Cu 2 O/ZnO p − n Junction Decorated with NiO x as a Protective Layer and Cocatalyst for Enhanced Photoelectrochemical Water Splitting

: Cuprous oxide (Cu 2 O) has attracted much interest as a photocathode for photoelectrochemical (PEC) water splitting because of its elemental abundance and the favorable band gap, but its poor stability in aqueous solutions hinders the practical PEC application. Compared to the mostly used TiO 2 and noble metal cocatalysts for coating the Cu 2 O photocathode, this work demonstrates a strategy to fabricate a noble metal-free photocathode. We construct a Cu 2 O/ZnO p − n junction photocathode decorated with the NiO x layer as both the protective layer and the hydrogen evolution reaction (HER) cocatalyst. The NiO x cocatalyst exhibits a small Tafel slope of 35.9 mV/ dec and a very low overpotential of 115 mV to drive a current of 10 mA/cm 2 , which are very close to the HER activity of the noble metal platinum. With decorated NiO x , the Cu 2 O/ZnO/NiO x photocathode exhibits signi ﬁ cantly improved stability and photocurrent density with a Faradaic e ﬃ ciency of H 2 gas evolution of 95 ± 4%, distinctly outperforming the Cu 2 O, Cu 2 O/ZnO, and Cu 2 O/ZnO/TiO 2 photocathodes. More-over, electrochemical impedance analysis evidenced that NiO x as a cocatalyst also facilitates the transfer of photogenerated electrons across the electrode/electrolyte interface for water reduction. This work demonstrates that NiO x is not only a stable protective layer against corrosion but also a highly active H 2 evolution cocatalyst. These ﬁ ndings provide new insights for the design of noble metal-free photocathodes toward solar fuel development.


INTRODUCTION
Photoelectrochemical (PEC) water splitting represents an attractive method to capture and store the immense energy of sunlight in the form of hydrogen, a clean chemical fuel. 1−5 To accomplish efficient solar energy conversion in the PEC water splitting cell, the semiconductor photoelectrode should meet the essential criteria including efficient sunlight absorption and carrier separation, high water splitting activity, and long-term stability. 6−10 In the past decades, n-type semiconductors such as TiO 2 , WO 3 , Fe 2 O 3 , and BiVO 4 have been extensively studied in the photoanode for PEC water oxidation. 11−18 However, there are relatively fewer stable and narrow band gap p-type candidates for the PEC water reduction to produce hydrogen fuel.
Cuprous oxide (Cu 2 O), an intrinsic p-type semiconductor material, has gained significant interest in PEC water reduction because of its elemental abundance and scalable synthesis techniques. 19−36 Moreover, Cu 2 O has a direct band gap energy of 2.0 eV, which enables a theoretical photocurrent density (j ph ) of 14.7 mA/cm 2 and a solar-to-hydrogen (STH) conversion efficiency of 18% under a standard 100 mW/cm 2 AM1.5G illumination. 19 However, the poor stability of Cu 2 O in aqueous solutions hinders its PEC water splitting application. 20 Therefore, the protective layer and the hydrogen evolution reaction (HER) cocatalyst are required to coat on the photocathode. 37,38 Recently, Paracchino and co-workers reported a highly active and stable Cu 2 O-based photocathode that was covered with a bilayer of 20 nm Al-doped ZnO and 10 nm TiO 2 and decorated with Pt nanoparticles as HER cocatalysts. 19,20 The p-Cu 2 O/n-ZnO junction facilitated charge separation, and the TiO 2 layer played as a protection layer to improve stability. Furthermore, using RuO 2 as the HER cocatalyst, Luo and co-workers reported the enhanced PEC performance of the Cu 2 O/AZO/TiO 2 /RuO 2 photocathode. 26 In most reported studies, a thin TiO 2 layer was employed to protect Cu 2 O against corrosion. However, one issue of using TiO 2 as the protective layer is that the photogenerated electrons accumulate in the TiO 2 layer and form Ti 3+ electron traps, thus resulting in a decreased photocurrent and low stability of the Cu 2 O/ZnO/TiO 2 photocathode. 20 Moreover, because of the low HER activity of TiO 2 , the noble metal cocatalysts such as Pt and RuO 2 are required to coat on TiO 2 to improve water reduction activity. In this work, we address these issues by a strategy to simultaneously protect Cu 2 O and enhance its PEC water reduction performance, wherein the decorated NiO x layer enables an increased stability and an enhanced water reduction reaction.
NiO is optically transparent to the visible sunlight because of its wide band gap of ∼3.5 eV, 39 which makes it a highly desirable protective layer for coating Cu 2 O as it minimizes sunlight loss. Unlike the precious metal cocatalysts of Pt and RuO 2 , the noble metal-free NiO cocatalyst has been identified as the choice of electrocatalytic material for HER in natural and alkaline electrolytes because of an optimal design stemming from their complementary bifunctional electrocatalytic activity. It has been reported that the NiO or NiO x coating on semiconductor photocathodes can improve the photocurrent and stability for water reduction. 40−43 To the best of our knowledge, there is no demonstration of the integrated NiO x material on a Cu 2 O/ZnO p−n junction photocathode for enhanced PEC water splitting.
Herein, with a coating of the NiO x layer as the protection layer and the HER cocatalyst, the noble metal-free Cu 2 O/ ZnO/NiO x photocathode promotes charge separation and transport for water reduction reaction. Notably, the outermost NiO x film performs dual function of protecting the Cu 2 O photocathode against corrosion and improving the HER reduction activity. In the PEC water-splitting cell, we demonstrate a significantly enhanced photocurrent and remarkably increased stability of the Cu 2 O/ZnO/NiO x photocathode. These results provide new insights into the development of noble metal-free photocathodes toward efficient solar hydrogen generation.  23 Then, the ∼250 nm Cu 2 O film was deposited at 400°C on the Au (10 nm)/ ITO substrate in an O 2 /Ar (7.5/42.5 sccm) atmosphere by reactive DC magnetron sputtering (Semicore Triaxis). 44 Cu (99.99%) was used as a target, and the presputtering time was 10 min to avoid any contamination. The power was fixed at 100 W. During sputtering deposition, the base pressure was kept below 5 × 10 −7 Torr, and the sample stage was rotated at a constant speed of 12 rpm. Then, a ∼50 nm Al-doped ZnO layer was deposited at 200°C on Cu 2 O in an Ar (50 sccm) atmosphere by radio frequency magnetron sputtering. The Al-doped ZnO (2 wt %) target was used with a power of 50 W. After Cu 2 O/ZnO deposition, we prepared the TiO 2 and NiO x layers as the outer protective layer for a comparison. and an Ag/AgCl (saturated KCl) electrode were used as the counter electrode and the reference electrode, respectively. Current density−potential (j−V) measurements were carried out at a scan rate of 30 mV/s with chopped illumination. The measured potential with respect to Ag/AgCl (V Ag/AgCl ) was converted to the potential versus reversible hydrogen electrode (V RHE ) using the following equation: V RHE = V Ag/AgCl + V 0 + 0.059 × pH, where V 0 is the potential of the Ag/AgCl reference electrode with respect to the standard hydrogen potential. The incident photon-to-current efficiency (IPCE) was measured at 0 V RHE under chopped illuminations of different wavelengths of light-emitting diodes (LEDs, 1.0 mW/cm 2 , spectral linewidth of 10 nm). The evolved H 2 gas was measured using a micro gas chromatograph (Agilent Technologies 490 Micro GC) at 0 V RHE under steady-state AM1.5G 100 mW/cm 2 illumination in 0.1 M NaPi solution, which was purged with high-purity Ar (99.999%) gas for over 30 min before the measurement. The Faradaic efficiency of H 2 gas evolution was determined by a comparison of the detected volume of H 2 gas and the calculated volumes of H 2 gas with a theoretical 100% Faradaic efficiency. Scanning electron microscopy (SEM) and energydispersive X-ray spectroscopy (EDXS) images were collected using an LEO 1550 Gemini instrument with an X-Max silicon drift detector (Oxford instruments). X-ray diffraction (XRD) was measured using a Philips MRD. X-ray photoelectron spectroscopy (XPS) spectra were recorded on a Kratos Ultra photoelectron spectrometer equipped.

RESULTS AND DISCUSSION
To compare the PEC performance of the Cu 2 O/ZnO photocathode with different protective layers, we prepared both the Cu 2 O/ZnO/TiO 2 and Cu 2 O/ZnO/NiO x photocathodes on the identical Cu 2 O/ZnO samples. The morphology of the Cu 2 O film was characterized by SEM images   Figure 1C compares the PEC performance of the Cu 2 O, Cu 2 O/ZnO, and Cu 2 O/ZnO/TiO 2 photocathodes in an Arpurged 0.1 M NaH 2 PO 4 (NaPi) electrolyte (pH = 5.0). The current density−potential (j−V) curves were measured under the chopped 100 mW/cm 2 AM1.5G illumination, so the dark and light current could be monitored simultaneously. Notably, the bare Cu 2 O photocathode exhibits a significant dark current below 0.25 V RHE , which is assigned to the reductive decomposition of the Cu 2 O material according to reaction 1 19 To check the stability of the Cu 2 O photocathode, the second scan was carried out. As shown in Figure S3A, the photocurrent behavior of the Cu 2 O photocathode substantially disappeared, leaving only a significant Cu 2 O reduction peak at around 0.2 V RHE . Reductive decomposition of the Cu 2 O electrode is visually observed by the formation of a black film (Cu) where the electrode is illuminated. As shown in Figure  1B, the XRD pattern displays that the diffraction peaks of Cu 2 O disappeared after PEC water splitting measurements. SEM images confirm that the nanostructure of the Cu 2 O layer has been destroyed after PEC measurements (Figures S1D,  S2D).
The Cu 2 O/ZnO photocathode also exhibits a large dark current and an unrepeatable j−V behavior under chopped illumination (Figures 1C, S3B). The XRD result shows reduced diffraction peaks of Cu 2 O after PEC measurements ( Figure 1B). The SEM images display a deteriorated surface of Cu 2 O/ZnO after PEC measurements, indicating that the ZnO layer was etched in the weak acid electrolyte during the PEC reaction ( Figures S1E, S2E).
With the deposition of 250 nm-thick TiO 2 , the Cu 2 O/ZnO/ TiO 2 photocathode shows a significantly reduced dark current ( Figure 1C). The XRD results of Cu 2 O/ZnO/TiO 2 show identical diffraction peaks before and after PEC measurements, while the surface morphology remains unchanged, indicating the protective effect of the TiO 2 layer (Figures 1B, and S1−2). However, the j−V curve of the Cu 2 O/ZnO/TiO 2 photocathode still shows a slight decrease in photocurrent in the second j−V scan ( Figure S3C). Considering that the structure and morphology of Cu 2 O/ZnO/TiO 2 remain unchanged with the protection layer of 250 nm TiO 2 , the photocurrent decay is attributed to the formation of the Ti 3+ electron traps in the TiO 2 layer during the PEC reaction. Paracchino et al. reported that the photogenerated electrons accumulate in the TiO 2 layer to form Ti 3+ electron traps, thus resulting in a decreased photocurrent and low stability of the Cu 2 O/ZnO/TiO 2 photocathode for PEC water splitting. 20 Moreover, TiO 2 has a low electrocatalytic activity for water reduction, and noble metal cocatalysts (such as Pt and RuO 2 ) are generally required to coat TiO 2 as described in previous studies. 19,26 To overcome these drawbacks of the Cu 2 O/ZnO/TiO 2 structure and noble metal cocatalysts, we employ the NiO x layer as both a protective and HER catalytic layer to fabricate a noble metal-free photocathode of Cu 2 O/ZnO/NiO x . The HER activity of NiO x has been demonstrated on other semiconductor photocathodes. 40−43 In this work, we develop a facile approach to prepare the NiO x layers. First, the 200 nm Ni layer was deposited on ITO substrates, followed by annealing at 400°C in air for 5, 15, 30, and 60 min. As seen in Figure 2A, XRD patterns of the annealed samples exhibit the peaks of NiO accompanied with small peaks of Ni, indicating that the deposited Ni was oxidized to a mixture of NiO and Ni (denoted as NiO x ). This result is further confirmed by the XPS measurement, as discussed below. Figure 2B,C shows the j−V curves and the corresponding Tafel plots of ITO, ITO/Ni, and ITO/NiO x with different annealing times. Compared to ITO or the Ni electrode, the NiO x samples exhibit a remarkably enhanced HER activity for water reduction. In particular, the NiO x annealed for 30 min shows the lowest overpotential (requiring a very low overpotential of 115 mV to drive 10 mA cm −2 ) and the smallest Tafel slope (35.9 mV/dec), which are very close to the HER activity of the Pt electrode measured under the same conditions (Tafel slope of the Pt electrode: 32.5 mV/dec). This result clearly demonstrates that the prepared NiO x exhibits as high HER activity as the welldeveloped noble metal HER cocatalysts. Therefore, we employ this NiO x preparation condition to fabricate the Cu 2 O/ZnO/ NiO x photocathode. The surface composition and chemical states of the NiO x layer were characterized by the XPS measurements. As shown in Figure S6, the Ni 2p spectrum displays the typical characteristics of the presence of Ni 0 and Ni 2+ species. The fitting peaks at 852.1, 856.8, and 858.6 eV are ascribed to the Ni 2p 3/2 and two satellite peaks of nickel metal (Ni 0 species), respectively. 45 The fitting peaks at 853.7 and 855.5 eV are attributed to two Ni 2p 3/2 peaks of Ni 2+ species (NiO), with three satellite peaks at high binding energy (861.3, 864.0, and 866.4 eV, respectively). 46,47 From the area of Ni 2P 3/2 peaks of Ni 0 and Ni 2+ , the composition of the NiO x catalyst prepared by annealing Ni at 400°C (in air, for 30 min) is estimated to be 16.8% of Ni and 83.2% of NiO. Additionally, the O 1s spectrum shown in Figure S6 exhibits two peaks at 529.3, 531.1, and 532.4 eV, which have been attributed to the lattice oxygen and the surface-adsorbed hydroxyl groups and water molecules, respectively.
The PEC water splitting measurements of Cu 2 O/ZnO/NiO x photocathodes were carried out in an Ar-purged 0.1 M NaPi electrolyte (pH = 5.0) under chopped AM1.5G 100 mW/cm 2 illumination. Figure 3C  The applied-bias-photon-to-current efficiency (ABPE) of the Cu 2 O-based photocathodes was derived using the equation: ABPE (%) = j ph × (E − E rev 0 )/P in × 100, where E rev 0 is 0 V RHE , E is the applied potential in V RHE , and P in is the power of the incident light (100 mW/cm 2 ). As shown in Figure 3D, the maximum ABPE of the Cu 2 O/ZnO/NiO x -200 nm photocathode is 0.07% at 0.19 V RHE , which is much higher than that of the Cu 2 O/ZnO photocathode (0.01% at around 0.1 V RHE ). Although the thinner NiO x (20 nm) gives rise to a higher maximum ABPE of 0.11% at 0.26 V RHE , the dark current is obviously increased because of the inefficient protection against corrosion. It is worthwhile mentioning that the applied potentials for the maximum ABPE are positive-shifted for the photocathodes with coating NiO x layers, indicating the reduced overpotentials for the H 2 -evolution reaction on NiO x . The incident photon-to-current efficiency (IPCE) of the Cu 2 O/ZnO/NiO x -200 nm photocathode was measured at 0 V RHE under chopped illuminations of different wavelengths of LEDs (1.0 mW/cm 2 , spectral linewidth of 10 nm) ( Figure  S7). The highest IPCE of 63.6% is obtained for the Cu 2 O/ ZnO/NiO x -200 nm photocathode at 450 nm.
The stability of the photocathodes and the evolved H 2 gas was measured under steady-state AM1.5G 100 mW/cm 2 illumination. As shown in Figure 3E, the chronoamperometry  Figure  3E). Meanwhile, the volumes of the evolved H 2 gas were measured by gas chromatography to evaluate the Faradaic efficiencies (η F ) of H 2 . As shown in Figure 3F,  25,26,32 this work demonstrates a facile approach to fabricate the noble metal-free photocathode using NiO x to replace both the TiO 2 and RuO 2 (Pt) layers (Table S1).
To understand the enhancement of the PEC performance of the Cu 2 O/ZnO/NiO x photocathode, EIS measurements were carried out in the frequency range of 1 to 10 5 Hz, at 0 V RHE , in 0.1 M NaPi (pH = 5.0), under AM1.5G 100 mW/cm 2 illumination. Figure 4A shows the Nyquist plots of the Cu 2 O, Cu 2 O/ZnO, Cu 2 O/ZnO/TiO 2 -250 nm, and Cu 2 O/ ZnO/NiO x -200 nm photocathodes. The EIS data were fitted using the equivalent circuits shown in the insets. As shown in Figure 4A, the Cu 2 O/ZnO photocathode exhibits a smaller diameter of the semicircle compared to Cu 2 O. The fitting results reveal that the value of charge-transfer resistance (R ct ) is decreased from 700 Ω cm 2 for Cu 2 O to 520 Ω cm 2 for Cu 2 O/ ZnO (Table S2). This can be explained by the built-in electric field in the Cu 2 O/ZnO p−n junction that promotes charge separation, thus reducing the charge transfer resistance ( Figure  4B). With a coating of 250 nm TiO 2 or 200 nm NiO x , the Nyquist plots show two semicircles, which can be fitted by the equivalent circuits shown in the right inset of Figure 4A. The fitting results show that Cu 2 O/ZnO/TiO 2 and Cu 2 O/ZnO/ NiO x have a similar value of charge-transfer resistance from the bulk to the photocathode surface, R ct,1 (R ct,1 = 100 Ω cm 2 for Cu 2 O/ZnO/TiO 2 and R ct,1 = 90 Ω cm 2 for Cu 2 O/ZnO/ NiO x ). However, Cu 2 O/ZnO/TiO 2 exhibits a rather large resistance of 2200 Ω cm 2 for the charge-transfer across the electrode/electrolyte interface (R ct,2 ). In contrast, the Cu 2 O/ ZnO/NiO x displays a significantly decreased resistance R ct,2 of 400 Ω cm 2 (Table S1). Figure 4B illustrates a schematic energy band diagram of the Cu 2 O/ZnO/NiO x photocathode, 19,20,48 where the built-in electric field in the Cu 2 O/ZnO p−n junction separates the photogenerated electron−hole pairs and sweeps the electrons to the NiO x catalyst for water reduction. Given the fact of similar thickness of TiO 2 and NiO x and the identical Cu 2 O/ZnO p−n junction underneath, the charge transfer resistance in the bulk of the Cu 2 O/ZnO/TiO 2 and Cu 2 O/ZnO/NiO x photocathodes is mainly affected by the built-in electric field and bulk recombination ( Figure 4B), which give rise to a similar value of R ct,1 . The remarkably reduced value of R ct,2 for Cu 2 O/ZnO/NiO x is mainly resulting from the presence of the high HER activity NiO x , as demonstrated in Figure 2B. This result further confirms that NiO x is not only a stable protective layer against corrosion but also a highly active H 2 evolution cocatalyst, distinctly outperforming the TiO 2 layer.
In summary, we have demonstrated a strategy of using NiO x as a protective layer and the HER cocatalyst to improve the stability and the efficiency of the Cu 2 O photocathode for H 2 evolution. We report a facile approach to fabricate the NiO x layer by Ni oxidation. The optimized NiO x layer exhibits a very low overpotential of 115 mV to drive 10 mA cm −2 and a small Tafel slope of 35.9 mV/dec, which are very close to the HER activity of the Pt electrode. With a decorated 200 nm NiO x layer on the Cu 2 O/ZnO p−n junction, the photocathode shows a synergetic enhancement of both the photocurrent and the Faradaic efficiency for H 2 evolution. The EIS results evidence that the NiO x layer significantly reduces the charge transfer resistance for water reduction. By comparison with the mostly used TiO 2 layer, this work unambiguously demonstrates the dual functions of NiO x as a protective layer and a  The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsaem.0c01198.
SEM of photocathodes before and after PEC measurements, PEC measurement results, and EDS and fitting results of the EIS data (PDF)