Characterization and Optimization of Silver-Modified In0.2Cd0.8S-Based Photocatalysts.

In this research, we performed scanning electrochemical microscopy to screen M x (In0.2Cd0.8)1-x S (M = V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, Ag, W, Ir, Pt, and Te) photocatalyst arrays for efficient photoelectrochemical reaction. Doping 30% Ag to form the Ag0.3(In0.2Cd0.8)0.7S electrode could result in the highest photocurrent, and also, the anode photocurrents were found to be 1 and 0.53 mA/cm2 under UV-visible and visible light, respectively, comparatively higher than that of the In0.2Cd0.8S electrode (0.45 and 0.25 mA/cm2). The highest incident photo-to-current conversion efficiency of the Ag0.3(In0.2Cd0.8)0.7S photocatalyst and In0.2Cd0.8S were found to be 64% (λ = 450 nm) and 57% (λ = 400 nm), respectively. The Mott-Schottky plots showed that In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photoelectrodes could exhibit a flat-band potential of -0.85 and -0.55 V versus Ag/AgCl, respectively. Based on these findings, the superior photocatalytic activity of the Ag0.3(In0.2Cd0.8)0.7S photoelectrode was mainly attributed to its high crystalline structure for efficient charge separation and reduction of charge recombination in the heterojunction of Ag0.3(In0.2Cd0.8)0.7S and Ag2S.

Introduction

As the global warming issue is becoming important, the utilization of solar energy for photoelectrochemical and photocatalytic splitting of water splitting into hydrogen and oxygen has attracted much attention due to the hydrogen as an alternative clean fuel.[1−3] In general, chalcogenide semiconductors could be considered as a good photocatalyst because of the suitable energy band corresponding to visible-light absorption. Among these chalcogenide semiconductors, CdS with a band gap of 2.4 eV was known as one of the promising photocatalysts for hydrogen production due to its sufficiently negative flat-band potential of −0.87 V and its good absorption capacity in the visible region of the solar spectrum.[4] For CdS, there is an issue of photocorrosion under prolonged irradiation and it is necessary to dope noble metals on its surface for efficient water splitting.

Previous reports showed that the Cd1–ZnS photocatalyst could achieve an apparent quantum yield (AQY) of 0.6%[5] and Cd0.1Cu0.01Zn0.89S could improve the AQY to 9.6% in the absence of Pt cocatalyst.[6] It was reported that a novel thermal sulfuration method could effectively increase the AQY to 10.23% for the Cd0.8Zn0.2S photocatalyst.[7] It was also suggested that doping Ni2+ ions into Cd1–ZnS solid solution could form a donor level above the valence band of Cd1–ZnS by still maintaining its conduction band. This tuning of the band structure could reduce its band gap and increase its visible-light absorption, greatly by improving the AQY to 15.9%.[8] The photocurrent of the Cd0.8Zn0.2S photoanode was three times higher than that of pure CdS, and the photocatalytic H2 evolution rate of Cd0.8Zn0.2S with 3 wt % Pt cocatalyst was observed as 3020 μmol g–1 h–1 under simulated solar light irradiation.[9] By applying zirconium–titanium phosphate in CdS–ZnS as a new composite material, a hydrogen production amount of 2142.7 μmol with an AQY of 9.6% under visible light was achieved.[10] In addition, a heterostructured ZnS–CuS–CdS composite photocatalyst could cause a high hydrogen production rate of 837.6 μmol g–1 h–1 under solar irradiation.[11] Hydrogen production from concentrated solar radiation was examined by handling the CdS–ZnS photocatalyst in a continuous flow reactor system and by presenting a wider spectrum capturing corresponding to 18% of the incident energy.[12] Recently, multinary copper-based chalcogenides photocatalysts, CuGaS2 and CuGaZnS, were reported to reinforce the charge separation and transfer for enhancing photocatalytic hydrogen evolution.[13,14] For the photovoltaic application, the modified WO3-based electrode was combined with a dye-sensitized solar cell to fabricate tandem cells, showing higher photocurrent density compared with the pristine WO3-based tandem cell.[15,16]

Combinatorial chemistry could provide an effective method for discovering and screening large numbers of diverse new materials by different combinations of specific building block atoms and molecules.[17] This method has been applied in searching for photocatalyst, gate dielectric and fuel cell catalyst materials.[18−20] The automated electrochemical synthesis was performed for screening diverse metal-doped tungsten oxides (Ni–W mixed oxide)[21] and Zn0.956Co0.044O[22] with the optimal solar hydrogen production. The other inkjet printing technique was also performed for screening various metal oxide patterns which could be used as photocatalysts.[23,24] A method, scanning electrochemical microscopy (SECM)[25−27] with an optical fiber, was applied for rapid screening of several metal oxide photocatalysts for water oxidation or hydrogen evolution. Some studies reported that by using an ultramicro-electrode tip connected to a xenon lamp, to replace an optical fiber for illuminating each spot on a photocatalyst array, the photocurrents could be obtained representing the activity of the photocatalyst.[28−34] Having considered all these, the aim of this work was to apply SECM for screening 15 kinds of metal-modified In0.2Cd0.8S-based photocatalyst arrays in Na2SO4/Na2SO3 solution. The suitable Ag(In0.2Cd0.8)1–S-based photocatalysts with optimum parameters were further characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), UV–visible spectroscopy (UV–vis), steady-state photoluminescence spectroscopy (PL), time-resolved PL spectroscopy, and electrochemical impedance spectroscopy (EIS). The flat potentials of the photocatalysts were measured by the Mott–Schottky method. The incident photon-to-current conversion efficiencies (IPCEs) in 0.1 M Na2SO4/Na2SO3 solution were recorded via the lock-in technique.

Results and Discussion

SECM for Screening Photocatalyst Arrays

Figure shows the SECM images of the metal-doped [(a) V, (b) Cr, (c) Mn, (d) Fe, (e) Co, (f) Ni, (g) Cu, (h) Zn, (i) Mo, (j) Ru, (k) Ag, (l) W, (m) Ir, (n) Pt, and (o) Te] In0.2Cd0.8S photocatalyst arrays at an applied potential of 0 V versus Ag/AgCl in 0.1 M Na2SO4/0.1 M Na2SO3 solution under UV–visible light illumination. Figure p presents the preprogrammed patterned photocatalyst arrays composed of In0.2Cd0.8/metal mixture precursors with a total volume of 15 drops for each spot, and the ratio represents the volumes of In0.2Cd0.8 and metal precursors (see Figure S1). The dark brown color in spot indicated the higher photocurrent, whereas the green color indicated the lower photocurrent. It was visually observed that only by doping a Ag metal precursor (Figure k) in InCl3/Cd(CH3COO)2 mixed precursor solutions could enhance the photocatalytic activity and photocurrent at a Ag composition of 30%. Doping the other 14 kinds of metal precursors indicated that the photocurrent was apparently decreased starting from 10% doping percentage.

Figure 1

SECM images of the (a) V-, (b) Cr-, (c) Mn-, (d) Fe-, (e) Co-, (f) Ni-, (g) Cu-, (h) Zn-, (i) Mo-, (j) Ru-, (k) Ag-, (l) W-, (m) Ir-, (n) Pt-, and (o) Te-modified In0.2Cd0.8S photocatalyst arrays at an applied potential of 0 V vs Ag/AgCl in 0.1 M Na2SO3/0.1 M Na2SO4 solution under UV–visible light illumination with a scan rate of 500 μm s–1. (p) Preprogrammed patterned photocatalyst composed of In0.2Cd0.8S/metal mixture precursors. Each patterned drop contains a total volume of 15 drops.

Photoelectrochemical Properties of an Electrode in Bulk Film

Figure a shows the cyclic voltammogram curves of In0.2Cd0.8S electrode with an applied potential in the range of −1.5 to 1.2 V versus Ag/AgCl in 0.1 M Na2SO4/Na2SO3 solution under dark, UV–visible and visible-light illumination. It is observed that the n-type semiconductor, In0.2Cd0.8S, electrode exhibits an enhanced anodic current under light illumination. The applied onset potential was about −0.97 V and the stable photocurrent was presented in the potential region between −0.5 and 0.5 V. The current for the In0.2Cd0.8S electrode significantly increased under light illumination due to the splitting of water, when the applied potential was more positive than 0.8 V. Thus, the potential at 0 V was set as the applied potential for further screening and photoelectrochemical measurements. Bulk Ag(In0.2Cd0.8)1–S film was prepared as the electrode to confirm the SECM results being applicable to large-scale films.

Figure 2

(a) Linear voltammetry of the In0.2Cd0.8S electrode under dark, UV–visible, and visible-light illumination. (b) Photocurrent of Ag(In0.2Cd0.8)1–S electrodes at various atomic percent of Ag and (c) chopped current time transient response of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S electrodes under UV–visible and visible-light illumination. (d) IPCE plots of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S electrodes calculated from the photocurrents at 0 V vs Ag/AgCl.

Figure b shows the photocurrent obtained from Ag(In0.2Cd0.8)1–S electrodes with different Ag volume percent doping illuminated under UV–visible and visible lights. The results show that the Ag0.3(In0.2Cd0.8)0.7S electrode formed by doping 30% Ag has the highest photocurrent under both UV–visible and visible lights. With the increasing percentage of Ag above 30%, there was a drastic decrease in the photocurrent, showing the consistent trend with the SECM screening results. Figure c represents the current–time transient responses of In0.2Cd0.8S and Ag(In0.2Cd0.8)1–S electrodes under chopped UV–visible and visible-light illumination. The anode photocurrents of the In0.2Cd0.8S electrode were 0.45 and 0.25 mA/cm2 at a bias of 0 V versus Ag/AgCl under UV–visible and visible-light illumination, respectively. For the Ag(In0.2Cd0.8)1–S electrode, the photocurrents could be increased to 1 and 0.53 mA/cm2 under UV–visible and visible-light illumination, respectively. These results show that the addition of Ag as the third metal component in photocatalyst could effectively enhance the photocurrent of Ag(In0.2Cd0.8)1–S twice than that of In0.2Cd0.8S electrodes. Figure d represents the IPCE plots of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S electrodes, which were measured in a 0.1 M Na2SO4/Na2SO3 solution at 0 V versus Ag/AgCl under monochromatic light irradiation. The IPCE was described in eq where iph is the photocurrent density (mA/cm2), λ is the wavelength (nm) of incident radiation, and Pin is the incident light power density (mW/cm2) at the selected wavelength. The In0.2Cd0.8S electrode displayed high IPCE values of 57 and 21% at the wavelength of 400 and 500 nm, respectively, whereas the Ag0.3(In0.2Cd0.8)0.7S electrode showed an increasing IPCE values to 61 and 42% at the corresponding wavelength, indicating a significant improvement of photocatalytic efficiency.

Figure a shows the UV–vis absorption spectroscopy of In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S, indicating that Ag0.3(In0.2Cd0.8)0.7S has the enhanced absorption than In0.2Cd0.8S in the UV and visible-light ranges. The direct band gaps of In0.2Cd0.8S, Ag0.3(In0.2Cd0.8)0.7S and Ag2S were determined by the intercept on the photon energy axis in the Kubelka–Munk plot (see Figure S2). Figure b shows the steady-state PL spectra of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S with the excitation wavelength of 450 nm according to the high absorbance value in UV–vis absorption spectroscopy. There are two main PL peak at 522.8 nm for In0.2Cd0.8S and 528.4 nm for Ag0.3(In0.2Cd0.8)0.7S. It can be seen that the PL emission intensity of Ag0.3(In0.2Cd0.8)0.7S is significantly decreased, indicating the suppressed of recombination of electrons and holes. This phenomenon could be resulted from the formation of a heterojunction of Ag0.3(In0.2Cd0.8)0.7S and Ag2S favoring charge transfer. We analyzed the charge recombination and obtained the carriers lifetimes by using time-resolved PL spectroscopy to monitor the PL decay (see Figure S3). The result shows Ag0.3(In0.2Cd0.8)0.7S with a slower PL decay than In0.2Cd0.8S and the carrier lifetimes are 11.2 and 14.0 ns for In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S, respectively. The longer decay time of Ag0.3(In0.2Cd0.8)0.7S is owing to the further reduction of bulk recombination caused by the heterojunction of Ag0.3(In0.2Cd0.8)0.7S and Ag2S.

Figure 3

(a) UV–vis absorption spectroscopy and (b) PL spectroscopy of In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S.

Figure a,b represents Mott–Schottky plots of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S bulk film electrodes at 3000 Hz in the dark condition (C–2 vs E, where C is the space charge capacitance of the semiconductor electrode). The flat-band potential could be estimated by the Mott–Schottky equation (eq ).where C is the space charge capacitance (F/cm2), q is the electronic charge, ε is the dielectric constant for the semiconductor, ε0 is the permittivity of free space (8.85 × 10–14 F/cm), and Nd is the carrier density (cm–3). V is the applied potential (Volts) and Vfb is the flat-band potential (Volts), k is the Boltzmann constant and T represents the temperature (K). The flat-band potential (Vfb) of the In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S estimated from intercepts of the Mott–Schottky plots were about −0.85 and −0.55 V versus Ag/AgCl, respectively. In addition, positive slopes of the curves confirmed that In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S films have electronic properties in n-type.

Figure 4

Mott–Schottky plots of (a) In0.2Cd0.8S and (b) Ag0.3(In0.2Cd0.8)0.7S photocatalyst at 3000 Hz.

We also performed EIS to study the interfacial properties between the photocatalyst electrodes and electrolyte solutions in the dark measured at 0 V versus Ag/AgCl. In Figure , EIS plots of both In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S present wide impedance arcs, which suggests that few charges could pass through the interface between the photoanode and electrolyte in the dark condition. Compared with the In0.2Cd0.8S electrode, the Ag0.3(In0.2Cd0.8)0.7S electrode presents a smaller impedance arc, representing a small interface transfer resistance which may be caused by the larger electrochemical active surface area.[15]

Figure 5

EIS plots In0.2Cd0.8S, and Ag0.3(In0.2Cd0.8)0.7S photocatalyst in the dark measured at 0 V vs Ag/AgCl.

Characterization of Photocatalyst Arrays

Figure shows the SEM images of Ag(In0.2Cd0.8)1–S (x = 0–0.9) and Ag2S spots in photocatalyst arrays. Without the addition of Ag precursor, the In0.2Cd0.8S film displayed a compact dense morphology as shown in Figure a. On doping 10% Ag precursor in InCl3/Cd(CH3COO)2 mixed solution, the Ag0.1(In0.2Cd0.8)0.9S film exhibited small granular aggregates as shown in Figure b. Larger granular aggregates were formed in the Ag(In0.2Cd0.8)1–S film with the increasing doping ratio of Ag. It was observed that the Ag0.3(In0.2Cd0.8)0.7S film [Figure d] could extensively represent the continuous aggregate structure covered with small granular particles. Table shows the actual compositions of each final Ag(In0.2Cd0.8)1–S and Ag2S films determined by energy-dispersive X-ray spectroscopy (EDX) which were very close to the elemental stoichiometry of the molar ratio.

Figure 6

SEM images (100 000×) of Ag(In0.2Cd0.8)1–S with various atomic percent of Ag: (a) x = 0, (b) x = 0.1, (c) x = 0.2, (d) x = 0.3, (e) x = 0.4, (f) x = 0.5, (g) x = 0.6, (h) x = 0.7, (i) x = 0.8, (j) x = 0.9 and (k) Ag2S and its spot position on photocatalyst arrays.

Table 1

Actual Composition of Ag(In0.2Cd0.8)1–S Spots in Arrays Measured by EDX

spot	actual atomic ratio (Ag:In:Cd:S)
In0.2Cd0.8S	(0:10:39:51)
Ag0.1(In0.2Cd0.8)0.9S	(5:8:36:51)
Ag0.2(In0.2Cd0.8)0.8S	(10:8:34:48)
Ag0.3(In0.2Cd0.8)0.7S	(16:8:28:48)
Ag0.4(In0.2Cd0.8)0.6S	(19:7:26:48)
Ag0.5(In0.2Cd0.8)0.5S	(25:5:21:49)
Ag0.6(In0.2Cd0.8)0.4S	(29:5:18:48)
Ag0.7(In0.2Cd0.8)0.3S	(36:3:12:49)
Ag0.8(In0.2Cd0.8)0.2S	(39:2:8:51)
Ag0.9(In0.2Cd0.8)0.1S	(45:1:5:49)
Ag2S	(51:0:0:49)

Figure shows the XRD profiles of the In0.2Cd0.8S, Ag(In0.2Cd0.8)1–S (x = 0.1, 0.3, 0.5, 0.7, and 0.9) films on fluorine-doped tin oxide (FTO)-coated glass substrates and standard reference of CdS (JCPDS #41-1049) and Ag2S (JCPDS #14-0072). The In0.2Cd0.8S exhibited several peaks at 2θ of 24.82, 26.54, 28.29, 43.79, and 47.90°, corresponding to 24.8° (100), 26.5° (002), 28.18° (101), 43.68° (110), and 47.84° (103) planes of wurtzite CdS (JCPDS #41-1049), respectively. The other two peaks at 37.70 and 51.49° were attributed from the FTO substrate (label as circles). The shift in the peaks could be explained based on the reason that both the In2S3 and CdS could be merged into each other with homogeneously distributed Cd and In atoms in a solid solution. For the Ag(In0.2Cd0.8)1–S film (x = 0.1, 0.3, 0.5, 0.7 and 0.9), main peaks of wurtzite structured CdS corresponding to (100), (002), (101), (110), and (103) planes (label as triangles) still remained in the same positions. In addition, there were some extra peaks at 31.55, 33.67, 37.10, and 37.74° (label as stars) which could be indexed to the crystal structure of Ag2S. When increasing the Ag doping amount, the intensities of (100) and (101) planes in wurtzite CdS are gradually decreased and are absent in Ag0.7(In0.2Cd0.8)0.3S and Ag0.9(In0.2Cd0.8)0.1S. Three peaks at 31.55, 37.10, and 37.74° corresponding to Ag2S phase observed in Ag0.3(In0.2Cd0.8)0.7S shows the lowest doping amount of Ag in 30%, and the enhanced peak intensities in high Ag doping amount suggest that the Ag-doped (In0.2Cd0.8)0.7S photocatalysts possessed lager grains and higher crystallinity.

Figure 7

XRD profiles of the In0.2Cd0.8S, Ag(In0.2Cd0.8)1–S (x = 0.1, 0.3, 0.5, 0.7, and 0.9) films and bare FTO substrate.

Figure a–d shows the chemical binding states of In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photocatalysts determined by XPS. The XPS spectra of S 2p (Figure a) for the In0.2Cd0.8S represented the binding energies of S 2p3/2 and S 2p1/2 at 161.2 (monosulfide) and 162.2 (disulfide),[35] respectively. For the Ag0.3(In0.2Cd0.8)0.7S, the binding energies of the chemical states of S 2p3/2 and S 2p1/2 were identical to those of the In0.2Cd0.8S. The XPS spectra of Ag0.3(In0.2Cd0.8)0.7S. Figure b shows two peaks at 367.2 eV (Ag 3d5/2) and 373.2 eV (Ag 3d3/2), corresponding to Ag2S.[35]Figure c shows that both In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photocatalysts could present two main peaks at 404.8 eV (Cd 3d5/2) and 411.4 eV (Cd 3d3/2), indicating the presence of the Cd2+ state in CdS.[36] The peak intensities of Cd 3d in Ag0.3(In0.2Cd0.8)0.7S were reduced significantly compared to that of intensities in In0.2Cd0.8S. The XPS spectra of both In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photocatalysts displayed two main peaks at 444.8 eV (In 3d5/2) and 452.2 eV (In 3d3/2) as shown in Figure d. In contrast to the Cd 3d intensity, the peak intensities of In 3d in Ag0.3(In0.2Cd0.8)0.7S were enhanced significantly compared to that of In0.2Cd0.8S.

Figure 8

XPS profiles of In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photocatalysts corresponding to (a) S 2p, (b) Ag 3d, (c) Cd 3d, and (d) In 3d.

Figure illustrates the charge excitation and transfer mechanisms based on the characterization results. Under the light irradiation, electrons in Ag0.3(In0.2Cd0.8)0.7S photocatalyst were excited to the conductive band and holes left in the valence band. The photoinduced electrons were rapidly transported to the external circuit through the FTO substrate while the holes were transported to Ag2S for reacting with SO32– and hydroxyl group to produce SO42– and oxygen. After doping the Ag metal ions to form Ag(In0.2Cd0.8)1–S solid solution, the enhanced photocatalytic activity was contributed to the higher photoabsorption in the visible region and higher crystallinity quality with fewer defects, which was beneficial for the photoinduced charge separation, and thus, the probability of charge recombination was reduced.

Figure 9

Schematic diagram of the charge-transfer process in a Ag0.3(In0.2Cd0.8)0.7S–Ag2S heterostructured photocatalyst.

Conclusions

The composition of M(In0.2Cd0.8)1–S (M = V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, Ag, W, Ir, Pt, and Te) photocatalyst arrays were screened by the SECM method to determine that doping 30% Ag to form the Ag0.3(In0.2Cd0.8)0.7S electrode could result in the highest photocurrent among these 15 kinds of metal-doped M(In0.2Cd0.8)1–S photocatalysts. The anode photocurrents of the Ag0.3(In0.2Cd0.8)0.7S electrode were found to be 1 and 0.53 mA/cm2 under UV–visible and visible light, respectively, which were comparatively higher than that of the In0.2Cd0.8S electrode (0.45 and 0.25 mA/cm2). The highest IPCE value of the Ag0.3(In0.2Cd0.8)0.7S photocatalyst and In0.2Cd0.8S were found to be 64% (λ = 450 nm) and 57% (λ = 400 nm), respectively. The Mott–Schottky plots showed that In0.2Cd0.8S and Ag0.3(In0.2Cd0.8)0.7S photoelectrode could exhibit a flat-band potential of −0.85 and −0.55 V versus Ag/AgCl, respectively. The XRD diffraction profile of Ag0.3(In0.2Cd0.8)0.7S films shows main peaks corresponding to (100), (002), (101), (110), and (103) of the wurtzite structured CdS and other small peaks at 31.55, 33.67, 37.10, and 37.74° indexed to the crystal structure of Ag2S. This study suggests that the superior photocatalytic activity of the Ag0.3(In0.2Cd0.8)0.7S photoelectrode was mainly attributed to its high crystallinity for promoting charge separation and reducing the probability of charge recombination.

Experimental Section

Precursor Solutions

Three stock solutions were prepared by dissolving 0.1 M InCl3 (Sigma-Aldrich), 0.1 M Cd(CH3COO)2 (Fisher), and 0.2 M CH4N2S (Showa) in water–glycerol (3:1) solutions. The In–Cd–S mixed precursors were prepared by mixing InCl3, Cd(CH3COO)2 and CH4N2S solutions in a volume ratio of 0.2:0.8:1. For the screening test, 15 different metal salts solutions (0.1 M), VCl2·xH2O, Cr(NO3)3·9H2O, Mn(NO3)2, Fe(NO3)3·9H2O, Co(NO3)2·6H2O, Ni(NO3)2·6H2O, CuSO4·5H2O, Zn(NO3)2·6H2O, (NH4)6Mo7O24·4H2O, RuCl3·xH2O, AgNO3, (NH4)6W12O39, IrCl4, and TeCl4 were prepared in water–glycerol (3:1) solutions. All chemicals were of reagent grade and used as received.

Screening of Photocatalyst Arrays

Prior to their use, the conducting substrates, FTO-coated glasses (15 × 30 × 2.2 mm3, Chang Shuan Electronics Co.) were sonicated and rinsed thoroughly with isopropanol and water. The photocatalyst arrays consisting of mixed salt solutions were deposited on FTO glasses using a piezo-based microarray dispenser (CHI model 1550, CH Instruments). The FTO substrate was placed under a picolitre piezodispenser (Micro Jet AB-01-60, Micro Fab Plano, TX), and the position was controlled by XYZ stepping motors in a preprogrammed pattern. The voltage pulse of 100 V was applied to the piezodispenser to eject the desired number of drops of precursor solutions onto the substrate. First, the InCl3/Cd(CH3COO)2 mixed precursor solutions were dispensed, and then, the thiourea was dispensed in the second time as the preprogrammed pattern (see Figure S1). Finally, the metal precursor solutions were dispensed the third time. The total volume of each spot in photocatalyst arrays remained constant. There was a special case for preparing the Ru(In0.2Cd0.8S)1– array which required a higher voltage pulse of 150 V due to the higher viscosity. The photocatalyst arrays containing different ratios of In/Cd mixed to metal precursor solutions were agitated for 5 min using an agitator and kept at 100 °C for 12 h in a furnace under an argon atmosphere to form M(In0.2Cd0.8)1–S films (M: metal).

Then, screening measurements were performed by SECM (CHI Instruments 900C) with an optical fiber as described in the previous report.[37] A 200 μm optical fiber (FIA-P200-SR, Ocean Optics) connected to a 150 W xenon lamp (SXE-150, Collimage International Co.) was attached to the tip holder of the SECM. The FTO substrate with photocatalyst arrays on the surface was placed in an SECM cell with its surface exposed at the bottom through an O-ring. A Pt wire and saturated Ag/AgCl electrode were used as the counter and reference electrodes, respectively. The electrolyte solution is 0.1 M Na2SO4/Na2SO3 (pH = 9.6), and the Na2SO3 was prepared as the sacrificial donor. The optical fiber was positioned perpendicular to the array surface at a distance of 100 μm and scanned across the surface at a speed of 500 μm/s. The substrate potential was held at 0 V versus Ag/AgCl, and a filter with a wavelength of 420 nm was used to block the UV light in the visible-light illumination experiments. The photocurrent produced during the scan was recorded and displayed as a two-dimensional image.

Photoelectrochemical Measurements of Electrodes in Bulk Films

According to the photocurrents of arrays in SECM results, a specific composition of mixed precursor solutions was determined for drop-casting on FTO substrates and annealing at 200 °C for 12 h forming the In0.2Cd0.8S bulk film as the photoelectrode (working electrode). The In0.2Cd0.8S working electrode, a reference electrode (Ag/AgCl) and counter electrode (Pt strip) were placed in a three-electrode cell which was filled with 0.1 M Na2SO4/Na2SO3 solution as the electrolyte. This three-electrode cell was irradiated by a Xe lamp to obtain the photocurrent curve. In addition, the flat-band potential and carrier density of the electrodes in bulk film were estimated from Mott–Schottky measurements performed by EIS. Additional characterizations were performed by XRD (D8, Bruker), UV–vis spectroscopy (V-650, JASCO), PL spectroscopy (F-7000, HITACHI), time-resolved PL spectroscopy, and XPS (Phi 5000 VersaProbe, ULVAC).