Multilayered Zn nanosheets as an electrocatalyst for efficient electrochemical reduction of CO2

Multilayered Zn nanosheets as an electrocatalyst for efficient electrochemical reduction of CO2

Journal of Catalysis 357 (2018) 154–162 Contents lists available at ScienceDirect Journal of Catalysis journal homepage: www.elsevier.com/locate/jca...

4MB Sizes 1 Downloads 196 Views

Journal of Catalysis 357 (2018) 154–162

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Multilayered Zn nanosheets as an electrocatalyst for efficient electrochemical reduction of CO2 Taotao Zhang a,b, Xianfeng Li a,c, Yanling Qiu a, Panpan Su a, Wenbin Xu a, Hexiang Zhong a,c,⇑, Huamin Zhang a,c,⇑ a b c

Division of Energy Storage, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China University of Chinese Academy of Sciences, Beijing 100049, China Collaborative Innovation Centre of Chemistry for Energy Materials (iChEM), Dalian 116023, China

a r t i c l e

i n f o

Article history: Received 4 August 2017 Revised 3 November 2017 Accepted 5 November 2017 Available online 24 November 2017 Keywords: CO2 reduction Zinc Electrocatalysis Nanosheets Selectivity

a b s t r a c t Electrochemical reduction of CO2 into useful fuels, when powered by renewable energy, is an ideal process for replacing fossil feedstocks and simultaneously decreasing CO2 emission. Developing inexpensive electrocatalysts for CO2 reduction to CO with high activity and selectivity is an important part of CO2 conversion. Zn as a low-cost metal is identified to be a promising electrocatalyst for CO2 conversion. Here, we report a Zn electrode composed of multilayered Zn nanosheets (MZnNSs) with high density of edge sites. The MZnNSs catalyst exhibited a maximal CO Faradaic efficiency about 86% at 1.13 V vs RHE, which is almost 9 times higher than that of bulk Zn foil. Density functional theory (DFT) calculations suggest that the improvement of the activity and selectivity of MZnNSs for CO2 reduction is attributed to its high density of edge sites. Ó 2017 Elsevier Inc. All rights reserved.

1. Introduction In the past two centuries, human society was taken into an unprecedented era of prosperity by the utilization of fossil fuels, which still account for a predominating proportion of the total energy source for worldwide consumption today [1,2]. However, the combustion of fossil fuels has significantly given rise to the level of carbon dioxide (CO2) in the atmosphere, which is directly related to global warming and climate changes, making a negative impact on the ecological environment [3]. In view of the inevitable depletion of fossil fuels and undesirable greenhouse gas effect, it is becoming critical to develop a sustainable way to recycle the overly produced CO2 into reusable carbon forms [4]. Generally, the feasible approaches of CO2 conversion to usable fuels mainly include chemical methods [5], photocatalytic reduction [6], electrochemical reduction [7] and biological means [8,9]. Among those techniques, electrochemical reduction of CO2 powered by renewable energy generation or nuclear energy is considered as a promising method. It is also an energy storage strategy for storing intermittent renewable electricity in energy dense car-

⇑ Corresponding authors at: Division of Energy Storage, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China. E-mail addresses: [email protected] (H. Zhong), [email protected] (H. Zhang). https://doi.org/10.1016/j.jcat.2017.11.003 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

bonaceous fuels [10], such as HCOOH, CH4, C2H4 and CO [11]. The product of CO mixed with byproduct H2 from CO2 electrochemical reduction can be utilized for downstream processing of syngas by using Fischer-Tropsch chemistry [12]. However, there are still many obstacles in the reduction of CO2 to CO. As CO2 is a thermodynamically stable molecule, the key step of CO2 reduction to intermediate species CO2 requires a high overpotential and suffers from poor reaction selectivity because of competitive proton reduction [13]. Thus, appropriate catalysts that can activate CO2 need to be developed to simultaneously increase the product selectivity and lower the overpotential [14]. During the past few decades, a number of metal catalysts have been evaluated for electrochemical CO2 reduction to CO. Typically, precious metals such as Au and Ag are the most active catalyst materials for efficient reduction of CO2 to CO, but they are too expensive to be commercialized at large scale [11]. Therefore, it is crucial to develop inexpensive metal catalysts that can convert CO2 to CO with high selectivity, efficiency and high activity [15]. Zn as an affordable and environmentally benign metal has recently attracted much attention for CO2 conversion applications, because Zn metal can act as an catalyst for chemical [16], photocatalytic [17] and electrochemical [18,19] reduction of CO2. In the field of electrochemical reduction of CO2, Zn has been identified to be a promising catalyst with a wide range of CO selectivity in terms of the Faradaic efficiency (3.3–63.3%) [20]. In addition, as

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

native surface oxides may be reduced to afford more active surface structures, surface oxidation treatment on the electrocatalysts is effective to improve the performance [21–25]. Therefore, the activity and product selectivity of a catalyst for CO2 reduction are dramatically influenced by its geometry, morphology and roughness [26]. Density functional theory (DFT) study has predicted that the edge sites of the catalysts are much more active for CO2 reduction to CO, while the corner sites are much more favorable for H2 evolution reaction, and this prediction has been validated by experimental results in Au and Ag catalysts [27,28]. Thus, the catalytic performance of Zn catalyst for CO2 reduction has a great chance to be improved by morphology design or surface modification. Here, we report a loose and porous structure Zn electrode consisting of multilayered Zn nanosheets (MZnNSs), which is prepared by an oxidation/reduction process, as shown in Scheme 1. During the reduction process of the ordered ZnO nanowires arrays formed by hydrothermal oxidation treatment of Zn foil, a large surface area and more edge sites are formed to offer an improved CO2 reduction kinetics. The prepared MZnNSs achieved a superior CO Faradaic efficiency about 86% with a CO partial current density above 5.2 mA cm 2 at moderately negative potential 1.13 V vs RHE, which is almost 9 times and 14 times higher than that of bulk Zn foil, respectively. Furthermore, the MZnNSs electrode showed a good stability with negligible deterioration during 7 h continuous operation. DFT calculations suggest that the high density of edge sites is benefit for the improvement of the activity and selectivity of MZnNSs for CO2 reduction. 2. Experimental section 2.1. Preparation of electrodes A piece of Zn foil (1.5  4.0 cm2, thickness 0.1 mm, 99.99%) was first mechanically polished to remove the native oxide layer. After rising with water/ethanol and drying under Ar, the Zn foil was put into a 100 mL Teflon-lined stainless-steel autoclave, which contained 25 mL 0.6 mM malic acid solutions. The autoclave was placed in an oven at 190 °C for 8 h and then cooled to room temperature. The Zn foil was taken out from the solution, rinsed with deionized water and then dried in an oven at 120 °C for 4 h to obtain the ordered ZnO nanowires array, which is denoted as OZnO. As control sample, the disordered oxidized Zn nanowires were prepared by the similar process with exception of that malic acid solution was replaced by deionized water, which is denoted as D-ZnO. The multilayered Zn nanosheets (MZnNSs) and RD-ZnO were obtained by reducing the O-ZnO and D-ZnO respectively at 0.63 V (vs RHE) for 30 min in CO2-bubbled 0.5 M NaHCO3 solution. 2.2. Materials characterizations The scanning electron microscopy (SEM) images were obtained on a JSM-7800F microscope. X-ray diffraction (XRD) patterns were collected from 30° to 90° in 2h at a scanning rate of 1° min 1 on a

155

DX-2700 X-ray diffractometer (Dandong Haoyuan Instrument Co.,) using the a Cu Ka radiation source at 40 kV and 30 mA (k = 0.154 nm). Transmission electron microscopy (TEM), high resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) images were recorded on a JEOL JEM2000EX (120 kV). X-ray photoelectron spectroscopy (XPS) measurements were carried out using an ESCALAB 250Xi spectrometer equipped with a nonmonochromatized Al Ka X-ray source. 2.3. Electrochemical measurements The electrochemical CO2 reduction measurements of the catalysts were conducted with a 2273 potentiostat (EG&G Instrument) through a three-electrode system. The working electrode was the as-prepared electrode (exposed area: 3 cm2). The counter electrode was a Pt sheet and the reference electrode was a Hg2Cl2/Hg/saturated KCl electrode (SCE), respectively. All the applied potentials are reported as reversible hydrogen electrode (RHE) potentials scale using E (vs. RHE) = E (vs. SCE) + 0.242 V + 0.0591 V  7.2. The electrochemical measurements were conducted in a gastight two-compartment electrochemical cell with a piece of proton exchange membrane Nafion 115 (Dupont, USA) as a separator. 0.5 M NaHCO3 saturated with CO2 and 0.1 M H2SO4 aqueous solution were adopted as cathode and anode electrolyte, respectively. During the CO2 reduction experiments, the cathodic electrolyte was stirred at the rate of 500 rpm and continuously saturated with CO2 gas at the flow rate of 60 ml min 1. The output gas was vented directly into the gas-sampling loop of a gas chromatograph (GC, Shimadzu GC-2014). Argon and nitrogen were used as the carrier gases. The hydrogen was quantified through a thermal conductivity detector (TCD) and CO was quantified by a flame ionization detector (FID). The liquid product was quantified by ion chromatography (ICS-1100, Dionex Corporation). The current densities are calculated based on geometric area (3 cm 2). The electrochemical active surface areas (EASC) were determined by measuring electrochemical double-layer capacitance [29]. ECSA = RfS, in which S stands for the geometric area (3 cm 2). The roughness factor Rf was estimated from the ratio of double-layer capacitance. The roughness measurement of the electrodes were conducted in N2 purged 0.5 M Na2SO4 aqueous solution at various scan rates. The scanning potential ranges from 1.33 to 1.23 V vs SCE. The current densities were obtained from the double layer charge/discharge curves at –1.28 V vs SCE. 3. Results and discussion 3.1. Characterizations of the prepared electrodes Zn foil (Fig. S1) was firstly treated under hydrothermal condition with the presence of malic acid to fabricate ZnO with morphology of relatively ordered nanowires. SEM images in Fig. 1a and b show that the hexagonal O-ZnO nanowires are generated and grown at some angle to the surface of Zn foil. Typical diameters of the obtained nanowires are in the range of 50–500

Scheme 1. Preparation process of the MZnNSs.

156

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

Fig. 1. (a) Low-resolution SEM and (b) high-resolution SEM images, (c) TEM image and (d) HRTEM image of O-ZnO. (e) Low-resolution SEM and (f) high-resolution SEM images of D-ZnO.

nm and ultra-thin ZnO nanobelts of the outer layer are formed by the transformation from zinc hydroxides. The TEM image of a typical O-ZnO nanowire in Fig. 1c shows that the length of the O-ZnO nanowire is about 1.5 lm. The corresponding HRTEM image in Fig. 1d exhibits that the observed lattice fringe distance of 2.61 Å corresponds to the spacing between the (0 0 2) planes of the hexagonal wurtzite ZnO. The corresponding SAED pattern indicates that the ZnO nanowire has single-crystalline structure and is grown along the [0001] direction. While, with the absence of malic acid under the same hydrothermal condition, the growth of ZnO nanowires on the zinc plate is disordered (D-ZnO, Fig. 1e and f), which indicates that the malic acid plays a key role for controlling the growth directions of the ZnO nanowires. To give crystallographic evidence of the formation of MZnNSs, X-ray diffraction was measured. Fig. 2 shows the XRD patterns of O-ZnO, D-ZnO and untreated Zn foil. The characteristic of Zn foil was matched with JCPDS files #04-0831 for metallic Zn. After hydrothermal oxidation, both the obtained O-ZnO and D-ZnO are in agreement with ZnO (JCPDS files #36-1451) and a small amount of metallic Zn is still present, indicating ZnO phase is formed on the surface of Zn foil after the hydrothermal treatment. When the OZnO and D-ZnO are used as catalysts for electrochemical CO2 reduction, the ZnO will be reduced to Zn before CO2 reduction to offer catalytic capacity [30]. Electrochemical reduction of O-ZnO and D-ZnO were performed in 0.5 M NaHCO3 purged with CO2 gas at 0.63 V for 30 min to obtain MZnNSs and RD-ZnO, and the

Fig. 2. XRD patterns of the as-prepared Zn-foil, O-ZnO, D-ZnO, MZnNSs and RDZnO. The peaks of standard Zn and ZnO samples are also presented.

cathodic peaks can be observed during the first 20–25 min, which is attributed to the reduction of ZnO to Zn metal (Fig. 3). The current densities become stable after 25 min, indicating the ZnO

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

Fig. 3. Electrode currents recorded during reduction of O-ZnO and D-ZnO at V vs. RHE in 0.5 M NaHCO3 purged with CO2 gas.

0.63

reduction is completed. After reduction, there is no cathodic peaks can be observed in the I-t curves for this two samples tested in the same conditions (Fig. S2). Combining with XRD patterns in Fig. 2, after electrochemical reduction, the characteristic of obtained MZnNSs is pure metallic Zn. However, the obtained RD-ZnO still

157

contains a small amount of ZnO, suggesting a small amount of oxides in RD-ZnO are surprisingly stable against reduction at the applied potential. After CO2 electrolysis, no peak of ZnO is observed in the XRD patterns of RD-ZnO (Fig. S3), suggesting that the reduction potential of some ZnO is shifted to more negative potential [31]. As shown in Fig. 4a, a different surface morphology of loose and porous structure is formed after reduction process for O-ZnO. It can be seen more clearly on a micrograph taken with a larger magnification in Fig. 4b that the highly porous structures are composed of continuous nanoparticles networks, and MZnNSs that have shared growth centers builds up the nanoparticles. This unique structure contributes to a large surface area and high porosity compared with smooth bulk Zn foil. The TEM image of the typical nanosheets in Fig. 4c shows that the two nanosheets are hexagonal with the diagonal distance of 80 and 200 nm. Fig. 4d shows the HRTEM image of the bigger nanosheet in Fig. 4c. The fringe spacing of nanosheet is approximately 2.31 Å, conforming to the atomic spacing of Zn (1 0 0) facets. The corresponding SAED pattern indicates that the sheet is composed of a single crystal structure and lies within the (0 0 2) plane. The measured angle of the two edges is 120°, which is well matched with the angle between (0 1 0) and (1 0 0) crystal face, meaning that the two edges point to the (0 1 0) and (1 0 0) facets. It can be inferred that the formation of MZnNSs may include the recombination of Zn atoms and the loss of oxygen atoms during the reduction of ZnO nanowire. The atom

Fig. 4. (a, b) SEM images, (c) TEM image and (d) HRTEM image of MZnNSs. The corresponding SAED pattern and atom model of the typical Zn nanosheet are illustrated. (e, f) SEM images of RD-ZnO.

158

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

model of the Zn nanosheet is presented in the insert illustration of Fig. 4d. The unique hexagonal nanosheets expose much more edge sites compared with corner sites. Comparably, the surface morphology of RD-ZnO obtained from reduction of D-ZnO is also a loose and porous structure (Fig. 4e). However, as shown in a larger magnification in Fig. 4f, the networks are composed of irregularly shaped nanowires, which is completely different form MZnNSs. This kind of structure exposed much lower density of edge sites compared with MZnNSs. To further gain insights into chemical structure changes on the surface of the catalyst during the preparation process, the corresponding X-ray photoelectron spectroscopy was performed (Fig. 5a). Fig. 5b shows the Zn 2p spectrum of MZnNSs, O-ZnO and Zn foil. The amount of Zn2+ increased after hydrothermal treatment, which is related to the transformation of Zn0 to Zn2+. While, the amount of Zn2+ decreased after electroreduction process, indicating the reduction of Zn2+ to Zn0.[32] Fig. 5c reveals the O 1s spectrum of the three electrodes that can be fitted to three typical peaks.[33] The peak at 530.2 eV is associated with O2 from ZnO. As expected, the intensity of this component is found to increase after hydrothermal oxidation and followed by a decreasing after electroreduction. The peak at 531.5 ± 0.1 eV is related to oxygen vacancy and the peak at 532.5 ± 0.1 eV is attributed to adsorbed oxygen, such as -OH. Based on XPS results, the surface of O-ZnO is dominated by adsorbed oxygen and less abundant of oxygen vacancy, which is owing to the oxidation process in hydrothermal condition. It should be emphasized that Zn is a kind of very active metal and can be easily oxidized by O2 under atmosphere.[18] Thus, O2 can still be detected on the surface of MZnNSs. Thus, the changes in zinc and oxygen state of XPS analysis are well consistent with former characterizations. 3.2. Electrocatalytic performance for CO2 reduction Since the MZnNSs catalyst with special nanostructure showed the most promising morphological properties for electrochemical CO2 reduction, electrocatalytic performance of the catalyst for CO2 reduction was investigated in aqueous solutions. The electrochemical behaviors of Zn-foil, RD-ZnO and MZnNSs for CO2 reduction were first evaluated in CO2 or N2 saturated 0.5 M NaHCO3 solution by linear sweep voltammetry test (Fig. 6). All electrodes exhibit higher current densities under CO2 atmosphere than N2 atmosphere, demonstrating that all the three electrodes have electrocatalytic activity towards CO2 reduction. The onset potentials (0.8 mA cm 2) for CO2 reduction on RD-ZnO and MZnNSs electrode

( 0.59 V vs RHE) are more positive than that of Zn foil electrode ( 0.72 V), indicating that CO2 can be more efficiently reduced by the RD-ZnO and MZnNSs catalyst than Zn foil electrode. To further examine the performance of Zn-foil, RD-ZnO and MZnNSs electrodes for electrocatalytic CO2 reduction, constantpotential electrolysis was conducted in CO2 saturated NaHCO3 solution (Fig. S4). The applied potentials for electrolysis were selected in the range of 0.83 to 1.33 V based on the onset potential for CO2 reduction on the as-prepared electrodes. The main product formed by the reduction of CO2 is CO and a small amount of formate is obtained (Fig. S5). And the byproduct formed by concurrent proton reduction reaction is H2. One way to represent the selectivity of CO2 reduction to a given product is by calculating Faradic efficiency. Fig. 7a illustrates the dependence of Faradaic efficiency for CO production on the applied electrolysis potential. A maximum CO Faradaic efficiency of 86% for MZnNSs electrode is achieved at moderately negative potential ( 1.13 V), which is almost 9 times higher compared to Zn foil (9.3%), because Zn foil is dominated by hydrogen evolution reaction (Fig. 7b). The total current densities for the three kinds of Zn electrodes under various potentials are compared and shown in Fig. 7c. Although the total current densities of RD-ZnO and MZnNSs electrodes at each potential are around 1.5 times as high as that of bulk Zn foil, the total current density for Zn foil is mostly contributed by hydrogen evolution reaction. Combining the results about the dependence of CO partial current density on applied potential presented in Fig. 7d, under the same potential 1.13 V, MZnNSs electrode presents a CO partial current density above 5.2 mA cm 2, approximately 14 times as high as that of Zn bulk foil, whereas the ECSA of MZnNSs is only 6.3 times as high as that of Zn bulk foil (Fig. S6, Table S1). The activity per unit ECSA of MZnNSs seems to be only 2.2 times as high as that of Zn bulk foil. However, at too negative potential with high current, the mass transfer will have a bigger influence on the electrode with a higher ECSA, leading to insufficient supply of electrolyte and CO2. Therefore, the surface area effect may be overestimated [18]. Actually, at more positive potentials ( 0.55 to 0.93 V vs RHE, Fig. 8), the CO partial current density of MZnNSs is 17–36 times as high as that of bulk Zn foil, suggesting MZnNSs electrode has much higher electrocatalytic activity for CO2 reduction. Compared with bulk Zn foil, the improved selectivity and activity of MZnNSs electrode towards CO2 reduction is supposed to be benefited from its special nanostructure with higher surface area and higher density of edge sites. Based on the surface morphology of the MZnNSs (Fig. 4), it is speculated that the hydrogen evolution reaction on the surface of MZnNSs catalyst is suppressed

Fig. 5. (a) XPS overall survey spectrum for Zn-foil, O-ZnO and MZnNSs. XPS spectra for as-prepared Zn foil, O-ZnO and MZnNSs: (b) Zn 2p and (c) O 1s.

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

Fig. 6. Linear sweep voltammetry of Zn-foil, RD-ZnO and MZnNSs in N2 (solid line) and CO2 (dotted line) saturated 0.5 M NaHCO3.

by its high density of edge sites. The maximum CO Faradaic efficiency of 46% for RD-ZnO electrode is achieved, which is a little lower than MZnNSs but much higher than Zn foil. It has been proved that ZnO would be reduced to Zn under the CO2 electrolysis condition by in-situ X-ray absorption studies [30]. However, the disorder ZnO nanowires on D-ZnO cannot be fully electroreduced at moderate negative potential (Fig. 2). The residual oxygen might be the source of the lowered current density and selectivity on RDZnO electrode, because the ZnO can be reduced in more negative

159

Fig. 8. Tafel plots of the CO productions for Zn foil, RD-ZnO and MZnNSs.

potentials during the CO2 electrolysis (Fig. S3), leading to insufficient utilization of current. What’s more, the surface structure of RD-ZnO contains lower density of edge sites than that of MZnNSs, leading to its lower CO selectivity. In recent years, some important results for the electrochemical reduction of CO2 on different Zn catalysts have been reported (Table 1. Electrochemical reduction of CO2 on Zn alloys [39] or bimetals [40] are not included). Among them, the Zn catalysts with morphology of high density of edge sites show high product selec-

Fig. 7. Comparison of (a) CO Faradaic efficiency, (b) hydrogen Faradaic efficiency, (c) total current density and (d) CO partial current density for Zn-foil, RD-ZnO and MZnNSs at 0.83 V to 1.33 V (vs RHE) in CO2 saturated 0.5 M NaHCO3 solution.

160

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

tivity of CO for CO2 reduction, while the Zn catalysts with morphology of particles or polycrystalline structure show preference for formate production. Presumably, the morphology and crystal structure of Zn catalysts have a strong influence on their selectivity for electrochemical CO2 reduction. It needs to be emphasized that the electrolyte [41], especially its pH value [42], can significantly influence the performance of the catalyst. For instance, a Zn nanoplates catalyst has been reported to show a 93% Faradaic efficiency of CO in aqueous NaCl solution, while a 57% Faradaic efficiency of CO in aqueous NaHCO3 solution (Table 1). In order to get a better comparsion with previous reports, the generally used NaHCO3 solution was chosen as electrolyte in this paper. In this case, the selectivity of MZnNSs in this work outperforms previously reported Zn electrodes for CO2 reduction to CO evaluated in bicarbonate electrolyte (Table 1). Additionally, as the applied potentials in previous reports (Table 1) have not been corrected with iR-drop, we use applied potentials for comparison. However, the iR-drop plays a role in determining the actual potential of electrochemical CO2 reduction. In order to assess the impact of iR-drop, the solution resistances of the three electrodes were measured, and corrected potentials were calculated and shown in Table S2. The potentials of the three electrodes become 0.02–0.12 V more positive after correction, meaning that the impact of iR-drop is much lower compared to the reports with high surface area Cu catalysts for CO2 reduction (0.48–0.78 V) [43,44], because these Cu catalysts are evaluated under low concentration (0.1 M) of bicarbonate electrolyte (high solution resistances) at high overpotentials (high total currents). Besides, the difference of corrected potentials of the three catalysts is little, because their total current values and solution resistances are close to each other. The corrected potentials of the MZnNSs and RD-ZnO are the same, and the corrected potential of Zn foil is 0.03 V more negative than the two catalysts, meaning a little bigger profit for Zn foil if the potentials are not corrected, so it does not influence the comparison of the performances for the three catalysts. 3.3. Mechanistic insight at CO2 reduction to CO To understand the kinetic mechanism for the improved activity and selectivity of the MZnNSs catalyst, we studied the thermodynamic pathway of CO2 reduction to CO on the basis of Tafel analysis (Fig. 8). Combining with previous reports [22,28,45], the pathway for electrochemical reduction of CO2 on MZnNSs catalyst

is illustrated in Fig. 9. Firstly, the gaseous CO2 is adsorbed on the surface of MZnNSs catalyst and then reduced to CO2 anion radical after obtaining an electron. Subsequently, the CO2 radical combines with a proton donated by the HCO3 to form COOH intermediate. After that, the COOH obtains another electron from the electrode to generate adsorbed CO and finally the CO is desorbed. The value of Tafel slope always indicates a possible ratedetermining step. It has been reported that theoretically [46], the Tafel slope of 118 mV dec 1 means that the rate-determining step is the one-electron reduction of CO2 to form CO2 , while 59 mV dec 1 suggests that the step of proton transfer from HCO3 to CO2 determines the reaction rate. As shown in Fig. 8, the Zn foil, RDZnO and MZnNSs electrodes give rise to Tafel slopes of 153, 132 and 129 mV dec 1 respectively, indicating the single-electron transfer to CO2 is the rate-determining step for CO2 reduction to CO on these Zn electrodes. Therefore, stabilization of the intermediate CO2 radical plays a key role in the overall reduction of CO2 to CO. What’s more, the RD-ZnO and MZnNSs electrode show smaller value of Tafel slope than Zn foil, demonstrating that RD-ZnO and MZnNSs electrode have improved CO2 reduction thermodynamic kinetics, which is coincident with the LSV analysis (Fig. 6). High density of edge sites and enlarged surface area of MZnNSs may favor the activation of CO2 to CO2 and stabilizing the intermediate CO2 species, thus facilitating CO production as well as suppressing the H2 generation. 3.4. Theoretical analysis Referentially, on other metal catalysts, such as Au and Ag, theoretical analysis has suggested that the edge sites are much more active for CO2 reduction to CO, while the corner sites favor hydrogen evolution reaction [27,28,47]. In order to explore the feasibility of such a claim in the case of metal Zn catalysts, the free energy changes of each proton–electron pair transfer in the mechanism of CO2 to CO on various model Zn surfaces were calculated by density functional theory (DFT) simulations (Fig. S7 and S8). For comparison, free energies for CO2 reduction to CO on Zn(1 0 1), Zn edge site and Zn corner site are calculated and shown in Fig. 10a. It can be observed for all the three sites that the first step for formation of COOH⁄ meets an uphill energy barrier. This obstacle can be overcome by supplying energy in the form of the overpotential required to form reaction intermediates. On Zn edge site at 0.59 V vs RHE, the required DG to form the COOH⁄ intermediate

Table 1 Comparison of electrochemical CO2 reduction performances for different Zn catalysts. The applied potentials are normalized as RHE. Zn catalysts

Electrolyte

Main product

Multilayered Zn nanosheets

0.5 M NaHCO3 (pH 7.2) 0.5 M NaHCO3 (pH 7.2) 0.1 M KHCO3 (pH = 6.83)a 0.5 M KHCO3 (pH 7.2) 0.5 M KCl (pH 3.9) 0.5 M NaHCO3 (pH 7.2) 0.5 M NaCl (pH 4.5) 0.25 M K2SO4 (pH 4.2) 0.5 M NaHCO3 (pH 7.2) 0.5 M KHCO3 (pH 7.2)

CO

Zn dendrites Zn foil Hexagonal Zn Hexagonal Zn Zn nanoplates Zn nanoplates Reduced nanoporous ZnO Zn nanoparticles Zn powder a

This is referred from another work [38].

Potential (vs RHE)

Faradaic efficiency (%)

Ref.

1.13 V

86

This work

CO

1.1 V

79

[18]

CO

1.1 V

79.4

[34]

CO

0.85 V

80

[35]

CO

1.05 V

95

[35]

CO

0.93 V

57

[19]

CO

1.09 V

93

[19]

CO

1.2 V

92

[30]

Formate

1.93 V

87

[36]

Formate

1.5 V vs OCV

78.5

[37]

161

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

Fig. 9. Schematic pathway for electrocatalytic CO2 reduction to CO on the surface of Zn electrode.

Fig. 10. Free energy diagrams for electrochemical reduction of (a) CO2 to CO and (b) protons to hydrogen on Zn(1 0 1), edge site and corner site at

is significantly lower than that on Zn(1 0 1) and corner site, suggesting a higher catalytic activity of edge site for CO2 reduction to CO. Besides, the edge site and corner site show lower binding energies with CO⁄ compared to Zn(1 0 1), which is favor for the product evolution rate. Thus, the order of the performance for CO2 reduction to CO is: edge site > corner site > Zn(1 0 1). The free energies for hydrogen evolution reaction on Zn(1 0 1), Zn edge site and Zn corner site are also calculated and shown in Fig. 10b. The first proton-coupled electron-transfer step on Zn(1 0 1), Zn edge site and Zn corner site is easy to overcome. However, the obstacle for hydrogen evolution reaction is the second protoncoupled electron-transfer step, as an uphill energy barrier can be observed. The edge site and corner site require much higher energy barrier than Zn(1 0 1) in the second proton-coupled electrontransfer step, indicating that both edge site and corner site are not favor for hydrogen evolution reaction compared to Zn(1 0 1). Nevertheless, the edge site is a little more optional for hydrogen evolution reaction than corner site. Thus, the order of the performance for hydrogen evolution reaction is: corner site < edge site < Zn(1 0 1). The results of calculated free energetics for edge site and corner on Zn metal are different from Au and Ag. However, it should be noted that the corner site shows 0.03 eV higher energy barriers than edge site for hydrogen evolution reaction, while 0.09 eV higher energy barriers than edge site for CO2 reduction to CO. Therefore, the order of the comprehensive performance for CO2 reduction is: edge site > corner site > Zn(1 0 1). Combining with previous report [35], Zn(0 0 2) facets are found to be favorable for hydrogen evolution reaction while Zn(1 0 1) facets are favorable for CO formation. In this context, the hexagonal Zn nanosheets are piled along with the same direction, mainly exposing their edges and only one top (0 0 2) facet (Fig. 4b). Therefore, the corner sites on the surface of MZnNSs seem to be negligible compared with edge sites. In comparison to bulk Zn foil and RD-ZnO, the MZnNSs catalyst has higher density of edge sites, which contributes to its much improved CO selectivity and CO partial current density.

0.59 V vs RHE.

3.5. Stability of MZnNSs electrode Apart from high catalytic activity, good stability also has crucial meanings for practical applications of the catalyst. The electrocatalytic durability of the MZnNSs catalyst was evaluated at a constant 1.13 V potential in CO2 saturated 0.5 M NaHCO3. Fig. 11 illustrates the CO Faradaic efficiency and total current density as functions of the time. During continuous 7 hours electrolysis program, the total current density of MZnNSs catalyst shows good stability without any significant deterioration. Besides, the Faradaic efficiency of CO experiences a slight decrease during the first 4 hours and then keeps stable at 82% till the 7th hour. The high surface area of the catalyst benefited from loose and porous structure prohibits the poisoning of catalyst surface from impurities or C contaminations, contributing to good catalytic stability for CO2 reduction, because higher surface area is supposed to have better ability to withstand this unwanted deposition [48]. Thus, MZnNSs catalyst with high catalytic activity and outstanding stability shows great potential for CO2 reduction to CO.

Fig. 11. Stability study of the MZnNSs electrode at saturated 0.5 M NaHCO3.

1.13 V (vs RHE) in CO2

162

T. Zhang et al. / Journal of Catalysis 357 (2018) 154–162

4. Conclusions In summary, we have demonstrated a nanostructured catalyst composed of multilayered Zn nanosheets, which is prepared by electrochemically reducing the ordered ZnO nanowires obtained from hydrothermal oxidation of Zn foil. The low-cost Zn catalyst exhibits enhanced activity and selectivity for electrocatalytic CO2 reduction to CO. Compared with Zn foil, the highest catalytic performance of MZnNSs catalyst is more than 9 times higher in CO Faradaic efficiency (86%) at 1.13 V vs RHE. DFT calculations suggest that the improvement in activity and selectivity is attributed to high density of edge sites for favoring CO2 reduction while suppressing H2 evolution. Notably, this catalyst shows good stability in long term operation owing to its special nanostructure. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (No. 21576255 and No. 21577141), the Youth Innovation Promotion Association, CAS (2015147) and the Outstanding Youngest Scientist Foundation, Chinese Academy of Sciences (CAS). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.jcat.2017.11.003. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

M.S. Dresselhaus, I.L. Thomas, Nature 414 (2001) 332–337. S. Chu, A. Majumdar, Nature 488 (2012) 294–303. T.R. Karl, K.E. Trenberth, Science 302 (2003) 1719–1723. M. Aresta, A. Dibenedetto, Dalton Trans. 28 (2007) 2975–2992. George A. Olah, Alain Goeppert, A.G.K.S. Prakash, J. Org. Chem. 74 (2009) 487– 498. P. Usubharatana, Dena McMartin, A. Veawab, P. Tontiwachwuthikul, Ind. Eng. Chem. Res. 45 (2006) 2558–2568. J.-P. Jones, G.K.S. Prakash, G.A. Olah, Isr. J. Chem. 54 (2014) 1451–1466. C. Stewart, M.-A. Hessami, Energy Convers. Manage 46 (2005) 403–420. M. Packer, Energy Policy 37 (2009) 3428–3437. D.T. Whipple, P.J.A. Kenis, J. Phys. Chem. Lett. 1 (2010) 3451–3458. H.-R.M. Jhong, S. Ma, P.J.A. Kenis, Curr. Opin. Chem. Eng. 2 (2013) 191–199. T. Takeshita, K. Yamaji, Energy Policy 36 (2008) 2773–2784. Q. Lu, J. Rosen, F. Jiao, ChemCatChem 7 (2015) 38–47. R. Kortlever, J. Shen, K.J. Schouten, F. Calle-Vallejo, M.T. Koper, J Phys. Chem. Lett. 6 (2015) 4073–4082. J. Medina-Ramos, J.L. DiMeglio, J. Rosenthal, J. Am. Chem. Soc. 136 (2014) 8361–8367.

[16] F. Jin, X. Zeng, J. Liu, Y. Jin, L. Wang, H. Zhong, G. Yao, Z. Huo, Sci. Rep. 4 (2014) 4503. [17] V.S.K. Yadav, M.K. Purkait, Sol. Energy 124 (2016) 177–183. [18] J. Rosen, G.S. Hutchings, Q. Lu, R.V. Forest, A. Moore, F. Jiao, ACS Catal. 5 (2015) 4586–4591. [19] F. Quan, D. Zhong, H. Song, F. Jia, L. Zhang, J. Mater. Chem. A 3 (2015) 16409– 16413. [20] Y. Hori, K. Kikuchi, S. Suzuki, Chem. Lett. 14 (1985) 1695–1698. [21] H. Won da, C.H. Choi, J. Chung, M.W. Chung, E.H. Kim, S.I. Woo, ChemSusChem 8 (2015) 3092–3098. [22] Y. Chen, C.W. Li, M.W. Kanan, J. Am. Chem. Soc. 134 (2012) 19969–19972. [23] C.W. Li, M.W. Kanan, J. Am. Chem. Soc. 134 (2012) 7231–7234. [24] Y. Chen, M.W. Kanan, J. Am. Chem. Soc. 134 (2012) 1986–1989. [25] C.H. Lee, M.W. Kanan, ACS Catal. 5 (2015) 465–469. [26] B. Kumar, J.P. Brian, V. Atla, S. Kumari, K.A. Bertram, R.T. White, J.M. Spurgeon, Catal. Today 270 (2016) 19–30. [27] W. Zhu, Y.J. Zhang, H. Zhang, H. Lv, Q. Li, R. Michalsky, A.A. Peterson, S. Sun, J. Am. Chem. Soc. 136 (2014) 16132–16135. [28] W. Zhu, R. Michalsky, O. Metin, H. Lv, S. Guo, C.J. Wright, X. Sun, A.A. Peterson, S. Sun, J. Am. Chem. Soc. 135 (2013) 16833–16836. [29] S. Gao, Y. Lin, X. Jiao, Y. Sun, Q. Luo, W. Zhang, D. Li, J. Yang, Y. Xie, Nature 529 (2016) 68–71. [30] X. Jiang, F. Cai, D. Gao, J. Dong, S. Miao, G. Wang, X. Bao, Electrochem. Commun. 68 (2016) 67–70. [31] H. Mistry, A.S. Varela, C.S. Bonifacio, I. Zegkinoglou, I. Sinev, Y.W. Choi, K. Kisslinger, E.A. Stach, J.C. Yang, P. Strasser, B.R. Cuenya, Nat. Commun. 7 (2016) 12123. [32] V. Perumal, U. Hashim, S.C. Gopinath, R. Haarindraprasad, W.W. Liu, P. Poopalan, S.R. Balakrishnan, V. Thivina, A.R. Ruslinda, PLoS One 10 (2015) e0144964. [33] M. Chen, X. Wang, Y.H. Yu, Z.L. b, L.S. Wen, X.D. Bai, C. Sun, R.F. Huang, L.S. Wen, Appl. Surf. Sci. 158 (2000) 134–140. [34] H.W.Y. Hori, T. Tsukamoto, O. Koga, Electrochim. Acta 39 (1994) 1833–1839. [35] H. Won da, H. Shin, J. Koh, J. Chung, H.S. Lee, H. Kim, S.I. Woo, Angew. Chem. Int. Ed. Engl. 55 (2016) 9297–9300. [36] T. Zhang, H. Zhong, Y. Qiu, X. Li, H. Zhang, J. Mater. Chem. A 4 (2016) 16670– 16676. [37] V.S.K. Yadav, M.K. Purkait, New J. Chem. 39 (2015) 7348–7354. [38] M. Ma, K. Djanashvili, W.A. Smith, Phys. Chem. Chem. Phys. 17 (2015) 20861– 20867. [39] T. Pardal, S. Messias, M. Sousa, A.S.R. Machado, C.M. Rangel, D. Nunes, J.V. Pinto, R. Martins, M.N. da Ponte, J CO2 Util. 18 (2017) 62–72. [40] G. Keerthiga, R. Chetty, J. Electrochem. Soc. 164 (2017) H164–H169. [41] J. Wu, F.G. Risalvato, F.S. Ke, P.J. Pellechia, X.D. Zhou, J. Electrochem. Soc. 159 (2012) F353–F359. [42] K.J.P. Schouten, E. Pérez Gallent, M.T.M. Koper, J. Electroanal. Chem. 716 (2014) 53–57. [43] K.P. Kuhl, E.R. Cave, D.N. Abram, T.F. Jaramillo, Energy Environ. Sci. 5 (2012) 7050–7059. [44] F.-S. Ke, X.-C. Liu, J. Wu, P.P. Sharma, Z.-Y. Zhou, J. Qiao, X.-D. Zhou, Catal. Today 288 (2017) 18–23. [45] J. Rosen, G.S. Hutchings, Q. Lu, S. Rivera, Y. Zhou, D.G. Vlachos, F. Jiao, ACS Catal. 5 (2015) 4293–4299. [46] Y.-C. Hsieh, S.D. Senanayake, Y. Zhang, W. Xu, D.E. Polyansky, ACS Catal. 5 (2015) 5349–5356. [47] S. Liu, H. Tao, L. Zeng, Q. Liu, Z. Xu, Q. Liu, J.L. Luo, J. Am. Chem. Soc. 139 (2017) 2160–2163. [48] H. Wang, Z. Han, L. Zhang, C. Cui, X. Zhu, X. Liu, J. Han, Q. Ge, J CO2 Util. 15 (2016) 41–49.