co), and α-Fe2O3/MOF-5(Ni) photoanodes, there are more nanoparticles surface-loaded and more neatly arranged on the surface of α-Fe2O3/MOF-5(Ni) photoanode (Figs. 1 and g and S2). This implies the more active sites on the surface of the α-Fe2O3/MOF-5(Ni) photoanode to participate in the PEC water oxidation. On account of the synergistic effect of Ni/Ru codoping and NH2:MOF-5 catalyst, the photocurrent density of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode (2.6 mA/cm2) at 1.23 VRHE is much higher than that of α-Fe2O3/MOF-5 (1.52 mA/cm2) photoanode (Fig. 4g). As shown in Fig. 4h, the photogenerated carrier recombination rate (16.3%) of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode is significantly lower than that of other α-Fe2O3 photoanodes, indicating that the introduction of MOF catalyst and Ni/Ru codoping can significantly reduce the hole accumulation on the surface of α-Fe2O3 photoanode and improve the surface reaction kinetics [41]. The α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode is among the reported α-Fe2O3 photoanodes with excellent PEC-WS performance in recent years (Fig. 4i and Tab. S2) [42,43,44,45,46,47,48,49].

3.3 Carrier dynamics analysis

Nyquist and Mott-Schottky plots are conducted to further investigate the intrinsic mechanism of the enhanced PEC-WS performance by introducing MOF catalyst and elemental doping. The four α-Fe2O3 photoanodes exhibit similar Nyquist plots at 1.23 VRHE. The α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode has the lowest charge transport resistance and the highest charge transfer efficiency (Fig. 5a and Tab. S3), which indicates that the incorporation of Ru into NH2:MOF-5(Ni) can improve the conductivity and promote the photogenerated hole transfer to the electrolyte for water oxidation reaction to further improve the PEC-WS performance [37]. The negative Efb displacement and maximum ND (7.48 × 1018) of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode in the Mott-Schottky plots indicate that the NH2:MOF-5 catalyst and Ni/Ru codoping can effectively improve the bulk conductivity and makes the band bending at the electrolyte interface more obvious, thereby inhibiting the photogenerated carrier recombination (Fig. 5b) [40]. By measuring the CV curves at different scan rates, the electrochemical bilayer capacitance (Cdl) of the photoanode is calculated from the CV curves at different scan rates, which is positively correlated with the electrochemically active surface area (Fig. 5 c and d). The α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode exhibits a high intrinsic OER activity (Figs. S12 and 5e), indicating that the loading of NH2:MOF-5(Ni)@Ru catalyst is conducive to accelerating the charge transfer and reducing the charge recombination in α-Fe2O3 photoanode [50]. As shown in Fig. 5f, the slope of LSV polarization curve for α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes is 176.8 mV/decade, which is significantly smaller than that of α-Fe2O3/MOF-5(Ni) (396.6 mV/decade) and α-Fe2O3/NH2:MOF-5(Ni) (321.2 mV/decade), indicating a highest electron transfer rate and best OER reaction kinetics for α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode [51]. The H2O2 as a hole scavenger is added to the electrolyte for PEC measurements, which further explores the effect on the increased photogenerated carrier transport efficiency [11, 27]. Fig. 5g shows the J-V curves of α-Fe2O3/MOF-5(Ni), α-Fe2O3/NH2:MOF-5(Ni), and α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes in the presence of H2O2 hole scavenger, and the photocurrent density relationship at 1.23 VRHE is consistent with that of α-Fe2O3 photoanodes without H2O2 hole scavenger. The photocurrent density of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode reaches 3.34 mA/cm2 and is much larger than that without H2O2 hole scavenger. This may be attributed to that the H2O2 hole scavenger can inhibit carrier recombination. Considering that the ηabs of the three photoanodes in the wavelength range are almost identical, the ηabs × ηsep curves directly reflect the ηsep relationship (Fig. 5h) [11, 27]. The ηtran of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode at 1.23 VRHE is 77.4%, which is greater than that of α-Fe2O3/MOF-5(Ni) and α-Fe2O3/NH2:MOF-5(Ni) photoanodes. The ηsep of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode at 1.23 VRHE is 26.5%, which is greater than that of α-Fe2O3/MOF-5(Ni) and α-Fe2O3/NH2:MOF-5(Ni) photoanodes. Compared with α-Fe2O3/MOF-5(Ni) and α-Fe2O3/NH2:MOF-5(Ni) photoanodes, the ηtran of α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode at 1.23 VRHE is increased by 17.7% and 11.7%, respectively (Fig. 5i). Therefore, the improvement of PEC-WS performance can be attributed to the significantly enhancement of photogenerated carriers separation and transfer efficiency. The mechanism diagram of photogenerated carrier transport in α-Fe2O3 photoanode modified with MOF catalyst is shown in Fig. S13. The PEC conversion is from an integration of photogenerated carrier generation, separation, recombination, and transfer by the five steps in the transport process: (1) photogenerated carrier generation, (2) carrier separation, (3) recombination in the bulk, (4) surface transfer, and (5) recombination on the surface. Through the above electrochemical and carrier transport analysis, it is found that MOF catalyst can effectively promote carrier separation and transfer (processes 2 and 4) and effectively reduce photogenerated carrier recombination (processes 3 and 5). It is the reason that MOF catalyst can improve the PEC conversion efficiency of α-Fe2O3 photoanode.

Fig. 5
figure 5

a Nyquist plots at 1.23 VRHE, b Mott-Schottky plots of α-Fe2O3/MOF-5(Ni), CV curves for c α-Fe2O3/MOF-5(Ni) and d α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes. e Evaluation of Cdl values by plotting the ∆j vs. scan rate, f Tafel plots, g J-V curves with H2O2, h ηabs × ηsep, and i ηtran of α-Fe2O3/MOF-5(Ni), α-Fe2O3/NH2:MOF-5(Ni), and α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanodes.

Full size image

4 Conclusion

In summary, we successfully synthesize α-Fe2O3 photoanodes modified with MOF-5 catalyst, electron donor group, and element codoping in three-dimensional nanostructures. Compared with the pure α-Fe2O3 photoanode, the photocurrent density of α-Fe2O3/MOF-5 photoanode is increased by 0.56 times, and the α-Fe2O3/NH2:MOF-5(Ni)@Ru photoanode increased by 1.68 times. The remarkable enhancement of PEC-WS performance can be attributed to that the MOF-5 catalyst with three-dimensional pore structures provides more active sites, elemental doping increases the conductivity, and the electron donor groups of –NH2 effectively prolong the lifetime of photogenerated holes. The synergistic effect of Ni/Ru codoping further promotes the photogenerated carrier separation and transfer. This work provides a new way for improving the photoelectrochemical conversion of oxide photoelectrode materials.