Electroreduction of CO₂ into Ethanol over an Active Catalyst: Copper Supported on Titania

Abstract

A simple, inexpensive, and novel method was used to prepare electrocatalysts from Cu supported on titanium dioxide (Cu/TiO₂). XRD, SEM, and TEM characterizations confirmed different loadings of Cu nanoparticles (NPs) on TiO₂. Cyclic voltammetry tests indicated that Cu/TiO₂ exhibited lower overpotential for CO₂ reduction than that of Cu NPs. Moreover, 40 wt % Cu/TiO₂ exhibited the highest faradaic efficiency for ethanol (FE_ethanol) of 27.4%, which is approximately 10-fold higher than that for Cu NPs (FE_ethanol = 2.7%). The 40 wt % Cu/TiO₂ electrocatalyst exhibits a stable current density of 8.66 mA/cm² over a 25 h stability test. The high efficiency towards CO₂ electroreduction to ethanol may be attributed to the synergistic effect of Cu and TiO₂ NPs. This work highlights the importance of compositional effect of NPs on their catalytic activities and provides a strategy for designing efficient catalysts for CO₂ electroreduction in the future.

1. Introduction

Large-scale consumption of fossil fuel has increased CO₂ concentration in the atmosphere. High atmospheric CO₂ concentrations are associated with greenhouse effect and climate change, which have gained increased research attention [1,2]. Once captured, CO₂ is a potentially useful feedstock if it can be converted into formate/formic acid, carbon monoxide, or more highly reduced hydrocarbon products. Electroreduction is a potential efficient method for CO₂ utilization and recycling, which can be performed at moderate temperature and atmospheric pressure and exhibits high product selectivity compared with other reduction methods, such as chemical, photochemical, and thermochemical methods. Besides, electrical energy could be provided by other green and sustainable energy sources, including wind, solar, and wave [3].

Hori et al. [4] performed pioneering works on electrochemical CO₂ reduction using transition metal electrocatalysts. In addition to metal and metallic oxide, alloys and other semiconductors have been studied as catalysts. Thus far, few studies have investigated the use of metal supported on metal oxides to form different composites, which have been reported to possess improved faradaic efficiencies (FEs) for CO₂ reduction compared to both parent metal and metal oxide. Besides, support materials have been designed to support and stabilize small nanoparticles (NPs) with high current density (J) and FE. For instance, Yu and his team [5] report that copper nanoparticle interspersed MoS₂ nanoflowers facilitate CO₂ electroreduction to hydrocarbon, such as CH₄ and C₂H₄ with high FE at low overpotentials. Sun et al. [6] demonstrated a new strategy of controlling synergistic effect between Cu and SnO₂ in the core/shell Cu/SnO₂ structure to achieve selective electrochemical reduction of CO₂ to CO with FE reaching 93% at −0.7 V vs. reversible hydrogen electrode (RHE) in 0.5 M KHCO₃.

Compared with the noble metals, such as Pd [7] and Au [8], Cu has been considered an ideal electrode material for CO₂ reduction because of its low cost and good catalytic effect. According to a previous report [9,10], it is the only metal electrocatalyst possessing the ability to form a substantial amount of hydrocarbons, which are regarded as ideal products due to their high energy density. Strasser et al. [11] used Cu NPs with different thicknesses as catalysts in electroreduction of CO₂ into CH₄ and C₂H₄; the investigated activity-selectivity-size relations provide novel insights into the CO₂ electroreduction reaction on nanoscale surfaces. Kuhl et al. [12] reported new insights into the electrochemical reduction of CO₂ on a metallic copper surface and observed a total of 16 different CO₂ reduction products, such as methanol, ethanol, formate, and methane.

Normally, titanium on its own has no significant catalytic activity towards CO₂ electroreduction. However, TiO₂ has shown some activity in both photocatalysis and electrocatalysis for CO₂ reduction [13,14,15]. When thin TiO₂ films or mixtures of TiO₂ and other metal and metal oxides are used as an electrocatalyst to catalyze CO₂ electroreduction, they show more excellent performance than TiO₂ NPs. For example, in the preparation of a CO₂ catalytic electrode, Monnier et al. [15] prepared TiO₂, TiO₂-Ru (or RuO₂), and TiO₂-Pt thin film electrodes by thermal deposition on titanium rods. Bandi [16] deposited metallic oxide mixtures (including RuO₂, TiO₂, MoO₂, Co₃O₄, and Rh₂O₃) on titanium foil. Cueto and Hirata [17] prepared a TiO₂-indium-tin oxide thin-film glass electrode via a fixed-potential bulk electrolysis process, in which 1-butyl-3-methylimidazolium tetrafluoroborate (BMImBF₄, an ionic liquid) was used as both solvent and supporting electrolyte. In addition, TiO₂ acts as a redox electron carrier that facilitates various reduction reactions, which include CO₂ conversion [18,19]. The TiO₂ surface assists in CO₂ adsorption to stabilize the intermediate of CO₂ electroreduction and reduces the overpotential [20,21]. In fuel cells, Pt-supported titanium dioxide exhibits improved electrocatalytic properties, such as increased Pt catalyst activity, stability, and durability [22]. To the best of our knowledge, however, no comprehensive research reports the synergistic effect of Cu and TiO₂ on the stability of the composite electrocatalysts for CO₂ reduction. Inspired by the initial results of Cu-based materials, we successfully synthesized TiO₂ as catalyst support for Cu in electroreduction of CO₂. Structural characterizations and electrochemical experiments were performed to compare different Cu loadings supported on TiO₂. Cu nanocatalysts supported on TiO₂ show high catalytic activity for CO₂ electroreduction to ethanol.

2. Results and Discussion

2.1. Performance Comparison of Different Cu-Supported TiO₂ Catalysts

Cu/TiO₂ catalysts were characterized through powder X-ray diffraction (XRD) analysis. Figure 1 (curves b–h) shows diffraction peaks at 43.3°, 50.4° and 74.1°, which are attributed to Cu (111), (200) and (220) reflections, respectively. The rest of peaks can be ascribed to the TiO₂, which were observed in Figure 1 (curves a–g), indicating that Cu and TiO₂ formed a composite electrocatalyst. Moreover, the peaks of Cu became evident in the XRD patterns with increasing Cu proportion. The presence of Cu and TiO₂ was confirmed by energy dispersive X-ray (EDX) analysis (Figure S1, Supplementary Materials (SM)); the results showed signals corresponding to copper, titanium, and oxygen. The values for actual Cu loading of the synthesized catalysts are determined through inductive coupled plasma (ICP) emission spectrometer (

Table 1. Cu composition of the synthesized catalysts ^a.

Catalyst	Cu loading (wt %)
	Intended	Actual a
5 wt % Cu/TiO2	5	4.68
10 wt % Cu/TiO2	10	8.52
20 wt % Cu/TiO2	20	18.39
40 wt % Cu/TiO2	40	37.12
60 wt % Cu/TiO2	60	53.06
80 wt % Cu/TiO2	80	73.02

^a Obtained by ICP.

). The actual loading values are consistent with the predicted values.

Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images were collected to observe morphological differences among Cu/TiO₂ catalysts (Figure 2 and Figure S2 in SM). The results indicated how Cu NPs are progressively loaded on the TiO₂ surface. The Cu/TiO₂ catalysts possess a three-dimensioned (3D) irregular nanoblock structure, with an average size of about 300 nm, and are covered with tiny NPs. Cu NPs gradually spread all over the TiO₂ surface, forming a massive and compact layer, with increasing Cu loadings. For few Cu loadings, TiO₂ was clearly covered by the sporadic Cu particles. Unexpectedly, the TEM images of 60 wt % and 80 wt % Cu/TiO₂ showed that Cu NPs are more prone to agglomerate than 40 wt % Cu/TiO₂, which stem from the increased loading of Cu, leading to the decline of active sites for CO₂ reduction. Predictably, the current density and FE on 60 wt % and 80 wt % Cu/TiO₂ are lower than those observed on 40 wt % Cu/TiO₂; this finding will be comprehensively described in the catalyst test section. Interestingly, without TiO₂ as support material, Cu NPs exhibited a large diameter, which varied between 20 and 100 nm. Thus, TiO₂ helps to create and stabilize small and well-dispersed Cu NPs, which allows more active sites-Cu NPs for CO₂ reduction.

2.2. Catalyst Activity in a Standard Three-Electrode Cell

Figure 3 shows the cyclic voltammograms (CVs) obtained with pure Cu NPs, TiO₂, 40 wt % Cu/TiO₂ and 40 wt % Cu/C, which was successfully synthesized, characterized by XRD, SEM, and TEM (Figures S3–S5, SM), coated on glass carbon electrode (d = 2 mm). When the electrolyte was bubbled with CO₂, reduction peaks associated with CO₂ reduction were observed. Comparison of Figure 3a,c,d, 40 wt % Cu/TiO₂ (−0.607 V vs. RHE) showed a 47 mV lower onset potential for CO₂ reduction than that on Cu NPs (−0.654 V vs. RHE) and a 143 mV lower onset potential than 40 wt % Cu/C (−0.750 V vs. RHE), respectively. The improvement of 40 wt % Cu/TiO₂ may be because of the adsorption and stabilization of CO₂ on the TiO₂ surface [23]. In order to better explain the phenomenon, TiO₂ alone as catalyst exhibited low onset potential of −0.644 V vs. RHE for CO₂ electroreduction (Figure 3b). The large overpotential of CO₂ reduction could be due to the barrier of the initial electron in forming a CO₂⁻ intermediate, which exhibits poor stability on most metal surfaces [4]. The CV of the 40 wt % Cu/TiO₂ catalyst was seen to have an earlier onset potential, which is in agreement with the better performance on the 40 wt % Cu/TiO₂ catalyst surface than that on the Cu NPs and 40 wt % Cu/C catalysts surface. In addition, for the pure Cu NPs electrode, as shown in Figure 3a, much higher reduction current density J_geo (based on geometric area of 0.0314 cm²) was observed for N₂ saturated solution than that for CO₂ saturated solution. It strongly indicates CO₂ reduction reaction was inhibited on the surface of Cu NPs in the entire potential range. The J_geo of 40 wt % Cu/C presents a slight increase under CO₂ atmosphere compared with N₂ atmosphere. However, in sharp contrast, for 40 wt % Cu/TiO₂ sample, rapid increment of the J_geo was observed under both conditions, in comparison with the J_geo curve (black line) in N₂ atmospheres, the CO₂-saturated solution showed larger J_geo in the entire potential range, which indicates 40 wt % Cu/TiO₂ sample exhibited obvious catalytic activity towards CO₂ reduction (Figure 3c). Besides, the maximum J_geo in CO₂ saturated solution of Cu NPs and 40 wt % Cu/C are 27.94 mA/cm² and 6.7 mA/cm² at −1.45 V vs. RHE, respectively, which are lower than the J_geo of 40 wt % Cu/TiO₂ (31.79 mA/cm²) at −1.45 V vs. RHE. This once again proves that 40 wt % Cu/TiO₂ shows remarkable performance of CO₂ electroreduction, which was reconfirmed by the electrolytic data in the following text.

2.3. The Electrocatalytic Activity of Various Cu/TiO₂

Electrolysis was performed in 0.2 M KI solution saturated with CO₂ at a constant cathodic potential of −1.45 V vs. RHE to evaluate the catalytic activity of Cu/TiO₂. The products were detected through GC analysis (Figure S6, SM). The results demonstrated that ethanol (with little propanol and trace methanol) is the main product, and we are only interested in liquid products. Figure 4 reveals a dramatic difference in the amount of Cu content in the overall CO₂ electrocatalytic reduction activity of Cu/TiO₂. In the case of bare TiO₂ catalyst, neither ethanol nor n-propanol were detected after electroreduction of CO₂. Comparatively, after loading of the Cu NPs, 40 wt % Cu/TiO₂ exhibited the optimal performance in terms of total FE (33.6%) among all the prepared samples. The total FE decreased to about 10−16% for loadings of 20, 60, and 80 wt % and to below 10% for loading of 10 wt %; the efficiency significantly decreased for loadings of 5 wt % and pure Cu NPs. For the sake of explaining the phenomenon, the TEM images of various Cu/TiO₂ were again compared and analyzed, and electrochemistry surface area (ECSA) test of various Cu/TiO₂ were carried out. As the TEM images in Figure 2 and histograms in Figure S7 (SM) show, the TEM images of 60 wt % and higher content of Cu samples suggest that Cu particles are more prone to agglomerate than 40 wt % sample, resulting in smaller surface area for these samples, and then the active surface of samples for CO₂ electroreduction decrease, which was consistent with the results of ECSA test for various Cu/TiO₂ (Figure S8 and Table S1, SM). Moreover, as seen from Figure 4 and Table S1 (SM), along with the mass ratio of Cu NPs in the Cu/TiO₂ samples increasing, the catalytic activity and ECSA revealed synchronous change. 40 wt % Cu/TiO₂ possesses a largest ECSA of 3.32 cm² in all various Cu/TiO₂ catalysts. Besides, 60 wt % Cu/TiO₂ displays lower catalytic activity than 80 wt %, due to the decrease of active surface, as confirmed by ECSA. It could be clearly seen that the catalytic activity of various Cu/TiO₂ had a close connection to the corresponding ECSA. In fuel cell applications, Ag supported on titanium dioxide exhibited the same properties [23]. For comparison, potentiostatic electrolysis was conducted by applying a potential of −1.45 V vs. RHE on 40 wt % Cu/C in 0.2 M KI solution saturated with CO₂, which only results in 8.4% FE for ethanol. As shown in Table S1 (SM), both ECSA and current density J_ECSA (based on ECSA) of 40 wt % Cu/C were smaller than those of 40 wt % Cu/TiO₂, which indicates that high performance for CO₂ electroreduction on Cu/TiO₂ may be ascribed to the synergistic effect between Cu and TiO₂. Furthermore, although 5 wt % Cu/TiO₂ and Cu NPs have a similar ECSA (Table S1, SM), the J_ECSA of 5 wt % Cu/TiO₂ was slightly larger than the one of Cu NPs. It was reconfirmed that the prepared composites with excellent activity for CO₂ conversion are obtained by the synergetic effect of Cu NPs and TiO₂.

2.4. Electrocatalytic Characterization

We also explored the performance of the catalyst amounts. The amounts of the catalyst loading—such as 0.5, 1, 2, 3, 4, and 5 mg·cm⁻²—for 40 wt % Cu/TiO₂ were painted on a carbon paper to form a cathode. As shown in

Table 2. FEs for ethanol and n-propanol with different amounts of 40 wt % Cu/TiO₂ deposited over carbon paper at −1.45 V vs. RHE ^a

Cu/TiO2 Amount (mg/cm2)	FEethanol (%) b	FEn-propanol (%) b
5	2.1	1.2
4	4.6	1.8
3	7.2	3.2
2	19.0	6.0
1	27.4	6.2
0.5	6.3	2.5

^a Electrolysis condition: electrolyte = 0.2 M KI, E = −1.45 V vs. RHE. ^b Obtained by GC.

, the catalyst amount of 1 mg·cm⁻² exhibited the optimal performance among all other amounts. Despite the increasing catalyst amount from 0.5 to 5 mg·cm⁻², the activity did not increase. The accessible surface area is not proportional to the corresponding catalyst loading. This finding is important because the resulting activity is correlated with the actual active surface area. Thus, to evaluate this finding, we performed SEM for analyzing the catalytic layer thickness and morphology of the as-prepared electrodes. The SEM images (Figure S9, SM) indicate how the Cu/TiO₂ NPs are progressively deposited on the carbon paper with increasing catalyst loading. Interestingly, for high loadings (3, 4, and 5 mg·cm⁻²), the Cu/TiO₂ NPs completely cover the porous structure of the carbon paper, forming a massive and compact layer, which results in Cu/TiO₂ NPs hardly contacting the catholyte and CO₂. For low loadings (0.5 mg·cm⁻²), the carbon fibers were uncovered by the Cu/TiO₂ NPs, and these fibers were deposited and formed clusters. Despite the clusters being relatively well accessible for the reactants, the amount of the electrocatalytic material is very low, thus the activities are also low. An intermediate situation was observed for medium loadings (1 and 2 mg·cm⁻²). In these electrodes, the carbon fibers were well covered with the Cu/TiO₂ NPs, and a massive and compact layer was not formed. The limit for the massive and compact layer formation is approximately 1 mg·cm⁻² metal loading. These observations are in accordance with the results shown in Table 2.

In order to evaluate the optimal potential, CO₂ reduction measurements on 40 wt % Cu/TiO₂ exhibiting its superior electrochemical characteristics was firstly carried out at the potential range from −1.05 to −2.05 V vs. RHE. Figure 5 shows the remarkable influence of cathodic potential on FE. As cathodic potential shifts from −2.05 V to −1.05 V vs. RHE, the total FE increased first, reached a maximum of 33.6% at −1.45 V vs. RHE, and decreased to 2.2% at −1.05 V vs. RHE. Similar results were found for the FE of ethanol, whereas it is almost insensitive to n-propanol. Presumably, the pathways for ethanol and propanol production are different, resulting in different FE tendencies. The total J_ECSA increased with decreasing potential. However, partial J_ECSA for ethanol is very different from the total J_ECSA and achieves the maximum at −1.45 V vs. RHE (Figure 5). This fact may be mainly caused by the competitive hydrogen evolution reaction, which becomes important for negative potentials that decrease the objective FEs.

To evaluate the stability of 40 wt % Cu/TiO₂ electrocatalysts, potentiostatic electrolysis measurements were carried out in continuous CO₂ electrochemical reduction at the optimal potential over a long time test. Figure 6a displays the J_ECSA vs. time curve at the corresponding potentials for 25 h. It can be clearly seen that the J_ECSA is sharply decreased at the beginning of about 25 min, and then keeps steady at a constant value (8.66 mA/cm²), indicating its excellent stability as well. Figure 6b shows the comparison of XRD curves for 40 wt % Cu/TiO₂ before and after 25 h reaction, of which the significant characteristic diffraction peaks for Cu NPs were represented by the dash lines. Similarly, there was not any obvious variation in the XRD curves for 40 wt % Cu/TiO₂, which offers powerful proof of the excellent stability of 40 wt % Cu/TiO₂. Furthermore, the comparison of SEM images of 40 wt % Cu/TiO₂ before and after 25 h reaction is presented in Figure S10 (SM). The negligible difference presented in the Figure S10 (SM) demonstrates the superior stability of 40 wt % Cu/TiO₂ samples in a prolonged electrochemical process.

3. Materials and Methods

3.1. Materials

Cu(NO₃)₂·3H₂O, sodium citrate and hydrazine hydrate (AR grade), TiO₂ (CP grade) were purchased from Sinopharm Chemical Reagent Co. (SCR, Shanghai, China) with 99% purity. Conductive carbon black (C, VXC72R) was purchased from Shanghai Hesen Electrical Co. (Shanghai, China) Nafion^® 117 solution (5%) and Nafion^® 117 membrane was obtained from Dupont (Wilmington, DE, USA).

3.2. Material Characterization and Product Measurement

Potentiostatic electrolysis and cyclic voltammetry were performed using a CHI 660 C electrochemical station (Shanghai Chenhua Instrument Co. Ltd., Shanghai, China). Crystal-phase XRD patterns of Cu/TiO₂ catalysts were recorded using an Ultima IV X-ray powder diffractometer (Kurary, Tokyo, Japan) equipped with Cu Kα radiation (k = 1.5406 Å). The SEM image of the samples was obtained using a Hitachi S-4800 field-emission scanning electron microscope (Tokyo, Japan). The average particle size and different morphologies of the samples were analyzed using a high-resolution Hitachi JEM-2100F transmission electron microscope (Tokyo, Japan). The values for actual Cu loading of the synthesized catalysts were determined on an inductively coupled plasma atomic emission spectroscopy (IRIS Intrepid II XPS, Waltham, MA, USA). Liquid-phase products were analyzed by gas chromatography instrument equipped with a flame ionization detector (SHIMADZU, GC-2014C, Tokyo, Japan). Analysis of the product (Figure S3, SM) and calculation of FE (Equation (S1), SM) are detailed in SM.

3.3. Material Synthesis

Citrate-protecting method was used to obtain different Cu/TiO₂ catalysts. For synthesis of 40 wt % Cu/TiO₂, Cu(NO₃)₂·3H₂O (0.761 g) was dissolved in the distilled water (60 mL). TiO₂ (0.300 g) was added to the solution and stirred for 40 min. Sodium citrate (1.471 M, 20 mL) dropwise was then added to the solution and stirred. Cu²⁺ was reduced by dropwise addition of hydrazine hydrate (50%, 5 mL) with vigorous stirring at room temperature. After gently stirring the solution overnight, the mixture was centrifuged, washed with distilled water, and dried in a vacuum oven at 80 °C for 4 h. Samples with Cu loadings of 5, 10, 20, 60, and 80 wt % were obtained by changing the amount of TiO₂, while pure Cu NPs was prepared without addition of TiO₂. Moreover, the prepared of 40 wt % Cu/C was used the same method, just replacing TiO₂ to C.

3.4. Electrode Preparation and Electrochemical Test

In electrochemical measurements, X wt % Cu/TiO₂ NPs were painted on a porous carbon paper (Toray, TGP-H60, Tokyo, Japan) to form the working electrodes. Potentiostatic electrolysis was carried out in a gas diffusion electrode cell, which contains a piece of Nafion^® 117 cation exchange membrane (H⁺ form) as a separator, platinum plate electrode as a counter electrode, and saturated calomel electrode (SCE) as a reference electrode. KI aqueous solution (0.2 M) was used as electrolyte and saturated with CO₂ by bubbling for 20 min before the electrolytic experiment. CO₂ was bubbled continuously during the experiment. The electrochemical behavior of the electrodes was tested in a three-electrode electrochemical cell, with a specific volume of catalytic ink coated on the glass carbon electrode (diameter = 2 mm) as the working electrode. A platinum plate electrode was employed as a counter electrode, and SCE was used as reference electrode. All these tests were performed at ambient temperature and pressure. All potentials were converted into the reversible hydrogen electrode (RHE) by the conversion

E(RHE) = E(SCE) + 0.059 × pH + 0.241.

(1)

4. Conclusions

The electroreduction of CO₂ to ethanol under ambient conditions was investigated on Cu/TiO₂ electrodes in a CO₂-saturated KI aqueous solution. By using TiO₂ as support, small and well-dispersed Cu NPs are loaded on the TiO₂ surface simply, which is beneficial for CO₂ electroreduction. Among various Cu/TiO₂ catalysts, 40 wt % (the optimal mass ratio) shows the highest electrocatalytic activity with the total FE, which is approximately eight-fold higher than the unsupported Cu NPs. The 40 wt % Cu/TiO₂ electrode shows the highest current density compared with other Cu/TiO₂ electrodes under similar experimental conditions. The prepared 40 wt % Cu/TiO₂ with excellent activity for CO₂ conversion is obtained by the high surface area effect and synergistic effect of Cu NPs and TiO₂. In addition, the 40 wt % Cu/TiO₂ possessed outstanding stability during a 25 h electrochemical process. It is anticipated that the Cu/TiO₂ composite electrocatalyst may provide new insights for transition metal and metal oxide particles applied in CO₂ conversion to useful chemicals.