CO2 is considered as the main greenhouse gas that contributes to global warming; therefore, finding an effective way to convert/store CO2 has attracted more and more attention [1, 2]. The main methods for reducing CO2 concentration in the atmosphere include CO2 capturing and storing it underground [3, 4] or convert it to added value chemicals [5,6,7]. Due to the stable chemical properties of CO2, it is necessary to use high temperature, high pressure, or catalyst to make it reactive. Considering the energy and economy cost, electrochemical conversion of CO2 under mild conditions is a promising strategy for decreasing the excess greenhouse gas and achieving an artificial carbon cycle [8,9,10]. However, the major difficulties in CO2 electrochemical reduction are the intrinsic stability of CO2, the lower potential of CO2 reduction reaction (CO2RR), and the low selectivity for reduction products [11]. Thus, it is urgent to develop CO2 reduction catalysts with high selectivity, good stability, and excellent activity.

Previous studies show several metallic electrodes, such as Au, Ag, Cu, Pd, and Sn, are attractive candidates for CO2RR [12]. Among them, copper is the only metal catalyst which is found to produce considerable C1–C3 hydrocarbon products and alcohols [13]. Au, which is a highly active catalyst towards CO2 electrochemical reduction, can produce CO from CO2 with high selectivity and low overpotential [11]. Except for Cu and Au, the other metal electrodes including Ag, Pd, and Sn primarily convert CO2 to CO or formate (HCOO) via a two-electron transfer pathway [14,15,16,17]. However, on the one hand, it is difficult to improve the selectivity and stability of Cu-based catalysts towards the electrochemical reduction of CO2 to C1–C3 hydrocarbon products. On the other hand, Au is highly selective for CO production, but its high cost and rare-earth abundance hinder its industrialization in CO2RR [18, 19]. The composites based on copper and gold are of great potential for CO2 electrochemical reduction. But most of the currently reported CuAu catalysts were synthesized via solvothermal method [20]. The morphology of the nanoparticles is difficult to control, and these particles tend to be easily oxidized and aggregated [21, 22]. Therefore, it is very important to develop a kind of gold and copper composite with controllable morphology, high stability, and high product selectivity for CO2 electrochemical reduction. Besides, it is reported that the metal-oxide interface could improve the electrocatalytical activity of the catalysts towards CO2RR [23].

Here in this paper, we report a surfactant-free Cu2O@Au nanocomposite in which Cu2O/Au interface was constructed for electrocatalytical reduction of CO2 in water. For comparison, the hollow cubic Au catalysts were prepared by dissolving Cu2O in Cu2O@Au catalysts in ammonia. The experimental results showed that the metal/oxide interface in Cu2O@Au catalyst could activate inert CO2 molecule and increase the FE of CO. The CO FE is 30.1% on Cu2O@Au electrode at − 1.0 V (vs. RHE) which is twice than that on Cu2O and Au electrodes prepared in this work. This result not only proved the metal-oxide interface could improve the electrocatalytical activity of the electrodes towards CO2RR, but also paved the way for metal-oxide catalyst synthesis.



Copper (II) trifluoroacetate (Cu (TFA)2, 98%), potassium trifluoroacetate (KTFA, 98%), and chloroauric acid (HAuCl4, 99.9%) were purchased from Sigma-Aldrich and used directly without any purification. All solutions were prepared with Milli-Q ultrapure water (Millipore ≥ 18.2 MΩ cm). Nitrogen (N2) (99.999%) and CO2 (99.999%) gases used in the experiment were purchased from Foshan MS Messer Gas CO., Ltd. Carbon paper which thickness is 0.3 mm was purchased from Hesen in Shanghai.

Preparation of Cu2O Nanocubes and Cu2O@Au

Cu2O nanocubes were synthetized according to the method reported in previous literature [24]. The cubic Cu2O nanoparticles were electrodeposited on a carbon paper (1 cm × 1 cm) using chronoamperometry at − 0.06 V (vs. SCE) for 1 h in 10 mM Cu (TFA)2 and 0.2 M KTFA solution. Before the Cu2O nanocubes electrodeposition, the carbon paper was washed by water and ethanol several times.

The preparation of Cu2O@Au composite was immersing Cu2O cubes into 2 mL HAuCl4 (1 mM) solution for 30 min at 277 K.

Preparation of Hollow Cubic Au

The as-prepared Cu2O@Au composite was immersed in 2 M ammonia water for 12 h at 277 K to remove Cu2O and retain hollow cubic Au on the carbon paper.


The morphologies and structures of nanomaterials were characterized by scanning electron microscopy (SEM, JEOL-6701F) equipped with an energy dispersive X-ray (EDX) detector system. The X-ray diffraction (XRD) patterns were recorded using Rigaku Ultima IV X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å) to study the compositions of the products.

Electrochemical Measurements of CO2

Electrochemical measurements were carried out with a CH Instruments 760D (Chenhua, Shanghai) and a three-electrode system. The CO2 electrochemical reduction was carried out in a two-compartment H-type cell with an Ag/AgCl and a platinum sheet (1 cm × 1 cm) used as reference and counter electrodes, respectively. Compensation for 85% iR drop was used in CO2 electrochemical reduction. In this work, all potentials reported in the CO2 electrochemical reduction were referenced against the reversible hydrogen electrode (RHE). The RHE used the following conversion: ERHE (V) = EAg/AgCl (V) + 0.197 V + (0.059 V × pH) [25]. A diagrammatic sketch of the H-type electrochemical cell is shown in Fig. 1. The two electrochemical cells were separated by a proton exchange membrane (Nafion 117, Sigma-Aldrich).

Fig. 1
figure 1

A diagrammatic sketch of the H-type electrochemical cell

Linear sweep voltammetry (LSV) experiments were performed in 0.1 M KHCO3 solution under N2 (99.999%) or CO2 (99.999%) atmosphere. Prior to LSV tests, N2 or CO2 were purged into the solution in H-type electrochemical cell for 20 min, respectively.

Before CO2RR experiments, the electrolyte solution was saturated for 20 min with CO2 and the pH of the 0.1 M KHCO3 solution was about 8.6. The CO2 electrochemical reduction was performed under potentiostatic conditions while the current and product concentration were monitored. The as-prepared materials were used as working electrodes. The CO2RR experiments were repeated thrice at each potential. Detection of CO2 reduction products were used by an online gas chromatography (GC Agilent, 7890B). A GC run was conducted every 930 s. The GC was equipped with two Plot-Q columns, a thermal conductivity detector (TCD), a flame ionization detector (FID), and a demethanizer with N2 (99.999%) as carrier gas. The contents of liquid products were neglected in this work. During the CO2RR experiments, CO2 was vented into the cathode electrolysis cell at a flow rate of 20 ml min−1 continuously.

$$ {i}_x=\frac{C_x\cdot q\cdot p}{RT}\cdot {n}_xF $$
$$ \mathrm{FE}\left(\%\right)=\frac{i_x}{i_{total}}\cdot 100 $$

The FE calculation equation is shown in Eqs. 1 and 2, in which itotal is the current density recorded by the potentiostat during CO2RR [26]. The partial current (ix) which is needed to generate each product (x = H2, CO, CH4, C2H4) is derived from Eq. 1. Cx is extracted from the calibration curve GC volume concentration of product x. nx is the number of reduced electrons required to produce x from carbon dioxide molecules. q is the gas flow rate, p is constant pressure, and T is the room temperature. R is the gas constant, and F is the Faradaic constant.

Results and Discussion


The morphologies and structures of as-prepared Cu2O and Cu2O@Au nanocubes characterized by SEM were shown in Fig. 2. The Cu2O nanocubes electrodeposited on the carbon paper had regular shapes and smooth surface (Fig. 2a). The average edge length of the Cu2O cubes was about 1 μm as shown in Fig. 2b. An appropriate reaction time and Au3+ solution concentration of GRR on Cu2O nanoparticles would produce Cu2O@Au nanostructures as shown in Fig. 2c and d.

Fig. 2
figure 2

The SEM images of Cu2O nanocubes (a, b), Cu2O@Au nanoparticles (c, d), and EDX of Cu2O@Au nanoparticles (e, f)

After Cu2O nanocubes were immersed in HAuCl4 (1 mM) solution for 30 min, the surface distribution of Au and Cu of Cu2O@Au composites was examined by EDX mapping shown in Fig. 2e and f. It showed that Au nanoparticles were uniformly distributed on the Cu2O nanocube surface. The GRR between Cu2O and HAuCl4 involves the evolution of an internal hollow core and surface precipitation of Au nanoparticles [27, 28].

As shown in Fig. 3, Cu2O in Cu2O@Au composites was removed and the retained Au nanoparticles inherit the cubic frame of the Cu2O@Au composites, after Cu2O@Au nanocubes were immersed in ammonia water for 12 h. The small Au nanoparticles in hollow cubic Au framework were about 20~30 nm in diameter.

Fig. 3
figure 3

The SEM images of hollow cubic Au (ac) of different magnification

XRD Analysis

The crystal structure of the as-prepared catalysts was investigated by XRD, and the diffraction patterns were shown in Fig. 4. The diffraction peak at 2θ = 54.51° belongs to the carbon paper. The diffraction peaks at 2θ = 36.46, 42.36, 61.44, and 73.55° are ascribed to the (111), (200), (220), and (311), respectively, of the Cu2O cube (JCPDS 78-2076). Four weak peaks at 2θ = 38.18, 44.39, 64.57, and 77.54° are assigned to the (111), (200), (220) and (311), respectively, of Au (JCPDS 04-0784) which replaced Cu2O on the carbon paper. Most of Cu2O were substituted by Au; thus, the diffraction peaks corresponded to Cu2O disappeared in XRD pattern of hollow cubic Au.

Fig. 4
figure 4

XRD patterns of (a) Cu2O cube, (b) hollow cubic Au, and (c) Cu2O@Au

CO2 Electrochemical Reduction Performance

The LSV curves of Cu2O cube, Cu2O@Au, and hollow cubic Au electrodes are shown in Fig. 5. The LSV experimental condition was obtained at a cathodic sweeping rate of 50 mV s−1 with N2-saturated or CO2-saturated 0.1 M KHCO3 solution. The current density of all the three samples under the N2 atmosphere is higher than that under CO2; this difference may be caused by the hydrogen evolution reaction (HER) on Cu2O cube, Cu2O@Au, and hollow cubic Au, i.e., with continuous flow of CO2 in the cathodic electrolytic cell, the surface of the electrode is covered by adsorbed CO molecules which will inhibit the HER on electrode surface and decrease the reduction current [29]. The current density of Cu2O@Au electrode in CO2-saturated 0.1 M KHCO3 solution is higher than Cu2O and hollow cubic Au electrodes as shown in Fig. 5d.

Fig. 5
figure 5

LSV curves obtained on a Cu2O cube, b hollow cubic Au, and c Cu2O@Au electrodes in N2-saturated (black solid line) and CO2-saturated (red dotted line) 0.1 M KHCO3 solutions. d LSV curves of the three samples in CO2-purged 0.1 M KHCO3 solutions

The electrochemistry method of amperometric it was used to evaluate the performance of CO2RR in 0.1 M KHCO3 solution at room temperature under atmospheric pressure. The potentials are set between − 0.7 and − 1.2 V for subsequent product determination. At different potentials, the FE of H2 and CO for CO2RR on Cu2O cubes have a significant difference, as shown in Fig. 6a, i.e., the FE of H2 is decreasing because the surface of Cu2O cubes is covered by CO molecules produced by CO2RR, and the HER is inhibited [30]. The FE of CH4 and C2H4 vary slightly in different potentials.

Fig. 6
figure 6

FE of a Cu2O cube catalyst, b Cu2O@Au catalyst, and c hollow cubic Au catalyst. d Comparison of FE for CO and H2 at − 1.0 V vs RHE on three catalysts

The FE of Cu2O@Au catalyst is shown in Fig. 6b. The FE of CO keeps upward trend with potential decreasing and reaches a maximum of 30.1%, at − 1.0 V (vs. RHE). The FE of H2 decreases from 56.7 to 45.6%. Compared with the Cu2O@Au catalyst, the maximum CO FE of hollow cubic Au catalyst is 16.3% at − 1.0 V (Fig. 6c). The CO FE of Cu2O@Au catalyst at − 1.0 V is about twice of hollow cubic Au catalyst at the same potential. Cu2O@Au composite shows superior catalytic activity for CO2 electrochemical reduction than Cu2O cube catalyst and hollow cubic Au catalyst, and it is related to the interfacial effect of metal oxides.

To understand the reaction mechanism on CO2RR to CO, we considered the following reaction steps:

$$ {\mathrm{CO}}_2\left(\mathrm{g}\right)+\ast +{\mathrm{H}}^{+}\left(\mathrm{aq}\right)+{\mathrm{e}}^{-}{\to}^{\ast}\mathrm{COOH} $$
$$ {}^{\ast}\mathrm{CO}\mathrm{OH}+{\mathrm{H}}^{+}\left(\mathrm{aq}\right)+{\mathrm{e}}^{-}{\to}^{\ast}\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}\left(\mathrm{l}\right) $$
$$ {}^{\ast}\mathrm{CO}\to \mathrm{CO}\left(\mathrm{g}\right)+\ast $$

Generally, Eq. 3 is perceived as the potential limiting step on CO2RR to CO [23]. The corresponding binding energy can be substantially lowered on the interface of Cu2O@Au, compared to the Cu2O cube surface or Au surface. In addition, the Eq. 4 and Eq. 5 are also facilitated at the Cu2O@Au interface. It indicates that the interfacial effect of metal oxides could enhance the CO2 adsorption and the electrochemical surface area [31, 32]. The Cu2O@Au catalyst consists of Cu2O and Au nanoparticles can supply a metal-oxide interface to activate inert CO2 molecules, enhance charge transfer efficiency, and increase FE of CO [33].

Compared to the mass transfer effect of hollow cubic Au catalyst composed by Au nanoparticles, the synergistic interactions of metal oxides fabricated by Cu2O cubes and Au nanoparticles are more advantageous to convert CO2 into CO by CO2 electrochemical reduction.

The FE comparison for CO and H2 at − 1.0 V vs RHE on Cu2O cube catalyst, Cu2O@Au catalyst, and hollow cubic Au catalyst is shown in Fig. 6d. The H2/CO ratio of these three catalysts is as follows: 3.9, 3.2, and 1.7. The Cu2O@Au catalyst production ratio of 1.7 by CO2 electrochemical reduction is closest to that of syngas (the mixture of CO and H2) ratio of 2 [34, 35]. The catalyst surface construct method and the proportion of product gases would contribute to design highly selective CO2RR catalysts.

The average current density of three catalysts, which were performed by amperometric it, is shown in Fig. 7. With the potential increase, it shows evidently increasing current densities of three catalysts as expected. The difference of the average total current density between hollow cubic Au (blue solid line) and Cu2O@Au (black solid line) expands at − 1.0 V. However, the difference of the average total current density between hollow cubic Au and Cu2O cube (red solid line) is not marked within − 0.7 to − 1.1 V. Consequently, we could conclude that the charge transfer efficiency of Cu2O@Au catalyst is higher than the other two catalysts.

Fig. 7
figure 7

The average total current density of the three catalysts for CO2 reduction at different potentials


In summary, surfactant-free and low Au loading electrodes for CO2 electrochemical reduction were prepared by electrodeposition and GRR. The Cu2O@Au catalyst shows superior catalytic activity for CO2RR than Cu2O cubes and hollow cubic Au catalyst due to the metal-oxide interface, i.e., the metal-oxide interface could activate the inert CO2 molecules absorbed on electrodes. For Cu2O@Au catalyst, it can convert CO2 to CO with a maximum FE of ~ 30.1% at − 1.0 V and is about twice of the other catalysts at the same potential. The produced gas of Cu2O@Au catalyst by CO2 electrochemical reduction has a H2/CO ratio of 1.7, which is close to the syngas ratio of Fischer–Tropsch process of 2. Based on these results, we can draw some conclusions that the Cu2O@Au catalyst fabricated by Cu2O cubes and Au nanoparticles could form a metal/oxide interface to activate inert CO2 molecules and this catalyst could be applied to syngas production by CO2 electrochemical reduction.