Noble Metal Modiﬁcation of CdS-Covered CuInS 2 Electrodes for Improved Photoelectrochemical Activity and Stability

: In this paper, e ﬃ cient and stable photoelectrochemical (PEC) hydrogen (H 2 ) evolution using copper indium sulﬁde (CuInS 2 ) thin ﬁlm electrodes was studied. Modiﬁcation with a cadmium sulﬁde (CdS) layer led to improved charge separation at the interface between CuInS 2 and CdS; however, the photocorrosive nature of CdS induced poor stability of the photocathode. Further surface coating with an electrodeposited Pt layer over the CdS-covered CuInS 2 photocathode prevented the CdS layer from making contact with the electrolyte solution, and enabled e ﬃ cient PEC H 2 evolution without appreciable degradation. This indicates that the Pt layer functioned not only as a reaction site for H 2 evolution, but also as a protection layer. In addition, it was found that surface protection using a noble metal layer was also e ﬀ ective for stable PEC carbon dioxide (CO 2 ) reduction when appropriate noble metal cocatalysts were selected. When Au or Ag was used, carbon monoxide was obtained as a product of PEC CO 2 reduction. sulfate (BaSO 4 ) as the reﬂectance standard. XPS were measured using an X-ray photoelectron spectrometer (AXIS-ULTRA, KRATOS, Stretford, UK). The C 1s peak (284.6 eV) was used as an internal energy reference in all the experiments. Mott–Schottky plots were recorded using an electrochemical analyzer equipped with a frequency response analyzer. The electrode potential was scanned at a frequency of 10 Hz in a 1 M Na 2 SO 4 aqueous solution (pH 6.0) that was purged with Ar prior to each experiment.


Introduction
Hydrogen (H 2 ) is a potential clean energy carrier that could replace fossil fuels and solve the associated environmental issues. Photoelectrochemical (PEC) water splitting using sunlight is a promising approach to producing H 2 without emitting carbon dioxide (CO 2 ), providing a sustainable future energy supply [1,2]. Since the first demonstration of PEC water splitting using a titanium dioxide (TiO 2 ) photoelectrode was reported by Fujishima and Honda [3], semiconductor materials have been extensively studied as photoelectrodes to achieve efficient solar-to-H 2 conversion [4][5][6][7][8][9][10].
Copper chalcopyrite compounds, formulated as Cu(In,Ga)(S,Se) 2 , have recently received considerable attention as promising candidates for photocathodes, owing to their high absorption coefficients (>10 4 M −1 cm −1 ), tunable narrow bandgaps (1.5-2.4 eV), and favorable conduction band positions with respect to the hydrogen evolution reaction (HER) [11][12][13][14][15][16]. Moreover, the PEC performance of copper chalcopyrite photocathodes has been reported to improve when combined with n-type semiconductor buffer layers such as cadmium sulfide (CdS), indium sulfide (In 2 S 3 ), and zinc sulfide (ZnS). This is due to the formation of a p-n junction, which facilitates charge separation at the interface [17][18][19][20]. However, to achieve efficient and stable PEC H 2 evolution, further improvements will be necessary to overcome the sluggish surface redox reactions and the instability of sulfide materials in aqueous solutions.
Noble metal cocatalysts such as Pt, Au, and Ag have been widely used to catalyze redox reactions efficiently [21][22][23][24]. Their loading not only facilitates the separation of photogenerated charges, but also  Figure 2a shows UV-vis-NIR absorption spectra of the CuInS2 and Pt0.88/CdS/CuInS2 films. As reported in the literature, the CuInS2 film showed intense absorption over a wide wavelength range from 300 to 850 nm [14,31,32]. The bandgap was calculated to be 1.46 eV using a Tauc plot ( Figure  2b), and the flat band potential was determined to be approximately 0.84 V using a Mott-Schottky plot (Figure 2c), indicating that the band position of the prepared CuInS2 film (Figure 2d) was in 38 Figure 2a shows UV-vis-NIR absorption spectra of the CuInS 2 and Pt 0.88 /CdS/CuInS 2 films. As reported in the literature, the CuInS 2 film showed intense absorption over a wide wavelength range from 300 to 850 nm [14,31,32]. The bandgap was calculated to be 1.46 eV using a Tauc plot (Figure 2b), and the flat band potential was determined to be approximately 0.84 V using a Mott-Schottky plot (Figure 2c), indicating that the band position of the prepared CuInS 2 film (Figure 2d) was in accordance with that previously determined by photoelectron spectroscopy [13,18]. The Pt 0.88 /CdS/CuInS 2 film showed an absorption spectrum similar to that of the CuInS 2 film, with an increase in absorption being observed at wavelengths above 820 nm (Figure 2a). The origin of this increased absorption is considered to be light scattering by the deposited Pt, which has often been observed for photocatalysts modified with Pt nanoparticles [33,34].  Figure 2a shows UV-vis-NIR absorption spectra of the CuInS2 and Pt0.88/CdS/CuInS2 films. As reported in the literature, the CuInS2 film showed intense absorption over a wide wavelength range from 300 to 850 nm [14,31,32]. The bandgap was calculated to be 1.46 eV using a Tauc plot ( Figure  2b), and the flat band potential was determined to be approximately 0.84 V using a Mott-Schottky plot (Figure 2c), indicating that the band position of the prepared CuInS2 film (Figure 2d) was in accordance with that previously determined by photoelectron spectroscopy [13,18]. The Pt0.88/CdS/CuInS2 film showed an absorption spectrum similar to that of the CuInS2 film, with an increase in absorption being observed at wavelengths above 820 nm (Figure 2a). The origin of this increased absorption is considered to be light scattering by the deposited Pt, which has often been observed for photocatalysts modified with Pt nanoparticles [33,34]. The morphological features of the synthesized films were investigated by SEM. Top-down and cross-sectional views of the CuInS2 film ( Figure S2) revealed that it consisted of a dense crystalline layer accompanied by plate-like particles on the surface, and its thickness was found to be about 1. The morphological features of the synthesized films were investigated by SEM. Top-down and cross-sectional views of the CuInS 2 film ( Figure S2) revealed that it consisted of a dense crystalline layer accompanied by plate-like particles on the surface, and its thickness was found to be about 1.3 µm. After modification with CdS, the surface became relatively smooth and porous (Figure 3a). In contrast, the formation of nanoparticles, with diameter of approximately 10 nm, was observed for the Pt-deposited samples ( Figure S3), and higher amounts of Pt deposition resulted in a more granular surface (Figure 3b-f). Furthermore, it was observed that Pt particles not only formed on the CdS grains, but also filled the pores of the CdS, suggesting that the CdS surface was entirely covered when depositing a sufficient amount of Pt.
This suggestion was supported by the XPS results. Figure 4a,b shows the XPS spectra of CdS/CuInS 2 and Pt/CdS/CuInS 2 films for the Cd 3d ( Figure 4a) and Pt 4f (Figure 4b) regions, respectively. As Pt loading was increased, the Cd 3d peaks significantly weakened while the Pt 4f peaks became stronger, with the Cd 3d peaks almost disappearing in the case of Pt 100 /CdS/CuInS 2 . On the basis of these results, it was anticipated that the CdS layer could be protected during PEC H 2 evolution by depositing Pt onto the surface. µm. After modification with CdS, the surface became relatively smooth and porous (Figure 3a). In contrast, the formation of nanoparticles, with diameter of approximately 10 nm, was observed for the Pt-deposited samples ( Figure S3), and higher amounts of Pt deposition resulted in a more granular surface (Figure 3b-f). Furthermore, it was observed that Pt particles not only formed on the CdS grains, but also filled the pores of the CdS, suggesting that the CdS surface was entirely covered when depositing a sufficient amount of Pt. This suggestion was supported by the XPS results. Figure 4a,b shows the XPS spectra of CdS/CuInS2 and Pt/CdS/CuInS2 films for the Cd 3d ( Figure 4a) and Pt 4f (Figure 4b) regions, respectively. As Pt loading was increased, the Cd 3d peaks significantly weakened while the Pt 4f peaks became stronger, with the Cd 3d peaks almost disappearing in the case of Pt100/CdS/CuInS2. On the basis of these results, it was anticipated that the CdS layer could be protected during PEC H2 evolution by depositing Pt onto the surface.

Photoelectrochemical Measurements
The PEC properties of the prepared electrodes were examined by linear sweep voltammetry measured under visible light irradiation. Although the bare CuInS2 electrode showed limited photoresponse ( Figure S4), a distinct cathodic photocurrent was generated after modification with both CdS and Pt. This is because the CdS layer deposited on CuInS2 possessed suitable band characteristics for the formation of a p-n junction (Figure 2d and Figure    This suggestion was supported by the XPS results. Figure 4a,b shows the XPS spectra of CdS/CuInS2 and Pt/CdS/CuInS2 films for the Cd 3d ( Figure 4a) and Pt 4f (Figure 4b) regions, respectively. As Pt loading was increased, the Cd 3d peaks significantly weakened while the Pt 4f peaks became stronger, with the Cd 3d peaks almost disappearing in the case of Pt100/CdS/CuInS2. On the basis of these results, it was anticipated that the CdS layer could be protected during PEC H2 evolution by depositing Pt onto the surface.

Photoelectrochemical Measurements
The PEC properties of the prepared electrodes were examined by linear sweep voltammetry measured under visible light irradiation. Although the bare CuInS2 electrode showed limited photoresponse ( Figure S4), a distinct cathodic photocurrent was generated after modification with both CdS and Pt. This is because the CdS layer deposited on CuInS2 possessed suitable band characteristics for the formation of a p-n junction (Figure 2d and Figure S5)

Photoelectrochemical Measurements
The PEC properties of the prepared electrodes were examined by linear sweep voltammetry measured under visible light irradiation. Although the bare CuInS 2 electrode showed limited photoresponse ( Figure S4), a distinct cathodic photocurrent was generated after modification with both CdS and Pt. This is because the CdS layer deposited on CuInS 2 possessed suitable band characteristics for the formation of a p-n junction (Figure 2d and Figure S5), facilitating extraction of excited electrons from the inside of CuInS 2 to the electrode surface, while Pt acted as a cocatalyst to reduce overpotential for H 2 evolution. Consequently, modification with these materials enabled the extraction of photoexcited electrons from CuInS 2 to the electrolyte. In contrast, the modification with only CdS made no increase in photocurrent ( Figure S4). Figure 5 shows polarization curves of the Pt-modified CdS/CuInS 2 electrodes with different Pt contents. As the amount of Pt was increased up to 0.88 µmol, a gradual improvement in photocurrent density and a positive shift in the onset potential were observed; in contrast, the increase in Pt loading had almost no effect on current generation under dark conditions ( Figure S6). However, further increases in Pt content above 0.88 µmol were detrimental to current density, as excess Pt loading limited light absorption by CuInS 2 . Thus, the largest photoresponse was observed for Pt 0.88 /CdS/CuInS 2 , with its photocurrent reaching 7.5 mA cm −2 at 0 V. This value is about half of the photocurrent recorded using the Pt/CdS/(CuInS 2 ) 0.81 (ZnS) 0.19 electrode which is a state-of-the-art CuInS 2 -based photocathode prepared by tuning the band position of the light absorber [12].
detrimental to current density, as excess Pt loading limited light absorption by CuInS2. Thus, the largest photoresponse was observed for Pt0.88/CdS/CuInS2, with its photocurrent reaching 7.5 mA cm −2 at 0 V. This value is about half of the photocurrent recorded using the Pt/CdS/(CuInS2)0.81 (ZnS)0.19 electrode which is a state-of-the-art CuInS2-based photocathode prepared by tuning the band position of the light absorber [12]. To confirm the H2 generation ability of the Pt/CdS/CuInS2 electrode, the amount of H2 evolved during photoelectrolysis was quantified. Figure 6 shows the typical time courses of H2 evolution over the Pt/CdS/CuInS2 photoelectrodes with an applied potential of 0 V and one-half of the electrons passing through the outer circuit (e -/2). Continuous H2 production was observed for all the photoelectrodes, and approximately 180 µmol of H2 generated over Pt0.88/CdS/CuInS2 after illumination for 120 min. The faradaic efficiency, defined as the ratio of the amount of H2 to that of e -/2, was calculated to be over 99% for all the Pt/CdS/CuInS2 photoelectrodes, indicating that no reduction reactions other than H2 evolution proceeded, irrespective of the amount of Pt. To confirm the H 2 generation ability of the Pt/CdS/CuInS 2 electrode, the amount of H 2 evolved during photoelectrolysis was quantified. Figure 6 shows the typical time courses of H 2 evolution over the Pt/CdS/CuInS 2 photoelectrodes with an applied potential of 0 V and one-half of the electrons passing through the outer circuit (e -/2). Continuous H 2 production was observed for all the photoelectrodes, and approximately 180 µmol of H 2 generated over Pt 0.88 /CdS/CuInS 2 after illumination for 120 min. The faradaic efficiency, defined as the ratio of the amount of H 2 to that of e -/2, was calculated to be over 99% for all the Pt/CdS/CuInS 2 photoelectrodes, indicating that no reduction reactions other than H 2 evolution proceeded, irrespective of the amount of Pt.
Catalysts 2020, 10, x 6 of 12  These results indicate that the stability of the Pt/CdS/CuInS2 photoelectrodes significantly depended on the Pt loading. Although rapid photocurrent decays were observed for Pt0.22/CdS/CuInS2 and Pt0.44/CdS/CuInS2, a significant improvement in the stability was observed at higher Pt loadings, with Pt100/CdS/CuInS2 retaining 99% of its photocurrent after photoelectrolysis. Photocurrent decay in CdS-modified photocathodes has previously been studied by several groups, and the phenomenon was attributed to the poor stability of CdS due to photocorrosion [28,35].   These results indicate that the stability of the Pt/CdS/CuInS 2 photoelectrodes significantly depended on the Pt loading. Although rapid photocurrent decays were observed for Pt 0.22 /CdS/CuInS 2 and Pt 0.44 /CdS/CuInS 2 , a significant improvement in the stability was observed at higher Pt loadings, with Pt 100 /CdS/CuInS 2 retaining 99% of its photocurrent after photoelectrolysis. Photocurrent decay in CdS-modified photocathodes has previously been studied by several groups, and the phenomenon was attributed to the poor stability of CdS due to photocorrosion [28,35]. experiments. These results indicate that the stability of the Pt/CdS/CuInS2 photoelectrodes significantly depended on the Pt loading. Although rapid photocurrent decays were observed for Pt0.22/CdS/CuInS2 and Pt0.44/CdS/CuInS2, a significant improvement in the stability was observed at higher Pt loadings, with Pt100/CdS/CuInS2 retaining 99% of its photocurrent after photoelectrolysis. Photocurrent decay in CdS-modified photocathodes has previously been studied by several groups, and the phenomenon was attributed to the poor stability of CdS due to photocorrosion [28,35].   were observed in some areas, whereas the Pt 088 /CdS/CuInS 2 electrode showed few changes other than the aggregation of Pt particles. The XPS spectra also showed evidence of the photocorrosion of CdS for the CdS/CuInS 2 electrode, with an increase in the Cu 2p and In 3d peaks being observed after photoelectrolysis (Figure 8b). In contrast, the Cu 2p and In 3d spectra of the Pt 0.88 /CdS/CuInS 2 electrode remained unchanged (Figure 8c). Therefore, Pt modification was found to be effective for both promoting PEC H 2 evolution and improving the stability of the CdS/CuInS 2 photocathode.
Catalysts 2020, 10, x 7 of 12 to be effective for both promoting PEC H2 evolution and improving the stability of the CdS/CuInS2 photocathode.  Figure 3a,e, respectively.) Comparison of XPS spectra of (b) CdS/CuInS2 and (c) Pt0.88/CdS/CuInS2 electrodes before and after 120 min of photoelectrolysis.

Photoelectrochemical CO2 Reduction
Finally, we examined PEC CO2 reduction using noble-metal-modified CuInS2 electrodes in CO2saturated NaHCO3 solutions. As was the case for PEC H2 generation, the photoresponse of the bare CuInS2 film was so small that CdS and metal cocatalysts were deposited to enhance photoreactivity. For this experiment, Au and Ag were used as cocatalysts in addition to Pt because electrochemical CO2 reduction using Au and Ag has shown high activity and selectivity for CO formation [36,37]. The loadings of all the cocatalysts were adjusted to 0.88 µmol based on the PEC H2 generation results achieved using the Pt/CdS/CuInS2 photocathode. Table 1 shows the results of PEC CO2 reduction  Figure 3a,e, respectively.) Comparison of XPS spectra of (b) CdS/CuInS 2 and (c) Pt 0.88 /CdS/CuInS 2 electrodes before and after 120 min of photoelectrolysis.

Photoelectrochemical CO 2 Reduction
Finally, we examined PEC CO 2 reduction using noble-metal-modified CuInS 2 electrodes in CO 2 -saturated NaHCO 3 solutions. As was the case for PEC H 2 generation, the photoresponse of the Catalysts 2020, 10, 949 7 of 12 bare CuInS 2 film was so small that CdS and metal cocatalysts were deposited to enhance photoreactivity. For this experiment, Au and Ag were used as cocatalysts in addition to Pt because electrochemical CO 2 reduction using Au and Ag has shown high activity and selectivity for CO formation [36,37]. The loadings of all the cocatalysts were adjusted to 0.88 µmol based on the PEC H 2 generation results achieved using the Pt/CdS/CuInS 2 photocathode. Table 1 shows the results of PEC CO 2 reduction conducted at −0.11 V under visible light irradiation. By adding metal cocatalysts, the photocurrent density increased in the order of Pt > Au > Ag; however, the photocurrent of Pt 0.88 /CdS/CuInS 2 was much smaller than that measured for PEC H 2 evolution. It is notable that deactivation during PEC CO 2 reduction can occur not only due to corrosion of CdS but also due to the poisoning of the cocatalyst. This photocurrent decrease was presumed to be caused by the poisoning of the Pt cocatalyst because CO molecules formed as the intermediate species occupied the active sites of Pt without being released into the solution [38]. In fact, product analysis indicated that a negligible amount of CO 2 reduction products was formed and that the faradic efficiency for H 2 evolution was almost 100%, although CO 2 reduction by photoexcited electrons generated in Pt 0.88 /CdS/CuInS 2 was thermodynamically favorable. Furthermore, XPS spectra measured after photoelectrolysis showed no increase in the peaks assigned to Cu and In, suggesting that CdS photocorrosion was also suppressed in this experiment ( Figure S7). When Au/CdS/CuInS 2 and Ag/CdS/CuInS 2 were employed for PEC CO 2 reduction, continuous CO production was observed, with the faradaic efficiencies for CO formation over Au/CdS/CuInS 2 and Ag/CdS/CuInS 2 being 6.7% and 3.5%, respectively. Although most of the photocurrent was still derived from H 2 evolution, the observed faradaic efficiencies for CO formation were relatively high compared to other relevant studies reported so far [39][40][41]. In addition, the CO 2 reduction photocurrent density (ca. 100 µA/cm 2 ) observed using Au/CdS/CuInS 2 was about 80-fold higher than that observed for CuGaS 2 , which is the only copper chalcopyrite compound reported for PEC CO 2 reduction [40]. SEM observation showed that both the Au and Ag cocatalysts had dendritic structures ( Figure S8), and similar-shaped electrocatalysts have been reported to convert CO 2 to CO with high selectivities of over 80% when overpotentials of 0.3-0.5 V are applied [42,43]. According to the conduction band minimum (CBM) potential of CdS [14,17], photoexcited electrons in the Au/CdS/CuInS 2 and Ag/CdS/CuInS 2 photocathodes were considered to possess reduction potentials of around −0.3 V, corresponding to 0.19 V of overpotentials. Recently, the CBM potential of CdS was found to be tunable by forming a solid solution with ZnS, and the introduction of ZnS shifted the CBM to a more negative potential [44]. Furthermore, a thin film of CdS-ZnS solid solution can be formed by the chemical bath deposition (CBD) method [45], which was used for deposition of the CdS layer in this study. Thus, the utilization of a CdS-ZnS solid solution layers in combination with appropriate amounts of Au and Ag cocatalysts would offer the possibility to achieve PEC CO 2 reduction with high selectivity and durability, and such studies are underway in our laboratory.

Preparation of CuInS 2 Film on Mo Foil
CuInS 2 thin film electrodes were synthesized using electrodeposition and subsequent sulfurization, as reported in the literature [14,31,32]. Prior to electrodeposition, Mo foil (Nilaco, Japan, 99.95%, 0.15 mm thickness) was cleaned by sonicating for 30 min in 0.1 M HCl. Electrodeposition was conducted using an electrochemical analyzer (HZ-5000, Hokuto Denko, Japan) and a single cell with a three-electrode setup consisting of the Mo foil, a Pt wire, and a silver/silver chloride (Ag/AgCl) electrode, which functioned as the working, counter, and reference electrodes, respectively. Metallic Cu and In layers were successively deposited onto the Mo foil under potentiostatic conditions. Cu deposition was conducted at −0.2 V vs. Ag/AgCl in an aqueous solution containing 50 mM CuSO 4 , 242 mM citric acid, and 150 mM trisodium citrate. Subsequently, In deposition onto the Cu layer was performed at −0.76 V vs. Ag/AgCl in an aqueous solution containing 30 mM InCl 3 ·4H 2 O, 36 mM citric acid, and 242 mM trisodium citrate. The amount of deposited Cu and In was controlled by adjusting the electric charge to 1042 and 1200 mC/cm 2 , respectively. The as-prepared Cu/In film was sealed in an evacuated Pyrex ampule with 5 mg of elemental sulfur and then sulfurized by heating at 600 • C for 10 min. Finally, the thus-obtained CuInS 2 films were etched by immersion in an aqueous KCN solution (10%) for 2 min to remove excess Cu x S.

Surface Modification of CuInS 2 Electrode
The CuInS 2 electrode was modified with CdS by the CBD method. The CuInS 2 electrode was immersed in an aqueous solution containing 12.5 mM CdSO 4 , 0.22 M thiourea, and 11 M NH 4 OH for 7 min at 60 • C. After the modification, the electrode was thoroughly rinsed with distilled water and dried at room temperature. CuInS 2 electrodes modified with CdS are hereafter denoted as CdS/CuInS 2 . The Pt cocatalyst was deposited on the CdS/CuInS 2 electrode by photoelectrodeposition. The electrode was immersed in 0.1 M Na 2 SO 4 solution (pH 4.0) containing 1 mM H 2 PtCl 6 , and the electrodeposition was carried out at −0.1 V vs. Ag/AgCl under illumination with a Xe lamp (LA-251 Xe; Hayashi Tokei). The amount of Pt was controlled to 0.22, 0.44, 0.66, 0.88, and 1.00 µmol by monitoring the electric charge passed through the electrode, and the electrode was then thoroughly rinsed with distilled water and dried at room temperature. Thus, the photoelectrode was prepared and its surface area was set to 0.7 cm 2 (0.7 cm × 1.0 cm). Hereafter, the electrodes thus obtained are denoted as Pt xxx /CdS/CuInS 2 , where XXX corresponds to the amount of electrodeposited Pt. HAuCl 4 (6 mM) and AgNO 3 (5 mM) were used as the electrolytes for the preparation of the Au-and Ag-deposited samples, respectively, with these being denoted as Au/CdS/CuInS 2 and Ag/CdS/CuInS 2 , respectively.

Characterization
XRD patterns were obtained using a PW-1700 system (PANalytical, Almelo, The Netherlands). The morphologies of the prepared films were observed by SEM (JSM-6500F, JEOL, Tokyo, Japan). Optical absorption spectra were measured in diffuse reflectance (DR) mode using a UV-vis-NIR spectrometer (UV-2600, Shimadzu, Kyoto, Japan) with barium sulfate (BaSO 4 ) as the reflectance standard. XPS were measured using an X-ray photoelectron spectrometer (AXIS-ULTRA, KRATOS, Stretford, UK). The C 1s peak (284.6 eV) was used as an internal energy reference in all the experiments. Mott-Schottky plots were recorded using an electrochemical analyzer equipped with a frequency response analyzer. The electrode potential was scanned at a frequency of 10 Hz in a 1 M Na 2 SO 4 aqueous solution (pH 6.0) that was purged with Ar prior to each experiment.

Photoelectrochemical Measurements
PEC measurements were performed using the same three-electrode setup as used for the preparation of the CuInS 2 electrodes. The electrolyte solutions used for H 2 evolution and CO 2 reduction were Ar-saturated 0.2 M KH 2 PO 4 (pH 6.0) and CO 2 -saturated 0.5 M NaHCO 3 (pH 6.8), respectively. Prior to the measurements, Ar or CO 2 gas was purged through the electrolyte for 20 min to remove dissolved oxygen. A Xe lamp with an optical filter (>420 nm, Y-44, Hoya) was used as a light source. All measured potentials were converted to the reversible hydrogen electrode (RHE) scale using the following equation: E RHE = E Ag/AgCl + 0.059 × pH + 0.199.
During the measurements, Ar or CO 2 gas was purged through the electrolyte at a flow rate of 20 mL min −1 and collected in a clean plastic bag (Smart Bag PA, GL Sciences) along with gaseous products. Evolved H 2 was quantified using a gas chromatograph (GC-8A, Shimadzu) equipped with a thermal conductivity detector (TCD). For the quantification of CO, a gas chromatograph (GC-8A, Shimadzu) equipped with a methanizer and a flame ionization detector (FID) was used.

Conclusions
In this study, we investigated the effect of Pt loading on the PEC H 2 -evolution activity and durability of CdS/CuInS 2 photocathodes. Structural and chemical characterization of the surface revealed that electrodeposited Pt covered the CdS surface, and the porous structure was filled with Pt by increasing the loading amount. This prevented the exposure of the CdS layer to the electrolyte. PEC experiments showed that the H 2 -evolution activity could be enhanced by Pt modification, resulting in both an increase in photocurrent density and a positive shift in the onset potential. As a result, efficient PEC H 2 evolution proceeded over the Pt/CdS/CuInS 2 photocathodes without appreciable degradation. In addition, the suppression of CdS corrosion using noble metals was also effective for PEC CO 2 reduction experiments, and CO generation was observed when Au or Ag was electrodeposited instead of Pt.

Conflicts of Interest:
The authors declare no conflict of interest.