Selective electrochemical CO₂ conversion to multicarbon alcohols on highly efficient N-doped porous carbon-supported Cu catalysts

Abstract

The selective and efficient electrocatalytic transformation of carbon dioxide (CO₂) to multicarbon alcohols (e.g., C₂H₅OH and C₃H₇OH) is a challenge in renewable and sustainable energy research. Herein, a series of hybrid catalysts consisting of Cu nanoparticles supported on N-doped porous carbon (Cu/NPC) were prepared. It was demonstrated that the selectivity for C₂H₅OH or C₃H₇OH could be tuned by introducing N-doped porous carbon materials as cocatalysts with different pyridinic N contents, which could in situ produce a reactive CO intermediate from CO₂. By varying the pyridinic N content, highly selective production of multicarbon alcohols was achieved using the Cu/NPC hybrid catalysts with a high faradaic efficiency for one pot production of multicarbon alcohols up to 73.3% at −1.05 V (vs. RHE). The faradaic efficiency for C₂H₅OH and C₃H₇OH was 64.6% and 8.7%, respectively. The pyridinic N species were likely the CO-producing sites and together with Cu catalytic sites acted cooperatively to produce C₂H₅OH and C₃H₇OH via a two-site mechanism for efficient CO₂ reduction to multicarbon alcohols. These findings provide novel guidance for the rational design of electrocatalysts and for tuning the catalytic activity and selectivity for multicarbon alcohol production from CO₂.

---

1. Introduction

The growing use of fossil fuels for energy leads to the accumulation of CO₂, a major greenhouse gas, which causes serious climate problems that may eventually harm human society.^1,2 As a potential remedy, CO₂ can be electrochemically converted to value-added products using renewable electricity, enabling both renewable energy storage and carbon recycling. Electrochemical CO₂ reduction can be powered by renewable energy sources and operated under ambient conditions with selective formation of products by controlling the reaction conditions.^3,4

Significant efforts have been made to develop efficient electrocatalysts that can selectively convert CO₂ in aqueous solutions over the competing hydrogen evolution reaction (HER) since the groundbreaking study by Hori et al.^5,6 Major progress in electrocatalytic CO₂ reduction has resulted in the production of gaseous carbon products including CO, CH₄, and C₂H₄, as well as liquid-phase products such as HCOO⁻.^7–10 However, the production of liquid-phase alcohols with multiple carbons (i.e., two or more carbons) is highly desired, as it may enable the synthesis of sustainable fuels that have high energy densities (21 MJ l⁻¹ for C₂H₅OH, 27 MJ l⁻¹ for C₃H₇OH) suitable for long-range and heavy freight transportation. In addition, the current market values of C₂H₅OH and C₃H₇OH make the process commercially attractive.¹¹ Unfortunately, the production of multicarbon alcohols via direct electrocatalytic CO₂ reduction remains challenging and is not economically viable due to the low catalytic activity and limited selectivity.^12,13 Therefore, improving the catalytic activity and selectivity toward products with greater economic value is an attractive technological target.

Among electrocatalysts for CO₂ reduction, Cu and Cu-based materials are the most promising for the conversion of CO₂ to multicarbon alcohols via multiple electron transfer reactions.^14–16 During CO₂ reduction, the adsorbed CO intermediate (*CO) is formed on a catalyst and is then further reduced and coupled to form multicarbon alcohols.¹⁷ However, insufficient surface coverage of *CO remains a challenge, preventing efficient and productive multicarbon alcohol generation.¹⁸ Although several methods are available to tune the reaction pathway for selective production of multicarbon alcohols, overcoming the previously mentioned limitations remains an issue. Li et al. demonstrated that Cu nanoparticles can catalyze direct CO reduction in 0.1 M KOH with faradaic efficiencies of up to 57% for the production of C₂H₅OH, CH₃COO⁻, and C₃H₇OH.¹⁹ Recently, the so-called “CO-insertion mechanism” in electrochemical CO₂ reduction reactions was proposed by several groups. For example, Ren et al. reported that C₂H₅OH can be produced using Cu_xZn catalysts, in which metallic Zn weakly binds to CO, producing an in situ source of mobile CO. The desorbed CO can subsequently diffuse to neighboring Cu sites to react with another C₁ intermediate, generating multicarbon alcohols.²⁰ Lum et al. showed that CO generated by a nearby CO-producing catalyst could be transferred to Cu, thus increasing the production of oxygenates (C₂H₅OH and C₃H₇OH).²¹ Furthermore, first-principles calculations have shown that all known products from CO₂ reduction on Cu catalysts, except formate, involve CO as the primary intermediate.^22–24 Thus, it is reasonable to expect that changes in the local CO concentration in the vicinity of the catalyst could affect product selectivity. Based on the above discussion, a feasible strategy to overcome the lack of CO solubility and insufficient CO coverage on the catalytic sites is to introduce a cocatalyst to facilitate CO formation in situ during electrocatalytic CO₂ reduction. Thus, a catalyst system can be prepared by coupling two sites, where one site can efficiently reduce CO₂ to CO, which then diffuses to a second site where it reacts to form longer carbon chain species via C–C coupling. For the CO-producing site, N-doped carbon materials have exhibited appealing catalytic performance for CO₂ reduction to CO with high selectivity. Theoretical and experimental studies revealed that pyridinic N species in N-doped carbon materials are efficient catalysts for CO production.^25–28

The design of catalysts with high activity for CO₂ reduction and high selectivity toward the formation of multicarbon alcohols could be produced using a combination of pyridinic N as a CO-producing site and Cu nanoparticles as a catalytic site where the latter could provide additional opportunities for C–C coupling between *C₁ or *C₂ intermediates and *CO. Therefore, highly dispersed Cu nanoparticles supported on N-doped porous carbon materials (Cu/NPC) were designed with different contents of pyridinic N species (e.g., Cu/NPC-700, Cu/NPC-800 and Cu/NPC-900; where the number is the pyrolysis temperature in °C for NPC synthesis). The results indicate that the selective CO₂ reduction to multicarbon alcohols (e.g., C₂H₅OH and C₃H₇OH) can be tuned by varying the pyridinic N content in the Cu/NPC hybrid catalysts. Furthermore, the introduction of N-doped porous carbon materials affected the size and electronic structure of the Cu catalyst and contributed toward efficient gas transport for improved accessibility to pyridinic N where CO₂ is adsorbed and reduced to CO. Supported by the results presented herein and previous studies, a two-site mechanism (pyridinic N as a CO-producing site and Cu as a catalytic site) was proposed to rationalize the efficient CO₂ reduction to C₂H₅OH and C₃H₇OH over the Cu/NPC hybrid catalysts. To the best of our knowledge, this is the first report of the introduction of N-doped porous carbon materials as a cocatalyst to catalyze CO formation in situ during electrocatalytic CO₂ reduction for the generation of multicarbon alcohols.

2. Experimental section

2.1 Chemicals and materials

Pyrrole (C₄H₅N, Acros Organics), ammonium persulfate ((NH₄)₂S₂O₈, Acros Organics), sodium citrate tribasic dihydrate (C₆H₅Na₃O₇·2H₂O, Sigma Aldrich), sulfuric acid (H₂SO₄, Sigma Aldrich), acetone (C₃H₆O, Sigma Aldrich), Triton X-100 (C₁₄H₂₂O(C₂H₄O)_n, Sigma Aldrich), ethylene glycol (C₂H₆O₂, JUNSEI Chemical Co.), and copper nitrate hydrate (Cu(NO₃)₂·xH₂O, Alfa Aesar) were purchased from the specified suppliers. Deionized water (DI, 18 MΩ cm, Millipore) was used for all of the experimental steps.

2.2 Catalyst preparation

Synthesis of porous carbon (PC). The preparation of PC was achieved via direct carbonization of sodium citrate tribasic dihydrate in a tube under an inert atmosphere filled with Ar at a temperature ramping rate of 3 °C min⁻¹ from room temperature to 800 °C and held at this temperature for 1 h. The resulting gray powder was then washed with 0.5 M H₂SO₄ aqueous solution to remove the residual inorganic impurities. The final product was collected by filtration and washed several times with distilled water and dried in a freezing dryer.

Synthesis of N-doped porous carbon (NPC). The N-doped porous carbon (NPC) materials were produced by coating the PC with polypyrrole via in situ polymerization of pyrrole. 40 mg of PC was dispersed in 40 mL of aqueous ethanol solution (volume ratio = 1 : 1) containing Triton X-100 (0.15 M) by using ultrasonic treatment for 0.5 h, followed by addition of the pyrrole monomer (2.2 mmol). Then 10 mL of distilled water that contained ammonium persulfate (1.3 mmol) was added dropwise into the suspension with continuous stirring for 4 h. The obtained black powder was filtered and washed with distilled water, followed by drying in a freezing dryer. Finally, polypyrrole coated PC (PPyPC) powder samples were heat-treated under an Ar atmosphere at different temperatures of 700, 800 and 900 °C for 2 h. The final products were denoted as NPC-700, NPC-800 and NPC-900, respectively.

Synthesis of copper nanoparticles on the carbon materials (Cu/PC or Cu/NPC). 40 mg of PC or NPC was suspended in 20 mL of ethylene glycol with ultra-sonication for 1 h. Then, a nominal amount (20 wt%) of Cu(NO₃)₂·xH₂O was added into the above mixture and stirred. Then, the resulting homogeneous suspension was placed in a microwave oven (KR-M202BMP, DAEWOO) at a power of 350 W for 3 min and cooled down to room temperature. The obtained black powder was washed several times using acetone and distilled water to remove the residue of ethylene glycol. Final products were then dried in a freezing dryer and denoted as Cu/PC, Cu/NPC-700, Cu/NPC-800 and Cu/NPC-900. In the meantime, the control samples with different Cu mass loadings of 10 and 30 wt% were also synthesized by adjusting the amounts of Cu(NO₃)₂·xH₂O and NPC-800 materials (denoted as Cu10/NPC-800 and Cu30/NPC-800). In addition, pure Cu nanoparticles were synthesized without adding carbon materials under the same experimental conditions.

2.3 Physicochemical characterization

The morphology and metal particle size of samples were investigated using scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images, as collected on a JEOL-JSN7500F instrument operated at 5 kV and a JEOL-JEM1011 instrument operated at 200 kV, respectively. X-ray diffraction (XRD) analysis data were collected on a Bruker-Advanced D8 diffractometer using Cu Kα (λ = 1.5418 Å) radiation operated at 40 kV and 100 mA. The Brunauer–Emmett–Teller (BET) surface area of samples was measured on a Quadrasorb SI (Quantachrome) by using N₂ adsorption/desorption isotherms at −196 °C; the pore size distribution curve was obtained from the Barrett–Joyner–Halenda (BJH) and Horvath–Kawazoe (HK) model. The CO₂ adsorption capacity was measured on a BELSORP-mini (MicrotracBEL). The content of Cu in the samples was measured by using an inductively coupled plasma atomic emission (ICP) spectrometer (Optima 7300DV). A high-performance X-ray photoelectron spectrometer (HP-XPS, VG multilab 2000) was used in order to obtain the XPS spectra of the surface elements. Fourier transform infrared (FT-IR) spectroscopy was carried out on a Frontier (PerkinElmer) with powder samples in a KBr pellet form. The XANES spectra of the Cu K-edge were obtained on the 7D beamline of the Pohang Accelerator Laboratory (PLS-II), in which ATHENA software was used for the XANES analysis. CO temperature-programmed desorption (CO-TPD) experiments were performed at 40 °C by injecting a gas mixture of 4% CO/He, and the desorbed CO was detected by using a mass spectrometer while the temperature was increased from 40 to 240 °C at a heating rate of 5 °C min⁻¹. Before the CO-TPD process, the samples were pre-treated at 200 °C with a gas flow of He at 30 mL min⁻¹ for 2 h. CO₂ temperature-programmed desorption (CO₂-TPD) experiments were carried out by injecting CO₂ gas for 1 h at 50 °C, followed by a He rinsing step for 1 h. The desorbed CO₂ was analyzed with a TCD detector while the temperature was increased from 50 to 300 °C at a heating rate of 10 °C min⁻¹. Before the CO₂-TPD experiment, the samples were pre-treated at 300 °C with a gas flow of He at 30 mL min⁻¹ for 1 h.

2.4 Electrochemical measurements

The electrochemical measurements were performed in an electrochemical reactor (H-type cell), composed of three electrodes, which were separated by Nafion 212 into two compartments (Fig. S1, ESI†). A graphite plate (geometric size, 1 cm × 1.5 cm) was used as the working electrode. The catalyst loading in the working electrode was 2 mg cm⁻². Pt coil and an Ag/AgCl electrode were used as the counter and reference electrodes, respectively. The potential was converted to a RHE using the following equation:

E (vs. RHE) = E (vs. Ag/AgCl) + 0.209 V + 0.0591 × pH

Linear sweep voltammograms (LSVs) were recorded in CO₂-saturated 0.2 M KHCO₃ aqueous solution at a scan rate of 20 mV s⁻¹. Electrochemical reduction of CO₂ was performed in 0.2 M KHCO₃ aqueous solution, which was stirred at a constant rate of 200 rpm, and CO₂ was purged for at least 0.5 h before the experiments. During the electrolysis, CO₂ was purged continuously into the cathodic compartment at a rate of 20 mL min⁻¹. The IR drop was compensated using the current interrupt mode. The gas and liquid products from the electrochemical CO₂ reduction were detected and analyzed by using a gas chromatograph (GC, Agilent 6890) and a high performance liquid chromatograph (HPLC, Waters e2695), respectively. Electrochemical impedance spectroscopy (EIS) was performed at −1.05 V (vs. RHE) in CO₂-saturated 0.2 M KHCO₃ aqueous solution.

3. Results and discussion

The carbon atoms in sodium citrate are crucial for the formation of carbon frameworks, while the sodium species contribute to the generation of pores. This template-free one-step process for the fabrication of porous carbon materials involves a simple pyrolysis process without several complex steps that use large amounts of corrosive chemical agents (i.e. KOH, NaOH, ZnCl₂, and phosphoric acid).²⁹ The prepared porous carbon (PC) was used as a precursor for the fabrication of N-doped porous carbon hybrid catalysts with deposition of Cu nanoparticles. To obtain various N-doped porous carbons, polypyrrole-coated porous carbon (PPyPC) materials were heat treated at 700 °C, 800 °C, and 900 °C, denoted as NPC-700, NPC-800, and NPC-900, respectively. Herein, polypyrrole was selected as the N source due to its facile and low-cost synthesis, as well as sufficient nitrogen content.³⁰ Finally, the N-doped porous carbons were impregnated with Cu nanoparticles (Cu/NPC-700, -800, and -900) via microwave-assisted polyol reduction.³¹ A schematic of the Cu/NPC hybrid catalyst fabrication is shown in Scheme 1. The physicochemical characteristics of the prepared Cu/NPC catalysts and their application in electrochemical CO₂ reduction will be discussed in detail.

Scheme 1 Schematic illustration of the overall fabrication process for Cu/NPC hybrid catalysts.

The microstructure and surface morphology of the prepared samples were investigated via SEM analysis. The PC exhibited a well-developed hierarchical porous structure with different types of open pores in multidirectional channels (Fig. 1a). Because of the polymerization process, polypyrrole coated PC (PPyPC) exhibited a relatively smooth surface without the porous structure of PC (Fig. S2, ESI†), indicating that the polypyrrole was preferentially formed as a uniform layer on the PC. After pyrolysis at different temperatures (700 °C, 800 °C, and 900 °C), the NPC materials maintained their original PC structures without notable changes (Fig. 1b–d), containing a myriad of macropores 50–500 nm in size. To further investigate the pore structure of the prepared carbon materials, N₂ adsorption/desorption isotherms of the four samples (PC, NPC-700, NPC-800, and NPC-900) were analyzed (Fig. S3, ESI†). The qualitative behavior of the isotherms clearly illustrated that all carbon materials had a similar pore structure (Fig. S3a, ESI†). Moreover, the relevant pore size analysis in Fig. S3b (ESI†) showed that all samples exhibited a uniform mesopore distribution, particularly from 2–20 nm, according to the Barrett–Joyner–Halenda (BJH) characterization. In addition, analysis of the N₂ desorption isotherm by the Horvath–Kawazoe (HK) model showed uniformly distributed micropores with sizes centered at approximately 0.4 nm (Fig. S3c, ESI†). The textural properties are summarized in Table S1 (ESI†) with BET surface areas and porosity. It has been widely reported that the porous structure in electrodes plays an important role in the overall CO₂ reduction reaction.^32–34 Therefore, the well-developed hierarchical porous structure should provide efficient gas transportation channels for improved accessibility to the N-doped sites.^33,35

Fig. 1 SEM image of (a) PC, (b) NPC-700, (c) NPC-800 and (d) NPC-900; the inset shows the high-magnitude SEM image.

The nature of the N species on the prepared sample surfaces was analyzed by XPS measurements. From the peak deconvolution of XPS N 1s spectra shown in Fig. S4 (ESI†), PPyPC showed a major peak at 400.0 eV, which was attributed to a neutral pyrrolic N (–N–), and a minor peak at 400.8 eV corresponding to a positively charged nitrogen (–N⁺–).³⁶ During pyrolysis, PPyPC was gradually transformed to N-doped carbonaceous materials, whereas most pyrrolic N in PPyPC was converted into various N species, yielding the N-doped carbon compounds. Thus, the NPC-700, NPC-800, and NPC-900 samples showed two major peaks at approximately 398.3 and 401.1 eV, which were assigned to pyridinic N and graphitic N, respectively, and minor peaks at 400.2 and 403.2 eV which were ascribed to pyrrolic N and oxidized N, respectively (Fig. 2a).³⁷ The total N contents with different N species in the NPC materials are summarized in Fig. 2b, suggesting that the pyrolysis temperature is highly correlated with the transformation behavior of the N species. With the increase of the pyrolysis temperature from 700 to 800 °C, the content of pyrrolic N is decreased whereas the contents of pyridinic N and graphitic N are increased due to the reconstruction of the N-doped carbon structure along with transformations of the N species. However, both the pyrrolic N and pyridinic N contents showed a decrease at temperatures greater than 800 °C with a simultaneous increase of the graphitic N content (Fig. 2b), which can be ascribed to the fact that the pyrrolic N and pyridinic N are not thermally stable compared to the graphitic N.^38,39 In addition, with increasing pyrolysis temperature from 700 to 900 °C, the total N content in the NPC materials decreased from 11.2 to 9.6 at%, which was attributed to the removal of relatively unstable C–N bonds.³⁵ The chemical features of PC, PPyPC, and NPC-800 were further investigated using FTIR spectroscopy, as shown in Fig. S5 (ESI†), to examine the functional groups at characteristic wavenumbers.⁴⁰ PPyPC showed distinct peaks at 1209 and 926 cm⁻¹, assigned to the stretching vibrations of C=N⁺–C and C–N⁺, respectively, and a peak at 1040 cm⁻¹ corresponding to the in-plane C–H vibrations. Notably, the successful incorporation of N atoms into the PC can be monitored by the disappearance of the characteristic peaks of polypyrrole in NPC-800. The PC exhibited a sharp peak at 1615 cm⁻¹ corresponding to the C=C group, and NPC-800 showed the same peaks for C=C/C=N groups at 1625 cm⁻¹. From Fig. 2c, the CO₂ uptake at 25 °C of the prepared samples followed the order: NPC-800 (2.26 mmol g⁻¹) > NPC-700 (1.62 mmol g⁻¹) > NPC-900 (1.20 mmol g⁻¹) > PC (0.45 mmol g⁻¹). In terms of the CO₂ adsorption capacity, micropores and pyridinic N are considered to be important factors.^41,42 However, because the micropore volume of all carbon materials was similar for the materials prepared herein, the difference in CO₂ adsorption capacity was attributed to the pyridinic N content. Therefore, the PC without pyridinic N showed a lower CO₂ adsorption capacity than the NPC materials (Table S2, ESI†). In addition, the hierarchical porosities likely facilitated gas transportation for improved accessibility to the pyridinic N site where CO₂ is adsorbed.^33,43

Fig. 2 (a) High resolution N 1s spectra of NPC-700, NPC-800 and NPC-900. (b) Summary of N atomic contents and relative concentrations. (c) CO₂ uptake isotherms at 25 °C.

TEM images with size distribution histograms of the Cu nanoparticles deposited on the PC and NPC materials are shown in Fig. 3. The three Cu/NPC hybrid catalysts exhibited significantly smaller particle sizes than Cu/PC. From the size distribution histograms, the average Cu particle sizes of Cu/PC, Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900 were approximately 12.1, 5.3, 5.2, and 5.4 nm, respectively, as estimated using 200 randomly selected particles from the TEM images. Despite using the same synthesis method, the difference in the Cu nanoparticle size between the PC and NPC materials arises from the N-doped sites in the NPC materials that facilitate the nucleation/growth of Cu.^44,45 The crystallographic data of the prepared samples were obtained via XRD, as shown in Fig. 4. The broad width and weak intensity peak centered at approximately 23.5° for the (002) plane of graphitic carbons (Fig. 4a) suggests a low degree of graphitization.³⁵ The presence of metallic Cu on the NPC materials (Cu/PC, Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900) was confirmed by the three strong Cu diffraction peaks at approximately 43°, 50°, and 74°, corresponding to the (111), (200) and (220) planes of the fcc crystal structure of metallic Cu. Meanwhile, the full width at half maximum (FWHM) of the Cu peaks of the Cu/NPC materials was relatively smaller than that observed for the Cu/PC sample (Fig. 4b), indicating smaller Cu nanoparticles with a higher degree of dispersion on the NPC materials.⁴⁴ The actual amounts of Cu on the PC and NPC materials were determined quantitatively using ICP spectrometry. The nominal Cu metal loading was fixed at 20 wt% and the actual Cu contents in the Cu/PC, Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900 samples were ca. 21.6, 19.7, 19.8, and 20.2 wt%, respectively.

Fig. 3 TEM images with particle size distribution of (a) Cu/PC, (b) Cu/NPC-700 (c) Cu/NPC-800 and (d) Cu/NPC-900 hybrid catalysts.

Fig. 4 XRD patterns of (a) PC, NPC-700, NPC-800 and NPC-900 and (b) Cu/PC and Cu/NPC-700, Cu/NPC-800 and Cu/NPC-900 hybrid catalysts.

Fig. S6a–c (ESI†) show the peak deconvolution of XPS N 1s spectra of Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900. Four types of N species were also detected: pyridinic, pyrrolic, graphitic, and oxidized N. In addition, even after the Cu nanoparticles were decorated on the NPC material, the pyridinic N content was still the highest for the Cu/NPC-800 sample (Fig. S6d and Table S3, ESI†). However, the Cu nanoparticles on the NPC materials induced changes in the N 1s spectra. The significant decrease in the relative graphitic N content suggested that most of the Cu nanoparticles were anchored on graphitic N sites. This is consistent with the results that showed that the NPC materials promoted the formation of Cu nanoparticles with smaller sizes and better dispersity due to their graphitic N species content with a relatively high electron density.^46–48 The slight increase in the relative percentage of pyridinic, pyrrolic, and oxidized N suggests that these N species are not associated with Cu nucleation and growth. The Cu 2p_3/2 XPS spectra of Cu/PC, Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900 are shown in Fig. 5a. In all samples, the Cu 2p_3/2 spectra were fitted with a major peak at 932.4 ± 0.3 eV associated with (Cu⁰ + Cu⁺).⁴⁹ Compared to that of Cu/PC, the binding energies of Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900 shifted to higher binding energies by 0.4–0.5 eV. This shift toward higher energy indicated that more electron transfer from the Cu metal to NPC materials occurred, which further strengthened the electronic interaction.⁴⁴ The Cu average particle sizes, Cu contents, and analytical results from the Cu 2p_3/2 spectra of all the prepared catalysts are summarized in Table S4 (ESI†).

Fig. 5 (a) High resolution Cu 2p_3/2 spectra of Cu/PC, Cu/NPC-700, Cu/NPC-800 and Cu/NPC-900 hybrid catalysts. (b) Cu K edge XANES spectra of reference samples, Cu/PC and Cu/NPC hybrid catalysts with (c) comparison in detail.

Because XANES is sensitive to the chemical valence state of the probed atoms, detailed local structural properties of the prepared samples were further investigated at the Cu K-edge. The normalized XANES spectra of commercial Cu₂O, CuO, Cu metal foil, and the prepared hybrid catalysts are shown in Fig. 5b. It is clear that the Cu K-edge XANES spectra of the Cu/NPC hybrid catalysts were similar to those of the Cu foil, indicating that the Cu nanoparticles on NPC-700, NPC-800, and NPC-900 mainly exist in a metallic state. The strong peak above the edge energy position (Fig. 5c) is closely related to the empty d states of Cu and the white line (WL) intensity can be used to measure changes in the local electronic structure of Cu.⁵⁰ Compared to the WL intensity of Cu/PC, the increased intensity of Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900 was attributed to increased density of empty Cu d states due to electron transfer to the NPC materials.⁵⁰ This phenomenon can be explained based on the previous XPS analysis results. The strong electronic interaction involving electron transfer between the Cu nanoparticles and NPC materials changes the electronic structure of Cu, which influenced the CO adsorption and C–C coupling during electrocatalytic CO₂ reduction.⁵¹ Additionally, to quantify the relative composition of Cu species in the Cu/PC and Cu/NPC hybrid catalysts, linear combination fitting with the XANES spectra was performed (Fig. S7 and Table S5, ESI†). All samples contained mainly metallic Cu rather than the oxidized Cu⁺ or Cu²⁺. This result agrees with the XRD analysis, where Cu⁰ was identified and no oxide species were detected. However, the proportions of Cu⁺ and Cu²⁺ in the Cu/NPC hybrid catalysts were relatively higher than those of Cu/PC, suggesting that the smaller metal nanoparticles were more easily oxidized on the surface of the catalysts than the larger nanoparticles. In addition, the relative compositions of Cu species were similar in all the Cu/NPC hybrid catalysts.

As shown in Fig. 6, electrochemical CO₂ reduction over the prepared samples was first estimated by LSV measurements in a CO₂-saturated 0.2 M KHCO₃ aqueous solution at a scan rate of 20 mV s⁻¹. It should be noted that the onset potentials of the Cu/NPC hybrid catalysts (Cu/NPC-700, Cu/NPC-800, and Cu/NPC-900) were more positive than those of the other samples. In addition, the Cu/NPC hybrid catalysts exhibited the largest reductive current densities over the entire potential range, suggesting that the integration of Cu nanoparticles and N-doped porous carbon materials facilitated the efficient reduction of CO₂ in the electrolyte solution. Chronoamperometry measurements were performed from −0.6 to −1.1 V (vs. RHE) in CO₂-saturated 0.2 M KHCO₃ aqueous solution using a H-type cell to further analyze the CO₂ electroreduction products. The faradaic efficiencies for various liquid/gas products as a function of applied potential on the PC, NPC-800, Cu/PC, and Cu/NPC-800 samples are shown in Fig. 7. The PC alone showed almost no catalytic activity for CO₂ reduction and instead the hydrogen evolution reaction (HER) was dominant because of incapability for catalyzing CO₂ electroreduction (Fig. 7a and Table S6, ESI†). For NPC-800, CO was observed as the major gas product and the faradaic efficiency of CO formation increased from −0.6 to −1.0 V (vs. RHE), reaching a maximum (61.8%) at −1.0 V (Fig. 7b and Table S6, ESI†). The same trend was observed for the NPC-700 and NPC-900 catalysts (Fig. S8 and Table S6, ESI†). As previously reported, pyridinic N is the most preferable and selective site for the CO production from CO₂ as it exhibits relatively weak binding energy for *CO and facilitates CO desorption.^25–28,52 As shown in Fig. S9 (ESI†), the change in CO₂ adsorption capacity as a function of pyridinic N content matched with the maximum faradaic efficiency for CO. In contrast to the NPC materials, Cu/PC showed a very broad product distribution comprising C₂H₄ and C₂H₅OH. In the applied potential range, Cu/PC produced C₂H₄ with a maximum faradaic efficiency of 30.2% at −1.0 V (vs. RHE) and C₂H₅OH with a maximum faradaic efficiency of 14.3% at −0.9 V (vs. RHE). Furthermore, trace C₃H₇OH was detected (Fig. 7c and Table S7, ESI†). Interestingly, all three Cu/NPC hybrid catalysts showed enhanced selectivities toward multicarbon alcohols (C₂H₅OH and C₃H₇OH) compared to Cu/PC (Fig. 7d, Fig. S10 and Table S7, ESI†). The maximum faradaic efficiencies for C₂H₅OH and C₃H₇OH were 64.6% and 8.7%, respectively, at −1.05 V (vs. RHE) for Cu/NPC-800. The partial current densities for the production of C₂H₅OH and C₃H₇OH (j_C2H5OH and j_C3H7OH) were analyzed to compare the performances of the Cu/PC and Cu/NPC hybrid catalysts (Fig. 8a and b). The observed trend was similar to that described for the faradaic efficiencies of C₂H₅OH and C₃H₇OH using the same catalysts. The values of j_C2H5OH and j_C3H7OH clearly indicate the superior catalytic activity of Cu/NPC-800 for multicarbon alcohol production over the entire potential range. In addition, among the Cu/NPC hybrid catalysts, Cu/NPC-800 showed the highest total current density at −1.05 V (vs. RHE), indicating the maximum faradaic efficiency for multicarbon alcohol generation (Fig. 8c). In Table S8 (ESI†), this novel class of hybrid catalysts, which is selective for multicarbon alcohols (C₂H₅OH and C₃H₇OH) over other C2–C3 products, was compared with the previously reported Cu-based catalysts.

Fig. 6 LSV curves of PC, NPC materials and Cu/NPC hybrid catalysts in CO₂-saturated 0.2 M KHCO₃ aqueous solution.

Fig. 7 Faradaic efficiencies of liquid and gaseous products for CO₂ reduction in CO₂-saturated 0.2 M KHCO₃ aqueous solution on (a) PC, (b) NPC-800, (c) Cu/PC, and (d) Cu/NPC-800 hybrid catalysts.

Fig. 8 Partial current densities of (a) C₂H₅OH and (b) C₃H₇OH production on Cu/NPC hybrid catalysts. (c) Total current densities measured at −1.05 V (vs. RHE). (d) Partial current densities of CO on NPC materials and Cu/NPC hybrid catalysts. (e) Proposed two-site mechanism (via CO-insertion route) scheme for electrocatalytic CO₂ reduction to multicarbon alcohols on Cu/NPC hybrid catalysts, indicating the CO-producing site (pyridinic N) and catalytic site (Cu nanoparticle).

The calculated faradaic efficiencies of the major reaction products over Cu/PC and Cu/NPC hybrid catalysts showed interesting selectivity for multicarbon alcohols. In the low negative potential region, CO production was significantly increased over the Cu/NPC hybrid catalysts. The faradaic efficiency of CO decreased monotonically with increasingly negative applied potential, whereas C₂H₅OH and C₃H₇OH production began and increased in the potential region, indicating that the *CO intermediate was intimately connected to C₂H₅OH and C₃H₇OH formation.^20,53,54 The aforementioned trends in C₂H₅OH and C₃H₇OH formation over the Cu/NPC hybrid catalysts suggest that the pyridinic N and Cu nanoparticles acted cooperatively to promote C₂H₅OH and C₃H₇OH formation. This cooperative catalysis was demonstrated by analyzing the faradaic efficiency of CO over these catalysts (Fig. 7d and Fig. S10a, b, ESI†). Otherwise, it could be anticipated that increasing faradaic efficiency of CO on the Cu/NPC hybrid catalysts as larger negative potentials were applied would be achieved due to the production of CO from pyridinic N (Fig. 7d and Fig. S10a, b, ESI†). Therefore, it was assumed that gaseous CO formed at the pyridinic N sites may further react to yield C₂H₅OH and C₃H₇OH. This speculation is corroborated by the negative shift in maximum faradaic efficiencies of C₂H₅OH and C₃H₇OH production by 150 and 50 mV over the Cu/NPC hybrid catalysts compared to those achieved using Cu/PC. Importantly, most of the CO produced at the pyridinic N sites of the NPC materials diffused toward Cu sites for further reaction, as indicated by the low CO partial current densities (Fig. 8d).

To understand the key factors influencing the catalytic performance for CO₂ reduction, the analytical and experimental results of prepared samples were compared (Tables S1–S5, ESI†). It was found that textural properties of NPC materials were similar to each other. Furthermore, the effects of the particle size, loading amount, composition and electronic properties of Cu nanoparticles were also negligible because they were similar for all Cu/NPC hybrid catalysts. In contrast, the pyridinic N content in the Cu/NPC hybrid catalysts has a critical effect on producing multicarbon alcohols, as shown by the fact that plenty of pyridinic N species in Cu/NPC-800 hybrid catalysts could directly contribute to the improvement of the catalytic activity and selectivity toward production of multicarbon alcohols from CO₂ reduction. In addition, the Cu catalysts supported on different carbon materials might explain why the selective production of multicarbon alcohols was enhanced on the Cu/NPC hybrid catalysts. It seems that the introduction of N-doped carbon materials is an important factor for the selective production of multicarbon alcohols because they not only affect the size and electronic structure of the Cu, but also promote the adsorption of CO₂ as well as CO production. The different catalytic performances of Cu/PC and Cu/NPC hybrid catalysts could be further supported through the EIS measurements. As shown in Fig. S11 (ESI†), the Nyquist plots revealed that the Cu/NPC-800 exhibited a smaller semicircle radius compared with that of Cu/PC, indicating accelerated charge transfer. This is associated with a lower activation energy⁵⁵ for the CO₂ reduction reaction on Cu/NPC-800, as determined by LSV measurements. These results suggest that the Cu/NPC-800 hybrid catalysts were highly active for CO₂ reduction due to the introduction of N-doped porous carbon materials.

Moreover, previous studies showed that the differences in CO binding energy are responsible for generating different products.^56,57 To investigate CO adsorption capabilities, CO-TPD measurement and voltammetric experiments were performed and the results are shown in Fig. S12 (ESI†). To clearly identify the desorbed CO signal, the mass-to-charge ratio (m/z) was set to 28 to detect the desorbed CO signal. The CO desorption signal appeared at the higher temperature for all the Cu/NPC hybrid catalysts as compared to the Cu/PC, suggesting the higher binding energies of CO over the Cu/NPC hybrid catalysts (Fig. S12a, ESI†). Furthermore, to investigate the CO adsorption strength under actual experimental conditions, voltammetric measurements were performed in a CO-saturated 0.2 M KHCO₃ aqueous solution with an applied potential of −0.12 V (vs. RHE) for 120 s (Fig. S12b, ESI†). This applied potential was chosen for a maximum amount of adsorbed CO before the occurrence of any electrochemical reactions. With the addition of N₂ gas after 120 s, the cathodic current density decrease was much slower for Cu/NPC-800 than that for Cu/PC. The CO-TPD and voltammetric results showed that the enhanced CO adsorption strength was caused by changes in the electronic properties of Cu,⁵¹ as indicated by the XPS and XANES spectra. In addition, the larger surface area of Cu nanoparticles with ∼5 nm sizes could lead to a higher chance of offering optimized geometries for CO adsorption.¹⁸ Thus, from these results, we could assume that change in CO adsorption capability on the Cu/NPC hybrid catalysts might tune the energetics of intermediate binding at the Cu site, affecting selectivity toward multicarbon alcohols.^54,58,59 This might explain why the production of multicarbon alcohols was favored over C₂H₄ on all Cu/NPC hybrid catalysts, whereas C₂H₄ was selectively produced over multicarbon alcohols on Cu/PC, although more information regarding the selectivity determining steps is necessary for mechanistic determination. To examine the adsorption behavior of CO₂ over Cu/NPC-700, Cu/NPC-800 and Cu/NPC-900, the CO₂-TPD measurements were performed as shown in Fig. S13a (ESI†). The first desorption peak could originate from physisorbed/chemisorbed CO₂, whereas the second peak was related to strongly chemisorbed CO₂.⁶⁰ The CO₂-TPD peak areas for all Cu/NPC hybrid catalysts followed the order: Cu/NPC-800 > Cu/NPC-700 > Cu/NPC-900 (Fig. S13b, ESI†). These results suggested a correlation between the CO₂ adsorption capacity and content of pyridinic N (Table S3 and Fig. S6d, S13, ESI†).

Generally, C₂H₄ is a favorable product of Cu-based catalysts over C₂H₅OH and C₃H₇OH via CO₂ reduction.^12,13 However, the product selectivity for multicarbon alcohols with a higher CO concentration was observed with other catalysts,^20,21,53 suggesting that a CO atmosphere is crucial for controlling product selectivity. Therefore, to assess the role of the CO concentration, the direct CO electroreduction on Cu nanoparticles (Cu NPs), Cu/PC, and Cu/NPC hybrid catalysts was performed at −1.05 V (vs. RHE) in 0.2 M KHCO₃ aqueous solution (Fig. S14, ESI†). As a control, Cu NPs prepared using the same synthetic method without carbon supports were prepared, for which the morphology and crystallographic information are provided in Fig. S15 (ESI†). It should be noted that the relative ratio of faradaic efficiencies of multicarbon alcohols to C₂H₄ significantly increased over the Cu nanoparticles and Cu/PC, whereas the Cu/NPC hybrid catalysts did not show substantial changes compared to CO₂ reduction (Fig. S14, ESI†). These results indicated that the favorable product could be changed from C₂H₄ to multicarbon alcohols by tuning the CO concentration, indicating that pyridinic N as the CO-producing site significantly affected the production of C₂H₅OH and C₃H₇OH during electrocatalytic CO₂ reduction. Thus, the CO population in accordance with the pyridinic N content of the Cu/NPC hybrid catalysts is a crucial factor for the selectivity toward multicarbon alcohols. Notably, the Cu/NPC hybrid catalysts showed a similar selectivity trend for C₂H₅OH and C₃H₇OH as a function of applied potential (Table S7, ESI†), suggesting that these alcohols may share common intermediates along their reaction pathways.⁵⁴ Based on the results presented herein and the recent literature, a two-site mechanism was proposed to rationalize the action of the Cu nanoparticles and pyridinic N species in N-doped porous carbon for the electrocatalysis of CO₂ to form multicarbon alcohols (Fig. 8e). The incoming CO₂ molecules bind to pyridinic N sites and are reduced to CO,^25–28 which adsorbs weakly at the pyridinic N sites, where it can later desorb. These steps are in agreement with product analysis (Fig. 7b, S8, and S9, ESI†). The desorbed CO from the pyridinic N sites subsequently diffuses toward the Cu nanoparticles, adsorbing onto the Cu surface where it can react with *C₁ or *C₂ intermediates.^20,61–67 These intermediates can be further converted to either C₂H₅OH or C₃H₇OH via C–C coupling. In this mechanism, the cooperative catalytic effect of the Cu nanoparticles and pyridinic N species is crucial for the production of C₂H₅OH and C₃H₇OH. As summarized in Fig. 9, it clearly demonstrates that the pyridinic N facilitates the transformation of CO₂ to CO near the catalytically active Cu site, changing the product selectivity toward oxygenated products.^53,64,68,69 This also explains why the faradaic efficiencies of multicarbon alcohols (C₂H₅OH and C₃H₇OH) increased while the faradaic efficiencies of H₂ and C₂H₄ decreased as a function of pyridinic N species content. Furthermore, the larger surface area of the small Cu nanoparticles (∼5 nm) with modified electronic properties enhanced the adsorption of CO, promoting C–C coupling (between *C₁ or *C₂ intermediates and *CO) for multicarbon alcohol production.

Fig. 9 Schematic of the relationship between the contents of pyridinic N in Cu/NPC hybrid catalysts and the corresponding faradaic efficiencies of multicarbon alcohols, C₂H₄ and H₂ in CO₂-saturated 0.2 M KHCO₃ aqueous solution at −1.05 V (vs. RHE).

The cooperative catalytic effect was further examined for CO₂ reduction using different Cu metal loadings of 10 and 30 wt% on the Cu/NPC-800 hybrid catalysts. The associated physicochemical characterization is provided in Fig. S16 and S17 (ESI†) and summarized in Table S9 (ESI†). Compared to Cu/NPC-800 with 20 wt% Cu loading, these catalysts exhibited smaller faradaic efficiencies toward multicarbon alcohols at −1.05 V (vs. RHE). For Cu10/NPC-800, the amount of Cu catalyst seems to be insufficient, but the catalytic activity toward CO formation is significantly increased with a larger pyridinic N species content in NPC-800 (Fig. S18a and Table S10, ESI†). In contrast, Cu30/NPC-800 showed increased faradaic efficiency for C₂H₄, but decreased efficiencies for C₂H₅OH and C₃H₇OH (Fig. S18b and Table S10, ESI†). This is likely due to overloading of the Cu nanoparticles. To maximize the cooperative catalytic effect, it is necessary for the Cu nanoparticles to stay close to or around the pyridinic N sites in the N-doped porous carbon structure. Thus, excessive loading of Cu may restrict the CO-producing ability of pyridinic N sites and attenuate the cooperative catalysis. Therefore, effective exposure of pyridinic N to the reactant and proper spacing with Cu nanoparticles are crucial for the CO₂ reduction catalysis to form multicarbon alcohols.

Another crucial factor for the design of an ideal CO₂ reduction catalyst is stability. Therefore, the stability test of Cu/NPC-800 hybrid catalysts was investigated at a constant applied potential of −1.05 V (vs. RHE) in CO₂-saturated 0.2 M KHCO₃ aqueous solution. As shown in Fig. 10, the Cu/NPC-800 revealed a quite stable performance in current density during 10 h. After the stability test, the catalyst was re-used in another CO₂ reduction test in a fresh 0.2 M KHCO₃ electrolyte under the same applied potential to estimate the faradaic efficiency for C₂H₅OH and C₃H₇OH. Total faradaic efficiencies of multicarbon alcohols were shown up to 71.1%, which is comparable to the initial value of 73.3% that was measured before the stability test. This result implied that the Cu/NPC-800 hybrid catalysts are relatively stable during the CO₂ reduction reaction without significant performance loss.

Fig. 10 Chronoamperometric measurements at a constant potential of −1.05 V (vs. RHE) in CO₂-saturated 0.2 M KHCO₃ aqueous solution.

4. Conclusions

In summary, a series of Cu/NPC hybrid catalysts were prepared with different pyridinic N contents for the selective production of multicarbon alcohols via electrocatalytic CO₂ reduction. The production of total multicarbon alcohols was maximized for the Cu/NPC-800 hybrid catalysts with 20 wt% Cu loading at −1.05 V (vs. RHE), achieving a total faradaic efficiency of 73.3% for C₂H₅OH and C₃H₇OH. The Cu/NPC-800 hybrid catalysts were also catalytically stable for the production of C₂H₅OH and C₃H₇OH for 10 h. We discovered that the selective production of multicarbon alcohols on the Cu/NPC hybrid catalysts was significantly influenced by the pyridinic N content in the N-doped porous carbon materials. High contents of pyridinic N facilitated CO₂ adsorption, and thus produced CO which efficiently diffused to neighboring Cu sites. Furthermore, suitable sized Cu nanoparticles with a modified electronic structure effectively catalyzed the production of *C₁ or *C₂ intermediates which subsequently reacted with CO. The strategy developed herein for cooperative catalysis between Cu nanoparticles and pyridinic N provides novel insights toward achieving the selective production of multicarbon alcohols via electroreduction of CO₂.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported by Basic Science Research Program (NRF-2019R1A2C2088174) through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT, and by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (20172010106300), and also by the Fundamental Research Program of the Korean Institute of Materials Science (Grant PNK6130).

References

1. Q. Lu, J. Rosen, Y. Zhou, G. S. Hutchings, Y. C. Kimmel, J. G. Chen and F. Jiao, Nat. Commun., 2014, 5, 3242–3248.
2. E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja, Chem. Soc. Rev., 2009, 38, 89–99.
3. C. Costentin, M. Robert and J. M. Saveant, Chem. Soc. Rev., 2013, 42, 2423–2436.
4. Z.-L. Wang, C. Li and Y. Yamauchi, Nano Today, 2016, 11, 373–391.
5. Y. Hori, Bull. Chem. Soc. Jpn., 1982, 55, 660–665.
6. Y. Hori, K. Kikuchi, A. Murata and S. Shin, Chem. Lett., 1986, 15, 897–898.
7. L. Jin, B. Liu, P. Wang, H. Yao, L. A. Achola, P. Kerns, A. Lopes, Y. Yang, J. Ho, A. Moewes, Y. Pei and J. He, Nanoscale, 2018, 10, 14678–14686.
8. Z. Weng, J. Jiang, Y. Wu, Z. Wu, X. Guo, K. L. Materna, W. Liu, V. S. Batista, G. W. Brudvig and H. Wang, J. Am. Chem. Soc., 2016, 138, 8076–8079.
9. S. Gao, Y. Lin, X. Jiao, Y. Sun, Q. Luo, W. Zhang, D. Li, J. Yang and Y. Xie, Nature, 2016, 529, 68–71.
10. X. Zheng, J. Han, Y. Fu, Y. Deng, Y. Liu, Y. Yang, T. Wang and L. Zhang, Nano Energy, 2018, 48, 93–100.
11. K. J. P. Schouten, F. Calle-Vallejo and M. Koper, Angew. Chem., Int. Ed., 2014, 53, 10858–10860.
12. C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 7231–7234.
13. A. Loiudice, P. Lobaccaro, E. A. Kamali, T. Thao, B. H. Huang, J. W. Ager and R. Buonsanti, Angew. Chem., Int. Ed., 2016, 55, 5789–5792.
14. W. Zhu, K. Zhao, S. Liu, M. Liu, F. Peng, P. An, B. Qin, H. Zhou, H. Li and Z. He, J. Energy Chem., 2019, 37, 176–182.
15. S. Shen, X. Peng, L. Song, Y. Qiu, C. Li, L. Zhou, J. He, J. Ren, X. Liu and J. Luo, Small, 2019, 15, 1902229.
16. D. Wakerley, S. Lamaison, F. Ozanam, N. Menguy, D. Mercier, P. Marcus, M. Fontecave and V. Mougel, Nat. Mater., 2019 DOI:10.1038/s41563-019-0445-x.
17. Z. Gu, H. Shen, L. Shang, X. Lv, L. Quan and G. Zheng, Small Methods, 2018, 2, 1800121.
18. J. Li, F. Che, Y. Pang, C. Zou, J. Y. Howe, T. Burdyny, J. P. Edwards, Y. Wang, F. Li, Z. Wang, P. D. Luna, C.-T. Dinh, T.-T. Zhuang, M. I. Saidaminov, S. Cheng, T. Wu, Z. Finfrock, L. Ma, S.-H. Hsieh, Y.-S. Liu, G. A. Botton, W.-F. Pong, X. Du, J. Guo, T.-K. Sham, E. H. Sargent and D. Sinton, Nat. Commun., 2018, 9, 4614.
19. C. W. Li, J. Ciston and M. W. Kanan, Nature, 2014, 508, 504–507.
20. D. Ren, B. S.-H. Ang and B. S. Yeo, ACS Catal., 2016, 6, 8239–8247.
21. Y. Lum and J. W. Ager, Energy Environ. Sci., 2018, 11, 2935–2944.
22. A. A. Peterson, F. Abild-Pedersen, F. Studt, J. Rossmeisl and J. K. Nørskov, Energy Environ. Sci., 2010, 3, 1311–1315.
23. H. Xiao, T. Cheng, W. A. Goddard and R. Sundararaman, J. Am. Chem. Soc., 2016, 138, 483–486.
24. W. Luo, X. Nie, M. J. Janik and A. Asthagiri, ACS Catal., 2016, 6, 219–229.
25. A. Vasileff, Y. Zheng and S. Z. Qiao, Adv. Energy Mater., 2017, 7, 1700759.
26. J. Wu, M. Liu, P. P. Sharma, R. M. Yadav, L. Ma, Y. Yang, X. Zou, X.-D. Zhou, R. Vajtai, B. I. Yakobson, J. Lou and P. M. Ajayan, Nano Lett., 2016, 16, 466–470.
27. J. Wu, R. M. Yadav, M. Liu, P. P. Sharma, C. S. Tiwary, L. Ma, X. Zou, X.-D. Zhou, B. I. Yakobson, J. Lou and P. M. Ajayan, ACS Nano, 2015, 9, 5364–5371.
28. S. Liu, H. Yang, X. Huang, L. Liu, W. Cai, J. Gao, X. Li, T. Zhang, Y. Huang and B. Liu, Adv. Funct. Mater., 2018, 28, 1800499.
29. M. Sevilla and A. B. Fuertes, J. Mater. Chem. A, 2013, 1, 13738–13741.
30. X. Yuan, X.-L. Ding, C.-Y. Wang and Z.-F. Ma, Energy Environ. Sci., 2013, 6, 1105–1124.
31. W. X. Chen, J. Y. Lee and Z. Liu, Chem. Commun., 2002, 2588–2589.
32. Y. Wang, P. Han, X. Lv, L. Zhang and G. Zheng, Joule, 2018, 2, 2551–2582.
33. X. Sun, L. Lu, Q. Zhu, C. Wu, D. Yang, C. Chen and B. Han, Angew. Chem., Int. Ed., 2018, 57, 2427–2431.
34. Y. Zhao, J. Liang, C. Wang, J. Ma and G. G. Wallace, Adv. Energy Mater., 2018, 8, 1702524.
35. H. Han, Y. Noh, Y. Kim, W. S. Jung, S. Park and W. B. Kim, Nanoscale, 2019, 11, 2423–2433.
36. H. Kim, Y. Kim, Y. Noh, S. Lee, J. Sung and W. B. Kim, ChemCatChem, 2017, 9, 1503–1510.
37. Z. Lin, G. H. Waller, Y. Liu, M. Liu and C.-P. Wong, Nano Energy, 2013, 2, 241–248.
38. H. J. Cui, H. M. Yu, J. F. Zheng, Z. J. Wang, Y. Y. Zhu, S. P. Jia, J. Jia and Z. P. Zhu, Nanoscale, 2016, 8, 2795–2803.
39. Z. Lin, G. Waller, Y. Liu, M. Liu and C.-P. Wong, Adv. Energy Mater., 2012, 2, 884–888.
40. H. Han, Y. Noh, Y. Kim, V. S. K. Yadav, S. Park, W. Yoon, S. Lee and W. B. Kim, ChemistrySelect, 2017, 2, 6260–6268.
41. M. Sevilla, J. B. Parra and A. B. Fuertes, ACS Appl. Mater. Interfaces, 2013, 5, 6360–6368.
42. G. P. Hao, W.-C. Li, D. Qian and A.-H. Lu, Adv. Mater., 2010, 22, 853–857.
43. H. Yang, Y. Wu, Q. Lin, L. Fan, X. Chai, Q. Zhang, J. Liu, C. He and Z. Lin, Angew. Chem., 2018, 130, 15702–15706.
44. R. Shi, J. Zhao, S. Liu, W. Sun, H. Li, P. Hao, Z. Li and J. Ren, Carbon, 2018, 130, 185–195.
45. M. V. Morales, E. Asedegbega-Nieto, B. Bachiller-Baeza and A. Guerrero-Ruiz, Carbon, 2016, 102, 426–436.
46. Y.-X. Yu, Phys. Chem. Chem. Phys., 2013, 15, 16819–16827.
47. C. Ma, X. Shao and D. Cao, J. Mater. Chem., 2012, 22, 8911–8915.
48. Z.-J. Jiang and Z. Jiang, Sci. Rep., 2016, 6, 27081.
49. P. Mondal, A. Sinha, N. Salam, A. S. Roy, N. R. Jana and S. M. Islam, RSC Adv., 2013, 3, 5615–5623.
50. R. Shi, M. Ren, H. Li, J. Zhao, S. Liu, Z. Li and J. Ren, Mol. Catal., 2018, 445, 257–268.
51. Z.-Q. Liang, T.-T. Zhuang, A. Seifitokaldani, C.-W. Huang, C.-S. Tan, Y. Li, P. D. Luna, C. T. Dinh, Y. Hu, Q. Xiao, P.-L. Hsieh, Y. Wang, F. Li, R. Quintero-Bermudez, Y. Zhou, P. Chen, Y. Pang, S.-C. Lo, L.-J. Chen, H. Tan, Z. Xu, S. Zhao, D. Sinton and E. H. Sargent, Nat. Commun., 2018, 9, 3828.
52. J. Wu, S. Ma, J. Sun, J. I. Gold, C. Tiwary, B. Kim, L. Zhu, N. Chopra, I. N. Odeh, R. Vajtai, A. Z. Yu, R. Luo, J. Lou, G. Ding, P. J. A. Kenis and P. M. Ajayan, Nat. Commun., 2016, 7, 13869.
53. C. G. Morales-Guio, E. R. Cave, S. A. Nitopi, J. T. Feaster, L. Wang, K. P. Kuhl, A. Jackson, N. C. Johnson, D. N. Abram, T. Hatsukade, C. Hahn and T. F. Jaramillo, Nat. Catal., 2018, 1, 764–771.
54. T.-T. Zhuang, Y. Pang, Z.-Q. Liang, Z. Wang, Y. Li, C.-S. Tan, J. Li, C. T. Dinh, P. D. Luna, P. L. Hsieh, T. Burdyny, H. H. Li, M. Liu, Y. Wang, F. Li, A. Proppe, A. Johnston, D.-H. Nam, Z.-Y. Wu, Y.-R. Zheng, A. H. Ip, H. Tan, L.-J. Chen, S.-H. Yu, S. O. Kelley, D. Sinton and E. H. Sargent, Nat. Catal., 2018, 1, 946–951.
55. Y. Zhou, F. Che, M. Liu, C. Zou, Z. Liang, P. D. Luna, H. Yuan, J. Li, Z. Wang, H. Xie, H. Li, P. Chen, E. Bladt, R. Quintero-Bermudez, T.-K. Sham, S. Bals, J. Hofkens, D. Sinton, G. Chen and E. H. Sargent, Nat. Chem., 2018, 10, 974–980.
56. Y. Lum and J. W. Ager, Nat. Catal., 2019, 2, 86–93.
57. A. Verdaguer-Casadevall, C. W. Li, T. P. Johansson, S. B. Scott, J. T. Mckeown, M. Kumar, I. E. L. Stephens, M. W. Kanan and I. Chorkendorff, J. Am. Chem. Soc., 2015, 137, 9808–9811.
58. T. A. Maark and B. R. K. Nanda, J. Phys. Chem. C, 2017, 121, 4496–4504.
59. H. Xiao, W. A. Goddard, T. Cheng and Y. Liu, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 6685–6688.
60. P. Yang, H. Zhuzhang, R. Wang, W. Lin and X. Wang, Angew. Chem., 2019, 131, 1146–1149.
61. Y. Hori, R. Takahashi, Y. Yoshinami and A. Murata, J. Phys. Chem. B, 1997, 101, 7075–7081.
62. E. L. Clark and A. T. Bell, J. Am. Chem. Soc., 2018, 140, 7012–7020.
63. S. Ma, M. Sadakiyo, R. Luo, M. Heima, M. Yamauchi and P. J. A. Kenis, J. Power Sources, 2016, 301, 219–228.
64. I. Ledezma-Yanez, E. P. Gallent, M. T. M. Koper and F. Calle-Vallejo, Catal. Today, 2016, 262, 90–94.
65. K. J. P. Schouten, Y. Kwon, C. J. M. Van der Han, Z. Qin and M. T. M. Koper, Chem. Sci., 2011, 2, 1902–1909.
66. H. Xiao, T. Cheng and W. A. Goddard, J. Am. Chem. Soc., 2016, 139, 130–136.
67. Y. Zheng, A. Vasileff, X. Zhou, Y. Jiao, M. Jaroniec and S.-Z. Qiao, J. Am. Chem. Soc., 2019, 141, 7646–7659.
68. D. Kim, J. Resasco, Y. Yu, A. M. Asiri and P. Yang, Nat. Commun., 2014, 5, 4948.
69. C. Genovese, C. Ampelli, S. Perathoner and G. Centi, J. Catal., 2013, 308, 237–249.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9gc03088c

---