Enhanced photoelectrochemical water splitting performance of α-Fe₂O₃ photoanodes through Co-modification with Co single atoms and g-C₃N₄

Abstract

The practical application of α-Fe₂O₃ in water splitting is hindered by significant charge recombination and slow water oxidation. To address this issue, a CoSAs–g-C₃N₄/Fe₂O₃ (CoSAs: cobalt single atoms) photoanode was fabricated in this study through the co-modification of CoSAs and g-C₃N₄ to enhance photoelectrochemical (PEC) water splitting. The coupling between g-C₃N₄ and α-Fe₂O₃ resulted in the formation of a heterojunction, which provided a strong built-in electric field and an additional driving force to mitigate charge recombination. Moreover, g-C₃N₄ served as a suitable carrier for single atoms, which effectively anchored CoSAs through N/C coordination. The highly dispersed CoSAs provided abundant active sites, which further promoted surface holes extraction and oxidation kinetics, resulting in higher PEC performance and photostability. This study indicates the benefits of these collaborative strategies and provides more efficient designs for solar energy conversion in PEC systems.

---

Introduction

Photoelectrochemical (PEC) water splitting technology converts renewable solar energy into storable chemical energy. Moreover, this technology is considered a promising strategy for addressing energy crises and mitigating environmental pollution.^1–4 Hematite (α-Fe₂O₃), a highly attractive photoanode material, has received significant attention in PEC water splitting owing to its narrow band gap (1.9–2.2 eV), which enables efficient absorption of visible light. Additionally, α-Fe₂O₃ exhibits high theoretical solar-to-hydrogen conversion efficiency (∼15.5%), earth-abundance, and stability.^5–8 However, the significant challenge in charge recombination hinders the practical application of α-Fe₂O₃.⁹ Currently, the reported current density of α-Fe₂O₃ is significantly below its theoretical value (12.4 mA cm⁻²), indicating significant potential for improvement.¹⁰ Therefore, it is crucial and challenging to further modify α-Fe₂O₃ to improve its efficiency in separating photogenerated carriers.

In recent years, single atom catalysts (SACs) have emerged as a promising area of study owing to their various advantages,¹¹ such as: (1) single metal atoms often serve as unsaturated coordination sites and are considered active centers; (2) strong metal–support interactions can increase the electron transfer between them, thereby regulating the electronic structure of metal atoms; (3) SACs can optimize the exposure of active sites and improve atomic utilization, leading to reduced catalytic costs. Owing to these advantages, SACs exhibit excellent activity and durability in numerous catalytic reactions, making them highly promising for applications in water splitting. For example, Bi et al. demonstrated the charge transfer and bond evolution between single atom Pt and C₃N₄ catalysts in photocatalytic water splitting.¹² Lee et al. reported the CoSAs–MoS₂/TiN nanorod (NR) electrocatalysts (CoSAs: cobalt single atoms). To the strong interaction between CoSAs and MoS₂, adjusted their electron density distribution, resulting in more catalytic active sites for reactant molecules. This contributed to the excellent overall hydrolysis performance of the electrocatalyst in pH universal electrolytes.¹³ Additionally, Corvini et al. reported that single atom Ru anchored to NiFe-layered double hydroxides significantly enhanced the oxygen evolution activity of BiVO₄.¹⁴ Transition metal Co exhibits unique physicochemical properties, making it a promising candidate material for SACs. Previous research has shown that Co SACs exhibit excellent electrocatalytic activity compared with Co, Fe, Ni, Cr, V, and Rh SACs.¹⁵ Considering that Co SACs have demonstrated effectiveness in improving catalytic activity,¹⁶ we suggest that incorporating (CoSAs) into α-Fe₂O₃ can further improve the PEC water oxidation performance by regulating charge distribution and increasing the reaction site. However, reports on CoSAs/Fe₂O₃ are scarce.¹⁷

Despite the excellent performance of SACs, the selection of a suitable carrier is crucial. Owing to the tendency of individual atoms to readily aggregate, only appropriate charge carriers can effectively and stably disperse these atoms. Currently, reports on CoSAs/Fe₂O₃ photoanodes are few, partly owing to the inadequacy of α-Fe₂O₃ as a suitable carrier. Therefore, introducing an additional suitable carrier medium into the photoanode system is crucial. The conjugated polymer graphite carbon nitride (g-C₃N₄) is considered a suitable candidate for a single atom carrier. g-C₃N₄ contains electron-rich N atoms, which can provide sufficient coordination sites and effectively anchor isolated metal atoms to nitrogen coordination centers, forming an M–N_x structure that serves as an active center for several catalytic reactions.^11,12,18 For example, Zhu et al. synthesized Co–g-C₃N₄/rGO SACs. Owing to the formation of Co–N and Co–3N coordination structures, Co–g-C₃N₄/rGO SACs exhibit higher hydrogen evolution activity compared with commercial Pt/C. Moreover, these compounds exhibit stability for up to 500 h at high temperatures.¹⁵ Additionally, g-C₃N₄, a non-metallic semiconductor, can form heterojunctions with other semiconductors, thereby facilitating charge migration and separation owing to its suitable band structure. This characteristic has broad potential applications in photocatalysis. For example, Huang et al. investigated the g-C₃N₄/Mn₂O₃/FTO p–n heterojunction as a photoelectrode for PEC water decomposition.¹⁹ Similarly, Fu et al. synthesized an Au/g-C₃N₄/TiO₂ nanotube array heterojunction as a photocatalyst for degrading o-chloronitrobenzene target pollutants.²⁰ These studies indicated that g-C₃N₄ has the potential to serve as a carrier for CoSAs and a semiconductor to form heterojunctions with α-Fe₂O₃, playing a mutually beneficial role.

Herein, a photoanode material, g-C₃N₄/Fe₂O₃ anchored with CoSAs (CoSAs–g-C₃N₄/Fe₂O₃), was designed and assembled to improve water oxidation performance. Scanning transmission electron microscopy (STEM) and X-ray absorption fine structure analyses indicated that Co atoms were securely anchored on g-C₃N₄ through N/C coordination bonds, thereby maintaining atomic isolation. The electrochemical results revealed that the strong built-in electric field and photovoltage generated at the g-C₃N₄/Fe₂O₃ interface effectively reduced the recombination of photogenerated charges. Furthermore, the dispersed CoSAs provided sufficient active sites for water oxidation reactions. The synergistic effect of these favorable factors facilitated efficient charge separation. The optimized CoSAs–g-C₃N₄/Fe₂O₃ photoanode exhibited a photocurrent density of 1.93 mA cm⁻² at 1.23 V_RHE, which was 3.22 times that of pure α-Fe₂O₃. The potential charge transfer pathways and PEC water oxidation mechanism of CoSAs–g-C₃N₄/Fe₂O₃ were extensively elucidated.

Experimental section

Synthesis of the α-Fe₂O₃ nanorod array (NRA) photoanode

The α-Fe₂O₃ NRA sample was prepared using our previously reported method.^21,22 Initially, 10 mL of aqueous solution containing 1.5 mmol FeCl₃·6H₂O and 1.5 mmol urea was transferred to a Teflon-lined stainless steel autoclave. Subsequently, a piece of cleaned FTO was placed in the aqueous solution. The autoclave was sealed and heated in an oven at 100 °C for 11 h. After the completion of the reaction and cooling to room temperature, the FTO substrate with the FeOOH film was removed, washed, and dried. The FeOOH film was then calcined in a muffle furnace at 550 °C for 2 h and at 700 °C for 10 min in an air atmosphere. Finally, the target α-Fe₂O₃ NRA grown on the FTO substrate was obtained and denoted as pure α-Fe₂O₃.

Synthesis of the g-C₃N₄/Fe₂O₃ NRA photoanode

The preparation process of g-C₃N₄/Fe₂O₃ is as follows: first, an 10 mL ethylene glycol solution containing a certain amount of melamine was dripped onto the α-Fe₂O₃ NRA photoanode, followed by drying in a muffle furnace at 220 °C to eliminate the ethylene glycol from the surface of α-Fe₂O₃. This procedure was repeated twice to increase the deposition amount of melamine. Finally, α-Fe₂O₃ NRA coated with melamine were annealed at 550 °C for 2 h in a tube furnace in an Ar atmosphere to produce the g-C₃N₄/Fe₂O₃ NRA photoanode. The thickness of the g-C₃N₄ layer was optimized by adjusting the amount of melamine added to the glycol solution. The resulting g-C₃N₄/Fe₂O₃ was labeled as g-C₃N₄/Fe₂O₃-x, with x representing the amount of melamine added to the ethylene glycol solution (e.g., g-C₃N₄/Fe₂O₃-0.1, g-C₃N₄/Fe₂O₃-0.2, and g-C₃N₄/Fe₂O₃-0.3 for 0.1, 0.2, and 0.3 g, respectively). The photoelectrochemical performance test revealed that g-C₃N₄/Fe₂O₃-0.2 exhibited the highest performance (Fig. S1†). Therefore, g-C₃N₄/Fe₂O₃ in the study denotes g-C₃N₄/Fe₂O₃-0.2, unless stated otherwise.

Synthesis of the CoSAs–g-C₃N₄/Fe₂O₃ NRA photoanode

CoSAs were anchored on the g-C₃N₄/Fe₂O₃ NRA surface as follows: initially, 0.88 g of Co(NO₃)₂·6H₂O was dissolved in 100 mL of deionized water and stirred for 30 min. Subsequently, g-C₃N₄/Fe₂O₃ NRA were immersed in the above solution and maintained at 50 °C for 5 h. Afterward, g-C₃N₄/Fe₂O₃ NRA were removed, washed several times with ethanol and deionized water, and dried at 60 °C. Finally, the sample was annealed in an Ar atmosphere at 400 °C for 2 h with a heating rate of 2 °C min⁻¹ to synthesize CoSAs–g-C₃N₄/Fe₂O₃ photoanode.

Results and discussion

Fig. 1a illustrates the schematic of the preparation process for the CoSAs–g-C₃N₄/Fe₂O₃ photoanode. Initially, melamine was uniformly deposited on the surface of α-Fe₂O₃ by dripping a glycol solution containing melamine.²⁰ Subsequently, α-Fe₂O₃ coated with melamine was thermally condensed in a tube furnace to produce the g-C₃N₄/Fe₂O₃ composite. Unlike g-C₃N₄/Fe₂O₃ formed through the spin-coating of g-C₃N₄ on the α-Fe₂O₃ surface, g-C₃N₄ produced during the thermal condensation process can form a strong chemical bond with α-Fe₂O₃. This prevented defects resulting from the mismatched layers in the heterostructure caused by the direct spin-coating of g-C₃N₄ on the α-Fe₂O₃ surface. These defects often served as sites for the undesirable recombination of photogenerated carriers. Moreover, the use of g-C₃N₄ as the substrate facilitated the anchoring of CoSAs onto g-C₃N₄/Fe₂O₃ through water bath deposition and calcination methods, resulting in the target CoSAs–g-C₃N₄/Fe₂O₃ NRA photoanode. Scanning electron microscopy (SEM) images (Fig. 1b–d) revealed that pure α-Fe₂O₃ exhibited a nanorod array structure, vertically grown on the FTO substrate. Conversely, the morphologies of g-C₃N₄/Fe₂O₃ and CoSAs–g-C₃N₄/Fe₂O₃ remained unchanged, indicating that g-C₃N₄ maintained an ultra-thin structure, and the dispersed CoSAs had no effect on the morphology of g-C₃N₄/Fe₂O₃. The X-ray diffraction (XRD) spectra of g-C₃N₄/Fe₂O₃ and CoSAs–g-C₃N₄/Fe₂O₃ (Fig. S2†) exhibited diffraction peaks corresponding to α-Fe₂O₃ (JCPDS No. 33-0664) and FTO.^5,23 No characteristic peaks associated with g-C₃N₄ and Co components were observed, likely owing to the low amount and poor crystallinity of g-C₃N₄ and the trace amount of dispersed CoSAs. These Co atoms did not aggregate into Co nanoparticles, which are crucial for producing prominent diffraction peaks.

Fig. 1 (a) Schematic of the CoSAs–g-C₃N₄/Fe₂O₃ NRA photoanode preparation process. Top-view SEM images of (b) α-Fe₂O₃; (c) g-C₃N₄/Fe₂O₃; (d) CoSAs–g-C₃N₄/Fe₂O₃.

To confirm the presence of g-C₃N₄ and CoSAs, the aberration-corrected annular bright-field STEM (AC-ABF-STEM) imaging was performed on CoSAs–g-C₃N₄/Fe₂O₃. The AC-ABF-STEM image (Fig. 2a) revealed an ultra-thin layer of g-C₃N₄ surrounding α-Fe₂O₃ NR, forming a close interface, suggesting the formation of a heterojunction between g-C₃N₄ and α-Fe₂O₃.^24,25 Additionally, numerous black dots corresponding to CoSAs were observed on g-C₃N₄ (marked by red circles), indicating the uniform dispersion of CoSAs without the presence of nanoparticles or clusters. The magnified AC-ABF-STEM images (Fig. 2b and c) revealed the poor crystallinity of the g-C₃N₄ layer, with an average thickness of ∼3 nm. The size of a single black dot corresponding to an isolated Co site was 0.15 nm. The high-resolution TEM (HRTEM) image (Fig. 2d) indicated a distinct interface structure with two lattice stripes. Particularly, the lattice stripe of 0.25 nm corresponded to the (110) crystal plane of α-Fe₂O₃.^26–28 Another lattice stripe, observed within a thin layer of poor crystallinity, exhibited a lattice spacing of 0.33 nm, corresponding to the (002) crystal plane of g-C₃N₄.²⁰ This strongly confirms that the ultra-thin layer with poor crystallization was g-C₃N₄, thereby validating the successful preparation of g-C₃N₄/Fe₂O₃ heterojunction. Moreover, TEM-EDS elemental mapping revealed a uniform distribution of Fe, O, C, N, and Co elements on NR (Fig. 2e), indicating the presence of the g-C₃N₄ layer and CoSAs. This confirms the successful formation of CoSAs–g-C₃N₄/Fe₂O₃ composite materials.

Fig. 2 (a)–(c) Atomic-resolution TEM images of CoSAs–g-C₃N₄/Fe₂O₃ NR (CoSAs are highlighted by red circles); (d) and (e) HRTEM image and EDS mapping of Co, Fe, O, C, and N for CoSAs–g-C₃N₄/Fe₂O₃ NR.

To analyze the surface chemical states, we performed X-ray photoelectron spectroscopy (XPS) characterization on α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃. Fig. 3a–d show the XPS spectra of C 1s, N 1s, O 1s, and Co 2p, for CoSAs–g-C₃N₄/Fe₂O₃, respectively. The XPS spectra of C 1s (Fig. 3a) exhibited three peaks corresponding to C–C (284.8 eV), C–N (286.3 eV), and N–C=N (288.4 eV).²⁰ The N 1s spectrum (Fig. 3b) featured two peaks at 399.7 and 401.4 eV, corresponding to N–(C)₃ and C–NH_x,²⁰ respectively. These results from the C 1s and N 1s spectra indicate the successful loading of g-C₃N₄ on the α-Fe₂O₃ surface, resulting in the formation of a g-C₃N₄/Fe₂O₃ heterostructure. The O 1s spectrum (Fig. 3c) exhibited three peaks at around 529.8, 531.5, and 533.1 eV, which were assigned to the Fe–O bond,¹⁰ oxygen vacancies within the α-Fe₂O₃ structure,²⁹ and surface hydroxyl groups, respectively.³⁰ The Fe 2p spectrum (Fig. S3†) featured two peaks at 724.4 and 710.6 eV, corresponding to Fe 2p_1/2 and Fe 2p_3/2, respectively, along with two satellite peaks at 732.7 and 718.2 eV, consistent with the reported values for Fe³⁺ in Fe₂O₃.³¹ Additionally, the Fe 2p_3/2 peak can be deconvoluted into three peaks at 712.2, 710.6, and 709.6 eV, indicating the characteristics of Fe³⁺. The Co 2p spectra (Fig. 3d) exhibited two main peaks corresponding to Co 2p_3/2 (781.5 eV) and Co 2p_1/2 (796.5 eV),¹⁵ along with satellite peaks, suggesting the successful anchoring of CoSAs on g-C₃N₄. Notably, after the loading of CoSAs and g-C₃N₄, the Fe 2p, O 1s, and N 1s peaks slightly shifted to higher binding energies, suggesting interactions between CoSAs, g-C₃N₄, and α-Fe₂O₃ (Fig. S4†).¹⁴

Fig. 3 High-resolution XPS spectra of (a) C 1s; (b) N 1s; (c) O 1s; (d) Co 2p, for CoSAs–g-C₃N₄/Fe₂O₃.

Furthermore, X-ray absorption near-edge structure (XANES) analysis was performed to investigate the oxidation state and coordination structure of Co atoms in CoSAs–g-C₃N₄/Fe₂O₃. The XANES spectra of CoSAs–g-C₃N₄/Fe₂O₃ (Fig. 4a) indicated the unique oxidation state of Co^δ+ (0 < δ < 2), which can be supported by the linear combination fitting (LCF) result (Fig. S5†). The extended X-ray absorption fine structure (EXAFS) spectra of CoSA–g-C₃N₄/Fe₂O₃ (Fig. 4b) exhibited a single dominant peak at 1.5 Å, corresponding to the first coordination shell of Co–N. Notably, the weaker peak intensity of the CoSAs–g-C₃N₄/Fe₂O₃ spectrum compared with CoPc suggested a lower Co–N coordination value and inferior crystal properties.³² Additionally, the absence of the characteristic Co–Co bond length (∼2.1 Å) in Co foils within CoSAs–g-C₃N₄/Fe₂O₃, indicated the lack of Co nanoparticles, consistent with atomic-resolution TEM data (Fig. 2a–c) and XRD pattern (Fig. S2†). The coordination of Co sites was further investigated through EXAFS curve fitting analysis (Fig. 4c), confirming that each Co atom was coordinated with 4.1 ± 0.2 N atoms at the first shell and 1.1 ± 0.4 C atoms at second shell according to the fitting results (Fig. S6 and Table S1†). Notably, unlike the reference samples, a strong peak of Co–N coordination in the wavelet transforms (WT) of Co K-edge EXAFS oscillations (Fig. 4d), directly confirmed the presence of CoSAs on the g-C₃N₄/Fe₂O₃ substrate. All results obtained from atomic-resolution STEM, HRTEM, XPS, and XANES confirmed the successful preparation of the CoSAs–g-C₃N₄/Fe₂O₃ photoanode.

Fig. 4 (a) XANES spectra; (b) R-space Co K-edge EXAFS spectra; (c) EXAFS R-space fitting curve; (d) WT-EXAFS signals of Co foil, CoO, CoPc and CoSAs–g-C₃N₄/Fe₂O₃.

The PEC water oxidation properties of the obtained photoanodes were measured using a three-electrode system. The linear sweep voltammetry (LSV) curves (Fig. 5a) revealed that pure α-Fe₂O₃ exhibited a photocurrent density of 0.59 mA cm⁻² at 1.23 V_RHE with an initial potential of 0.775 V_RHE. Upon coupling α-Fe₂O₃ with g-C₃N₄, the photocurrent density of g-C₃N₄/Fe₂O₃ increased to 1.4 mA cm⁻², accompanied by a negative initial potential shift of 37 mV. This shift suggests that the formed g-C₃N₄/Fe₂O₃ heterojunction promoted rapid charge transfer. Moreover, after anchoring CoSAs on g-C₃N₄, CoSAs–g-C₃N₄/Fe₂O₃ exhibited a photocurrent density of 1.93 mA cm⁻². This indicates that CoSAs provided additional active sites for water oxidation, which effectively inhibited surface charge recombination and promoted interfacial charge transfer. Compared with g-C₃N₄/Fe₂O₃, the initial potential of CoSAs–g-C₃N₄/Fe₂O₃ (0.58 V_RHE) decreased by 158 mV, indicating a significant contribution from dispersed CoSAs to the acceleration of oxygen evolution reaction (OER) kinetics. This finding was consistent with transient photocurrent test results (Fig. 5b). α-Fe₂O₃ exhibited a large transient photocurrent spike owing to the significant occurrence of charge recombination in the bulk and surface of the material.^33,34 The coupling of α-Fe₂O₃ with g-C₃N₄, significantly suppressed the transient peak. Additionally, the gradual increase in the photocurrent density of g-C₃N₄/Fe₂O₃ can be attributed to the built-in electric field generated by the g-C₃N₄/Fe₂O₃ heterojunction, which facilitated the directional separation of photogenerated charges.^1,35,36 Compared with g-C₃N₄/Fe₂O₃, CoSAs–g-C₃N₄/Fe₂O₃ exhibited a smaller transient spike, which almost disappeared, while the photocurrent further increased. This phenomenon indicates that the dispersed CoSAs cocatalyst promoted sluggish water oxidation kinetics, thereby significantly facilitating charge transfer and separation. Furthermore, the incident photon-to-current efficiency (IPCE) of the sample was calculated to evaluate its solar energy conversion efficiency. At a wavelength (λ) of 350 nm, α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃ exhibited IPCE values of 14.5, 26.8, and 40.9%, respectively, indicating a gradually increasing trend (Fig. 5c). Regarding the influence of light collection efficiency, charge separation efficiency, and injection efficiency on IPCE, the IPCE result strongly confirmed the synergistic effect of g-C₃N₄ and CoSAs in enhancing the light absorption, promoting charge separation, and improving hole injection efficiency of the pure α-Fe₂O₃ photoanode. The applied bias photon-to-current conversion efficiency (ABPE) results exhibited a similar trend (Fig. 5d). The α-Fe₂O₃ photoanode exhibited an ABPE value of 0.05% at 1.06 V_RHE, which was significantly lower than those of g-C₃N₄/Fe₂O₃ (0.1% at 1.05 V_RHE) and CoSAs–g-C₃N₄/Fe₂O₃ (0.18% at 1.03 V_REH). The increase in ABPE values and the negative shift of the peak position indicated that the introduction of g-C₃N₄ and CoSAs can enable efficient charge separation in the low potential range, facilitating better energy conversion.^1,14,22

Fig. 5 (a) LSV curves; (b) transient photocurrent measurements at 1.23 V_RHE; (c) IPCE curves; (d) ABPE curves of α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃ photoanodes.

The Mott–Schottky (M–S) plots, open circuit photovoltage (OCP) curves, and the potential required to generate a photocurrent density of 1 mA cm⁻² were used to further investigate the influence of g-C₃N₄ and Co SAs on the PEC performance of α-Fe₂O₃. All photoanodes exhibited positive slopes, reflecting their n-type semiconductor characteristics (Fig. 6a).^29,37 The α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃ photoelectrodes featured carrier densities of 3.4 × 10¹⁷, 1.1 × 10¹⁸, and 1.2 × 10¹⁸ cm⁻³, respectively. The increase in carrier concentration can be attributed to the surface modifications introduced by g-C₃N₄ as a heterojunction and CoSAs as a cocatalyst. These modifications significantly inhibited the recombination process of the bulk and surface charge carriers, facilitating the easier release and injection of more photogenerated holes into the electrolyte. Consequently, this led to improved charge separation efficiency and water oxidation performance. Additionally, the successive negative shift in the flat potential of α-Fe₂O₃ (0.62 V_RHE), g-C₃N₄/Fe₂O₃ (0.49 V_RHE), and CoSAs–g-C₃N₄/Fe₂O₃ (0.45 V_RHE) indicated increased photovoltage, enhanced charge mobility, and improved water oxidation kinetics. Generally, a higher the ΔV_OC (ΔV_OC = OCP_dark − OCP_light) indicates a stronger inherent electric field and a greater driving force for carrier separation.^14,38 The α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃ photoanodes exhibited ΔV_OC values of 0.206, 0.225, and 0.229 V_RHE, respectively (Fig. 6b). The g-C₃N₄/Fe₂O₃ composite exhibited a higher ΔV_OC than α-Fe₂O₃, indicating the formation of a strong built-in electric field resulting from the coupling of g-C₃N₄ and α-Fe₂O₃. This increased band bending provided an additional driving force for charge separation and effectively suppressed carrier recombination.^39,40 Upon the introduction of CoSAs, the ΔV_OC of CoSAs–g-C₃N₄/Fe₂O₃ further increased, indicating that CoSAs can further reduce carrier recombination and promote charge separation. The potentials required for α-Fe₂O₃, g-C₃N₄/Fe₂O₃, and CoSAs–g-C₃N₄/Fe₂O₃ to generate a photocurrent density of 1 mA cm⁻² were 1.36, 1.17, and 1.03 V_RHE, respectively (Fig. 6c). The significant decrease in overpotential indicated that the synergistic effect of the g-C₃N₄/Fe₂O₃ heterojunction and CoSAs can improve charge separation and injection efficiency, thereby significantly enhancing PEC water oxidation performance. To accurately assess the effect of g-C₃N₄ and CoSAs on the PEC properties of α-Fe₂O₃, the bulk charge separation efficiency (η_bulk) and surface charge separation efficiency (η_surface) were independently calculated. This calculation was performed using the LSV data and the integrated photocurrent density (J_abs) (Fig. S7†) with Na₂SO₃ as the sacrificial agent. The η_bulk of pure α-Fe₂O₃ considerably increased from 11.5% to 20.6% with the introduction of g-C₃N₄, and further increased to 26.3% with CoSAs (Fig. 6d). Notably, both the incorporation of g-C₃N₄ as a heterojunction and the introduction of CoSAs as a cocatalyst can reduce the recombination of bulk photogenerated electron–hole pairs and promote charge separation, with the construction of heterojunctions playing a more critical role. Additionally, the η_surface of α-Fe₂O₃ significantly increased from 40.3% to 58.8% with the introduction of g-C₃N₄ and further to 91.3% with the addition CoSAs. The highest charge injection efficiency achieved through the introduction of CoSAs can be attributed to the dispersed CoSAs, which effectively extracted holes from the surface of g-C₃N₄/Fe₂O₃ and provided more active sites to deoxidize water (Fig. 6e).⁴¹ Electrochemical impedance spectroscopy (EIS) was used to examine the charge transfer resistance at the photoanode interface. All Nyquist plots were fitted using the equivalent circuit (inset of Fig. 6f). The relevant fitting results are summarized in Table S2.† The g-C₃N₄/Fe₂O₃ composite (337.2 Ω) exhibited a significantly lower charge transfer resistance than α-Fe₂O₃ (634.3 Ω), indicating the heterojunction interface facilitated rapid interfacial charge transfer (Fig. 6f). With the anchoring of CoSA cocatalysts, CoSAs–g-C₃N₄/Fe₂O₃ exhibited the lowest resistance (269.9 Ω), signifying the role of the CoSA cocatalyst in improving the efficiency of surface charge transfer and hole injection. All these results confirm that the synergistic effect of CoSAs and heterojunction fabrication can significantly improve the overall carrier density, charge separation, and injection efficiency, leading to a significant enhancement in PEC water splitting performance.

Fig. 6 (a) M–S plots; (b) OCP transient decay curves; (c) potential (V_RHE) versus time for the photoelectrodes at 1 mA cm⁻²; (d) η_bulk; (e) η_surface; (f) EIS of α-Fe₂O₃, g-C₃N₄/Fe₂O₃, CoSAs–g-C₃N₄/Fe₂O₃ photoanodes.

Stability is a vital parameter for evaluating the performance of photoelectrodes. Therefore, we investigated the water oxidation stability of CoSAs–g-C₃N₄/Fe₂O₃, and the relevant results are shown in Fig. 7. After continuous irradiation for 3.8 h (Fig. 7a), the photoanode exhibited good stability and acceptable attenuation of photocurrent density. The XRD (Fig. 7b), SEM (Fig. 7c), and HRTEM (Fig. 7d) images of the photoanodes remained consistent with those obtained before testing. This indicates that the crystal structure and array morphology of the surface CoSAs–g-C₃N₄/Fe₂O₃ electrode remained unchanged.

Fig. 7 (a) Photocurrent density stability of CoSAs–g-C₃N₄/Fe₂O₃ photoanodes at 1.23 V_RHE; (b) XRD pattern; (c) SEM; (d) HRTEM of CoSAs–g-C₃N₄/Fe₂O₃ photoanodes after long-term stability tests.

To elucidate possible charge transfer pathways and water oxidation mechanisms, the band positions of α-Fe₂O₃ and g-C₃N₄ were accurately examined. The energy band location of CoSAs–g-C₃N₄/Fe₂O₃ was calculated and plotted based on the results of UV-Vis spectroscopy (Fig. S8†) and UPS. The α-Fe₂O₃ and g-C₃N₄ samples exhibited band gap values of 2.07 and 2.64 eV, respectively (Fig. 8a). The UPS results of the α-Fe₂O₃ and g-C₃N₄ samples are illustrated in Fig. 8b and c. The work function (Φ) indicated the energy of the Fermi level regarding the vacuum level, while |E_VBM| was calculated using eqn (S5) and (S6).† The α-Fe₂O₃ and g-C₃N₄ samples featured Φ values of 3.82, and 2.67 eV, respectively. Additionally, α-Fe₂O₃ and g-C₃N₄ exhibited |E_VBM| values of 5.49 and 4.88 eV, respectively. Moreover, the inset in Fig. 8b and c shows the band alignment diagrams derived from α-Fe₂O₃ and g-C₃N₄, respectively. With the use of the band arrangement diagrams obtained from α-Fe₂O₃ and g-C₃N₄, Fig. 8d illustrates the schematic of the CoSAs–g-C₃N₄/Fe₂O₃ band location indicating the formation of a typical type-II heterojunction between α-Fe₂O₃ and g-C₃N₄. Therefore, the potential water oxidation mechanism of CoSAs–g-C₃N₄/Fe₂O₃ is illustrated in Fig. 8e. Upon contact between α-Fe₂O₃ and g-C₃N₄, band bending occurred, thereby establishing the Fermi level equilibrium, and forming heterojunction. Upon illumination, the photoelectrons of α-Fe₂O₃ and g-C₃N₄ were excited from the valence band (VB) to the conduction band (CB), respectively. In the presence of the built-in electric field formed at the interface between α-Fe₂O₃ and g-C₃N₄, photogenerated charges underwent rapid separation and transfer, effectively inhibiting electron–hole pairs recombination. Photoelectrons transitioned from the CB of g-C₃N₄ to the CB of α-Fe₂O₃ and then traveled through the external circuit to the Pt electrode, facilitating the reduction of water to produce hydrogen. Furthermore, the VB holes in α-Fe₂O₃ rapidly migrated to the VB of g-C₃N₄, where they were further captured by the highly dispersed CoSAs. Owing to the strong oxidation of these photogenerated holes, Co^δ+ can be oxidized to Co²⁺/Co³⁺, which oxidized water to produce oxygen. Moreover, the reduction of Co²⁺/Co³⁺ ions to Co^δ+, enabled the efficient separation of photogenerated electrons and holes.

Fig. 8 (a) Calculated band gap values (based on UV-Vis absorption spectra) of α-Fe₂O₃ and g-C₃N₄. Ultra UV photoelectron spectroscopy (UPS) and work function of (b) α-Fe₂O₃ and (c) g-C₃N₄; (d) energy band diagram of the g-C₃N₄/Fe₂O₃ heterojunction; (e) Schematic illustration of electron–hole separation of CoSAs–g-C₃N₄/Fe₂O₃ photoanode for PEC water splitting.

Conclusions

We successfully prepared a novel CoSAs–g-C₃N₄/Fe₂O₃ photoanode with excellent PEC performance by coupling g-C₃N₄ with α-Fe₂O₃ and anchoring CoSAs on g-C₃N₄. Compared with pure α-Fe₂O₃, the g-C₃N₄/Fe₂O₃ heterostructure significantly inhibited the recombination of photogenerated charges. The incorporation of CoSAs into the photoanode structure further improved the charge separation efficiency and injection efficiency of α-Fe₂O₃, thereby promoting reaction kinetics. The synergistic effect of Co SAs and g-C₃N₄ contributed to the excellent PEC performance of CoSAs–g-C₃N₄/Fe₂O₃. The optimized CoSAs–g-C₃N₄/Fe₂O₃ photoanode exhibited a maximum photocurrent density of 1.93 mAcm⁻² at 1.23 V_RHE, which was 3.22 times that of pure α-Fe₂O₃, with a negative shift in initial potential by 195 mV. This study provides a promising approach for developing efficient and stable single atom photoanodes for PEC water splitting applications.

Data availability

All experimental details and characterisation data can be found in the ESI.†

Author contributions

J. W., L. J. and W. L. designed the experiment and wrote the manuscript. J. W., B. H., H. D. and N. W. conducted the experiments. J. W., X. D., M. L. and H. C. analysed data. All authors discussed the results at all stages and participated in the development of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

The authors are grateful for the financial support from the Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant No. ZKNUC2020042), Key Scientific and Technological Project of Henan Province (Grant No. 232102240044, 242102230101), Natural Science Foundation Project of Henan Province (Grant No. 232300420400).

References

1. Z. Peng, Y. Su, I. Ennaji, A. Khojastehnezhad and M. Siaj, Chem. Eng. J., 2023, 477, 147082.
2. W. Bai, H. Li, G. Peng, J. Wang, A. Li and P. F.-X. Corvini, Appl. Catal., B, 2024, 352, 124023.
3. R. T. Gao, L. Liu, Y. Li, Y. Yang, J. He, X. Liu, X. Zhang, L. Wang and L. Wu, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2300493120.
4. R. T. Gao, T. N. Nguyen, T. Nakajima, J. He, X. Liu, X. Zhang, L. Wang and L. Wu, Sci. Adv., 2023, 9, eade4589.
5. L. Wang, J. Zhu and X. Liu, ACS Appl. Mater. Interfaces, 2019, 11, 22272–22277.
6. J.-B. Pan, X. Liu, B.-H. Wang, Y.-A. Chen, H.-Y. Tan, J. Ouyang, W. Zhou, S. Shen, L. Chen, C.-T. Au and S.-F. Yin, Appl. Catal., B, 2022, 315, 121526.
7. L. K. Dhandole, T. S. Koh, P. Anushkkaran, H.-S. Chung, W.-S. Chae, H. H. Lee, S. H. Choi, M. Cho and J. S. Jang, Appl. Catal., B, 2022, 315, 121538.
8. J. Bai, R.-T. Gao, X. Guo, J. He, X. Liu, X. Zhang and L. Wang, Chem. Eng. J., 2022, 448, 137602.
9. H. Chai, S. Wang, X. Wang, J. Ma and J. Jin, ACS Catal., 2022, 12, 3700–3709.
10. H. Chai, L. Gao, P. Wang, F. Li, G. Hu and J. Jin, Appl. Catal., B, 2022, 305, 121011.
11. Y. Wang, X. Huang and Z. Wei, Chin. J. Catal., 2021, 42, 1269–1286.
12. L. Zhang, R. Long, Y. Zhang, D. Duan, Y. Xiong, Y. Zhang and Y. Bi, Angew. Chem., Int. Ed., 2020, 59, 6224–6229.
13. T. L. L. Doan, D. C. Nguyen, S. Prabhakaran, D. H. Kim, D. T. Tran, N. H. Kim and J. H. Lee, Adv. Funct. Mater., 2021, 31, 2100233.
14. Y. Sun, H. Li, Y. Hu, J. Wang, A. Li and P. F.-X. Corvini, Appl. Catal., B, 2024, 340, 123269.
15. X. Liu, Y. Deng, L. Zheng, M. R. Kesama, C. Tang and Y. Zhu, ACS Catal., 2022, 12, 5517–5526.
16. D. A. Kuznetsov, Z. Chen, P. V. Kumar, A. Tsoukalou, A. Kierzkowska, P. M. Abdala, O. V. Safonova, A. Fedorov and C. R. Muller, J. Am. Chem. Soc., 2019, 141, 17809–17816.
17. P. Liao, J. A. Keith and E. A. Carter, J. Am. Chem. Soc., 2012, 134, 13296–13309.
18. J. Büker, X. Huang, J. Bitzer, W. Kleist, M. Muhler and B. Peng, ACS Catal., 2021, 11, 7863–7875.
19. Y. Zheng, Q. Ruan, J. Ren, X. Guo, Y. Zhou, B. Zhou, Q. Xu, Q. Fu, S. Wang and Y. Huang, Appl. Catal., B, 2023, 323, 122170.
20. S. Xin, X. Ma, J. Lu, G. Zhang, S. Huo, M. Gao, P. Xu, W. Liu and W. Fu, Appl. Catal., B, 2023, 323, 122174.
21. J. Wu, M. Qi, G. Wang, B. Yu, C. Liu, W. Hou and W. Liu, ACS Sustainable Chem. Eng., 2020, 8, 5200–5208.
22. J. Wu, M. Meng, X. D. Du, M. Li, L. Jin and W. Liu, Inorg. Chem., 2024, 63, 6192–6201.
23. S.-S. Yi, J.-M. Yan and Q. Jiang, J. Mater. Chem. A, 2018, 6, 9839–9845.
24. G. Yang, Y. Li, H. Lin, X. Ren, D. Philo, Q. Wang, Y. He, F. Ichihara, S. Luo, S. Wang and J. Ye, Small Methods, 2020, 4, 202000577.
25. X. Zhou, T. Wang, L. Zhang, S. Che, H. Liu, S. Liu, C. Wang, D. Su and Z. Teng, Appl. Catal., B, 2022, 316, 121614.
26. Z. Masoumi, M. Tayebi, M. Kolaei, A. Tayyebi, H. Ryu, J. I. Jang and B. K. Lee, ACS Appl. Mater. Interfaces, 2021, 13, 39215–39229.
27. J. Wu, P. Huang, H. Fan, G. Wang and W. Liu, ACS Appl. Mater. Interfaces, 2020, 12, 30304–30312.
28. T. Wang, X. Long, S. Wei, P. Wang, C. Wang, J. Jin and G. Hu, ACS Appl. Mater. Interfaces, 2020, 12, 49705–49712.
29. J. Lin, X. Han, S. Liu, Y. Lv, X. Li, Y. Zhao, Y. Li, L. Wang and S. Zhu, Appl. Catal., B, 2023, 320, 121947.
30. K. Kang, C. Tang, J. H. Kim, W. J. Byun, J. H. Lee, M. H. Lee, H. Zhang and J. S. Lee, ACS Catal., 2023, 13, 7002–7012.
31. S.-S. Yi, B.-R. Wulan, J.-M. Yan and Q. Jiang, Adv. Funct. Mater., 2019, 29, 1801902–1801910.
32. X. Liu, L. Zheng, C. Han, H. Zong, G. Yang, S. Lin, A. Kumar, A. R. Jadhav, N. Q. Tran, Y. Hwang, J. Lee, S. Vasimalla, Z. Chen, S. G. Kim and H. Lee, Adv. Funct. Mater., 2021, 31, 2100547.
33. F. Li, J. Li, L. Gao, Y. Hu, X. Long, S. Wei, C. Wang, J. Jin and J. Ma, J. Mater. Chem. A, 2018, 6, 23478–23485.
34. R. Chong, Y. Du, Z. Chang, Y. Jia, Y. qiao, S. Liu, Y. Liu, Y. Zhou and D. Li, Appl. Catal., B, 2019, 250, 224–233.
35. Y. Zhang, Y. Huang, S. S. Zhu, Y. Y. Liu, X. Zhang, J. J. Wang and A. Braun, Small, 2021, 17, 2100320.
36. J. Liu, W. Chen, Q. Sun, Y. Zhang, X. Li, J. Wang, C. Wang, Y. Yu, L. Wang and X. Yu, ACS Appl. Energy Mater., 2021, 4, 2864–2872.
37. S. Zhang, Z. Liu, D. Chen and W. Yan, Appl. Catal., B, 2020, 277, 119197.
38. J. H. Kim and J. S. Lee, Adv. Mater., 2019, 31, 1806938.
39. G. Yang, S. Li, X. Wang, B. Ding, Y. Li, H. Lin, D. Tang, X. Ren, Q. Wang, S. Luo and J. Ye, Appl. Catal., B, 2021, 297, 120268.
40. H. Zhang, D. Li, W. J. Byun, X. Wang, T. J. Shin, H. Y. Jeong, H. Han, C. Li and J. S. Lee, Nat. Commun., 2020, 11, 4622.
41. R. T. Gao, J. Zhang, T. Nakajima, J. He, X. Liu, X. Zhang, L. Wang and L. Wu, Nat. Commun., 2023, 14, 2640.

---

Footnote

† Electronic supplementary information (ESI) available: Supplementary figures; LSV, XRD, XPS, EXAFS, PL, UV-Vis, the calculated current density flux and integrated current density (J_abs), dielectric loss, and leakage current density, EXAFS fitting parameters at the Co K-edge for various samples, fitting data of the EIS. See DOI: https://doi.org/10.1039/d4sc03442b

---