Sequential Multilayer Design with SnO₂-Layer Decoration for Inhibiting Photocorrosion of Cu₂O Photocathode

Abstract

The Cu₂O-based photocathode has been widely applied in photoelectrocatalytic hydrogen evolution and carbon dioxide reduction systems. However, the poor stability of Cu₂O caused by photocorrosion highly restricts the application. In this work, a multilayer configuration is designed as Cu₂O/ZnO/SnO₂ via sequential depositions of electrodeposition and spin-coating. The liquid-phase epitaxial growths of the Cu₂O and ZnO layers are achieved by sequential electrodepositions on a FTO/Au substrate. The decoration of a uniform SnO₂ layer onto Cu₂O/ZnO is realized by a SnO₂ QDs coating and provides dual functions for boosted electron transfer and surface reaction. The protection of the SnO₂ layer is fulfilled by the inhibition of Cu⁺ transformation, resulted from the compact covering of SnO₂ QDs onto the exposed surface of the Cu₂O and ZnO layers. Consequently, the enhanced photocurrent density and improved stability are obtained for Cu₂O/ZnO/SnO₂ compared to bare Cu₂O and Cu₂O/ZnO sample photocathodes. The necessary role of SnO₂ QDs serving as electron transfer and protection layers studied in this work reveals the remarkable potential in the modification of other vulnerable electrode materials.

1. Introduction

Cuprous oxide (Cu₂O) semiconductor with an intrinsic p-type characteristic is a promising candidate for photoelectrocatalytic systems in terms of H₂ evolution and CO₂ reduction [1,2,3,4]. Due to the suitable band level being more negative than −0.7 V (vs. RHE), it is thermodynamically favorable to apply a p-type Cu₂O photocathode to accomplish the reduction potentials of H⁺ to H₂ and CO₂ to C₁ products (i.e., CO, CH₄, CH₃OH, HCHO, HCOOH) in the range of −0.2~0.2 V (vs. RHE) [5,6,7,8]. As estimated, a peak photocurrent density of 14.7 mA·cm⁻¹ and a solar-to-hydrogen conversion efficiency of 18% could be obtained theoretically by the Cu₂O photocathode under a simulated AM1.5G illumination. Nonetheless, the optimal performance is constrained by the unsatisfactory aspect ratio of light penetration depth (~10 μm) and diffusion length of minority carriers (<200 nm) within the electrodeposited Cu₂O layer [9,10]. In addition to the mismatched ratio, the poor stability in an aqueous solution is perceived as the predominant limitation, which originated from the inevitable redox reactions of monovalent Cu⁺, as the corresponding potentials lie in the forbidden band (Cu⁺ reduction and oxidation approximately at 0.8 V and 0.65 V above valence band, respectively) [11,12]. The decoration of a Cu₂O photocathode with a conformal overlayer coating, including an electron transfer layer (ETL) and a protection layer (PL), is essential for tackling the aforementioned two obstacles through the promotion of electron diffusion mobility and the inhibition of photocorrosion. The n-type oxide semiconductors (i.e., ZnO, Ga₂O₃, TiO₂) have been applied as an effective ETL for efficient electron migration behaviors in a Cu₂O photocathode [13,14,15,16,17,18]. The construction of a buried p-n junction in the interface induces the downward band bending between conduction band levels, hence boosting electron extraction and suppressing excessive electron accumulation within Cu₂O layer [15,19,20,21]. The buried p-n junction contributes to a significantly earlier onset potential and higher photocurrent density for overlayered Cu₂O photocathodes [16]. Moreover, PL introduction (i.e., TiO₂) provides a protective shielding by cutting off water permeation for the reduced probability of physical contact between the Cu₂O surface and H₂O molecule for further inhibition of photocorrosion [22]. In addition, it is indispensable to decorate PL as the outmost layer due to the vulnerability of the ETL (i.e., ZnO is unstable in acid and strong alkali environments) [23,24]. Numerous efficient multilayer configurations have been proposed for the Cu₂O photocathode (e.g., ZnO/TiO₂, AZO/TiO₂, Ga₂O₃/TiO₂, ZnO/NiO_x, etc.) and are obtainable by variable deposition techniques, including atomic layer deposition (ALD), magnetron sputtering deposition (MSD), electrodeposition, and spin-coating methods [25,26,27,28,29,30,31,32,33,34].

Due to the decisive role of the growth uniformity in fulfilling the ETL and PL functions of deposited overlayers, the compact deposition of the ETL and PL with a uniform distribution via a facile and economical route is necessary for constructing efficient and stable multilayer Cu₂O-based photocathodes. As a prevalent selection of PL, TiO₂ layers are usually decorated onto Cu₂O via ALD and MSD techniques which are costly and complicated [35,36]. Nonetheless, the quality of ALD- and MSD-deposited TiO₂ layers is highly sensitive to the surface to be coated, which requires pre-deposited layers (i.e., ZnO, Ga₂O₃) as buffer layers. It has been reported that the incomplete coverage of the TiO₂ layer via ALD from titanium tetraisopropoxide and H₂O fails to realize the dual function of PL and ETL in the direct contact (Cu₂O/TiO₂/Pt), leading to a considerable electron loss in the heterogeneous interface compared to bare Cu₂O (Cu₂O/Pt) [15]. In terms of the wet-chemistry electrodeposition method, the TiO₂ layer derived from the dehydration of the deformable TiO_xH_y layer avoids the imperfect contact of Cu₂O/TiO₂ and decreases the interfacial recombination sites [22]. However, the extreme pH values of the Ti precursor solution restrict the universal application of the electrodeposited TiO₂ layer. Based on the above considerations, it is appealing to seek an alternative PL for TiO₂, which could be decorated onto Cu₂O-based photocathodes feasibly and economically. Herein, n-type SnO₂ has extraordinary optoelectronic properties, which have been extensively used in perovskite solar cells as ETL [37,38]. Compared to the TiO₂ layer, the SnO₂ layer exhibits superior electron transport mobility, a lower n-type resistivity, a higher transparency, and a lower conduction band level for better electron injection. The good chemical stability of SnO₂ unlocks a potential to serve as PL for Cu₂O-based photoelectrodes [39,40,41]. Based on the consideration of compact coating, it should be more suitable to apply SnO₂ quantum dots (QDs) with relatively smaller particle size distribution for PL formation [42,43]. In addition, the SnO₂ QDs layer could be fabricated via a simple solution method at a low-temperature [44]. The fabrication of the SnO₂ QDs layer in mild conditions is more adaptable for vulnerable substrates and unrestricted by the surface quality of substrates in contrast to the TiO₂ layer.

In this work, the tailored decoration of SnO₂ QDs layer is introduced to construct a Cu₂O/ZnO/SnO₂ photocathode with a multilayer design for the promotion of photocurrent and improvement of stability. The sequential liquid-phase epitaxial growths of the Cu₂O layer and ZnO layer have been achieved via controlled electrodeposition methods, which are assigned as an absorption layer and ETL. The uniform coating of SnO₂ QDs leads to the compact decoration of SnO₂ PL and covers the exposed surface of Cu₂O and ZnO nanoparticles. SnO₂ QDs at the outmost surface also provide more active sites for accelerated interfacial reduction. The dual functions of a SnO₂ QDs layer decorated onto a Cu₂O/ZnO multilayer in terms of better electron transfer and considerable protection are demonstrated, which facilitates the elevated photoelectrocatalytic performances with improved stability.

2. Results and Discussions

2.1. Observation of SnO₂ QDs

Colloidal SnO₂ QDs solution was obtained via a facile and reproducible room-temperature wet-chemistry method. The steady hydrolysis–dehydration–oxidation reactions of the SnCl₂ precursor were assisted with thiourea for accelerating and stabilizing reactions. The observation of SnO₂ QDs was realized by high-resolution transmission electron microscopy (HR-TEM). As displayed in Figure 1(ai,aii), the particle size distribution of SnO₂ QDs was uniform with the estimated average size of 2~3 nm. The lattice information was acquired by a selected-area electron diffraction (SAED) pattern (Figure 1b) and lattice fringe (Figure 1c). The diameter of the recorded diffraction ring (Line #1) was measured as 5.874 nm⁻¹ from the line distribution of intensity, as shown in Figure 1d, and the corresponding interplanar spacing was calculated as 3.40 Å. The direct measurements of lattice fringe (Line #2 and Line #3) were conducted specifically for the single dot (Figure 1c) and the corresponding spacing of the lattice facet was measured as 1.354 nm and 1.356 nm along Line #2 and Line #3 across four intervals, respectively (Figure 1(ei,eii)). The average facet spacing was calculated as 3.39 Å, in agreement with the parameter that resulted from SAED patterns, which was ascribed to the (110) crystal facet of rutile-phase SnO₂ (d = 3.34 Å) [44]. The element distribution (Figure 1f) verified the detection of the Sn element. Based on the above results from HR-TEM characterization, SnO₂ QDs with a uniform size distribution were determined. The smaller particle size of SnO₂ was considered suitable and necessary for the compact layer deposition of the SnO₂ layer.

2.2. Determination of Multilayer Design

The illustration displayed the multilayer design for the Cu₂O/ZnO/SnO₂ sample via sequential layer depositions (Figure 2a), including the Cu₂O layer and ZnO layer assembly via electrodeposition and the SnO₂ layer assembly via spin-coating. The colors of the photocathodes were observed as purple, tawny, and tan for the Cu₂O, Cu₂O/ZnO, and Cu₂O/ZnO/SnO₂ samples, respectively. The corresponding top-view observations of the morphological microstructures of each layer were shown in Figure 2b–e.

The uniform deposition of Au nanoparticles provides an appropriate platform for the epitaxial growth of the Cu₂O layer and improves the conductivity. The favorable growth of Cu₂O along [111] orientation was prioritized on Au-sputtered substrate, leading to the high-quality Cu₂O layer coating with the exposed corner of Cu₂O nanocrystals (Figure 2c and Figure S2a). The exposed edges of the corners were hundreds of nanometers, and the surface around the corners was regarded as surface active sites. The higher density of the Cu₂O {111} facets was bound to improve the surface reduction kinetics of the Cu₂O-based photocathode [45,46,47]. After the assembly of the ZnO layer onto the Cu₂O sample, the colors of the photocathodes turned tawny, which implied that the electrodeposited ZnO layer was thin and transparent. As shown in Figure 2d, the ZnO layer was composed of dispersive ZnO nanoparticles with a size distribution ranging from 50 nm to 100 nm. Spherical ZnO nanoparticles were embellished onto the exposed sharp corners of Cu₂O nanocrystals. During the electrodeposition of the ZnO layer, the reduction in p-BQ molecules was predominant in the cathodic reactions, which caused the increase in pH value [48]. Subsequently, Zn(OH)₂ was precipitated from Zn²⁺ ions with the increased pH in the local micro-environment and loaded onto the Cu₂O surface. The relatively smaller size of ZnO nanoparticles (50~100 nm) compared to Cu₂O nanocrystals (~500 nm) resulted from the decomposition of Zn(OH)₂ species as the system was heated to 80 °C [49,50]. The structural retention of Cu₂O nanocrystals with sharp corners indicated no erosion of Cu₂O in a ZnO precursor solution as electrolyte (mildly acidic). Furthermore, the SnO₂ layer was assembled by a tight and uniform coating of SnO₂ QDs onto ZnO nanoparticles, as observed in Figure 2e. Specifically, the exposed surface of the ZnO nanoparticles were embellished by SnO₂ QDs with a smaller size, which was similar to the plum pudding type. In summary, the anticipated multilayer design was successfully obtained via sequential layer deposition steps.

The multilayer design was further determined by EDX mapping in the cross-section view, as shown in Figure 3. From the direct observation in Figure 3b, the Cu₂O/ZnO/SnO₂ multilayer were epitaxially grown onto FTO/Au substrate at the right side. The depth of the Cu₂O bulk layer was measured as 800 nm (Figure 3(civ)), which was directly deposited onto the Au layer (Figure 3(cii)). The signals of Cu and Au were apparently distinguished from the signals of Si and Sn in the FTO region (Figure 3(ci,cii)). In contrast, the relatively smaller depths of the ZnO and SnO₂ layers with the smaller particle sizes led to the weaker signals of Zn (Figure S3) and Sn (Figure 3(cii)) on the surface of the Cu₂O layer [51]. The signal of Cu away from the bulk Cu₂O layer (in the left side of concentrated Cu signals in purple), observed in Figure 3(civ), might be ascribed to Zn distribution because of the adjacent element numbers of Cu and Zn. Furthermore, the evidence of the existence of ZnO and SnO₂ was also provided by AFM results (Figure 3d and Figure S4). The dispersive small nanoparticles were observed, indicating the decoration of ZnO and SnO₂ layers onto the Cu₂O surface. Therefore, the anticipated multilayer structure in the sequence of the Cu₂O, ZnO, and SnO₂ layers onto FTO/Au substrate was preliminarily demonstrated, as displayed in the schematics of Figure 2(aii). The configuration of the sample photocathode was clearly presented in Figure 3(ai). The conductor Cu wire and connection Ag paste were encapsulated by UV curving adhesive, which was separated from the deposition region. The deposition region of the FTO/Au substrate was covered by the assembly of Cu₂O (red cubes), ZnO (blue spheres), and SnO₂ (brown spheres), as is vividly illustrated in Figure 3(ai).

The physicochemical properties of Cu₂O, Cu₂O/ZnO, and Cu₂O/ZnO/SnO₂ sample photocathodes were characterized by spectroscopic measurements of XRD and XPS. The characteristic diffraction peaks of FTO, Au, and Cu₂O species were clearly pointed out, as displayed in Figure 4a, according to the standard cards (#99-0024, #01-1172, and #77-0199), respectively. Particularly, the main peak of the Cu₂O layer located at 36.5° was assigned to the (111) facet [10]. The deposited Au layer as a growth substrate induced the controlled growth orientation of Cu₂O along the [111] direction [47]. Due to the similarity of the XRD peaks, the peak signals of ZnO and SnO₂ might be involved in the Cu₂O and FTO peaks, respectively. The supplementary information of surface chemical states was offered by XPS spectra, as shown in Figure 4b–d. Due to the fact that XPS measurement is a surface-sensitive detection technique, the detection depth normally lies in the range of several nanometers. For the detection of the Cu₂O layer, which was beneath the ZnO and SnO₂ layers, an etching process by Ar ions was conducted to remove the surface species and to expose the Cu₂O bulk layer with an etching depth of 200 nm before detecting Cu signals [22]. Zn and Sn XPS signals were detected without Ar etching pre-treatment. The characteristic peak located at 569 eV corresponded to Cu⁺ states in Cu LM2 spectra (Figure 4b). In Zn 2p spectra, the doublet peaks of Zn 2p_1/2 and 2p_3/2 located at 1022.0 eV and 1044.0 eV, respectively, indicated the coating of the ZnO layer on the Cu₂O surface [49]. In Sn 3d spectra, the doublet peaks of Sn 3d_3/2 and 3d_5/2 were detected at 486.5 eV and 495.0 eV, respectively, which verified the decoration of SnO₂ at the outmost surface [52,53]. The similarity between the Cu₂O/ZnO (B.) and Cu₂O/ZnO/SnO₂ (B.) sample photocathodes indicated the maintenance of chemical states and compositions in the multilayer structure with SnO₂ layer deposition. The spin-coating step with mild conditions of suitable pH value and annealing temperature avoided destroying the ZnO and Cu₂O layers.

2.3. Functions of SnO₂ Layer Decoration

Based on the above results from microscopic and spectroscopic characterizations, it could be conservatively concluded that the Cu₂O/ZnO/SnO₂ multilayer photocathode had been assembled and determined. The measurements of photoelectrochemical performances in terms of photocurrent and stability were conducted in LSV and I-t experiments, as shown in Figure 5 and Figure S5. In terms of the response of cathodic current under the irradiation of simulated solar light (Figure 5a), the Cu₂O/ZnO/SnO₂ photocathode exhibited a superior performance compared to the pristine Cu₂O and Cu₂O/ZnO photocathodes. The decorations with the ZnO and SnO₂ layers led to the elevated photocurrent density compared to the pristine Cu₂O photocathode (Figure S5a), which verified the similar function of the SnO₂ layer to the ZnO layer as the ETL. The larger improvement in the photocurrent density of the Cu₂O/SnO₂ photocathode compared to the Cu₂O/ZnO photocathode indicated the better ability of electron extraction and migration for the SnO₂ layer [42,43]. In addition, a considerable dark current was found in the relatively negative range of applied potentials (0.1~0.2 V vs. RHE), and a significant photocurrent was obtained in the relatively positive range (0.5~0.7 V vs. RHE) for sample photocathodes with SnO₂ decoration, which indicated that the SnO₂ layer could serve as the surface catalyst to reduce the overpotential [52]. Therefore, the SnO₂ layer displayed the dual functions as ETL and active catalytic sites to accelerate electron transfer and interfacial reduction reaction.

The comparisons of reaction stability among sample photocathodes were carried out by I-t experiments (Figure 5b). For the pristine Cu₂O photocathode, a significant decrease in the response of photocurrent density by almost 90% indicated extremely poor stability of bare Cu₂O by severe photocorrosion. The Cu₂O/ZnO photocathode also displayed a decreasing tendency in photocurrent density by almost half, which implied that the ETL function of the ZnO layer helped to suppress photocorrosion to an extent. The color of the Cu₂O/ZnO photocathodes changed from tawny to blackish green, apparently in contrast to the Cu₂O/ZnO/SnO₂ sample which has a slight color change (Figure S6). The decoration with the SnO₂ layer largely inhibited photocorrosion as the stable photocurrent was obtained during constant irradiation (30 min) for the Cu₂O/ZnO/SnO₂ photocathode. The inhibition of photocorrosion by the SnO₂ layer was revealed by the XPS comparison result, shown in Figure 6(ai,aii). In the comparison of the Cu₂O/ZnO photocathodes, Cu₂O/ZnO (B.) and Cu₂O/ZnO (A.) represented samples before and after stability tests, respectively. The emergence of a new peak at 568.0 eV in Cu LM2 spectra, corresponding to the XPS signal of Cu⁰ state, was found for the Cu₂O/ZnO (A.) sample [22]. The photocorrosion mechanism could be elucidated as the transformation from Cu⁺ to Cu⁰ species in the Cu₂O layer contributed to the decreased performance. On the contrary, almost no change was found in Cu LM2 spectra for the Cu₂O/ZnO/SnO₂ (B.) and Cu₂O/ZnO/SnO₂ (A.) samples, indicating the protection of the SnO₂ QDs layer to inhibit the generation of Cu⁰ species. For both the Cu₂O/ZnO and Cu₂O/ZnO/SnO₂ samples, surface ZnO layers were relatively stable according to the similarity of Zn 2p signal (Figures S7b and S8), which implied the stable chemical state of ZnO in the reaction. In addition, the SnO₂ layer also displayed chemical stability, as shown in Figure S7a. The structural change of Cu₂O photocorrosion was also reflected by microscopic observations (Figure 6(bi–biii),(ci–ciii)). The Cu₂O layer inside the Cu₂O/ZnO (A.) sample showed more density of tiny cavities in the Ar etching region (Figure 6(ci)) compared to the Cu₂O/ZnO (B.) sample. The transformation of Cu⁰ species in photocorrosion caused structural destruction and formed cavities in the Cu₂O layer. For the ZnO layer without decoration of the SnO₂ layer, the exposed surface of ZnO nanoparticles became rugged, implying the structural change in the ZnO layer (Figure 6(cii)). The structural deterioration of the ZnO layer resulted in the permeation of water molecules and the direct contact of the Cu₂O layer and water molecules, leading to an increased possibility of photocorrosion. However, no obvious morphological change was observed for the Cu₂O/ZnO/SnO₂ sample in the Ar etching region (Figure 6(bi)) and decoration region (Figure 6(bii,biii)). In contrast, the application of the Al₂O₃ layer as a similar candidate for inert protection layers failed to enhance photocurrent density (Figure S5b) and improve stability (Figure S9a), which might be ascribed to the imperfect covering of the Al₂O₃ nanoparticles coating. The difference between the SnO₂ layer and Al₂O₃ layer deposited by spin-coating was derived from uniformly coating the SnO₂ QDs in perfect contact with the surface of Cu₂O/ZnO, whereas coating Al₂O₃ nanoparticles with a similar size to ZnO nanoparticles still remained the interspace. Special Janus bilayers of SnO₂ QDs were constructed by the sequential deposition of R-SnO₂ and O-SnO₂, which was modified by reduction and oxidation treatments before spin-coating. The Janus bilayers structures of R-SnO₂/O-SnO₂ and O-SnO₂/R-SnO₂ both diminished the dark current and further improved stability compared to merely the SnO₂ layer (Figure S9b), which implied the surface reconstruction of SnO₂ by redox modifications. Nonetheless, the comprehensive elucidation of the Janus bilayers and the contribution from surface Sn states of reduced and oxidized Sn species were necessary, which needs further deep study in future research.

3. Experimental Procedures

3.1. Preparation of Samples

Blank FTO/Au sample: The commercial FTO (F-doped tin oxide) glass (length: 18 mm, width: 12 mm, thickness: 2 mm) was applied as the substrate for the further fabrication of photoelectrode samples. Before the layer fabrication, successive purification procedures of 10 min ultrasonication in acetone, ethanol, and pure water for three cycles were conducted to remove the residual organic pollutants on the FTO surface followed by a N₂ purge. A Au film with a thickness of 100 nm was coated onto the conductive surface of purified FTO via vacuum evaporation to obtain FTO/Au glass. The FTO/Au photoelectrode was assembled by FTO/Au glass and Cu wire covered with an insulating layer. The connection between the conductive surface of FTO/Au glass and the exposed section of the Cu wire was constructed by the adhesion of conductive Ag paste. The connection region was subsequently packaged by UV curving adhesive followed by UV illumination. The UV-curable encapsulant was supposed to cut off the contact between the Ag-paste connection and electrolyte solution, which excluded the interference of the exposed Ag paste and Cu wire for photoelectrochemical measurements. Apart from the packaged region, the exposed region was measured as 15 mm × 12 mm for the following decoration of the multilayer.

Cu₂O sample: The fabrication of the Cu₂O layer was achieved via the electrodeposition method according to previous reports. The Cu solution containing 160 mL pure water, 3M lactic acid, 0.5M K₂HPO₄, and 0.2M Cu₂SO₄·5H₂O was prepared with vigorous stirring. The pH of the Cu solution was subsequently adjusted to around 12 by a 2M KOH solution and the initial Cu solution with a clear sky-blue color turned into a turbid solution and finally became a clear navy-blue color during the dropwise addition of the KOH solution. The pH-adjusted Cu precursor solution with a volume of approximately 600 mL was used freshly for Cu₂O layer deposition. The electrodeposition was conducted in a water bath (30 °C) with galvanostatic mode (−0.1 mA·cm⁻²) in a two-electrodes cell using blank FTO/Au substrate and Pt wire as the working electrode and counter electrode, respectively. After Cu₂O electrodeposition for 2 h, the deposited photoelectrode was immersed in pure water and dried by a N₂ purge, which was designated as the Cu₂O sample.

Cu₂O/ZnO sample: The fabrication of the ZnO layer was achieved via the electrodeposition method according to previous reports. The Zn solution containing 40 mL dimethyl sulfoxide (DMSO), 0.1M p-benzoquinone (p-BQ), and 0.05M Zn(NO₃)₂·6H₂O was prepared with vigorous stirring. The ZnO layer deposition was conducted in the fresh Zn precursor solution with galvanostatic mode (−0.125 mA·cm⁻²) heated by an oil bath (95 °C) in the same two-electrodes cell using the Cu₂O sample as the working electrode. After ZnO electrodeposition for 7 min, the deposited photoelectrode was immersed in pure water and dried by a N₂ purge, which was designated as the Cu₂O/ZnO sample. It needs to be highlighted that the deposited side of the working electrode should be perpendicular to the flow vector in the stirring solution for steady mass transportation in both the Cu₂O and ZnO electrodepositions with a constant stirring rate of 200 rpm. Bubbling treatment by nitrogen gas for 30 min was introduced before the Cu₂O and ZnO electrodepositions, and nitrogen flow was maintained during the electrodepositions to remove residual dissolved oxygen.

Cu₂O/ZnO/SnO₂ sample: The synthesis of SnO₂ QDs was achieved via the wet-chemistry solution method according to the previous report. The colloidal SnO₂ QDs solution containing 30 mL pure water, 0.15M SnCl₂, and 0.15M thiourea (CH₄N₂S) was obtained with vigorous stirring for 48 h. The initial milky yellow suspension turned into a clear yellow colloidal SnO₂ QDs solution with sufficient exposure to air. The pH was further adjusted to around 9 by 2M KOH as the acid SnO₂ QDs solution would etch the surface of Cu₂O and ZnO. The fabrication of the SnO₂ layer was implemented via the spin-coating method. The pH-adjusted SnO₂ QDs solution (0.15M, 200 μL) was dropped onto the Cu₂O/ZnO surface followed by spinning at 400 rpm for 3 s and 3000 rpm for 60 s. After annealing (160 °C, 2 h) and a N₂ purge, the Cu₂O/ZnO/SnO₂ sample was finally constructed.

Cu₂O/SnO₂ sample: The synthetic procedures were the same as that of the Cu₂O/ZnO/SnO₂ sample except for the substitution of the Cu₂O sample for the Cu₂O/ZnO sample as the working electrode for the SnO₂ layer deposition.

Cu₂O/ZnO/Al₂O₃ sample: The synthetic procedures were the same as that of the Cu₂O/ZnO/SnO₂ sample except for the substitution of the Al₂O₃ solution for the colloidal SnO₂ QDs solution in the spin-coating step. The Al₂O₃ solution (0.15M) was prepared by rigorously mixing Al₂O₃ nanopowders with pure water.

Cu₂O/ZnO/R-SnO₂/O-SnO₂ sample: The synthetic procedures were the same as that of the Cu₂O/ZnO/SnO₂ sample except for the substitution of the R-SnO₂ and O-SnO₂ QDs solutions for the pristine SnO₂ QDs solution in the spin-coating step. The R-SnO₂ and O-SnO₂ QDs solutions were then reduced and oxidized by the SnO₂ QDs solution after mild reduction and oxidation treatments, respectively. The reduced R-SnO₂ was obtained by the addition of fresh 0.1M NaBH₄ solution (10 mol%) into pristine colloidal SnO₂ QDs solution (clear yellow, pH = 9) after thoroughly stirring for 4 h. Similarly, the oxidized O-SnO₂ was obtained by the addition of fresh 0.1M H₂O₂ solution (10 mol%) into pristine colloidal SnO₂ QDs solution (clear yellow, pH = 9) after thoroughly stirring for 4 h. In the spin-coating step, the reduced SnO₂ QDs solution (100 μL) was dropped onto the Cu₂O/ZnO surface followed by spinning at 400 rpm for 3 s and 3000 rpm for 60 s. After annealing (160 °C, 2 h) and a N₂ purge, the Cu₂O/ZnO/R-SnO₂ sample was obtained. Subsequently, a similar spin-coating step was repeated for the Cu₂O/ZnO/R-SnO₂ sample using the oxidized SnO₂ QDs solution (100 μL) to obtain the Cu₂O/ZnO/R-SnO₂/O-SnO₂ sample finally.

Cu₂O/ZnO/O-SnO₂/R-SnO₂ sample: The synthetic procedures were the same as that of the Cu₂O/ZnO/R-SnO₂/O-SnO₂ sample except for the reverse order of dropping the oxidized and reduced SnO₂ in the spin-coating step. For the Cu₂O/ZnO/O-SnO₂/R-SnO₂ sample, the spin-coating of the oxidized SnO₂ QDs solution (100 μL) was followed by the spin-coating of the reduced SnO₂ QDs solution (100 μL).

3.2. Characterizations

The compositions of the multilayer were characterized by X-ray diffraction (XRD) pattern spectra (Ultima IV system, Rigaku, Tokyo, Japan) using Cu-Kα radiation as an X-ray source. The diffraction range (2θ) was set from 10° to 80° at 5°/min scanning speed. The surface chemical states of the multilayer were determined by X-ray photoelectron spectra (XPS) (AXIS UltraDLD, Kratos, British) with monochromatic Mg-Kα radiation. The calibration by C 1s peak at 284.6 eV was carried out for all binding energies of XPS peaks. The light absorption ability was reflected by ultraviolet–visible (UV–vis) diffuse reflection spectra (DRS) (UV-1800, Shimadzu, Kyoto, Japan) fitted by an integrating sphere assembly. The microstructure and morphology were observed from the imaging of field-emission scanning electron microscopy (SEM) (JCM-6000, JEOL, Tokyo, Japan) and field-emission transmission electron microscopy (FE-TEM) (TALOS F200X G2, FEI, Waltham, MA, USA). The surface morphology was scanned by atomic force microscopy (AFM) (Dimension XR, Bruker, Karlsruhe, Baden-Württemberg, Germany).

3.3. Photoelectrochemical Measurements

All electrochemical measurements were conducted on an electrochemical workstation (Modulab XM PhotoEchem, Princeton Applied Research, Oak Ridge, Tennessee, USA) in a three-electrode cell configuration. A sodium borate buffer (pH = 10.6) was prepared by mixing a Na₂CO₃ (0.05M, 57 mL 05M, 57 mL) solution and a Na₂B₄O₇ (0.05M, 6 mL 6 mL) solution followed by the addition of 0.1M 4-Hydroxy-TEMPO used as the electrolyte. The presence of 4-Hydroxy-TEMPO as the sacrificial agent accelerates surface redox reaction rates of photocathodes. A Pt filament and an Ag/AgCl (4 M KCl) electrode were selected as the counter electrode and reference electrode, respectively. The potentials measured against the Ag/AgCl reference electrode could be converted to potentials against a reversible hydrogen electrode (RHE) according to the following equation. A solar simulator was applied as the light source, and the corresponding irradiation intensity was adjusted to 100 mW·cm⁻². The sample photocathodes were irradiated from the back side. Photoelectrochemical measurements of linear sweep voltammetry (LSV) and chronopotentiometry (I-t) were conducted to determine the photocurrent density and reaction stability of sample photocathodes. The conditions of LSV and I-t measurements were front-side irradiation from a simulated sunlight source and continuous steady N₂ bubbling in a buffer solution containing sacrificial agents of 4-Hydroxy-TEMPO (pH = 10.6). LSV measurements were conducted with scanning applied potentials from +0.75 V to +0.05 V (vs. RHE) in a scanning rate of 20 mV/s and chopped light irradiation with an interval time of 1.5 s. I-t measurements were conducted with +0.3 V (vs. RHE) and chopped light irradiation with an interval time of 30 min.

$$\mathrm{E} \left(\mathrm{vs}. \mathrm{RHE}\right)=\mathrm{E} \left(\mathrm{vs}. \mathrm{Ag}/\mathrm{AgCl}\right)+0.1976+0.0591\times \mathrm{pH}$$

4. Conclusions

In this work, a tailored multilayer design was applied for the Cu₂O/ZnO/SnO₂ photocathode via sequential deposition steps, including epitaxial growth of the Cu₂O and ZnO layer via electrodeposition and decoration of the SnO₂ QDs layer via spin-coating. The multilayer configuration was determined and demonstrated via top-view and cross-section SEM imaging, which was further supported by XPS detection. Compared to the Cu₂O/ZnO photocathode, the decoration of the SnO₂ layer exhibited dual functions for accelerating electron extraction and transfer and boosting surface reactions. Furthermore, the SnO₂ layer in the form of SnO₂ QDs coating facilitated to suppress photocorrosion of Cu⁺ species, superior to the decoration of the inert Al₂O₃ layer. The protection function of the SnO₂ QDs layer was derived from the inhibition of Cu⁺ transformation and the perfect covering for the exposed surface of the Cu₂O and ZnO layers, which were determined by morphological observation and XPS detection in terms of surface and Ar-etching regions. On the whole, key findings in this work could provide inspiration for the potential application of selective SnO₂ QDs layer decoration on other photoelectrocatalytic and electrocatalytic electrode systems.