Boosting Electrocatalytic CO 2 Reduction with Conjugated Bimetallic Co/Zn Polyphthalocyanine Frameworks

Open AccessCCS ChemistryRESEARCH ARTICLES8 Jun 2022Boosting Electrocatalytic CO2 Reduction with Conjugated Bimetallic Co/Zn Polyphthalocyanine Frameworks Nan Li, Duan-Hui Si, Qiu-jin Wu, Qiao Wu, Yuan-Biao Huang and Rong Cao Nan Li State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 University of Chinese Academy of Science, Beijing 100049 Google Scholar More articles by this author , Duan-Hui Si State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 Google Scholar More articles by this author , Qiu-jin Wu State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 Google Scholar More articles by this author , Qiao Wu State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 Google Scholar More articles by this author , Yuan-Biao Huang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 University of Chinese Academy of Science, Beijing 100049 Google Scholar More articles by this author and Rong Cao *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002 University of Chinese Academy of Science, Beijing 100049 Fujian Science and Technology Innovation Laboratory for Optoelectronic Information of China, Fuzhou, Fujian 350108 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.022.202201943 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Development of high-efficiency electrode materials for the electrochemical CO2 reduction reaction (CO2RR) with high current density and selectivity compatible with industry is an important but significant challenge. Herein, we describe a facile strategy to enhance the selectivity and current density by regulating the local electron density of the cobalt site in a series of stable, conjugated, bimetallic Co/Zn polyphthalocyanine frameworks CoxZnyPPc with an AB stacking model under alkaline aqueous conditions. When adjusting the ratio of Co and Zn to 3:1, the optimal Co3Zn1PPc exhibits an industry-compatible CO partial current density of 212 mA cm−2 at −0.9 V versus reversible hydrogen electrode in a flow cell, which is 1.7 and 9.1 times that of the single metal polyphthalocyanine CoPPc and ZnPPc, respectively. Co3Zn1PPc shows a high CO Faraday efficiency of more than 90% in a wide operating potential window of −0.3 to −0.9 V. In-depth experimental and theoretical analysis revealed that introduction of electron-rich Zn atoms modified the electron density of the active Co center, placing Co in the electron-rich region and weakening the bonding strength with the reaction intermediate, thereby improving the CO2RR performance. These results clarify the interaction mechanism of dual metal sites at the atomic level and provide a new avenue for the design of electrocatalysts with potential in industrial applications. Download figure Download PowerPoint Introduction The electrochemical CO2 reduction reaction (CO2RR) using renewable energy, electricity at room temperature, and atmospheric pressure is believed to be a promising approach to mitigate greenhouse emission and produce value-added products including CO, HCOOH, CH4, CH3OH, C2H4, and C2H5OH.1–15 CO, the two-electron transfer product, is one of the most important products from this reaction and can be used in the Fischer–Tropsch industrial process.12,16 However, due to the strong thermodynamic stability and chemical inertness of CO2, the CO2RR usually suffers from poor selectivity and low current density, and has difficulty meeting the requirements of commercial applications.17 In recent years, great effort has been devoted to the development of various electrode materials for the conversion of CO2 to CO. These include gold- and silver-based metal materials18–20 and metal-nitrogen active (M–Nx) sites anchored in M–N–C catalysts.7,8,16–40 However, the high cost and scarcity of the noble metal-based electrodes limits their industrial applications. Although the single M–Nx active sites have been identified as a type of effective CO2RR catalyst, their large-scale preparation and characterization is difficult. To resolve the aforementioned problems, the development of new types of electrocatalysts is necessary. The activity of catalysts for the CO2RR is closely related to the local electron density of the metal center atoms. Specifically, the formation of an electron-rich region on the active center can enhance the adsorption energy of the active site with a CO2 molecule, accelerate the electron transfer from the active center atom to the adsorbed intermediate, and thus enhance the reaction kinetics of CO2RR.41 There are very few reports, however, on designing electrocatalysts based on this consideration for CO2RR. Co phthalocyanine (CoPc) has the ability to activate CO2, but the selectivity and current density still must be improved to meet the industrial level.42,43 Improving the electrocatalytic performance by constructing a CoPc-based conjugated polymer to obtain a larger plane to improve the electron transfer efficiency is feasible.44,45 The CO2RR catalyst activity can be further promoted by fabricating a diatomic metal system to introduce electron-rich metals to regulate the electron density of the active center.46–49 In particular, the binding energy with the reaction intermediate can be tailored by the redistribution of electrons between two adjacent different metal species. The electron-rich environment makes more electrons transfer from Co atoms to the adsorbed intermediates, thereby improving the reaction kinetics of the CO2RR process. This would significantly enhance, to an industrial level, the current density of the Co polyphthalocyanine framework-based electrode material for CO2RR. It is known that the electronegativity of Zn (1.65) is lower than that of Co (1.90), and this should result in an electron-rich Zn atom donating its electrons more easily to Co centers in bimetallic materials. Herein, we report our design and preparation of a series of robust bimetallic Co- and Zn-based polyphthalocyanine framework electrocatalysts, termed CoxZnyPPc, which enhance the activity of the CO2RR in alkaline electrolytes producing highly selective and industry-compatible current density by introducing electron-rich Zn atoms to increase the local electron density of the Co centers. The Co center is in an electron-rich environment, and this could be favorable for the adsorption of reactant molecules and transfer of electrons from Co atoms to the adsorbed intermediates. Compared with the cobalt polyphthalocyanine framework (CoPPc) and zinc polyphthalocyanine framework (ZnPPc), the optimal bimetal electrocatalyst Co3Zn1PPc exhibits significantly better performance in the conversion of CO2 to CO with a high partial current density of 212 mA cm−2 and a turnover frequency (TOF) of 28,310 h−1 at −0.9 V versus reversible hydrogen electrode (RHE, all potentials mentioned here refer to the RHE). The long-term stability of Co3Zn1PPc can be expressed as a constant current density of ∼220 mA cm–2 at –1.1 V for 9 h and a high CO selectivity of about 90%. Theoretical analysis revealed that the d-band of the Co site of CoZnPPc has more electrons than that of CoPPc, suggesting that the electrons from Zn are being successfully transferred to Co. Consequently, the CO2 adsorption ability of the Co centers in CoZnPPc is enhanced, and the bonding with the reaction intermediate is weakened. Thus, the lower free energy of the rate-determining step (RDS), the *COOH formation, on the Co site of CoZnPPc (0.83 eV) as opposed to that of CoPPc (0.91 eV) can be achieved, thereby improving the CO2RR performance. Experimental Methods Materials and chemicals All reagents and chemicals were obtained commercially and used without further purification. ZnCl2, CoCl2·6H2O, urea, and (NH4)2Mo2O7 were purchased from Aladdin (Shanghai, China). NH4Cl was purchased from Xilong Scientific Co., Ltd. (Shantou, China). Pyromellitic dianhydride was purchased from Adamas Reagent, Ltd. (Shanghai, China). Acetone and methanol were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Deionized water was supplied by a UPT-I-5T ultrapure water system (18.3 MΩ cm). Synthesis of CoPPc CoPPc was prepared by a solid phase synthesis method in a muffle furnace.50 Briefly, a mixture of CoCl2·6H2O (1.20 g), urea (4.10 g), NH4Cl (1.00 g), (NH4)2Mo2O7 (0.025 g), and pyromellitic dianhydride (2.10 g) was ground into a powder that was placed in a 250 mL crucible and heated in a muffle furnace at 220 °C for 3 h with a ramp rate of 2 °C min−1. After cooling to room temperature, the product was washed with water, acetone, and methyl alcohol and finally, dried in vacuo at 70 °C. Synthesis of ZnPPc The ZnPPc was prepared by the method that was used for CoPPc, except that CoCl2·6H2O was replaced with ZnCl2 (0.70 g). Synthesis of CoxZnyPPc Co3Zn1PPc, Co4Zn1PPc, and Co5Zn1PPc were prepared by the method that was used for CoPPc, except that CoCl2·6H2O was replaced with a mixture of CoCl2·6H2O and ZnCl2 (nCoCl2·6H2O + nZnCl2 = 5.00 mmol): CoCl2·6H2O (1.07 g, 4.50 mmol) and ZnCl2 (0.07 g, 0.50 mmol) for Co3Zn1PPc; CoCl2·6H2O (0.95 g, 4.00 mmol) and ZnCl2 (0.136 g, 1.00 mmol) for Co4Zn1PPc; and CoCl2·6H2O (0.71 g, 3.00 mmol) and ZnCl2 (0.273 g, 2.00 mmol) for Co5Zn1PPc. Physicochemical characterization Scanning electron microscopy (SEM) images were obtained using a JSM-6700F microscope (JEOL, Japan) at 10 kV. Transmission electron microscopy (TEM) images were recorded by a FEI Tecnai 20 microscope (FEI, United States) at 200 kV. The energy-dispersive system (EDS) of samples were performed with a Titan Cubed Themis G2 300 high-resolution transmission electron microscope (FEI, United States) operated at 200 kV. Powder X-ray diffraction (PXRD) patterns were recorded on a Miniflex 600 diffractometer (Rigaku, Japan) using Cu Kα radiation (λ = 0.154 nm) and the beamline BL14B1 station at the Shanghai Synchrotron Radiation Facility. The Fourier transform infrared (FT-IR) spectra were measured using VERTEX 70 (Bruke, Germany). UV–vis absorption spectra were measured on a PerkinElmer Lambda 900 UV–vis–NIR spectrophotometer (PerkinElmer, United States). The CO2 sorption isotherm were measured using a Micromeritics ASAP 2020 instrument (Micromeritics, United States). The content of Co and Zn in the solid samples was measured by inductively coupled plasma atomic emission spectroscopy on an Ultima2 analyzer (Jobin Yvon, France). X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB 250Xi X-ray photoelectron spectrometer (FEI, United States). The data of X-ray absorption spectra at the Co K-edge (7709 eV) and the Zn K-edge (9659 eV) were measured at the beamline BL14W1 station of the Shanghai Synchrotron Radiation Facility, China. Gas chromatography (GC) measurements were performed on an Agilent 7820A gas chromatogram equipped with a flame ionization detector (FID) and thermal conductivity detector (TCD). Nuclear magnetic resonance (NMR) spectroscopy was performed on a Bruker-Biospin AVANCE III spectrometer (Bruker, Germany), operating at 400 MHz for 1H NMR. The Nyquist plots were obtained by electrochemical impedance spectroscopy measurements, which were conducted in the frequency range of 100 kHz to 50 mHz. Electrocatalytic CO2RR measurements H-type cell Electrochemical experiments were implemented in a designed H-type electrochemical cell with two compartments separated by an anion-exchange membrane of Nafion-117. Each compartment contained 70 mL of electrolyte (0.5 M KHCO3 made from ultrapure water). The working electrode was fabricated by coating 80 μL of catalyst ink onto a carbon fiber paper electrode of 1 cm × 1 cm. The homogeneous ink was prepared with 5 mg of catalyst and 2 mg ketjenblack (an electroconductive carbon black) dispersed into 500 μL of isopropanol containing 40 μL Nafion solution (5 wt %). Electrochemical measurements were performed in a three-electrode cell using a Ag/AgCl electrode as the reference electrode and Pt gauze as the counter electrode. Before the CO2 electrochemical reduction, the electrolyte in the cathodic compartment was purged with CO2 gas for at least 30 min to achieve a CO2-saturated solution (pH 7.3). Linear sweep voltammetry (LSV) was performed from −0.5 to −2.0 V versus Ag/AgCl in CO2-saturated 0.5 M KHCO3 electrolyte at a scan rate of 10 mV s−1. All potentials in the studies were converted to the potential versus RHE according to the equation E (vs RHE) = E (vs Ag/AgCl) + 0.1989 V + 0.059 × pH. CO2 gas was delivered at an average rate of 30 mL min−1 (at room temperature and ambient pressure) and routed into the gas sampling loop (0.8 mL) of a gas chromatograph. The gas phase composition was analyzed by GC every 15 min. The separated gas products were analyzed by a TCD (for H2) and an FID (for CO). Flow cell Electrochemical measurements with high current densities were performed in a flow cell consisting of a gas diffusion electrode (GDE), an anion exchange membrane, and a Pt plate anode. Catalyst ink (80 μL) was loaded onto a 0.4 cm × 2 cm carbon paper to create a GDE. The homogeneous ink was prepared using 5 mg of catalyst and 2 mg ketjenblack dispersed into 500 μL of isopropanol containing 40 μL Nafion solution (5 wt %). A Ag/AgCl electrode acted as the reference. A 1 M KOH aqueous solution, used as the electrolyte, was circulated through the anode side using a peristaltic pump. CO2 gas was supplied to the cathode side at a constant flow rate of 30 mL min−1 monitored by a flow controller. LSV was performed the same way as in the H-type cell. The liquid products were analyzed subsequently by quantitative 1H NMR using dimethyl sulfoxide (DMSO) as an internal standard. The Faraday efficiency of a gas product was calculated by the equation: F E = P V R T × ν N F × 10 − 6 ( 3 m / mL ) I × 60 ( s / min ) where v (vol %) is the volume concentration of certain gas product in the exhaust gas from the cell (GC data); V is the gas flow rate measured by a flow meter, and is typically 30 mL min−1; I is the total steady-state cell current; N is the electron transfer number for product formation; F is the Faraday constant, 96,485 C mol−1; R is the universal gas constant, 8.314 J mol−1 K−1; P is 1 atmosphere, 1.013 × 105 Pa; and T is room temperature, 298.15 K. The TOF for a given product was calculated from the formula: TOF = I product / N F ω m cat / M Co 3600 where Iproduct is the partial current for certain product; N is the number of electrons transferred for product formation, which is 2 for CO; mcat is the catalyst mass in the electrode, g; ω is the Co loading in the catalyst; and MCo is the atomic mass of Co, 58.93 g mol−1. Computational Methods All intermediates were optimized by density functional theory (DFT) using the Vienna Ab initio Simulation Package (VASP).63–66 Pseudopotentials were performed by the projector-augmented wave67,68 and Perdew–Burke–Ernzerh exchange-correlation functional.69 The convergence criteria, the cutoff energy of plane wave basis, and the threshold for force were set to 1 × 10−5 eV, 500 eV, and −0.05 eV·Å−1, respectively. The van der Waals correction was adopted by Grimme (DFT+D3).70 The spin polarization scheme was considered in all calculations. The CoPPc and Co3Zn1PPc surfaces (21.3 Å × 21.3 Å) were modeled by three-layer slab, the top layer and the intermediate layers were fully relaxed, and the other layers were fixed. To avoid the interactions between repeated slabs, a 17 Å vacuum region was built. The brillouin zone integration was accomplished with a 2 × 2 × 1 Monkhorst–Pack k-point mesh. The Gibbs free energies of intermediates were calculated at 298.15 K by the VASPKIT program.71 The free energy of H+ ions was corrected by the concentration dependence of the entropy: ΔG(pH) = −kTln[H+ ] = kTln10 × pH.72 Furthermore, the effect of a bias was considered by ΔG(U) = −eU, where U indicates the electrode potential.72 Results and Discussion As shown in Figure 1a, the phthalocyanine-based conjugated polymer CoxZnyPPc was synthesized using a mixture of urea, pyromellitic dianhydride, CoCl2·6H2O, and ZnCl2 by a condensation reaction in a muffle furnace at 220 °C. For comparison, the single metal-based phthalocyanine polymers CoPPc and ZnPPc were also prepared using a similar method. The synchrotron radiation PXRD pattern of the as-synthesized Co3Zn1PPc is similar to that of CoPPc, showing that the layered AB-stacking model is the dominant crystal structure (Figures 1b and 1c and Supporting Information Figures S1 and S2 and Tables S1 and S2).73 This is a rare example of an AB-stacking structure that is different from most 2D covalent organic frameworks, which have an AA-stacking model.74 Such an AB-stacking model usually lowers the energy, thus leading to more stable frameworks than AA-stacking structures, and facilitates the CO2RR performance under harsh conditions. The SEM and TEM images of Co3Zn1PPc showed a 2D sheet-like morphology, corresponding to the pattern in the PXRD, which also indicated that the Co3Zn1PPc network has a 2D structure (Figure 2a and Supporting Information Figure S3). As shown in the atomic force microscopy (AFM) images and their height profiles in Figures 2b and 2c, ultrathin 2D Co3Zn1PPc nanosheets with a thickness of ca. 4.68 nm were successfully prepared by an exfoliation method with high-frequency sonication, which is close to two layers of Co3Zn1PPc nanosheets, proving that Co3Zn1PPc is indeed a two-dimensional sheet-like structure. No metal nanoparticles were observed in the TEM and scanning TEM (STEM) images (Figure 2e), indicating the Co and Zn elements were atomically dispersed in these materials. Aberration-corrected high-angle annular dark-field STEM (AC HAADF-STEM) was conducted to confirm the atomic dispersion of the Co and Zn atoms in the typical Co3Zn1PPc. Since the electron density of metal is much higher than that of the non-metals C and N, the appearance of the individual bright spots with high density reveals that the Co and Zn atoms are dispersed as single atoms in Co3Zn1PPc (Figure 2d). The STEM mapping images demonstrate that Co, Zn, N, and C are uniformly distributed spatially throughout the material (Figure 2e), which suggests that the Co and Zn elements were atomically dispersed within in the frameworks of CoxZnyPPc. Figure 1 | (a) Illustration of the preparation of CoxZnyPPc (monolayer and AB-stacking layered structures on the right side of the arrow) by solid phase synthesis. (b) PXRD patterns for CoPPc and ZnPPc, synchrotron radiation experimental PXRD patterns for Co3Zn1PPc, and simulated patterns with AA-stacking and AB-stacking modes for CoPPc. (c) The topology of the AB-stacking modes. Download figure Download PowerPoint The test results of ICP-AES ( Supporting Information Table S3) proved that a series of bimetallic catalysts with Co and Zn ratios of 3:1, 4:1 and 5:1 were successfully synthesized. Subsequently, spectroscopic characterization of Co3Zn1PPc was performed to identify the formation of a conjugated polymer structure with CoPc and ZnPc units. The UV–vis spectra of Co3Zn1PPc, CoPc, and ZnPc ( Supporting Information Figure S6) show the typical Q bands in the visible region (600–800 nm) and the B bands in the ultraviolet region (300–500 nm). Compared with the CoPc and ZnPc monomers, a red shift of the Q band and a blue shift of the B band for Co3Zn1PPc indicates the conjugated polymeric nature and strong electronic interaction between the metal centers and the phthalocyanine moieties in Co3Zn1PPc.74 In comparison with CoPc and ZnPc monomers, the Q-band intensity of Co3Zn1PPc was greatly weakened due to the formation of a conjugated square structure. The formation of Co3Zn1PPc was also confirmed by FT-IR spectroscopy ( Supporting Information Figure S7). The peak at 1238 cm−1 which is related to the stretching of the C–O band of the pyromellitic dianhydride disappears. At the same time, the typical characteristic peaks of the phthalocyanine skeleton vibration are observed at 1453, 1309, 1155, 1090, 900, 755, and 725 cm−1, indicating the formation of the polyphthalocyanine. The 13C cross-polarization with magic angle spinning spectra of CoPPc, ZnPPc, and Co3Zn1PPc has four peaks at 169.6, 166.1, 136.7, and 115.9 ppm, which were attributed to the polyphthalocyanine skeleton ( Supporting Information Figure S8). The Raman spectrum of Co3Zn1PPc shows several typical bands of CoPc monomer at 685, 750, 1138, 1445, and 1541 cm−1 ( Supporting Information Figure S9), which proves that Co3Zn1PPc has the basic skeleton of a phthalocyanine. Figure 2 | (a) TEM image and (b) AFM image of Co3Zn1PPc. (c) The height profile follows the white line in (b). (d) AC HAADF-STEM images of Co3Zn1PPc, partial of single Co and Zn atoms are highlighted by red circles. (e) STEM image of Co3Zn1PPc with the corresponding elemental mapping of C, N, Co, and Zn, respectively. Download figure Download PowerPoint To confirm the state of the atomically dispersed metal species in Co3Zn1PPc, XPS was performed. The XPS spectrum in Figure 3a confirms the presence of Co, Zn, N, and C elements in Co3Zn1PPc, in agreement with the EDS mapping analysis (Figure 2e). The high-resolution N 1s spectrum of Co3Zn1PPc in Figure 3b can be deconvoluted into a peak of M–N (M = Co, Zn) at 399.8 eV and a peak of imine-N at 398.2 eV, corresponding to the phthalocyanine skeleton. The high-resolution Co 2p spectrum in Figure 3c shows two main peaks of 2p3/2 and 2p1/2 at 795.74 and 780.65 eV,74,75 indicating the Co(II) is anchored by the phthalocyanine skeleton. Similarly, as shown in Figure 3d, the Zn 2p spectrum can be divided into the Zn(II) 2p3/2 and 2p1/2 peaks at 1021.51 and 1044.63 eV, respectively.76 As shown in Figure 3c, the Co(II) of CoPPc exhibits a positively shifted binding energy of 0.22 eV in comparison with that of Co3Zn1PPc. Similarly, compared with that of Co3Zn1PPc, a negatively shifted binding energy of 0.75 eV for the Zn(II) of ZnPPc is observed (Figure 3d). It can be reasonably inferred that Zn atoms donated their electrons to Co atoms in the 2D conjugated square networks in Co3Zn1PPc, which would endow the electron-rich Co sites with greater nucleophilicity and a stronger bond with Lewis acidic CO2 molecules. Thus, Co3Zn1PPc can contribute to CO2 adsorption and the charge transfer between catalyst and adsorbate, and this was evidenced by the results of the CO2 isotherm adsorption experiments. As shown in Supporting Information Figures S10–S14, the CO2 adsorption capacity of the bimetallic Co/Zn polyphthalocyanines is significantly larger than those of the CoPPc and ZnPPc. The Co3Zn1PPc has the highest CO2 adsorption capacity of 44.06 cm3 g−1 at 298 K, which is about 2.5 times that of CoPPc (16.90 cm3 g−1) and 6 times that of ZnPPc (6.80 cm3 g−1) when tested at the same temperature. Figure 3 | (a) Wide XPS spectrum of Co3Zn1PPc. High-resolution XPS spectra of (b) N 1s, (c) Co 2p for Co3Zn1PPc and CoPPc, and (d) Zn 2p for Co3Zn1PPc and ZnPPc. Download figure Download PowerPoint To further confirm the electronic state of the atomically dispersed metal species and the coordination environment, measurements of X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) were conducted. As shown in Figure 4a, the Co K-edge XANES profile (red line) shows the edge energy value of Co3Zn1PPc located between Co foil and CoO, confirming that the average oxidation state of Co lies between Co(0) and Co(II).77,78 Fourier transformation of the EXAFS analysis shows that both Co3Zn1PPc and CoPc possess a major peak at ∼1.47 Å, which corresponds to the Co–N bonds. At the same time, the absence of a peak near 2.15 Å suggests that there is no metal Co–Co bond in Co3Zn1PPc (Figure 4b), which is consistent with the XPS (Figure 3c) and TEM results (Figure 2a). The fitting results show that the coordination number of Co sites in Co3Zn1PPc is ∼4.04, which is similar to the four-coordinated structure of CoPc (Figure 4c and Supporting Information Table S4). The Zn K-edge XANES profile (red line, Figure 4d) exhibits that Co3Zn1PPc is located between the Zn foil and ZnO, suggesting that the valence state of Zn in Co3Zn1PPc is positively charged, but below +2, consistent with the results of the Zn 2p XPS in Figure 3d.76 The valence state of Zn in Co3Zn1PPc is much higher than that of the ZnPc (Figure 4d, inset spectra), indicating the electron was transferred from Zn. Figure 4 | (a) Normalized Co K-edge XANES spectra of the Co3Zn1PPc, CoPc, CoO, Co3O4, and Co foil. (b) Fourier-transform EXAFS spectra of Co3Zn1PPc, CoPc, and Co foil. (c) The corresponding EXAFS fitting curves of Co3Zn1PPc. (d) Normalized Zn K-edge XANES spectra of the Co3Zn1PPc, ZnPc, ZnO, and Zn foil. (e) Fourier-transform EXAFS spectra of Co3Zn1PPc, ZnPc, ZnO, and Zn foil. (f) The corresponding EXAFS fitting curves of Co3Zn1PPc. Download figure Download PowerPoint Figure 5 | Electrocatalytic CO2RR performance in a flow cell system. (a) Schematic of a gas-fed flow cell configuration. (b) LSV curves of CoPPc, ZnPPc, and Co3Zn1PPc in 1 M KOH electrolyte under CO2. (c) CO partial current density and (d) Faraday efficiencies of CO for CoPPc, ZnPPc, and Co3Zn1PPc. The error bars represent the standard deviation of the independent measurements. (e) Turnover frequency of CoPPc and Co3Zn1PPc. (f) Comparison of jCO values between Co3Zn1PPc with other reported electrocatalysts evaluated in a flow cell.35,51–62 Download figure Download PowerPoint The Fourier transform of the EXAFS analysis shows that both Co3Zn1PPc and ZnPc have a peak at approximately 1.58 Å, which corresponds to the Zn–N bond (Figure 4e). The absence of the peak near 2.26 Å proves that there is no Zn–Zn bond in Co3Zn1PPc. The fitting results show that the coordination number of the Zn site in Co3Zn1PPc is about 4.09, which is similar to that in the four-coordination structure of ZnPc (Figure 4f and Supporting Information Table S5). Based on these results, we conclude that the Co and Zn species in Co3Zn1PPc exist with Co–N4 and Zn–N4 coordination bonds. To explore the effect of the introduction of the Zn atom in CoxZnyPPC on the CO2RR, an H-type cell containing two compartments with a three-electrode system separated by a proton-exchange membrane in a 0.5 M KHCO3 solution saturated with CO2 was constructed. GC and 1H NMR spectroscopy were utilized to detect the gaseous products and liquid products, respectively. Notably, compared with ZnPPc which produced H2 as the major gas in a wide potential range from −0.6 to −0.9 V, CoPPc afforded CO with a high CO Faraday efficiency (FECO) of more than 90% at −0.70 and −0.80 V ( Supporting Information Figure S16). This indicates that Co centers, not Zn atoms, were the active sites promoting the CO2RR. Interestingly, as shown in Supporting Information Figures S15 and S16, all the bimetallic CoxZnyPPc materials showed higher FECO than those of CoPPc, with >90% between −0.60 to −0.90 V, suggesting that the introduction of an electron-rich Zn atom improved the selectivity of the CO2RR. In view of the excellent electrocatalytic CO2RR performance of the bimetallic CoxZnyPPc materials in the H-cell, we used a flow cell equipped with a GDE (Figure5a) to enhance the activity of the CO2RR by increasing mass transfer to the active sites. Because the robust metal polyphthalocyanines have good alkali resistance ( Supporting Information Figure S17), the CO2RR experiments were performed in the typical KOH aqueous electrolytes. As shown in Supporting Information Figure S18, the current densities of CoPPc, ZnPPc, and CoxZnyPPc in a CO2 atmosphere were higher than those in the Ar saturated electrolyte, indicating that their activities mainly originated from CO2RR. As shown in the LSV in Figure 5b and Supporting Information Figure S19, the current densities of a