Enhancing the Photoelectrochemical Performance of a Superlattice p–n Heterojunction CuFe₂O₄/ZnFe₂O₄ Electrode for Hydrogen Production

Abstract

A p–n heterojunction film consisting of p-type CuFe₂O₄ and n-type ZnFe₂O₄ was fabricated in this study. The n-type ZnFe₂O₄ film was deposited on a stainless steel substrate using the spray pyrolysis method, after which a top layer of p-type CuFe₂O₄ thin film was deposited and annealed. Characterization techniques, such as X-ray diffraction, scanning electron microscopy, UV–Vis diffuse reflectance spectroscopy, and photoluminescence, confirmed the formation of a superlattice p–n heterojunction between CuFe₂O₄ and ZnFe₂O₄. Photoelectrochemical measurements were conducted to investigate the photoelectrochemical properties of the samples, resulting in a photocurrent of 1.2 mA/cm² at 1.5 V (vs. Ag/AgCl) under illumination from a 100-watt LED light source. Utilizing the p–n junction of CuFe₂O₄/ZnFe₂O₄ as a photoanode increased the hydrogen production rate by 30% compared to that of the dark measurement. This enhancement in performance was attributed to the potential barrier at the p–n heterojunction interface, which improved the separation of photoinduced electron–hole pairs and facilitated a more efficient charge transfer. Additionally, coating the stainless steel electrode with this ferrite sample improved both the corrosion resistance and the stability of hydrogen production over extended operation times.

1. Introduction

Producing hydrogen as an alternative source of clean energy has become a significant challenge [1,2]. Photoelectrochemical cells (PECs) have been extensively studied since it was discovered that TiO₂ can split water under UV irradiation [3,4,5,6]. However, these PEC cells have limited efficiency due to the use of UV radiation, which accounts for only 4% of the total solar irradiation spectrum. To improve hydrogen production under solar light, which contains a large portion (43%) of visible radiation [7,8], there is a need to develop a new photocatalyst with a smaller energy gap in the visible range, chemical stability, low electron–hole recombination rate, and nontoxic properties.

Magnetic spinel ferrites (MFe₂O₄; M = Cu, Ni, Cd, Ca, Zn, etc.) have emerged as promising materials for hydrogen production [9,10]. These ferrites exhibit high thermal and chemical stability [11,12,13] and possess a narrow energy gap in the visible range [14,15,16], making them suitable for photocatalysis and applications involving solar water splitting [17,18,19,20]. Combining p-type ferrites with n-type semiconductors to form p–n heterojunctions, such as CuFe₂O₄/CdS, CuFe₂O₄/SnO₂, CuFe₂O₄/TiO₂, CaFe₂O₄/WO₃, CaFe₂O₄/BiVO₄, and CsFe₂O₄/Fe₂O₃, has improved PEC performance compared to bare n-type materials [21,22,23,24,25,26,27]. The enhanced photocatalytic activity is attributed to the separation of photogenerated electron–hole pairs by the internal potential barrier at the p–n heterojunction interface. However, the challenge lies in the chemical and physical structure mismatch between the p-type and n-type materials, resulting in low PEC cell efficiency due to defects in the interface.

In this paper, a superlattice heterojunction between p-type CuFe₂O₄ and n-type ZnFe₂O₄ is investigated as a photoanode for hydrogen production. Using these ferrites offers several advantages: (1) they have the same chemical and crystal structure, (2) their lattice mismatch is minimal, (3) their energy gaps are in the visible range, and (4) they are expected to generate a high internal potential at the p–n junction interface.

Throughout this study, we employed various techniques to characterize the structural and optical properties of the samples. Additionally, linear sweep voltammetry (LSV) was used to evaluate the performance of the PEC cell, and the hydrogen production rate was monitored over an extended period of operation time.

2. Sample Preparation

2.1. Preparation of CuFe₂O₄ and ZnFe₂O₄ Films on Stainless Steel Sheets

CuFe₂O₄ and ZnFe₂O₄ films were fabricated on stainless steel 304 plates using the spray pyrolysis technique. Initially, the stainless steel sheets were cut into 5 cm × 5 cm pieces. To increase the surface area, the plates were etched with 1200 grinding papers for 5 min. Then, the plates underwent ultrasonic cleaning in ethanol and deionized water for 10 min each. Finally, the stainless steel plates were dried using a hot air path to prepare them for the next steps. For depositing the desired metal ferrites, Fe(NO₃)₃·7H₂O, Cu(NO3)2·5H₂O, and Zn(NO₃)₂·6H₂O salts from Merck India with 98% purity were utilized as sources of Fe, Cu, and Zn ions, respectively. All chemicals were from the molar ratios of Fe:Zn and Fe:Cu, which were maintained at 2:1. The solvent used in this study was a mixture of water and ethanol with a volume ratio of 3:1. The solution was sprayed on stainless steel plates using an ultrasonic nebulizer at a temperature of 350 °C. Finally, the deposited ferrite films were annealed using air heated to 500 °C for 2 h. The thickness of the prepared films was controlled by the deposition time. The film thickness was estimated using the weight difference method, revealing that a deposition time of 30 min resulted in films with a thickness of 2 μm.

2.2. Preparation of CuFe₂O₄/ZnFe₂O₄ P–N Junction

One of the annealed Zn ferrite films with a thickness of 2 μm served as the substrate for depositing the Cu ferrite film using the same steps, except that the deposition time was reduced to 15 min, and the annealing temperature was set to 450 °C for 1 h. The thickness of the Cu ferrite film was approximately 200 nm. Confirmation of the CuFe₂O₄/ZnFe₂O₄ p–n junction formation was achieved by measuring the electric resistance in forward bias (indicating low resistance) and reverse bias (indicating high resistance).

2.3. Characterization Techniques

In this study, the crystal structure and phase identification were investigated using the X-ray diffractometer Philips model (PW-1710, Philips, Amsterdam, The Netherlands) with Cu-Kα radiation (λ = 1.5406 Å). The surface morphologies of the thin films were examined using field emission scanning electron microscopy. To determine the bandgap energy value, the thin films were examined using a UV–Vis model JASCO 670 spectrophotometer (Jasco, Tokyo, Japan) within the wavelength range of 200–800 nm. The electrochemical performances of the deposited films were assessed in a standard three-electrode cell at room temperature. The anode electrode was prepared for each of the thin films, a graphene electrode served as the counter electrode, and an Ag/AgCl reference electrode was employed.

Electrochemical impedance spectroscopy (EIS) and Mott–Schottky plots were carried out using a TEGAM LRC Meter Model 3550 (TEGAM, Geneva, OH, USA). Cyclic voltammetry (CV) studies of the samples were conducted in a 0.3 M NaOH aqueous electrolyte using a specially designed linear sweep voltammetry (LSV) device. This LSV device was calibrated using an electrochemical workstation (Potentiostat/Galvanostat PARSTAT 2273 model, Princeton Applied Research, Oak Ridge, TN, USA). The hydrogen production rate was measured by connecting the generator to an inverted measuring cylinder for a specific period to quantify the production rate.

3. Results and Discussion

3.1. X-ray Diffraction Measurements

Figure 1 presents the X-ray diffraction (XRD) patterns obtained for the prepared samples. The XRD patterns of the bare stainless steel exhibit characteristic peaks corresponding to the stainless steel’s 304 phase. Specifically, the peaks observed at 34°, 51°, and 74° can be attributed to the face-centered cubic (FCC) structure of stainless steel 304 [28]. For the ferrite samples, the XRD patterns indicate the formation of a single-spinel cubic-phase structure on the stainless steel substrate. By applying Bragg’s law and the Williamson–Hall equation to the three most intense peaks, various parameters, such as the lattice parameter, crystallite size, and microstrain, were calculated and are listed in

Table 1. Lattice parameter, crystallite size, and microstrain of the prepared samples.

Sample	Lattice Parameter (Å)	Crystallite Size (nm)	Microstrain
ZnFe2O4	8.425	23	0.005
CuFe2O4	8.384	39	0.002
CuFe2O4/ZnFe2O4	8.416	47	0.011

.

It is evident that the lattice parameter of Zn ferrite (a_Zn = 8.425 Å) is greater than that of Cu ferrite (a_Cu = 8.384 Å) [29,30]. This difference can be attributed to the larger ionic radius of Zn²⁺ (0.78 nm) compared to that of Cu²⁺ (0.65 nm). Furthermore, the crystallite size of Cu ferrite is larger than that of Zn ferrite, indicating a higher degree of crystallinity in Cu ferrite.

Additionally, the lattice parameter of the CuFe₂O₄/ZnFe₂O₄ sample was between a_Zn and a_Cu but leaned toward a_Zn. However, the microstrain showed a significant increase. These results confirm the presence of a superlattice structure consisting of Cu ferrite film on Zn ferrite. The increased microstrain can be attributed to the lattice mismatch (∆a/a_Zn = (a_Zn − a_Cu)/a_Zn = 1%) between Cu ferrite and Zn ferrite. Finally, the crystallite size of this sample is greater than that of both Cu ferrite and Zn ferrite, indicating an enhancement in the crystallinity of the Cu ferrite film when grown on Zn ferrite, with the Zn ferrite crystals acting as seeds for the growth of the Cu ferrite film.

3.2. Optical Absorption and Photoluminescence

The diffuse reflection method was employed to determine the optical bandgap of the prepared samples. Figure 2 illustrates the reflection coefficient (R) of all the investigated samples in the UV–Vis range.

The absorption coefficient (α) can be estimated using the following Kubelka–Munk equation [31]:

α = (1 − R)²/2R

The direct energy gap of the prepared sample can be determined using Tauc’s relation [32], expressed as follows:

(αhν)² = A(hν − E_g)

Here, A represents a constant, hν denotes the photon energy, and E_g represents the optical direct energy gap.

Figure 3 displays Tauc’s relation between (αhν)² and hν. The energy gap can be determined by identifying the intersections of the lines with hν axis. It was observed that the energy gap of Zn ferrite was 2.49 eV, while that of Cu ferrite was 2.04 eV (inset figure). These results are consistent with previous reports [33,34,35]. For the CuFe₂O₄/ZnFe₂O₄ sample, it is crucial to note the presence of two transition states corresponding to Cu ferrite and Zn ferrite. It is well established that CuFe₂O₄ is a p-type semiconductor with a smaller energy gap than ZnFe₂O₄, which is an n-type semiconductor. Consequently, a space charge region and a potential barrier formed at the interface between the two ferrites to achieve Fermi-level continuity, and the energy gap for CuFe₂O₄ on the ZnFe₂O₄ layer decreased to 1.78 eV. This decrease could be a result of the increase in the grain size of CuFe₂O₄, as it was grown on ZnFe₂O₄ [36].

The presence of this barrier potential at the interface can reduce the recombination rate between electrons and holes, thereby influencing the photoluminescent intensity. Therefore, to determine the optical properties, photoluminescence (PL) measurements were conducted.

Figure 4 displays the PL emission spectra obtained at an excitation wavelength of 325 nm (E_exc. = 3.8 eV), which exceeds the energy gaps of Zn ferrite and Cu ferrite. It is evident that the PL signal is stronger for Cu ferrite than for Zn ferrite, indicating a faster recombination rate of electron–hole pairs in Cu ferrite. Conversely, a slight decrease in the PL signal was observed for the p–n junction sample. This result can be attributed to the presence of a strong barrier potential at the CuFe₂O₄/ZnFe₂O₄ interface, as mentioned earlier.

3.3. Electrochemical Impedance Spectroscopy

The basic photoelectrochemical cell (PEC) parameters of fabricated photoelectrodes, including CuFe₂O₄, ZnFe₂O₄, and CuFe₂O₄/ZnFe₂O₄, were investigated under dark conditions using electrochemical impedance spectroscopy (EIS). EIS measurements were conducted with a voltage amplitude of 50 mV over the frequency range of 1.0 Hz–100 kHz. The resulting Nyquist plots (Figure 5) display distinct semicircles, indicating the charge transfer process at the electrode–electrolyte interface, while the semicircle diameter represents the charge transfer resistance.

Notably, the CuFe₂O₄ sample exhibited the largest diameter, indicative of slow charge transfer at the electrode–electrolyte interface, whereas the ZnFe₂O₄ sample displayed a smaller radius, signifying rapid and effective charge transfer [36,37]. Remarkably, the CuFe₂O₄/ZnFe₂O₄ electrode showed the smallest diameter of the semicircle, indicating a higher charge transfer rate, i.e., smaller charge transfer resistance and, consequently, a smaller recombination rate, suggesting a faster charge transfer and more efficient electron–hole pair separation. This finding suggests that the formation of a barrier potential at the CuFe₂O₄/ZnFe₂O₄ interface enhances electron–hole charge transfer.

Consequently, the CuFe₂O₄/ZnFe₂O₄CuFe₂O₄/ZnFe₂O₄ heterostructure holds promise as a potential photoanode for high-efficiency solar hydrogen generation due to its fast charge separation and transport capabilities.

The Mott–Schottky analysis was conducted to explore the band structure alignment of CuFe₂O₄ and ZnFe₂O₄ samples. Their valence and conduction band positions were assessed using the flat-band potential (V_fb) and bandgap values (E_g). Figure 6 depicts the Mott–Schottky (M-S) curves for CuFe₂O₄ and ZnFe₂O₄ samples, revealing a negative slope for CuFe₂O₄ (indicating a p-type semiconductor) and a positive slope for ZnFe₂O₄ (indicating an n-type semiconductor) [38,39]. The flat-band potential (V_fb) was determined by extrapolating the plot of 1/C² versus the potential to the x-axis. For ZnFe₂O₄ and CuFe₂O₄, the flat-band potentials were −0.74 eV and 0.62 eV vs. Ag/AgCl, respectively. As the difference between the flat-band potential and the bottom/top edge of the conduction/valence band is negligible for n-type/p-type materials, we regard the flat-band potential as the conduction/valence band edge for n-type and p-type semiconductors, respectively [39]. From the flat-band positions and the band energy gaps, the band structure alignment between ZnFe₂O₄ and CuFe₂O₄ could be estimated.

Figure 7 depicts the proposed electronic band structure of the p–n junction between CuFe₂O₄ and ZnFe₂O₄ under the light condition of our experiment. Strong band bending occurs, leading to the generation of a high barrier potential at the interface. Due to the small thickness of the CuFe₂O₄ layer on ZnFe₂O₄, some photons are absorbed within the bulk of ZnFe₂O₄ (E_g1), while others are absorbed at the surface of CuFe₂O₄ (E_g2).

3.4. Linear Sweep Voltammetry (LSV) Measurements

A three-electrode configuration was employed using a photoelectrochemical cell (PEC) to investigate the photocurrent density–voltage (J-V) characteristics. The experimental setup is depicted in Figure 8. The modified stainless steel plate served as the working electrode (WE), graphite was used as the counter electrode (CE), and Ag/AgCl served as the reference electrode (RE). J-V measurements were conducted in an alkaline medium (0.3 M NaOH) using a 100-watt white LED source and a concentrated lens under both dark and illuminated conditions. All of the electric components of the experiment (power supply and LED light source) were supplied by a 500-watt solar panel system. The scanning speed rate was fixed at 10 mV/s.

Figure 9 presents the J-V curves for all samples. It is evident that both CuFe₂O₄ and ZnFe₂O₄ electrodes exhibited an onset potential of approximately 0.9 volts. Additionally, CuFe₂O₄ achieved its maximum current at 0.95 volts, while ZnFe₂O₄ reached its maximum current at 1.01 volts. The decrease in current beyond a certain anode potential is attributed to the onset of electrode oxidation at specific potentials (cathodic potential). Considering that CuFe₂O₄ possesses two oxidation–reduction mechanisms (Fe²⁺ $\stackrel{}{\leftrightarrow }$ Fe³⁺ and Cu⁺¹ $\stackrel{}{\leftrightarrow }$ Cu²⁺), and ZnFe₂O₄ has only one (Fe²⁺ $\stackrel{}{\leftrightarrow }$ Fe³⁺), it can be concluded that the corrosion resistance of ZnFe₂O₄ is superior to that of CuFe₂O₄.

Moreover, the effect of illumination on the current density was more pronounced for ZnFe₂O₄ compared to CuFe₂O₄. This can be explained in terms of the recombination rate between the generated electron–hole pairs under light illumination. As discussed in the section regarding photoluminescence (PL) and electrochemical impedance spectroscopy (EIS) measurements, the PL signal and radius of the semicircles in Nyquist plots for CuFe₂O₄ were significantly greater than those for ZnFe₂O₄, indicating a higher recombination rate in CuFe₂O₄, which in turn affected the photogenerated current. The increase in photocurrent for both Cu ferrite and Zn ferrite was attributed to the enhanced photoconductivity of the samples.

For the p–n junction sample, it is noteworthy that the starting potential was 1.05 volts, and the anodic voltage was 1.57 volts, indicating improved corrosion resistance. Indeed, the performance remains lower than the other samples until reaching 1.44 V (vs. Ag/AgCl) due to the altered bandgap of the CuFe₂O₄ layer. This alteration causes the band binding for the holes to be greater than that of the electrons (see Figure 7). In other words, the separation rate of generated holes on the Zn ferrite side is faster than the separation rate of generated electrons on the Cu ferrite side, increasing the likelihood of recombination between the holes and electrons. At an applied voltage of 1.44 V (vs. Ag/AgCl), it appears that the rate of electron separation becomes sufficiently fast to overcome recombination with holes. Furthermore, the CuFe₂O₄ material itself might hinder changes despite the improved charge separation. Additionally, the effect of illumination on the generated current was more pronounced for the p–n junction sample compared to the other electrodes. These results can be explained by considering the suggested p–n junction electronic band diagram, which demonstrates the generation of an internal electric potential. This facilitates the transfer of photogenerated holes to CuFe₂O₄, while more photogenerated electrons can be transferred to the conductive substrate via the ZnFe₂O₄ layer. Furthermore, the presence of electron trapping at the interface reduced the recombination rate between the photogenerated electrons and holes, leading to an enhancement in photoelectrochemical performance.

Based on these findings, it is evident that the p–n junction sample, which acts as a photoanode, offers significant advantages for hydrogen generation, particularly under illuminated conditions.

3.5. Evaluation of Hydrogen Production Rate

In this part of the study, the hydrogen production rate (HPR) for the samples was evaluated under both dark and illuminated conditions at a fixed voltage of 1.5 volts (vs. AgCl/Ag) over a period of 10 h. Figure 10 illustrates the relationship between HPR and time. It is apparent that the steady state of production is reached after approximately 1 h of the experiment, and the production rate remains relatively constant throughout the duration of the experiment. This indicates a high corrosion resistance of the modified electrodes. Furthermore, a significant increase in HPR was observed under light illumination.

The HPR under illumination exhibited an enhancement of approximately 30% compared to the measurements taken in the dark. The mechanism of hydrogen production under illumination is depicted in Figure 11.

To examine the effect of photoelectrolysis over 10 h on the crystal and surface structures of the CuFe₂O₄/ZnFe₂O₄ p–n junction, XRD and SEM patterns were compared before and after the hydrogen production experiment. Figure 12 and Figure 13 show the XRD chart and the SEM images before and after the experiment, respectively. The XRD chart indicates that the structure remains a single crystal ferrite even after 10 h of experimentation, demonstrating the chemical stability of the sample during prolonged hydrogen production. Furthermore, the SEM images revealed minimal changes in surface morphology; the sample became smoother, except for some etching, particularly at the grain boundaries. This study highlights the good chemical and physical stability of the investigated sample during hydrogen production.

SEM images show a highly rough surface structure, which results from the spray pyrolysis method of preparation, affecting the quality of the prepared samples. For future studies, a new method of thin film preparation, such as atomic layer deposition (ALD), may be used to enhance PEC cell performance.

4. Conclusions

The implementation of a CuFe₂O₄/ZnFe₂O₄ heterojunction demonstrated a significant improvement in the hydrogen production rate, achieving an enhancement of approximately 30%. This enhancement can be attributed to the presence of an interfacial barrier potential, which effectively reduces electron–hole recombination rates. Consequently, the photocatalytic effects were more pronounced, leading to improved performance at the electrolyte–electrode interface. Additionally, coating stainless steel with this ferrite compound resulted in enhanced corrosion resistance. The investigated sample exhibits good chemical and physical stability after a relatively long period of hydrogen production.