Ultrasonication-assisted liquid-phase exfoliation enhances photoelectrochemical performance in α-Fe₂O₃/MoS₂ photoanode

Graphical abstract

Highlights

• •Ultrasonication-assisted synthesis of MoS₂ nanosheets improved PEC performance.
• •Liquid phase exfoliation was used for MoS₂ nanosheets with 6 nm thickness.
• •Photocurrent of α-Fe₂O₃ photoanode significantly increased by MoS₂ heterojunction.
• •α-Fe₂O₃/8-MoS₂ showed a thinner space charge layer and reduced flat band potential.
• •p-n junction decreased resistance due to the reduced recombination rate of charge carriers.

Abstract

This study successfully manufactured a p-n heterojunction hematite (α-Fe₂O₃) structure with molybdenum disulfide (MoS₂) to address the electron–hole transfer problems of conventional hematite to enhance photoelectrochemical (PEC) performance. The two-dimensional MoS₂ nanosheets were prepared through ultrasonication-assisted liquid-phase exfoliation, after which the concentration, number of layers, and thickness parameters of the MoS₂ nanosheets were respectively estimated by UV–vis, HRTEM and AFM analysis to be 0.37 mg/ml, 10–12 layers and around 6 nm. The effect of heterojunction α-Fe₂O₃/MoS₂ and the role of the ultrasonication process were investigated by the optimized concentration of MoS₂ in the forms of bulk and nanosheet on the surface of the α-Fe₂O₃ electrode while measuring the PEC performance. The best photocurrent density of the α-Fe₂O₃/MoS₂ photoanode was obtained at 1.52 and 0.86 mA.cm⁻² with good stability at 0.6 V vs. Ag/AgCl under 100 mW/cm² (AM 1.5) illumination from the back- and front-sides of α-Fe₂O₃/MoS₂; these values are 13.82 and 7.85-times higher than those of pure α-Fe₂O₃, respectively. The results of electrochemical impedance spectroscopy (EIS) and Mott-Schottky analysis showed increased donor concentration (2.6-fold) and decreased flat band potential (by 20%). Moreover, the results of IPCE, ABPE, and OCP analyses also supported the enhanced PEC performance of α-Fe₂O₃/MoS₂ through the formation of a p–n heterojunction, leading to a facile electron–hole transfer.

1. Introduction

The immense global consumption of fossil fuels has led to serious environmental impacts around the world. Many studies have aimed to address these issues by investigating the production of renewable energy resources such as hydrogen gas by using photoelectrochemical (PEC) water splitting technology [1], [2], [3], [4], [5], [6]. This emerging technology can produce more renewable energy by directly harvesting solar energy, leading to reduced air emissions from greenhouse gases. PEC water splitting requires enough energy to cause a reaction in semiconductor materials with valence and conduction band positions that are appropriate for generating hydrogen and oxygen [7], [8]. Semiconductors for PEC should be efficient, environmentally friendly, and nontoxic end products, and they should enable cost-effective energy conversion [9], [10], [11], [12], [13], [14]. Hematite (α-Fe₂O₃) satisfies all these requirements and also has a proper band gap (E_g = 2.1 eV) making it a promising candidate for water oxidation [15], [16], [17], [18], [19], [20]. However, bare α-Fe₂O₃ has some limitations in terms of PEC performance, including fast electron-hole recombination, short hole diffusion length (typically 2–4 nm), and low conductivity [2], [21], [22], [23]. Various methods can be used to overcome these drawbacks, such as the application of doping elements to improve electron transport [24] the inclusion of a heterojunction with other semiconductors to enhance charge separation [25], [26], [27], [28], [29] the use of nanostructures to generate more electron–hole pairs, the addition of co-catalysts on the surface to promote charge migration, and other methods to control morphology [30], [31], [32]. The p–n junction between a p-type and an n-type semiconductor can enhance the efficiency of the photoanode by not only decreasing the charge recombination rate but also improving the electron–hole separation. This heterojunction can reduce the recombination rate by facilitating electron and hole transfer via an internal electric field [30], [33]. The surface-to-volume ratio of a material specifies how much of it is exposed to the environment which is a crucial factor that significantly affects chemical reactions. With a higher surface area-to-volume ratio, more reactant can be in contact with the material, ultimately leading to a faster reaction. Two-dimensional (2D) materials tend to be more reactive to their bulk counterparts, and as they contain unique properties, they have the potential to reveal intriguing new phenomena [34], [35], [36]; transition-metal dichalcogenides (TMDs) are one of the most noteworthy examples of the type of material. In the type MX₂, M is a transition metal atom and X is a chalcogen atom. The unique combination of a direct bandgap, favorable mechanical and electronic properties, and strong spin–orbit coupling makes them more attractive for applications in optoelectronics, electronics, and energy harvesting materials [37], [38]. Molybdenum Disulfide (MoS₂) is one of the most attractive and widely used TMDs; it has desirable properties, such as a low coefficient of friction, high chemical stability, and thermal stability, and can thus enhance visible light absorption with a proper band edge [39], [40]. The discovery of ultrasonication-mediated liquid-phase exfoliation techniques for the separation of single layers in 2D materials has made it possible to study the properties of MoS₂ flakes. When changing from a bulk material to a 2D material, the bandgap length is modified from an indirect bandgap of 1.4 eV to a direct bandgap of 2 eV. This change in the bandgap makes it an extremely interesting material for a wide range of applications [41], [42], [43], [44], [45]. The use of MoS₂ could also improve the separation of photogenerated electron-hole pairs in the α-Fe₂O₃ photoanode [46], [47].

In this study, we synthesized MoS₂ nanosheets using a liquid exfoliation method. A p–n junction of α-Fe₂O₃/MoS₂ was fabricated on FTO substrate with two simple hydrothermal and drop casting processes. Further, the effect of the ultrasonication process on PEC performance was investigated for α-Fe₂O₃/MoS₂ electrodes prepared with and without the ultrasonication process. The best PEC performance of α-Fe₂O₃/MoS₂ was achieved by optimizing the amount and the thickness of the MoS₂ in the p–n heterojunction. We confirmed that the p–n junction effect of MoS₂ in α-Fe₂O₃ improved the PEC performance. The α-Fe₂O₃/MoS₂ heterojunction had 240% improved PEC efficiency compared to pristine α-Fe₂O₃, which can be attributed to the enhanced charge transfer to the electrolyte solution caused by the reduced charge recombination rate. The analysis results reveal that this heterojunction increased the donor concentration (N_D) and decreased the flat band potential (V_fb), space charge layer (W_SCL), and charge transfer resistance (R_ct).

2. Experimental section:

2.1. Chemicals and reagents:

Iron (III) chloride hexahydrate (FeCl₃·6H₂O), Sodium nitrate (NaNO₃), Hydrochloric acid (HCl), and ethanol (C₂H₅OH) were purchased from Daejung Chemical & Metal Co., Ltd (Korea). Molybdenum (IV) sulfide (bulk of MoS₂) was procured from Sigma Aldrich. Deionized water was used to prepare all aqueous solutions.

2.2. Preparation of α-Fe₂O₃ photoanode

The PEC cells were manufactured using FTO (Fluorine Doped Tin Oxide Coated Glass, 25 mm × 25 mm × 2.2 mm, ~7 O/sq) as the substrate and washed with deionized water, ethanol, and acetone in equal volumes for 15 min, then dried for subsequent use. Hematite (α-Fe₂O₃) thin films were manufactured through a hydrothermal method, typically using an aqueous solution of 1 M sodium nitrate (NaNO₃) and 0.15 M ferric chloride (FeCl₃·6H₂O) as precursor. After mixing these solutions, hydrochloric acid (HCl) was added dropwise to regulate the pH of the mixture to 1.5. The cleaned FTO was placed at the bottom of a Teflon container, and 80 ml of the solution was added into this container. Then, the autoclave was placed in an oven at 100 °C for 6 h. After the autoclave had cooled down, the coated FTO were taken out, and the residues on the surfaces of the samples were washed off with deionized water. The yellowish layer that was coated on the FTO was β-FeOOH. To transform the β-FeOOH into α-Fe₂O₃ and to prepare the α-Fe₂O₃ photoanode, the FTO coated with β-FeOOH was placed in a furnace in air at 550 °C for 4 h, as shown in Fig. 1.

Fig. 1.

Preparation processes of pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂ photoanodes.

2.3. Preparation of 2D-MoS₂ nanosheet by ultrasonication

The MoS₂ nanosheets were synthesized via a liquid exfoliation method, as shown in a schematic in Fig. 1. First, 300 mg of MoS₂ powder was added into the solution prepared by combining 45 ml Ethanol with 55 ml water as a solvent. The exfoliated MoS₂ nanosheets precursor was sonicated continuously for five days. The dispersions were centrifuged at 3500 rpm for 60 min to separate MoS₂ nanosheets. Then, the supernatant containing thin MoS₂ nanosheets were collected from the top of the solution.

2.4. Preparation of α-Fe₂O₃/MoS₂ thin film

The p-n heterojunction between these two semiconductors of α-Fe₂O₃ and MoS₂ was manufactured through a simple and low-cost drop casting method. We dropped 100 μl of liquid exfoliation MoS₂ nanosheets solution on to the surfaces of the α-Fe₂O₃ photoanodes_, then placed them in a furnace for 5 min at 450 °C. We repeated the drop casting process under the same conditions four, eight, and 12 times, and the resulting samples were labeled α-Fe₂O₃/4-MoS_2, α-Fe₂O₃/8-MoS₂, and α-Fe₂O₃/12-MoS₂, respectively. Finally, all the heterojunction samples of the α-Fe₂O₃/MoS₂ photoanodes were calcined in a furnace for 2 h at 450 °C.

2.5. Characterization measurement

The morphologies and structures of the samples were characterized by a scanning electron microscope (SEM, Model Quanta 250 FEG) and a transmission electron microscope (TEM, JEOL, JEM-2100F). Structure and crystallinities of the different catalytic films were characterized by X-ray diffraction (XRD). The X-ray source was Bruker D8Advance with monochromatic Cu Kα radiation (λ = 1.5406 Å) in the 2θ range of 10–55°. Raman spectra was used with the laser line at 785 nm as the excitation source at room temperature (Bruker, model: Senteraa 2009, Germany). The chemical states of the component elements in the film samples were investigated using a Thermo Scientific Sigma Probe spectrometer with a monochromatic AlKα source (photon energy 1486.6 eV), a spot size of 400 μm, an energy step size of 1.0 eV, and a pass energy of 200 eV. The optical properties of the photoanodes were measured by Perkin Elmer UV–Vis-NIR model Lambda 950 and photoluminescence (PL) spectra (Edinburgh F-4600 NF900 (FLS920) fluorescence spectrophotometer) at 400 nm excitation.

2.6. Photoelectrochemical (PEC) measurement

A standard three-electrode structure was used for PEC measurements. The pure α-Fe₂O₃ and modified α-Fe₂O₃/MoS₂ heterojunction photoelectrodes were used as working electrodes, a platinum wire was used as a counter electrode, and Ag/AgCl was used as a reference electrode in 1 M NaOH electrolyte. The PEC performance was measured under 100 mW/cm² (AM 1.5) an illumination from a 300 W Xe lamp from the front- and back-sides of the photoanodes in the voltage range of −0.7 to 0.7 V (vs. Ag/AgCl). EIS was measured with the usage of potentiate by the identical electrode formation. IPCE measurements were also taken using different filters in this system at 0.6 V (vs. Ag/AgCl) under 100 mW.cm⁻² (AM 1.5) illumination to investigate the conversion ratio of the incident photon to electron.

3. Results and discussion

3.1. Structural characterization

One method used to investigate the van der Waals interactions in 2D materials is atomic force microscopy (AFM). The optical absorption of MoS₂ can be seen from the near-infrared (NIR) to the ultraviolet (UV) region because of its electronic band structure. Points A, B, and C shown in Fig. 2a imply the existence of MoS₂ nanosheets. The accumulation of layers slightly influences d-orbitals in the metal atom and causes shifts in the A and B exciton peaks. By increasing the number of layers, the energy difference between excitons B and A increases (ΔE = E_B − E_A) [48]. The C peak occurs in various energies for monolayered versus multilayered materials. The C transitions are attributed to the van der Waals interactions, which require the p-orbitals of the S atoms. In our work, the C peak in MoS₂ is centered at ~397 nm, while the B and A transitions are at ~610 nm and ~670 nm, respectively, with a difference of ~59 nm between their excitons.

Fig. 2.

(a) UV–Vis and (b) AFM images and height profiles for MoS₂ nanosheets.

These results indicate a significant contribution of the few-layered nanosheets [1], [47]. The thicknesses of the MoS₂ nanosheets are estimated using Eq. (1) [49].

$$\mathrm{T}\mathrm{h}\mathrm{i}\mathrm{c}\mathrm{k}\mathrm{n}\mathrm{e}\mathrm{s}\mathrm{s}(\mathrm{N}(\mathrm{n}\mathrm{m}))=2.3\times {10}^{36}\times exp(-58444/{\lambda }_A)$$ (1)

In this equation, λ_A is the wavelength of the absorbance light at point A in Fig. 2a. The thickness of the MoS₂ nanosheets exfoliated by an ultrasonication process was find to be around 6 nm, which represents the few layers of nanosheets used.

The AFM results, specifically the deposited height measurement, can provide a further estimate of the number of MoS₂ nanosheets layers. The average thickness measured from the height profile diagram was estimated to be approximately 4.5–6 nm (Fig. 2b). Further, the length (Eq. 1 in the Supporting information) and concentration (Eq. 2 in the Supporting information) of the MoS₂ nanosheets on each photoanode were estimated to be 104.74 nm and 0.37 mg/ml, respectively (details in the Supporting information).

Fig. 3a shows TEM images of typical MoS₂ nanosheets prepared using a sonication process in the liquid exfoliation method with a lateral size of 100–150 nm. Fig. 3b and c shows HRTEM images of the MoS₂ layers, which are well accumulated with an interlayer interval of 0.26 nm, corresponding to the (1 0 0) lattice planes of the hexagonal MoS₂ phase.

Fig. 3.

(a and b) HRTEM images of MoS₂ nanosheets exfoliated by ultrasonication process. (c and d) High-resolution TEM images of 10–12 layer MoS₂ nanosheet. The inset in (d) is the corresponding SAED pattern. (e) Schematic diagram showing the role of ultrasonication in the exfoliation of MoS₂.

HRTEM analysis in the boundary areas is a common method used to directly distinguish the numbers of layers of the MoS₂ nanosheet [50], [51]. As shown in Fig. 3c, 10–12 dark and bright patterns can be seen for the exfoliated MoS₂ nanosheet, indicating that the MoS₂ was accumulated with 10–12 single layers, which is consistent with the AFM results.

The high-resolution TEM image in Fig. 3d shows that the lattice structure of the MoS₂ nanosheet was not damaged during the sonication process. The electron diffraction pattern shown in the inset in Fig. 3d indicates that the MoS₂ nanosheet has an acceptable crystallinity for the synthesis process.

The role of ultrasonication on the liquid-phase exfoliated MoS₂ nanosheets is clearly demonstrated in Fig. 3e. The direct ultrasonication of bulk MoS₂ in organic solvents or aqueous surfactant solutions can be used to avoid changes in the chemical and electronic properties of the MoS₂ nanosheets while maintaining the advantages of liquid-phase exfoliation. Ultrasonication with sound waves produces cavities or tiny bubbles in the liquid. With continues application, these bubbles implode, thus giving off a burst of energy, which is then used for exfoliation of the MoS₂ nanosheets. When the MoS₂ is exfoliated, the organic solvent molecules (or the surfactant molecules, in case of an aqueous dispersion) stabilize the MoS₂ nanosheets, in turn preventing reaggregation. This means that the interaction between MoS₂ nanosheets and solvent molecules is stronger than the interlayer forces present between the MoS₂ sheets of the bulk crystal [52], [53], [54].

The morphologies of the pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂ electrodes were analyzed by SEM, and the results are shown in Fig. 4(a–d). The optimized sample of the heterojunction α-Fe₂O₃/MoS₂ confirms that it has the same morphology as the pure α-Fe₂O₃ sample. The α-Fe₂O₃/8-MoS₂ nanorods in Fig. 4g are denser, and they appear as rice-shaped rods that are both longer and wider [55]. As shown in Fig. 4c and d, the thicknesses of the pristine α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ thin films are around 430 and 500 nm, respectively. The concentration of MoS₂ on the surface of the thin films affects the light harvesting of the electrodes. Moreover, concentration is an important parameter reflecting charge separation and migration in PEC efficiency. The photocurrent density of the α-Fe₂O₃/MoS₂ thin films decreases when loading more MoS₂ on the surface of the photoelectrode in question. In the high resolution TEM image of the α-Fe₂O₃ electrode shown in Fig. 4e, the identified lattice space is 0.23 nm, which is attributed to the lattice plate (1 1 0); this was also detected in the XRD analysis results (see Fig. 5a). The thickness of the MoS₂ increased with the dropping of more MoS₂ nanosheets on the surface of α-Fe₂O₃ photoanodes with increased repetitions of drop casting from four to eight to 12 times, respectively denoted as 4-MoS₂, 8-MoS₂, and 12-MoS₂ (see Table 1). The evolutions of 0.7 nm, 3.33 nm, and 7.6 nm of the MoS₂ nanosheets on the α-Fe₂O₃ surface were detected in α-Fe₂O₃/4-MoS₂ (Fig. 4f), α-Fe₂O₃/8-MoS₂ (Fig. 4g), and α-Fe₂O₃/12-MoS₂ (Fig. 4h), respectively. We found that the eight times-repeated drop casting of MoS₂ on the α-Fe₂O₃ electrode (α-Fe₂O₃/8-MoS₂) produced a thickness of around 3.33 nm and also exhibited the best photocurrent density and PEC performance of all the samples (see Fig. 7a).

Fig. 4.

Top-view (a and b) and cross-sectional (c and d) FE-SEM images of pure α-Fe₂O₃ (a and c) and α-Fe₂O₃/MoS₂ (b and d). HRTEM images of (e) α-Fe₂O₃, (f) α-Fe₂O₃/4-MoS₂, (g) α-Fe₂O₃/8-MoS₂, and (h) α-Fe₂O₃/12-MoS₂. The thickness of the MoS₂ layer was controlled by changing the time spent on the drop casting of MoS₂ nanosheets with the same concentration on the FTO/α-Fe₂O₃ electrode.

Fig. 5.

(a) X-Ray diffraction peaks and (b) Raman spectra and XPS spectra for (c) Fe 2p (d) O1s (e) Mo3d, and (f) S2p for pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂.

Table 1.

Amount of MoS₂ nanosheets on the photoelectrodes.

Samples name	Time of drops	Volume of MoS2 precursor (0.1 ml per each drop cycle)	Concentration of MoS2 nanosheets in precursor (mg/ml)	Concentration of MoS2 nanosheets in each photoelectrode (mg/ml)	Amount of MoS2 nanosheets in photoelectrode’s area (mg/cm2) (FTO area = 25 mm × 25 mm)
α-Fe2O3-4-MoS2	4	4 × 0.1 = 0.4	0.37	0.148	0.023
α-Fe2O3-8-MoS2	8	8 × 0.1 = 0.8	0.37	0.296	0.47
α-Fe2O3-12-MoS2	12	12 × 0.1 = 1.2	0.37	0.44	0.71

Fig. 7.

(a) Linear scan voltammetry, (b) photocurrent response, (c) photocurrent stability at 0.6 V (vs. Ag/AgCl), (d) photocurrent density vs. potential linear of α-Fe₂O₃ and α-Fe₂O₃/MoS₂ at potentials from −0.6 to 0.6 V (vs. Ag/AgCl) under 100 mW cm⁻² back- and front-side illumination, and (e) schematic of the pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂ light illumination sides. The electrolyte was a 1 M NaOH.

X-ray diffraction (XRD) analysis was used to investigate the crystallographic structures of the α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ electrode, and the results are shown in Fig. 5. Since FTO was used as the substrate, its XRD peaks were detected in all samples (PDF 99-0024). The peaks around 34, 37, 62, and 64° respectively corresponded to the 1 0 4, 1 1 0, 2 1 4, and 3 0 0 planes of the α-Fe₂O₃ structure (PDF 02-0915) in both samples. These results confirm that there was no restructuring in the α-Fe₂O₃ photoanode after the addition of the p-n junction with MoS₂ nanosheets. The diffraction peaks at 14.4° and 33.5°,which respectively corresponded to the (0 0 2) and (1 0 1) planes of MoS₂ (PDF 73-1508), prove the presence of MoS₂ coated on the α-Fe₂O₃ film as heterojunction layers [46], [56].

The Raman spectra of the photoanodes display a dominant band at about 1318 cm⁻¹, which corresponds to the existence of hematite in both structures (Fig. 5b). The band at (148,383, 403 cm⁻¹) further supports the formation of dual α-Fe₂O₃/MoS₂ nanocomposites. In addition, the Raman plot around 1318 cm⁻¹ shifts to a higher wavelength after the loading of MoS₂. The red shift for α-Fe₂O₃/8-MoS₂ can be observed, which is attributed to the improved crystallinity of the material, therefore indicating that the material lattice was not compressed [57].

X-ray photoelectron spectroscopy (XPS) spectra were taken to better investigate the elemental compositions of the α-Fe₂O₃/MoS₂ thin films (Fig. 5c–f). The survey spectrum indicates that Fe, O, and Mo elements coexisted on the surface, while the C 1s peak at 284.8 eV is attributed to adventitious carbon from the XPS equipment (see Fig. S2). The high resolution XPS spectrum with all deconvoluted peaks of Fe 2p shows two peaks at 709.43 and 723.72 eV, which are respectively attributed to Fe 2p_3/2 and 2p_1/2, and these are characteristic of Fe²⁺ in α-Fe₂O₃. Moreover, the distance of 13.8 eV between Fe 2p_1/2 and Fe 2p_3/2 represented the Fe³⁺ ions in α-Fe₂O₃ [58]. Fig. 5d shows the binding energies of the two distinguishable diffraction O 1s peaks at 529.97 eV (Fe–O) and 532.04 eV (–OH). The oxygen peaks in the α-Fe₂O₃/MoS₂ sample are higher than those in pure α-Fe₂O₃, which may indicate the presence of MoO₃ as a side product of the formation of MoS₂ on the α-Fe₂O₃. Two peaks are shown in the modified sample at 232.46 and 235.54 eV, which respectively correspond to Mo 3d_3/2 and Mo 3d_5/2 and confirm the existence of Mo⁶⁺ cations in MoS₂. The energy separation between Mo 3d_5/2 and Mo 3d_3/2 is estimated to be 3.08 eV (<3.3 eV), which is indicative of the presence of MoS₂ [59].

S 2p_1/2 and 2p_3/2 peaks appear at around 164 and 163 eV, respectively. In a higher binding energy, the SO₄²⁻ peak shows MoO₃, which formed as a side product during the calcination process of MoS₂ on the hematite substrate [43]. Thus, the XPS data further confirm the existence of MoS₂ in α-Fe₂O₃/MoS₂ film, which is consistent with the XRD and EDX results (see Fig. S1).

Fig. 6a shows the optical properties and band-gap energies of α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂. To evaluate the effect of MoS_2, the light absorption ability and photon-to-charge conversion efficiency of the α-Fe₂O₃-based samples are considered. From the Tauc Plots, we can estimate the band gaps of the prepared α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ films.

$$\text{F}={\left(1-\text{R}\left(\text{h}v\right)\right)}^2/2\text{R}$$ (2)

Fig. 6.

(a) UV–Vis reflectance spectra and band gap energies, (b) photoluminescence (PL) spectrum analysis for α-Fe₂O₃ and α-Fe₂O₃α-Fe₂O₃/MoS₂, and (c) mechanism for the p–n heterojunction of α-Fe₂O₃/MoS₂.

The band gaps [60] estimated using Eq. (2) for α-Fe₂O₃ and α-Fe₂O₃/MoS₂ films were 2.09 eV and 2.05 eV, respectively. These results confirm that the α-Fe₂O₃/8-MoS₂ electrode is a suitable photocatalytic material due to its high visible light response [58].

Photoluminescence (PL) spectrum analysis can be used to investigate the charge trapping and recombination rate for photogenerated electron-hole pairs in the photocatalyst. The PL intensity of α-Fe₂O₃/MoS₂ thin film is less than that of pure α-Fe₂O₃ between 420 nm and 600 nm (Fig. 6b). These data indicate the fast electron–hole transfer between α-Fe₂O₃ and MoS₂, which is attributed to the p–n junction present in the α-Fe₂O₃/MoS₂ photoelectrode [30].

Fig. 6c shows the electron–hole transfer mechanism of the p–n heterojunction in the α-Fe₂O₃/MoS₂ photoelectrode. The band bending achieved equilibrium of the Fermi level when MoS₂ was in contact with α-Fe₂O₃ in the p-n junction. Following visible light illumination, the electrons obtained from electron-hole separation were transferred from the valence band (VB) to the conduction band (CB) of α-Fe₂O₃ and MoS₂. Due to the difference in energy levels, an electron was transferred from the CB of MoS₂ to the CB of α-Fe₂O₃. The holes were simultaneously driven by the electrostatic field from the VB of α-Fe₂O₃ to the VB of MoS₂, and they reacted with OH to generate O₂ gas. The electrons were led to the Pt electrode via the FTO substrate, and this process produced hydrogen gas. Hence, this inner electrostatic field formed in the p-n heterojunction can increase the extraction of the photo-induced carriers at the space charge region of the p-n junction, and can also gain more O₂ and H₂ generation by facilitating the electron-hole separation pathway.

3.2. Effect of MoS₂ nanosheets on the PEC activity of α-Fe₂O₃ photoanode

Linear scan voltammetry (LSV) and chopped LSV of the pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂ electrodes were taken under continuous and on-off cycling between −0.6 and 0.6 V vs. Ag/AgCl in 1 M NaOH solution (pH ~ 12), as shown in Figs. 7a and S3, respectively. After starting light illumination, the photoanode photocurrent density reached ~0.11 mA cm⁻² at 0.6 V vs. Ag/AgCl for the pure α-Fe₂O₃ electrode. The photocurrent density significantly increased after the deposition of MoS₂ nanosheets with the highest value of 0.63 mA cm⁻² obtained at 0.6 V vs. Ag/AgCl for α-Fe₂O₃/8-MoS₂. This behavior could be attributable to the heterojunction between MoS₂ and α-Fe₂O₃, which improved the charge separation and reduced the recombination of the photogenerated electrons and holes. However, further increases the amount of MoS₂ loaded on the surface of α-Fe₂O₃ gradually decreased the photocurrent density, because excessive MoS₂ may block the light absorption of α-Fe₂O₃ [58], [59]. Moreover, after the deposition of MoS₂ nanosheets, the onset potential significantly shifts from 0.25 V for α-Fe₂O₃ to a lower applied voltage of 0.1 V. Further, the chopped plot shown in Fig. S3 is consistent with the photocurrent density of LSV presented in Fig. 7a. This indicates that the α-Fe₂O₃ photoanode underwent a quick reaction to the light and promptly generated photocurrent from zero to its equilibrium value. Chronoamperometry scans were measured for 10 cycles (Fig. 7b) under intermittent irradiation (30 min light on and 15 min light off) in 1 M NaOH solution, and these confirm the photocurrent responses. The fast and uniform reactions to each light on/off interval of the two electrodes reflect the good producibility of the samples. The pristine α-Fe₂O₃ has a low photocurrent, which implies a low quantum efficiency of α-Fe₂O₃. By contrast, the α-Fe₂O₃/MoS₂ heterostructure shows a higher photocurrent response, which in turn has a positive effect on the decreased recombination for photogenerated charges. As shown in Fig. 7c, the photocurrent stabilities of different α-Fe₂O₃/MoS₂ films were obtained by continuous illumination after 10 on-off cycles. The photoelectrodes with the p-n heterojunction showed stable photocurrent, even in 30-min light illumination, without any significant downturn.

The photocurrent density depends on many conditions, including the direction of light illumination; back-side illumination significantly increases photocurrent density compared to front-side illumination (Fig. 7d). The identified photocurrent for α-Fe₂O₃/8-MoS₂ thin film reached ~1.52 mA.cm⁻², which was 13.8 and 1.77 times more than that for pure α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ in front-side illumination, respectively. The schematic picture shown in Fig. 7e depicts the generation of electrons near the FTO, which led to better collection and transfer for electrons in the back-side illumination of α-Fe₂O₃/8-MoS₂. In addition, the absorption of photons was more effective for back-side photocurrent due to the reduced electron–hole recombination [61], [62].

The effect of MoS₂ thickness on the charge transfer process at the α-Fe₂O₃/MoS₂/electrolyte was elucidated using electrochemical impedance spectroscopy (EIS). The EIS was studied at the frequency range from 100 kHz to 0.1 Hz in a 1 M NaOH electrolyte in front of 100 mW/cm² (AM 1.5) illumination. The Nyquist plots (Fig. 8a) were fitted with the RC circuit modeling analysis results, which were obtained using ZSimpWin software, and the results are summarized in Table S1. As shown in Fig. 8b, the equivalent circuit includes the resistance for FTO/α-Fe₂O₃ (R_S), the space charge capacitance of the bulk α-Fe₂O₃ (C₁), the space charge capacitance of the α-Fe₂O₃/MoS₂/electrolyte or α-Fe₂O₃/electrolyte interface (C₂), the resistance of the inside of bulk α-Fe₂O₃ (R₁), and the charge transfer resistance at the α-Fe₂O₃/MoS₂/electrolyte (R₂). The onset point in the EIS plot shows the sheet resistance (R_S) in the PEC cells. The R_S values for α-Fe₂O₃/MoS₂ were unchanged after the addition of MoS₂ nanosheets. The α-Fe₂O₃/8-MoS₂ had the lowest R₁ and highest C₁ values among all samples. The fact that the α-Fe₂O₃/8-MoS₂ photoelectrode had the lowest R₁ value may be related to the improved extraction of holes from α-Fe₂O₃ due to the thin MoS₂ layer in the p–n heterojunction, which can decrease the recombination rate in α-Fe₂O₃/8-MoS₂ photoelectrodes. However, α-Fe₂O₃/12-MoS₂ showed the lowest C₁ and highest R₁ values, which confirms the effect of thickness on the recombination in the samples and these results are is consistent with the TEM results shown in Figs. 4e to 3h. The reaction among water and holes from the electrode caused water oxidation, and the resistance at the interface of the electrolyte/photoanode (R₂) is the highest resistance among all PEC cells. The resistance value at the interface (R₂) of α-Fe₂O₃/8-MoS₂ with the electrolyte was only 14% that of α-Fe₂O₃ (see Table S2). This significant reduction can be attributed to the reduced recombination rate and facile hole extraction through the p–n heterojunction of α-Fe₂O₃/8-MoS₂. The highest R₂ and lowest C₂ values of α-Fe₂O₃/12-MoS₂ with the thickness of 7.6 nm (see Fig. 4h) implies less water oxidation reaction because of the ineffective hole extraction through the amorphous thick MoS₂ layer. Thus, the resistance in the amorphous thick MoS₂ layer was increased with the addition of more MoS₂ layers. The reduced hole extraction suggests that the surface of the photoanode could have been blocked [63], [64].

Fig. 8.

(a) Nyquist plot, (b) equivalent circuit, (c) Mott–Schottky plot, (d) IPCE, (e) open circuit potential (ΔOCP), and (f) ABPE plots for pure α-Fe₂O₃ and α-Fe₂O₃/MoS₂ electrodes. The supporting electrolyte was a 1 M aqueous solution of NaOH.

Further, the donor concentration and the flat band potential, which are key factors of photoelectrode performance, were obtained from the Mott-Schottky plot shown in Fig. 8c. These represent the charge transport time factor and the electron mobility corresponding to the electron recombination. The Mott-Schottky data were plotted using Eq. (3) [65], [66]:

$$\frac{1}{C^2}=\frac{2}{\epsilon {\epsilon }_0{A}^{2}e{N}_D}\left(V-V_{\mathit{fb}}-\frac{K_{B}T}{e}\right)$$ (3)

The donor concentration N_D and the flat band potential V_Fb can be respectively obtained from the slope and the intercept on the potential axis of the Mott-Schottky plot (See Table S2). The slopes of the Mott-Schottky curves were significantly reduced after the p-n junction of the α-Fe₂O₃/MoS₂. Table S2 shows the outcome of N_D in α-Fe₂O₃/8-MoS₂ photoelectrodes (2.07E+27), which has a 2.5-fold higher donor concentration than that of the pure α-Fe₂O₃ (8.11E+26), leading to increased charge carrier transfer and improved PEC performance. Moreover, the qualitative p-type nature of MoS₂ (See Fig. S4) for approving the formation of p–n junctions can be evaluated by the negative slope of Mott–Schottky measurements [67].

The incident-photon-to-current-efficiencies (IPCE) quantitatively reflect the activity of the photoelectrode at different wavelengths. The IPCE of α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ were measured under 100 mW cm⁻² at a bias voltage of 0.6 V (vs. Ag/AgCl) in 1 M NaOH, and they were calculated by placing the appropriate values in Eq. (4):

$$\mathit{IPCE},\%=\left[\frac{1240}{\lambda }\times \frac{J_{\mathit{light}}}{P_{\mathit{light}}}\right]\times 100$$ (4)

The steady-state photocurrent density at the wavelength (λ) of incident light (J_light) and the light power at the specific wavelengths (P_light) were replaced with our measured data. The IPCE values of the photoanodes were in the range from 350 to 600 nm (see Fig. 8d), and they are consistent with the light absorption spectrum results of α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ (see Fig. 6a). The highest IPCE value (7.6%) was detected at around 325 nm for α-Fe₂O₃/8-MoS₂, which is 7.5 times higher than that of the pure α-Fe₂O₃ electrode (1%). This improvement reflects the successful conversion of the absorbed photons to photocurrent, which is confirmed by the LSV result (See Fig. 7a).

To further elucidate the production behavior of the photogenerated electrons, open current potential (ΔOCP) analysis was conducted (Fig. 8e). In the figure, higher ΔOCP values represent the higher production of photogenerated electrons and the improved conductivity [68]. As shown in Fig. 6c, the α-Fe₂O₃/MoS₂ conductivity increased after loading MoS₂ on the surface of α-Fe₂O₃ eight times, and this was attributed to the higher production of photogenerated electrons as well as the substantially reduced electron-hole recombination on the surface of the electrodes.

The applied bias photon-to-current conversion efficiency (ABPE) under different applied potentials was calculated from LSV using Eq. (5), and the diagram is shown and interpreted in Fig. 8f.

$$\mathrm{A}\mathrm{B}\mathrm{P}\mathrm{E}(\%)=\left(\frac{\mathrm{J}\left(\frac{\mathrm{m}\mathrm{A}}{{\mathrm{c}\mathrm{m}}^2}\right)\times \left(1.23-\mathrm{V}\mathrm{b}\mathrm{i}\mathrm{a}\mathrm{s}\right)\left(\mathrm{V}\right)}{\mathrm{P}\left(\frac{\mathrm{m}\mathrm{W}}{{\mathrm{c}\mathrm{m}}^2}\right)}\right)$$ (5)

In this equation, V is the applied potential; P is the power density of the incident light, 100 mW/cm² (AM 1.5); and J is the net photocurrent density at different potentials [69], [70]. The ABPE values of all the α-Fe₂O₃ heterojunctions with MoS₂ are higher than those of the pure α-Fe₂O₃ over the measured potential range in a 1 M (pH = 12) aqueous solution of NaOH. The highest percentage for ABPE was around 0.132% at 0.97 V (RHE) for the α-Fe₂O₃ 8layer MoS₂, which is 30.69 times higher than that of pure α-Fe₂O₃ (0.0043%). Therefore, the α-Fe₂O₃/8-MoS₂ photoanode shows the highest photoelectrochemical properties.

The MoS₂ precursor was prepared with and without ultrasonication, then coated on the α-Fe₂O₃ nanorods to fabricate heterojunction. SEM micrographs were taken from the surfaces of the α-Fe₂O₃/8-MoS₂ thin films, and the images show clear corroborated morphology, which means the ultrasonication process turned partially bulk MoS₂ into few-layer nanosheets with substantially fewer defects. From the top-view SEM images, it is easy to identify that the α-Fe₂O₃-bulk MoS₂ nanorods were aggregated and blocked the surface (Fig. 9a). However, α-Fe₂O₃-exfoliated MoS₂ further confirmed the uniformity of the nanorods (Fig. 9b), which may have increase the light harvesting and facilitating the charge transfer in the α-Fe₂O₃-exfoliated MoS₂ electrode, as revealed in the LSV measurement (Fig. 9c). Therefore, the results show that the MoS₂ nanosheets significantly improve the photoelectrochemical performance of α-Fe₂O₃/MoS₂ photoanode. It is known that the ultrasonication process reduces defects and improves the efficiency of the liquid-phase exfoliation method [71].

Fig. 9.

SEM images (a) after loading MoS₂ precursor which prepared using the bulk powder, (b) after loading MoS₂ nanosheets which prepared by ultrasonication liquid-phase exfoliation technique. (c) LSV for pure α-Fe₂O₃ and α-Fe₂O₃/8-MoS₂ (with ultrasonication and without ultrasonication) electrodes. The supporting electrolyte was a 1 M aqueous solution of NaOH.

4. Conclusion

The heterojunction of α-Fe₂O₃ with exfoliated two-dimensional MoS₂ nanosheets, prepared through the ultrasonication process showed an improved PEC performance caused by the facile electron–hole transfer due to the formation of an electrostatic field. The optimized α-Fe₂O₃/MoS₂ with a thin 8-MoS₂ layer (3.3 nm thickness), labeled α-Fe₂O₃/8-MoS₂, showed photocurrent densities of 0.86 and 1.52 mA.cm⁻² at 0.6 V (vs. Ag/AgCl) in front- and back-side illumination, respectively, which were 7.85 and 13.81 times higher than those of pure α-Fe₂O_3, respectively. The enhanced PEC performance of the α-Fe₂O₃/MoS₂ heterojunction was due to the improved production efficiency of photo-generated electron-hole pairs as well as the decreased resistance caused by the reduced recombination rate of charge carriers and facile hole extraction through the p–n junction. Moreover, the ultrasonication process influenced the photoelectrochemical performance of α-Fe₂O₃/MoS₂ photoanode and increased both light harvesting efficiency and electrical conductivity due to the improved efficient charge transport through the p-n junction.

CRediT authorship contribution statement

Zohreh Masoumi: Investigation, Methodology, Data curation, Writing - original draft. Meysam Tayebi: Conceptualization, Writing - review & editing. Byeong-Kyu Lee: Supervision, Writing - review & editing, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: