Introduction

Greenhouse gases have recently garnered great attention from the viewpoint of global warming and climate change. Since CO2 gas is a greenhouse gas, CO2 reduction is required to address the global issue of increasing CO2 emission [1]. To this end, many researchers have undertaken CO2 reduction and conversion to energetic compounds by chemical, biological, photochemical, inorganic, and electrochemical approaches [2,3,4]. Among these approaches, the electrochemical reduction of water-dissolved CO2 in the presence of a metal catalyst can generate CO, HCOO−, and hydrocarbons with high current efficiency and high reaction selectivity [5,6,7]. The products obtained from CO2 depend on the metal species used as the catalyst. Cu as CO2 reduction reaction (CO2RR) catalyst generates hydrocarbons including C1, C2, and C3 products [7,8,9]. Au, Ag, Pd, and Zn catalysts provide high selectivity for CO production and HCOO− is the major product obtained with Sn, Hg, and Pb catalysts [5, 6, 10]. In the presence of other metal catalysts, hydrogen evolution reaction (HER) occurs preferentially compared to CO2RR [5]. This product selectivity is due to the binding energy of CO2 and its reaction intermediates to the metal species, as confirmed by density functional theory (DFT) studies [7, 10,11,12]. Binding energy was also affected by the crystal face and valence of metals such as Cu and Cu2O [13,14,15,16]. In other words, for highly efficient and selective CO2RR, we need to design catalysts considering the binding energy of CO2 and it reaction intermediates to the catalyst surface.

CO2RR activity of Cu electrode has attracted attention over the year because it can produce hydrocarbons like methane, ethylene, methanol, and ethanol with a relatively high efficiency. In polycrystalline Cu, CO formation competes with HCOO− formation at applied potentials less than − 0.9 V vs. reversible hydrogen electrode (RHE). Hydrocarbons formation starts at − 0.9 V and ethylene is formed first, followed by methane. At applied potentials more negative than − 0.9 V, methane formation dominates over ethylene formation, and the formation of ethanol, methanol, and C3 products starts. The HER competes directly with CO2RR but decreases with increasing applied potential [8, 9]. On the other hand, in Cu single-crystal electrode, product selectivity of CO2RR depends on the Cu crystal surface [13], suggesting that changes in deficiency of surface, step, and edge have a significant effect on product selectivity [9]. In addition, Cu2O possesses high selectivity for ethylene formation due to the significant inhibition of methane formation [15, 16]. At present, the factor responsible for changing C1 and C2 selectivity is not evident; however, the aforementioned results yield useful knowledge in terms of selectivity control.

DFT calculations have established that the selectivity of CO2RR is enhanced by using alloys [17,18,19]. Alloying controls the binding energy of the CO2RR reactive species to the electrode surface by changing the metal species and content ratio. Kortlever et al. reported that altering the composition of Pd and Pt in the Pd x Pt(100-x)/C catalyst led to a major change in HCOO− selectivity [20]. Among Cu-based alloy catalysts, oxide-derived Cu x Zn prepared by electroplating possesses high selectivity for ethanol formation with faradaic efficiency (FE) of ~ 30%, as reported by Ren et al. [21]. Meanwhile, CuInO2 prepared by electrophoretic deposition showed increased CO selectivity compared to Cu2O [22]. Moreover, Cu-Sn catalyst achieved CO production FE of ~ 90% by changing the deposition amount of Sn on oxide-derived Cu surface. This results in suppressed H2 and HCOO− formation, as confirmed by DFT calculations [23]. However, the cause for increasing selectivity due to the formation of Cu-Sn alloy is not clear. The dependence of selectivity on alloy composition was also not clarified.

When CO2RR was performed on commercial Sn electrode, FE of HCOO− production was over 91% in 0.1 M KHCO3 at − 1.8 V vs. Ag/AgCl [24]. Moreover, for Sn electroplated on Cu plate, FE of HCOO− varied with Sn electrodeposition thickness and it was 91% at − 1.4 V vs. SCE [25]. As mentioned earlier, Sn has high selectivity toward HCOO−. This phenomenon suggested that the surface oxide of Sn plays an important role in HCOO− formation. It has been reported that Sn etched with HBr showed significantly low FE for HCOO− production compared with untreated Sn electrode in 0.5 M NaHCO3 at − 0.7 V vs. RHE [26]. The relation between the valence of Sn and CO2RR activity is unclear because Sn is susceptible to oxidation.

In this study, we focused on an electrode coated with electrodeposited Cu-Sn alloy for CO2 electroreduction catalysts. There are many reports on co-deposition of two metal catalysts to improve selectivity, but we control reaction products from the standpoint of alloy structures like solid solutions or intermetallic compounds. Specifically, we proposed to control product selectivity of CO and HCOO− by changing the crystal structure and alloy composition ratio of Cu and Sn, respectively. The Cu-Sn alloy electrode was prepared using an electroplating method and the composition ratio was adjusted by altering the molar ratio of the Cu-Sn electrolyte. Surface characteristics of the Cu-Sn catalysts were analyzed using scanning electron microscopy (SEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and auger electron spectroscopy (AES). Our study shows that the Cu-Sn alloy can control product selectivities of CO and HCOO− by altering the alloy composition. The effect of alloy composition on the selectivity for CO2RR was also discussed.

Experimental

Electrodeposition of Cu-Sn Alloy

The Cu-Sn alloy catalysts were prepared by the electrodeposition method. All the prepared catalyst electrodes used Cu plate (T=0.2 mm, Nillaco Corp.) as the substrate. The Cu plate was washed three times by ultrasound in acetone and then was similarly rinsed with ultrapure water. It was then immersed in 1 M HCl for 30 s and dried by N2 followed by rinsing with ultrapure water before use as the working electrode (WE). The electrolyte consisted of 0.5 M K4P2O7 (97%, Sigma-Aldrich Co. LLC), 0.05 M C6H14N2O7 (99.0%, Sigma-Aldrich Co. LLC), and varying concentrations of CuSO4·5H2O (99.5%, Wako Ltd.) and SnSO4 (96%, Wako Ltd.), such that the total concentration of metals was 0.2 M dissolved in ultrapure water. The electrodeposition on the Cu substrate was carried out using a two-electrode system where the counter electrode (CE) was a Pt plate. To study the effect of Cu-Sn alloy catalysts on CO2 reduction activity, we employed Cu and Sn electrodeposited electrodes. The electrolyte for Cu electrodeposition comprised 0.5 M K4P2O7, 0.2 M C4H12N2O5 (99.0%, Sigma-Aldrich Co. LLC), and 0.2 M CuSO4·5H2O dissolved in ultrapure water. The Sn electrodeposition electrolyte comprised 0.5 M K4P2O7, 0.2 M SnSO4, 1 g/L PEG 6000 (Wako Ltd.), and 0.6 mL/L formalin. The prepared catalysts were washed with ultrapure water and dried N2. The electrodeposition conditions for various Cu-Sn alloys, Cu, and Sn on Cu substrate are shown in Table 1.

Table 1 Electrodeposition condition of various catalysts. The composition ratio of each Cu-Sn alloy was decided by AES analysis
Full size table

Characterization of Prepared Catalysts

To determine the composition ratio of the prepared Cu-Sn catalysts before and after CO2 electroreduction, they were analyzed by AES (JEOL Ltd., JAMP-7810). The X-ray source used an LaB6 filament. The surface morphology of the prepared catalysts was analyzed before and after CO2 reduction using SEM (JEOL Ltd., JSM-6701). The crystal face and structure of the Cu-Sn alloy was evaluated before and after CO2 reduction by thin-film XRD (MAC Science, MO3XHF22) technique. The X-ray source used Cu Kα (λ = 0.154 nm, 30 mA) and the scan angle recorded at 2θ value between 25° and 80° in steps of 0.2°. The chemical bonding states of Cu 2p and Sn 3d were analyzed before and after CO2 reduction by XPS (Shimadzu Co., KRATOS AXIS-NOVA) with Al Kα as the X-ray source. The obtained binding energies of Cu 2p and Sn 3d orbitals were calibrated with the C–C bond of carbon contamination as a reference at 284.2 eV.

Electrochemical Reduction of CO2

A custom-made two-compartment electrolysis cell made from acrylic plastic was used (Fig. 1). The Pt plate and Ag/AgCl (3.3 M KCl) were used as the CE and reference electrode (RE), respectively. The cathode was separated from the anode by a cation-exchange membrane (Nafion NRE-212, Sigma-Aldrich Co. LLC). The electrolyte (23 mL) used was 0.1 M KHCO3 (99.5%, Wako Ltd.) solution saturated with CO2 gas (99.995%, JFP) for 40 min. The CO2 gas flowed continuously during the CO2 electroreduction at a flow rate of 5 mL/min. The electrolyte was stirred by a stirring bar.

Fig. 1
figure 1

Schematic of the electrolysis cell used for CO2 reduction. WE, working electrode; RE, reference electrode; CE, counter electrode

Full size image

In the electrochemical measurements, the potential and current were controlled or measured using a potentiostat (Hokuto Denko Corp., HZ-7000) and the IR drop was corrected. In this study, the Ag/AgCl was converted to a reversible hydrogen electrode (RHE) at all potentials using the following formula: ERHE = EAg/AgCl + 0.210 + (0.0591 × pH). Here, the pH of CO2-saturated 0.1 M KHCO3 solution was 6.8. The electrochemical CO2 reduction was performed using controlled potential chronoamperometry for 1800 s and the applied potential was in the range from − 0.69 to − 1.09 V vs. RHE. The FE of the products from CO2 reduction was calculated based on Eq. (1).

$$ \mathrm{FE}=\frac{znF}{Q}\times 100 $$
(1)

Where, z is the number of electron transfers needed for CO2 reduction, n is the number of moles from the results of the quantitative analyses, F is the Faraday constant (96,485 C/mol), and Q is the amount of electricity.

The gaseous products formed by CO2 electroreduction and the introduced CO2 gas were collected by a connected gas bag (Smart Bag PA, GL Sciences Inc.) and the head space of the electrolysis cell. The collected gas was analyzed by gas chromatography (GC, Shimadzu Co., Tracera). We used Micropacked ST (Shinwa Chemical Industries Ltd.) as the GC column and He (Iwatani Corp.) used as the carrier gas. The HCOO− content in the liquid products obtained after CO2 reduction was analyzed by anion chromatography (IC, Shimadzu Co., Prominence HIC-SP) after dilution with ultrapure water. The column employed for IC was Shim-pack IC-SA3 (Shimadzu Co.) and the eluting solution was 3.6 mM Na2CO3 (99.5%, Wako Ltd.) solution. The prepared electrodes were used for electrochemical CO2 reduction just once. In this study, the FE of gaseous and liquid products was the mean amount of time spent on the CO2 reduction. Additionally, FE of the products was obtained by averaging the results of the experiments three times with CO2 reduction on each electrode.

Electrochemical Impedance Spectroscopy

The electrochemical impedance spectroscopy (EIS) was carried out under the same conditions as those used for CO2 electroreduction condition at − 0.89 and − 1.09 V vs. RHE in 0.1 M KHCO3 solution employed by the potentiostat. Before applying the AC amplitude potential, the cathode was kept at − 0.89 or − 1.09 V for 120 s to become stabilized to the current through the sample. The AC amplitude potential and measurement potential were set at 100 mV and − 0.89 or − 1.09 V, respectively, and the measurement frequency was scanned in a range of 100 kHz to 500 mHz.

Results and Discussion

Characterization of Cu-Sn Alloy Catalysts

The composition ratio of Cu and Sn in the prepared Cu-Sn catalysts was determined before and after CO2 reduction using AES measurement (Table 2). In this paper, the prepared Cu-Sn catalysts are described as Cu87Sn13, Cu76Sn24, and Cu55Sn45, based on AES results. The Cu and Sn deposited electrodes are described as Cu100 and Sn100, respectively. AES Cu signal on the Sn100 electrode originated from the Cu substrate or diffusion of the Cu to deposited Sn [27, 28]. Any detected element apart from Cu and Sn was considered contamination from ambient air or oxide on the metal surface. Table 2 shows that the relative composition ratio of oxygen increased with increasing ratio of Sn due to Sn being more susceptible to oxidation than Cu. After CO2 reduction at − 0.89 V vs. RHE for 30 min, the Cu-Sn ratio of the Cu-Sn catalysts slightly increased due to diffusion of Cu and Sn upon applying the reduction potential, as suggested by the SEM and XRD results after CO2 reduction [27, 28].

Table 2 Relative composition ratios on the surfaces of the prepared catalysts, as per AES measurements
Full size table

Figure 2a–j shows the SEM images of electrodeposited Cu and Sn on the Cu substrate. Cu100 and Sn100 comprise distributed particles with size of ~ 300 nm. The Cu-Sn catalysts, except Cu55Sn45, have uniformly dense structures and particle size decreased with increasing Cu concentration. Cu55Sn45 showed a unique surface morphology and crystal growth was complicated. After CO2 reduction at − 0.89 V vs. RHE for 30 min, the macroscale morphologies of all the catalysts remained unchanged. The Cu100 revealed particle abrasion, but particle size did not change before and after CO2 reduction. On the other hand, Sn100 exhibited cohesion and spalling of particles throughout the surface. Correspondingly, Cu55Sn45 and Cu87Sn13 were examined for deposition of small particles on the catalyst surface.

Fig. 2
figure 2

SEM images of the prepared catalysts before CO2 reduction: a Cu100; b Cu87Sn13; c Cu76Sn24; d Cu55Sn45; e Sn100 deposited on Cu plate. SEM images of the prepared catalysts after CO2 reduction: f Cu100; g Cu87Sn13; h Cu76Sn24; i Cu55Sn45; j Sn100 deposited on Cu plate

Full size image

The crystalline structures of the Cu-Sn catalysts deposited on Cu plate were analyzed by XRD before and after CO2 reduction (Fig. 3). All the catalysts exhibiting Cu diffraction peak originated from the substrate. Figure 3a shows that the main orientations of polycrystalline Cu100 and Sn100 exhibited Cu (111) and Sn (200), respectively. The Cu-Sn alloys show peaks that are different from those shown by elemental Cu or Sn [29,30,31,32]. For the Cu87Sn13, Cu and Sn formed solid solutions, inducing an increase in lattice spacing compared with Cu (111). The Cu76Sn24 catalyst exhibited crystalline structures of both solid solutions and intermetallic compounds, which was identified as Cu3Sn. Meanwhile, Cu55Sn45 comprised Cu3Sn and SnO2 [33, 34]. After CO2 reduction at − 0.89 V vs. RHE for 30 min, Cu-Sn alloy catalysts showed changes in X-ray diffraction peaks. These structural changes are attributed to the diffusion of Cu and Sn upon the application of a potential [27, 28].

Fig. 3
figure 3

XRD patterns of the prepared catalysts before CO2 reduction. a XRD patterns of the prepared catalysts after CO2 reduction in 0.1 M KHCO3 at − 0.89 V vs. RHE. b Crystal structures of Cu87Sn13 were identified as solid solutions. Cu76Sn24 consists of Cu3Sn as intermetallic compounds and solid solutions. For Cu55Sn45, Cu3Sn and SnO2 were identified as components

Full size image

The chemical bonding states of Cu 2p and Sn 3d in the alloys were analyzed by XPS before and after CO2 reduction (Fig. 4). The chemical state of Cu on Cu100 and the Cu-Sn alloy catalysts was CuO (934 eV). For Sn100, the Cu 2p peak is attributed to the substrate and the diffusion of Cu to Sn; this is similar to the AES result. For Sn 3d, Cu87Sn13 and Cu76Sn24 showed the same oxidation state (485.6 eV), consisting of a high percentage of SnO. On the other hand, the Sn 3d peaks of Cu55Sn45 and Sn100 shifted to high binding energy (486.0 eV), consisting of higher percentages of SnO2 than in Cu87Sn13 and Cu76Sn24 [35, 36]. Hence, the difference in the oxidation state of Sn might affect the selectivity for CO2RR. Apart from Cu100, the other samples showed very little difference in the Sn 3d and Cu 2p spectra before and after CO2 reduction at − 0.89 V vs. RHE for 30 min. The Cu state in Cu100 became Cu0 and CuO upon reduction from the Cu2O state during CO2 reduction. These results clearly suggest that the changes in morphology and crystal structure were not affected by the chemical bonding states of Cu and Sn.

Fig. 4
figure 4

XPS spectra. a Sn 3d spectra of Cu87Sn13, Cu76Sn24, Cu55Sn45, and Sn before reduction. b Sn 3d spectra after reduction. c Cu 2p spectra Cu100, Cu87Sn13, Cu76Sn24, Cu55Sn45, and Sn before reduction. d Cu 2p spectra after reduction. The broken line indicates the position of each chemical bonding state: SnO2, SnO, CuO, Cu2O, and Cu

Full size image

CO2 Reduction on Cu-Sn Alloy Catalysts

The CO2 electroreduction on the prepared catalysts was conducted in a custom-made electrolysis cell (Fig. 1) for 1800 s at applied potentials ranging from − 0.69 to − 1.09 V vs. RHE. The FE values for products by CO2 electroreduction are shown in Fig. 5. The product distribution obtained with the Cu deposition catalyst was similar to that for a reported Cu electrode and distinct from an oxide-derived Cu electrode (Fig. 5a) [7, 8, 16, 37]. FE for H2 decreased with increasing applied potential; production of CH4 and C2H4 was confirmed at − 0.79 V vs. RHE and it showed FE of 36% at − 1.09 V. FE values for CO and HCOO− presented the same tendency, FE for HCOO− being higher than that for CO at all applied potentials. These results showed that the Cu electrodeposition electrode leads to poor selectivity toward CO and HCOO−. On the other hand, the Sn electrodeposition catalyst showed high FE for HCOO− (Fig. 5e), in agreement with other research on CO2 electroreduction with Sn electrode [24, 25]. Sn100 exhibited similar FE for H2 and HCOO− at − 0.69 V; FE for H2 decreased with increasing applied potential, while FE for HCOO− remarkably increased to 87.5% at − 1.09 V. As mentioned earlier, it is evident that the selectivity of CO production on the Sn100 electrode is poor.

Fig. 5
figure 5

Faradaic efficiency of CO2 electroreduction products as a function of potential. a Cu100; b Cu87Sn13; c Cu76Sn24; d Cu55Sn45; e Sn100 deposited on Cu plate. Electrolyte, 0.1 M KHCO3

Full size image

The prepared Cu-Sn alloy catalysts exhibited different selectivities for CO2RR. All Cu-Sn alloys can suppress HER, suggesting that the presence of Sn within the alloy by binding strongly to the surface H [23]. In Cu87Sn13, the selectivity of CO production was increased and maximum FE for CO was 59.5% at − 0.99 V vs. RHE (Fig. 5b). However, the FE for CO did not change significantly with increasing applied potential. Cu87Sn13 showed a decreasing trend for H2 generation with increasing potential. HCOO− production shows a trend opposite to that for H2 formation, and exhibited 12.3% FE at − 1.09 V. Cu87Sn13 produced very little hydrocarbons compared to Cu100, and its maximum FE was 1.2% at − 1.09 V. From the differences in selectivity, it is clear that the crystal structure significantly affected the selectivity for CO2RR. More specifically, the formation of solid solutions affects the binding energy or binding pattern of CO2 and its reaction intermediates. Additionally, when CO formed mainly, it may be suspected that the binding energy against surface CO (*CO) gets lowered as compared to that of Cu. This decrease in binding energy is considered to depend on the lattice spacing of Cu (111) or changing surface electron density, or both.

The CO2 reduction behavior of Cu76Sn24 was intermediate between those of Cu87Sn13 and Cu55Sn45 (Fig. 5c). From − 0.69 to − 0.89 V, the FE for the products showed trends similar to Cu87Sn13. At over − 0.89 V, HCOO− production increased dramatically, and it has nearly the same FE as that of CO at − 1.09 V. H2 formation decreased with increasing applied potential, same as for other electrodes. These characteristics suggest that Cu76Sn24, which consists of Cu3Sn and solid solutions, was present at different reactive sites for CO2 reduction on applying potential. At a low applied potential, the Cu site that originated in Cu3Sn and solid solutions involved CO production. On the other hand, at a high applied potential, HCOO− production occurring from the Sn site originated from Cu3Sn. The FE for CO of Cu87Sn13 and Cu76Sn24 suggested that CO production was closely related to the Cu lattice spacing and electron density.

The product distribution using the Cu55Sn45 alloy electrode was close to that obtained with the Sn100 electrode (Fig. 5d). This suggests that the Cu3Sn structure has no effects on the selectivity for CO2RR. Another possible factor affecting CO2RR could be the chemical binding state of Sn within the alloys. The crystal structures of SnO2 and Sn were observed by XRD analysis of Cu55Sn45 (Fig. 3). Additionally, the XPS results (Fig. 4) show that the binding energy of Sn 3d, attributed to SnO x , was the same as that of Sn100; however, it was different from that of Cu87Sn13 and Cu76Sn24. In fact, the FE for HCOO− depends on the oxidation state of Sn [26]. From these results, it is believed that the oxidation state of Sn, rather than the crystal structure of alloy at Cu55Sn45, strongly affects the formation of HCOO−.

Figure 6 shows the total current density and partial current density of H2, CO, and HCOO− during CO2 reduction against the applied potentials. The current density was calculated from the geometric surface area (2.75 cm2). The j CO and j HCOO − values of each electrode showed the same trend as did the FE. We discuss the relationship between the selectivity, current density, and charge transfer resistance (CT resistance) in the following sections.

Fig. 6
figure 6

Total current density for CO2 electroreduction as a function of potential on Cu100, Cu87Sn13, Cu76Sn24, Cu55Sn45, and Sn100. a Partial current density of the products from CO2 reduction on each catalyst: b H2; c CO; d HCOO−

Full size image

EIS Measurements on Cu-Sn Alloy

EIS measurements were performed under the same conditions as those for the CO2RR at − 0.89 and − 1.09 V vs. RHE; the obtained results were illustrated using the Cole-Cole plot. According to Fig. 7a and c, the low value in the x-axis indicates the solution resistance, and the diameter of the semicircle indicates the CT resistance (Fig. 7b, d). The CT resistance was found to be closely related to the current density; the catalysts with low CT resistance showed an increase in j total, except for Cu100.

Fig. 7
figure 7

Cole-Cole plot for Cu100, Cu87Sn13, Cu76Sn24, Cu55Sn45, and Sn100 during electrochemical CO2 reduction in 0.1 M KHCO3 at a − 0.89 and c − 1.09 V vs. RHE. Charge transfer resistance of each electrode at b − 0.89 and d − 1.09 V vs. RHE

Full size image

Cu100exhibited high current density despite the large CT resistance, because it showed a high FE for multi-electron reduction products, including CH4 and C2H4 (Fig. 5a). The CT resistance of Cu55Sn45 and Sn100, which had a high j HCOO −, was lower than that of the other catalysts at − 0.89 and − 1.09 V. This suggested that the resistance of the reaction pathway for the generation of HCOO− from CO2 was low; that is, CO2RR is not routed through the adsorption of CO2 and its intermediates on Cu55Sn45 and Sn100. At − 1.09 V, the CT resistance of Cu55Sn45 was lower than that of Sn100 and this result corresponded to the highest j HCOO −. Cu76Sn24 exhibited the same CT resistance as Sn100 at − 0.89 and − 1.09 V, but the j total, j CO, and j HCOO − values were different from those of Sn100. This result suggests that the Sn sites originating from Cu3Sn possessed high selectivity for HCOO− and decreased CT resistance. The CT resistance of Cu87Sn13 at both applied potentials was larger than that of the other Cu-Sn alloy, but the j total value was smaller. Based on these results, we speculated that CO2RR on solid solutions progresses through the CO2 reaction intermediates binding to the catalyst surface. Cu87Sn13 exhibited a larger CT resistance as compared to Cu76Sn24 at both potentials, probably because of the increase in the amount of adsorbed intermediates as compared to Cu87Sn13.

CO2RR Mechanism for Each Cu-Sn Alloy

CO2RR mechanism on Cu surface has been elucidated employing DFT calculations [18, 38,39,40]. Briefly, CO2 first adsorbs and reduces on Cu surface by accepting an electron and proton, and then forms surface HOCO (*HOCO), which is adsorbed on the electrode. HCOO− is generated to desorb *HOCO at this stage, and with the progression of the reduction reaction, adsorbed *CO and H2O are generated. Similarly, CO is generated by the desorption of *CO, and adsorbed surface HCO (*HCO) is generated by the progress of the reduction reaction. Furthermore, with the progression of the reduction reaction, the *HCO intermediate becomes reduced and generates CH4 and CH3OH. Generation of C2 compounds require dimer formation from intermediates like *CO [18]. The prepared Cu100 electrode is considered to perform according to the above mechanism.

CO2RR on Cu87Sn13 is suggested to behave the same as the adsorbed phase of Cu100 by EIS (Fig. 7). However, Cu100 and Cu87Sn13 have different selectivities for CO2 reduction and the main product on Cu87Sn13 was CO, because Cu87Sn13 was a solid solution which consists of Cu and Sn. It seems that formation of solid solutions increases the lattice spacing of Cu (111) and changes surface electron density. This surface characteristic change might lower the binding energy of the reaction intermediate *CO. Furthermore, Cu87Sn13 inhibits further *CO reduction at − 1.09 V, thereby enhancing CO formation [7].

Meanwhile, the proposed reaction pathway for HCOO− formation on Sn involved a reaction with CO2 to generate surface H (*H) [38, 41]. In this study, the Sn100 electrode might react according to the abovementioned pathway, because Sn100 exhibits a lower CT resistance than Cu100 (Fig. 7). Among the prepared alloy catalysts, Cu55Sn45 follows the same mechanism because its CO2RR selectivity and EIS results are similar to those of Sn100. These results suggest that the active site of CO2RR is Sn, i.e., the surface oxidation state of Sn, rather than the crystal structure of Cu55Sn45, is the major factor contributing to CO2RR.

In contrast, Cu76Sn24 exhibited a CT resistance similar to that of Sn100 and Cu55Sn45, but the selectivity for CO2RR was different from theirs. Discussion on Cu76Sn24 is difficult because the Cu and Sn sites in Cu3Sn affect CO2RR selectivity (Fig. 5c). The EIS results suggested that surface Sn affects the adsorbed state and the reaction pathway.

Conclusion

We prepared Cu-Sn alloy catalytic electrodes through the simple electrodeposition method for electrochemical CO2 reduction. The Cu-Sn alloy catalysts can control the selectivity for CO and HCOO− formation from CO2 by changing the crystal structure. In addition, these catalysts strongly inhibit HER compared to Cu due to the presence of Sn. Cu87Sn13 showed high selectivity toward CO formation, such that its faradaic efficiency was 60% at − 0.99 V vs. RHE differing from electrodeposited Cu catalysts. The surface analyses revealed that solid solutions play an important role in CO formation upon CO2 reduction, because solid solution formation weakens the binding energy between the Cu-Sn alloy and the reaction intermediate *CO. On the other hand, Cu55Sn45 showed 89.5% of FE for HCOO− at − 1.09 V; this result was nearly the same as that for Sn100 (87.5%). These results indicate that increasing surface SnO2 promotes HCOO− rather than the presence of intermetallic compounds. The Cu76Sn24 alloy exhibited properties intermediate between Cu87Sn13 and Cu55Sn45 at − 1.09 V, i.e., 37.4% FE for CO and 40.8% FE for HCOO−. This result suggested that there may be two active sites: Cu and Sn in Cu3Sn, due to which the selectivity behavior of CO2RR differs depending on applied potential.