Improvement of selectivity for CO₂ reduction by using Cu₂ZnSnS₄ electrodes modified with different buffer layers (CdS and In₂S₃) under visible light irradiation

Abstract

Cu₂ZnSnS₄ (CZTS) electrodes modified with different n-type buffer layers (CdS and In₂S₃) were used as photocathodes for CO₂ reduction under visible light irradiation (420 < λ < 800 nm) in aqueous media. Compared to a bare CZTS electrode, CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃), by which a p–n heterojunction between CZTS and the n-type buffer layer is formed, showed a significant increase in the photocurrent assigned to CO₂ reduction. In addition, product selectivity of CO₂ reduction was improved by surface modification with an n-type buffer layer: that is, selective CO₂ reduction into CO was achieved by using a CdS/CZTS electrode, while HCOOH selectivity was observed over an In₂S₃/CZTS electrode. In this study, we investigated the photoelectrochemical properties of CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃) in conjunction with the structural and optical properties, and we investigated their activity for PEC CO₂ reduction.

---

Introduction

Carbon dioxide (CO₂) is the thermodynamically stable product of most fossil fuel combustion processes and is a significant contributor to the greenhouse effect. The accumulation of CO₂ in the atmosphere has a great impact on climate change and could threaten the environment and eventually the worldwide economy. The conversion of CO₂ into useful products and chemicals is thus a very attractive research area. Photoelectrochemical (PEC) reduction of CO₂ to value-added products has been widely studied as a possible technique for the mitigation of atmospheric CO₂.^1–3 Over the past few decades, tremendous work on PEC CO₂ reduction has been done by using semiconductor electrodes in water electrolysis as both an electron donor and a proton source,^4,5 a technique is currently called artificial photosynthesis. One way to achieve artificial photosynthesis is to directly convert CO₂ into organic fuels over p-type semiconductor electrodes (photocathodes) in combination with n-type semiconductor electrodes (photoanodes) for oxygen generation from water without applying an external voltage under sunlight illumination. There have been many studies on high-efficiency photoanodes such as WO₃,^6,7 TaON,^8,9 Fe₂O₃,^10,11 and BiVO₄,^12,13 while there have been few reports on photocathodes with a high selectivity for CO₂ reduction because a semiconductor electrode generally shows preferential H₂ production in an aqueous solution.

Cu₂ZnSnS₄ (CZTS) is one of the promising p-type semiconducting materials for PEC devices due to its ability to absorb visible light (optical band gap of ca. 1.5 eV) and its relative stability in aqueous environments.^14–16 In addition, it is made of cheap, non-toxic, earth-abundant elements and can be easily produced on a large scale. PEC CO₂ reduction utilizing a CZTS electrode was first reported in 2011 by Arai et al., who carried out a study in which CO₂ was reduced to formate with a high selectivity (current efficiency > ca. 80%) by using a Ru complex-modified CZTS electrode under visible light irradiation (400 < λ < 800 nm).¹⁷ In the absence of a Ru complex, no formate was produced, indicating that the active center of CO₂ reduction might be a Ru complex. The activity for CO₂ reduction was also improved by the insertion of Se into the CZTS electrode, presumably due to an improvement of carrier mobility.

We were inspired by their work and became interested in exploring the potential of CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃) for efficient PEC CO₂ reduction reaction. Since the deposition of n-type buffer layers (CdS, In₂S₃) over the p-type CZTS layer results in the formation of a depletion region at the solid–solid interface (p–n heterojunction), which can also assist in the extraction photo-generated electrons from CZTS toward the n-type buffer layers, the photocurrent is enhanced.^18–20 Moreover, colloidal CdS has been studied as a visible light responsive photocatalyst that showed activities for CO₂ reduction to form CO with high quantum yields from DMF including sacrificial reagents.²¹ Li et al. reported that In₂S₃ exhibited visible-light activity toward photocatalytic reduction of CO₂ in the presence of water vapor into renewable CH₄.²² Based on these observations, it was expected that enhancement of the photocurrent and improvement of selectivity for CO₂ reduction is anticipated by using CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃). However, as far as we know, there has been no report on PEC CO₂ reduction by using n-type buffer layer-modified CZTS electrode. The present study was carried out from this standpoint.

We fabricated a CZTS electrode by the sol–gel spin-coating method on a molybdenum-coated glass substrate. Different n-type buffer layers, such as CdS and In₂S₃, were deposited on the CZTS electrode by the chemical bath deposition (CBD) method. The bare CZTS electrode exhibited a cathodic photocurrent under visible light irradiation, and the photocurrent reached ca. −100 µA cm⁻² at 0 V vs. reversible hydrogen electrode (RHE) in CO₂-saturated NaHCO₃ solution. The cathodic photocurrent was enhanced by deposition of n-type buffer layers due to efficient charge carrier separation as a result of the creation of a p–n heterojunction, leading to PEC reaction with the primary products being hydrogen, carbon monoxide, and formate. In this study, PEC properties of the CZTS electrode were investigated in conjunction with structural and optical properties, and the contributions of n-type buffer layers (CdS, In₂S₃) to the CO₂ reduction property were investigated.

Experimental

Preparation of Cn₂ZnSnS₄ electrode

The Cu₂ZnSnS₄ electrode was fabricated by the sol–gel and spin-coating method using the following procedure.²³ First, 0.4 mol L⁻¹ Cu(CH₃COO)₂·H₂O (Wako, 99.0%), 0.25 mol L⁻¹ Zn(CH₃COO)₂·2H₂O (Wako, 99.9%), 0.2 mol L⁻¹ SnCl₂·2H₂O (Wako, 99.9%) and 1.6 mol L⁻¹ SC(NH₂)₂ (Wako, 98%) were dissolved in 2-methoxyethanol (10 mL, Wako, 99.0%) containing monoethanolamine (Sigma-Aldrich, >99.0%). The concentration of each metal ion coincides with the generally reported data for high-efficiency solar cells, and an excess amount of thiourea was used to complex metal ions and alleviate the loss of sulfur during the annealing process. To prevent the formation of cracks in the precursor thin film during the sulfurization process, additive monoethanolamine is necessary in the precursor solution. Approximately 70 µL of the precursor solution was dropped on a Mo-coated soda-lime glass (Mo/glass, 18 mm × 20 mm in size, GEOMATEC Co., Ltd.) and then the excess precursor solution was removed by spin-coating (3000 rpm, 30 s). The electrode was dried in air at 200 °C for 5 min on a hotplate. At this time, the precursor solution is considered to be formed metal–thiourea–oxygen complexes, which is stable in air.²³ After this process had been repeated for a maximum of 5 times, the electrode was calcined at 560 °C for 1 h in a sulfur/N₂ atmosphere. It should be noted that the number of times of spin-coating was optimized in this study (refer Fig. S1 in ESI†). To remove impurity phases such as Cu_2−xS over the CZTS film, (NH₄)₂S etching processes were performed for 6 h. Then a bare CZTS electrode was obtained.

Surface modification with n-type buffer layers

The n-type buffer layers (CdS, In₂S₃) were deposited on the CZTS electrode by the chemical bath deposition (CBD) method using the following procedure. The CZTS electrode was immersed in 0.125 mol L⁻¹ Cd(CH₃COO)₂·2H₂O (Wako, 98.0%), 2.8 mol L⁻¹ SC(NH₂)₂ (Wako, 98%), and NH₃ aqueous solution at 65 °C for 9 min to yield CdS-modified CZTS (CdS/CZTS). For In₂S₃-modified CZTS (In₂S₃/CZTS), the CZTS electrode was immersed in an aqueous solution containing 0.025 mol L⁻¹ InCl₃·4H₂O (Wako, 99.9%), 0.1 mol L⁻¹ CH₃CSNH₂ (Wako, 98.0%), and 0.1 mol L⁻¹ CH₃COOH (Wako, 99.7%) at 65 °C for 15 min. After the CBD techniques, the electrodes were annealed at 200 °C and 100 °C, respectively, and then CdS/CZTS and In₂S₃/CZTS electrodes were fabricated.

Characterization

The crystalline phases were characterized by using a powder X-ray diffraction (XRD) instrument (MiniFlex II, Rigaku Co.) with CuKα (λ = 1.5418 Å) radiation (cathode voltage: 30 kV, current: 15 mA). Raman spectroscopy measurements were performed by using a JASCO NRS 5100 laser Raman spectrophotometer. Absorption properties of the CZTS electrode were determined by using the diffuse reflection method with a UV/VIS/NIR spectrometer (V570, JASCO Co., Japan) attached to an integral sphere at room temperature. Mott–Schottky analysis was carried out by using an electrochemical analyzer (604D, ALS Co.) with a bare CZTS electrode, a platinum electrode, an Ag/AgCl electrode and 0.1 M Na₂SO₄ + NaOH solution (pH = 9.5) used as a working electrode, counter electrode, reference electrode and electrolyte, respectively. X-ray photoelectron spectroscopy (XPS) measurements were performed by using a Kratos AXIS Nova spectrometer (Shimazu Co.) with a monochromatic Al Kα X-ray source. The binding energy was calibrated by taking the carbon (C) 1s peak of contaminant carbon as a reference at 284.6 eV.

Photoelectrochemical (PEC) measurement

PEC performance of the CZTS electrode was investigated in a three-electrode configuration using an Ag/AgCl reference electrode and a Pt coil counter electrode. The electrolyte was CO₂-saturated 0.1 M NaHCO₃ solution. Linear sweep voltammetry and chronoamperometry measurements were carried out by using an automatic polarization system (HSV-110, Hokuto Denko Co.) under a Xe lamp equipped with an L-42 cut-off filter (SCF-50s-42L, SIGMAKOKI Co., Ltd.) and IR cut-off filter (420 < λ < 800 nm). The scan rate for the linear sweep voltammetry was 10 mV s⁻¹.

Analysis of products

PEC CO₂ reduction was performed in a gastight three-electrode configuration cell with a double compartment in which a CZTS electrode, Pt coil and silver–silver chloride (Ag/AgCl) electrode were used as a working electrode, counter electrode and reference electrode, respectively. The electrolyte solution was 0.1 M NaHCO₃, which was purged with CO₂ gas for ca. 40 min prior to the start of measurement. After CO₂ bubbling, the cell was sealed and irradiated by visible light irradiation (Xe lamp, 420 nm < λ < 800 nm, 100 mW cm⁻²) for 1 h. Evolved H₂, CO gas was detected by an on-line gas chromatograph (GC) with a thermal conductivity detector (Agilent Technology Co. MicroGC) equipped with an MS-5A column. He gas was used as the carrier gas. The liquid CO₂ reduction product, formic acid (HCOOH), was detected by using single-channel ion chromatography (ICS900, Thermo Fisher Scientific Inc.), and other products including methanol (CH₃OH), ethanol (C₂H₅OH), and formaldehyde (HCOH) were detected by gas chromatography (G-3500, Hitachi Co.) with a DB-WAXetr column (122-7332, Agilent Co.).

Results and discussion

Structural and optical properties of the CZTS electrode

Fig. 1(a) shows the XRD pattern of the CZTS electrode together with the ICDD standard file of the tetragonal phase of Cu₂ZnSnS₄ (JCPDS 00-026-0575). An intense peak corresponding to the bottom Mo substrate was observed at 40.02°, while the other diffraction peaks coincided with the reference pattern corresponding to (112), (200), (220), (312), (224) planes of CZTS, indicating that CZTS crystal could be synthesized by the sol–gel and spin-coating method. Fig. 1(b) shows a Raman spectrum of the CZTS electrode under 532 nm laser illumination. There is an intense peak centered at 333 cm⁻¹ along with two weak peaks located at 288 cm⁻¹, and 368 cm⁻¹; these signals are in agreement with those of the kesterite CZTS structure,^24–26 and no other peaks were found to be binary phases such as hexagonal Cu_2−xS (475 cm⁻¹), Sn₂S₃ (304 cm⁻¹), ZnS (352 cm⁻¹) and cubic Cu₂SnS₃ (303 cm⁻¹ and 356 cm⁻¹).²⁷ However, SnO₂ phase was observed at 663 cm⁻¹.²³ Therefore, our fabricated CZTS electrode consists of the major phase of kesterite CZTS and minor phases of SnO₂. Although a slightly weak peak corresponding to hexagonal MoS₂ phase was observed at 400 cm⁻¹,^23,27 the MoS₂ phase plays an important role in the characteristics of the CZTS electrode, as described later.

Fig. 1 (a) XRD pattern of the CZTS electrode together with that of a reference pattern. (b) Raman spectrum of the CZTS electrode under 532 laser illumination.

Fig. 2 shows top and cross-section SEM photographs of the CZTS electrode. The flat and tight structure of the CZTS film was observed to have good adhesion to the Mo substrate. The thickness of the CZTS layer was estimated to be ca. 400 nm. Moreover, a MoS₂ layer was observed at the interface between the CZTS layer and Mo substrate. Formation of the MoS₂ layer was inevitable in the sulfurization process of CZTS, which may facilitate electrical quasi-ohmic contact and improve the adhesion of CZTS to the Mo back contact, but leads to high series resistance and accordingly deteriorates the device efficiency if not thin enough. In this study, we optimized the thicknesses of the CZTS layer and MoS₂ layer by adjusting the spin-coating and sulfurization processes, respectively.

Fig. 2 Top (left) and cross-sectional (right) SEM photographs of the CZTS electrode.

Fig. 3 shows the diffuse reflectance spectrum of the CZTS electrode. CZTS is a direct energy gap material. The optical absorption of CZTS could be converted to a Tauc plot, (αhv)ⁿ vs. hv, where n = 2 for a direct bandgap, by extending the tangential lines to the abscissa. The band gap was estimated to be ca. 1.5 eV, which is well consistent with the reported ones in ref. 18.

Fig. 3 Tauc plot for the CZTS electrode. Inset figure shows a picture of our fabricated CZTS electrode.

Structural and optical properties of buffer layer-modified CZTS electrode

Surface modifications of the CZTS electrode with n-type buffer layers were performed by using the CBD method. Fig. 4 shows top and cross-section SEM photographs for CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃). Compared to the bare CZTS, the surface morphology was appreciably changed. The surface of the CdS/CZTS electrode was composed of porous grains with sizes on the order of several micrometers. In contrast, the In₂S₃/CZTS electrode surface was composed of a billowing thin layer. As seen in the cross-section SEM photographs, the thicknesses of CdS and In₂S₃ buffer layers were estimated to be ca. 50 nm and ca. 60 nm, respectively. For clear confirmation of the formation of In₂S₃ and CdS layers, the XRD measurement has been done. However, no diffraction peaks of In₂S₃ and CdS was observed, suggesting that their thicknesses were too thin to be detected by XRD analysis. We therefore carried out the Raman spectroscopy measurement for CdS/CZTS and In₂S₃/CZTS electrodes together with that for CdS/Mo and In₂S₃/Mo substrates. As shown in Fig. 5(a), a relatively sharp peak detected from CdS/CZTS/Mo and CdS/Mo substrates at 299 cm⁻¹ was considered to be the longitudinal optical mode Raman reflection of CdS in accordance with literature studies.^28,29 This results indicated that the CdS layer could be formed on the CZTS electrode by CBD method. On the other hand, broad peaks were observed at 304 and 368 cm⁻¹ in In₂S₃/Mo substrate (see Fig. 5(b)). These peaks were attributed to β-In₂S₃ phase,^30,31 implying that crystalline In₂S₃ could be formed on the CZTS electrode by CBD method. It should be noted that Raman peak of In₂S₃ was buried under that of CZTS in the case of the In₂S₃/CZTS/Mo.

Fig. 4 Top and cross-sectional SEM photographs of (a), (b) the CdS/CZTS electrode and (c), (d) the In₂S₃/CZTS electrode.

Fig. 5 Raman spectra for (a) CdS/Mo glass and CdS/CZTS/Mo glass and (b) In₂S₃/Mo glass and In₂S₃/CZTS/Mo glass.

Fig. 6(a) show the UV-vis spectra of CdS and In₂S₃ layers which fabricated on the ITO substrate by CBD method. The optical absorption of CdS layer could be converted to a Tauc plot, (αhv)ⁿ vs. hv, where n = 2 for a direct bandgap, by extending the tangential lines to the abscissa. On the other hand, a straight plot was obtained for the In₂S₃ layer by applying n = 1/2 to the Tauc plot due to its indirect bandgap material. From intersects of the linear portions of these (αhv)ⁿ vs. hv curves with the photon energy axis (see Fig. 6(b) and (c)), the optical bandgap of CdS and In₂S₃ buffer layers used in this study were determined to be 2.48 and 2.30 eV, respectively.

Fig. 6 (a) UV-vis spectra for CdS and In₂S₃ layers on the ITO substrate prepared by chemical bath deposition method. (b) Tauc plot for direct transition of CdS. (c) Tauc plot for indirect transition of In₂S₃.

Photoelectrochemical property of the CZTS electrode

Fig. 7 shows current–potential curves for bare CZTS, CdS/CZTS and In₂S₃/CZTS electrodes in 0.1 M NaHCO₃ solution purged with CO₂ gas with irradiation of visible light (420 < λ < 800 nm). The bare CZTS electrode exhibited a cathodic photocurrent in response to irradiation of incident light, and the photocurrent density reached ca. −160 µA cm⁻² at −1.0 V vs. Ag/AgCl. The cathodic photocurrent was increased by ca. 2 times after deposition of n-type buffer layers (CdS, In₂S₃) because of efficient charge separation in the depletion region of the p–n heterojunction between p-type CZTS and n-type buffer layers. Deposition of Pt contributed to further enhancement of the cathodic photocurrent due to improvement of the surface reaction rate (see Fig. S2 in ESI†); however, the observed photocurrent was attributed to water reduction reaction with the primary product being hydrogen, as will be discussed later. No CO₂ reduction product was observed in Pt/CZTS, Pt/CdS/CZTS, and Pt/In₂S₃/CZTS. Thereafter, we therefore studied PEC CO₂ reduction by using bare CZTS, CdS/CZTS and In₂S₃/CZTS electrodes.

Fig. 7 Current–potential curves in 0.1 M NaHCO₃ solution under chopped visible light irradiation (420 < λ < 800 nm, 100 mW cm⁻²) for CZTS, CdS/CZTS and In₂S₃/CZTS electrodes.

Fig. 8 shows time courses of the cathodic photocurrent from bare CZTS, CdS/CZTS and In₂S₃/CZTS electrodes in 0.1 M NaHCO₃ solution purged with CO₂ gas at −1.0 V vs. Ag/AgCl under visible light irradiation (420 < λ < 800 nm). For the bare CZTS electrode, the cathodic photocurrent rapidly increased when it was exposed to visible light, and then the photocurrent remained constant with time, implying that the CZTS electrode is stable upon PEC reaction. On the other hand, the cathodic photocurrent for CdS/CZTS and In₂S₃/CZTS electrodes showed a slight increase up to about 20 min and was relatively stable over the observed time course. A similar tendency was reported by Moriya et al. and by Septina et al., who showed that the CdS buffer layer in Pt/CdS/CuGaSe₂ and Pt/CdS/Cu(In,Ga)S₂ was dissolved in the electrolyte upon PEC reaction due to photocorrosion of CdS, and the cathodic photocurrent in the electrodes may have increased as the CdS buffer layer became thinner.^32,33 Based on above reports, we studied the XPS spectra of bare CZTS, CdS/CZTS and In₂S₃/CZTS electrodes before and after PEC reaction. For the bare CZTS electrode (see Fig. 9), there was almost no difference in the XPS spectra for Cu 2p, Zn 2p and S 2p before and after PEC reactions, whereas Sn 3d_3/2 XPS spectrum was slightly changed: a relatively weak peak was observed at 496.4 eV before PEC reaction, and this peak was attributed to Sn(IV) species in SnO₂. After PEC measurement, the SnO₂ peak was disappeared, suggesting that a part of the cathodic photocurrent of bare CZTS electrode was consumed in reducing SnO₂ impurity phase. For the CdS/CZTS electrode (see Fig. 10(a)), major peaks at 405.1 eV and 411.9 eV were observed before PEC measurement, and these peaks were attributed to typical values of Cd 3d_5/2 and 3d_3/2, respectively. After PEC reaction, both of peak intensities were decreased, suggesting that the CdS buffer layer became thinner, similar to previous reports described above.^32,33 For the In₂S₃/CZTS electrode (see Fig. 10(b)), major peaks at 444.5 eV and 452.1 eV were observed before PEC measurement, and these peaks were attributed to typical values of In 3d_5/2 and 3d_3/2, respectively. These peak profile of the In 3d XPS spectrum was clearly changed after PEC reaction; small shoulder peaks located at 443.3 eV and 450.9 eV were newly observed. These peaks were identified to be metal indium In(0). These results indicated that In(III) species in the In₂S₃/CZTS electrode was gradually reduced to In(0) species by PEC reaction. We speculated that the metal indium In(0) contributed to photocurrent response. That is, metal indium In(0) acts as a co-catalyst of the In₂S₃/CZTS electrode, which leading to an enhancement of the cathodic photocurrent.

Fig. 8 Time courses for the photocurrents of CZTS, CdS/CZTS and In₂S₃/CZTS electrodes in aqueous 0.1 M NaHCO₃ solution at −1.0 V vs. Ag/AgCl under visible light irradiation (420 < λ < 800 nm, 100 mW cm⁻²).

Fig. 9 XPS spectra for Cu 2p, Zn 2p, Sn 3d and S 2p before and after photoelectrochemical measurements of the bare CZTS electrode.

Fig. 10 XPS spectra for Cd 3d and In 3d before and after photoelectrochemical measurements of the CdS/CZTS and In₂S₃/CZTS electrodes.

Finally, we examined PEC CO₂ reduction over bare CZTS, CdS/CZTS and In₂S₃/CZTS electrodes at −1.0 V Ag/AgCl in CO₂-saturated 0.1 M NaHCO₃ solution under visible light irradiation (420 < λ < 800 nm, 100 mW cm⁻²). The results are shown in Table 1, where all of these values are averages of three times (for detail, see the Table S1 in ESI†). For the bare CZTS electrode, H₂, CO and HCOOH were detected as products and no formation of CH₄, CH₃OH and C₂H₅OH was observed. The thermodynamic potential for PEC reduction of CO₂ to CO and HCOOH in the presence of protons is generally explained by the following equation:

CO₂ + 2H⁺ + 2e⁻ → HCOOH + H₂O (−0.61 V vs. NHE)

CO₂ + 2H⁺ + 2e⁻ → CO + H₂O (−0.53 V vs. NHE)

Table 1 Photoelectrochemical CO₂ reduction by using the CZTS electrode^a

Entry	Sample	CO2	Coulomb/C	Faradaic efficiency/%
				H2	CO	HCOOH	Total
a Irradiated at visible light (420 < λ < 800 nm, 100 mW cm^−2) for 1 h in aqueous 0.1 M NaHCO3 solution at −1.0 V vs. Ag/AgCl. b Under Ar-purged Na2SO4 aqueous solution.
1	CZTS	○	0.9	74.3	1.4	2.4	78.1
2	CZTS	✗^b	0.5	62.7	0	0	62.7
3	CdS/CZTS	○	1.5	75.0	7.4	1.1	83.5
4	In2S3/CZTS	○	1.0	72.0	3.5	4.6	80.1
5	Pt/CdS/CZTS	○	6.0	67.0	0	0	67.0
6	Pt/In2S3/CZTS	○	6.0	73.4	0	0.3	73.7

---

The conduction potential of the CZTS electrode was estimated to be −1.25 V vs. RHE by Mott–Schottky analysis (see Fig. S3 in ESI†), and the CO₂ reduction reactions into CO and HCOOH were therefore thermodynamically favorable. However, the calculated faradic efficiencies for H₂, CO and HCOOH were approximately 74.3%, 1.4% and 2.4% (the average of three times), respectively, indicating that most of the cathodic photocurrent of the bare CZTS electrode was derived from H₂ evolution. These results indicated that water reduction reaction preferentially occurred over the CZTS electrode in the aqueous solution. In contrast, CO evolution was increased by depositing a CdS buffer layer over the CZTS electrode, and its faradic efficiency increased up to 7%, implying that selectivity for CO₂ reduction was improved by using the CdS/CZTS electrode. It should be noted that no CO evolution was observed when PEC measurement was performed in Ar-saturated Na₂SO₄ aqueous solution (also N₂-saturated NaHCO₃ aqueous solution) or in a dark condition, suggesting that the evolved CO gas was derived from PEC reduction of purged CO₂ gas over the CdS/CZTS electrode. In addition, the Pt-modified CdS/CZTS electrode exhibited H₂ evolution with trace CO and trace HCOOH in CO₂-saturated NaHCO₃ aqueous solution under visible light irradiation. Considering that Pt shows high activity for H₂ evolution with a low overpotential, it is highly possible that water reduction reaction preferentially occurred over the Pt/CdS/CZTS electrode rather than CO₂ reduction. These results indicate that the CdS buffer layer plays an important role in selective PEC CO₂ reduction into CO. Indeed, Yanagida et al. proposed that CO evolution over CdS nanoparticles was caused by adsorption of CO₂ molecules at the CdS surface with sulfur vacancies through bidentate-type adsorption of CO₂, resulting in selective CO formation.²¹ Therefore, our observation is reasonable. On the other hand, the In₂S₃/CZTS electrode generated a relatively large amount of HCOOH compared with that generated by the bare CZTS electrode. Its origin is currently unknown. Since the surface of In₂S₃/CZTS electrode contained metal indium based on the results of XPS spectra (refer to Fig. 10(b)), it is difficult to identify whether the CO₂ reduction reaction occurs at the In₂S₃ buffer layer or metal indium. Indeed, metal indium is well-known to exhibit a high CO₂ selectivity for HCOOH formation in the research field of electrochemistry. In addition, Li et al. reported that In₂S₃ contributed to visible-light activity toward photocatalytic reduction of CO₂ into CH₄ without any other products.²² On the basis of those reports, PEC CO₂ reduction by using an In₂S₃/CZTS electrode may occur on the In₂S₃ surface or metal indium, and thus is likely show selective CO₂ reduction into HCOOH. Thus, CZTS electrodes modified with n-type buffer layers (CdS and In₂S₃) showed a strong photocurrent response and selectivity for CO₂ reduction, which was strongly dependent on the kind of buffer layer: that is, CO selectivity was improved by deposition of a CdS buffer layer, while HCOOH selectivity was improved by deposition of an In₂S₃ buffer layer.

Conclusions

We fabricated a CZTS electrode by the sol–gel and spin-coating method on a Mo/glass substrate and performed photoelectrochemical CO₂ reduction under visible light irradiation (λ > 420 nm). The CZTS electrode exhibited a cathodic photocurrent under visible light irradiation; however, H₂ evolution (faradaic efficiency > ca. 74%) preferentially occurred over the bare CZTS electrode in the CO₂-saturated NaHCO₃ solution. By depositing an n-type buffer layer (CdS and In₂S₃) over the CZTS electrode, enhancement of photocurrent response and improvement of selectivity for CO₂ reduction were achieved: that is, CO selectivity was improved by deposition of a CdS buffer layer, and its faradaic efficiency was increased up to ca. 7%. In contrast, HCOOH selectivity was improved by deposition of an In₂S₃ buffer layer, and its faradaic efficiency was increased up to ca. 5%. The improvement of selectivity for CO₂ reduction was related to the kind of n-type buffer layer, and the adsorption of CO₂ molecules at the n-type buffer layer plays an important role in selective PEC CO₂ reduction. Further research is underway to test this via the deposition of surface co-catalysts that can facilitate PEC CO₂ reduction. This study could be applied to improve the selectivity for CO₂ reduction by using other Cu–chalcogenide photocathodes.

Acknowledgements

The authors thank Prof. Shigeru Ikeda of Konan University and Prof. Tetsuya Haruyama for their valuable contributions to this study. This work was supported by a grant from Advanced Catalytic Transformation program for Carbon utilization (ACT-C), Japan Science and Technology Agency (JST).

References

1. M. Halmann, Nature, 1978, 275, 115–116.
2. M. Schreier, P. Gao, M. T. Mayer, J. Luo, T. Moehl, M. K. Nazeeruddin, S. D. Tilley and M. Grätzel, Energy Environ. Sci., 2015, 8, 855–861.
3. G. Sahara, R. Abe, M. Higashi, T. Morikawa, K. Maeda, Y. Ueda and O. Ishitani, Chem. Commun., 2015, 51, 10722–10725.
4. S. Sato, T. Arai, T. Morikawa, K. Uemura, T. M. Suzuki, H. Tanaka and T. Kajino, J. Am. Chem. Soc., 2011, 133, 15240–15243.
5. T. Arai, S. Sato, T. Kajino and T. Morikawa, Energy Environ. Sci., 2013, 6, 1274–1282.
6. K. Sivula, F. L. Formal and M. Grätzel, Chem. Mater., 2009, 21, 2862–2867.
7. J. A. Seabold and K. S. Choi, Chem. Mater., 2011, 23, 1105–1112.
8. M. Higashi, K. Domen and R. Abe, Energy Environ. Sci., 2011, 4, 4138–4147.
9. M. Higashi, K. Domen and R. Abe, J. Am. Chem. Soc., 2012, 134, 6968–6971.
10. A. Kay, I. Cesar and M. Grätzel, J. Am. Chem. Soc., 2006, 128, 15714–15721.
11. K. Sivula, R. Zboril, F. Le Formal, R. Robert, A. Weidenkaff, J. Tucek, J. Frydrych and M. Grätzel, J. Am. Chem. Soc., 2010, 132, 7436–7444.
12. K. Sayama, A. Nomura, T. Arai, T. Sugita, R. Abe, M. Yanagida, T. Oi, Y. Iwasaki, Y. Abe and H. Sugihara, J. Phys. Chem. B, 2006, 110, 11352–11360.
13. Y. H. Ng, A. Iwase, A. Kudo and R. Amal, J. Phys. Chem. Lett., 2010, 1, 2607–2612.
14. D. Yokoyama, T. Minegishi, K. Jimbo, T. Hisatomi, G. Ma, M. Katayama, J. Kubota, H. Katagiri and K. Domen, Appl. Phys. Express, 2011, 3, 101202.
15. L. Rovelli, S. D. Tilley and K. Sivula, ACS Appl. Mater. Interfaces, 2013, 5, 8018–8024.
16. F. Jiang, Gunawan, T. Harada, Y. Kuang, T. Minegishi, K. Domen and S. Ikeda, J. Am. Chem. Soc., 2015, 137, 13691–13697.
17. T. Arai, S. Tajima, S. Sato, K. Uemura, T. Morikawa and T. Kajino, Chem. Commun., 2011, 47, 12664–12666.
18. N. Guijarro, M. S. Prévot and K. Sivula, J. Phys. Chem. Lett., 2014, 5, 3902–3908.
19. P. Wang, T. Minegishi, G. Ma, K. Takanabe, Y. Satou, S. Maekawa, Y. Kobori, J. Kubota and K. Domen, J. Am. Chem. Soc., 2012, 134, 2469–2472.
20. C. Yan, F. Liu, N. Song, B. K. Ng, J. A. Stride, A. Tadich and X. Hao, Appl. Phys. Lett., 2014, 104, 173901.
21. H. Fujiwara, H. Hosokawa, K. Murakoshi, Y. Wada, S. Yanagida, T. Okada and H. Kobayashi, J. Phys. Chem. B, 1997, 101, 8270–8278.
22. H. Li, Y. Gao, Y. Zhou, F. Fan, Q. Han, Q. Xu, X. Wang, M. Xiao, C. Li and Z. Zou, Nano Lett., 2016, 16, 5547–5552.
23. Z. Su, K. Sun, Z. Han, H. Cui, F. Liu, Y. Lai, J. Li, X. Hao, Y. Liu and M. A. Green, J. Mater. Chem. A, 2014, 2, 500–509.
24. M. Altosaar, J. Raudoja, K. Timmo, M. Danilson, M. Grossberg, J. Krustok and E. Mellikov, Phys. Status Solidi A, 2008, 205, 167–170.
25. P. A. Fernandes, P. M. P. Salomé and A. F. da Cunha, Thin Solid Films, 2009, 517, 2519–2523.
26. K. Wang, O. Gunawan, T. Todorov, B. Shin, S. J. Chey, N. A. Bojarczuk, D. Mitzi and S. Guha, Appl. Phys. Lett., 2010, 97, 143508.
27. P. A. Fernandes, P. M. P. Salomé and A. F. da Cunha, J. Alloys Compd., 2011, 509, 7600–7606.
28. O. Zelaya-Angel, F. L. Castillo-Alvarado, J. Avendailo-Lopez, A. Escamilla-Esquivel, G. Contreras-Puente, R. Lozada-Morales and G. Torres-Delgado, Solid State Commun., 1997, 104, 161–166.
29. I. O. Oladejia, L. Chowa, J. R. Liub, W. K. Chub, A. N. P. Bustamantec, C. Fredricksena and A. F. Schultea, Thin Solid Films, 2000, 359, 154–159.
30. K. Kambas, J. Spyridelis and d. M. Balkan, Phys. Status Solidi B, 1981, 105, 291–296.
31. E. Kärber, K. Otto, A. Katerski, A. Mere and M. Krunks, Mater. Sci. Semicond. Process., 2014, 25, 137–142.
32. M. Moriya, T. Minegishi, H. Kumagai, M. Katayama, J. Kubota and K. Domen, J. Am. Chem. Soc., 2013, 135, 3733–3735.
33. W. Septina, Gunawan, S. Ikeda, T. Harada, M. Higashi, R. Abe and M. Matsumura, J. Phys. Chem. C, 2015, 119, 8576–8583.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra22546b

---