Revealing the Reaction Mechanism of C-C Coupling on Cu-based Tandem Catalysts towards CO2 Reduction

Cu-based tandem catalysts have widely been exploited to improve the catalytic performance for multi-carbon (C2+) products towards CO2 electroreduction. Nevertheless, the underlying reaction mechanism for tandem catalytic system remains ambiguity. Herein, we unraveled the relationship between the behavior of adsorbed CO intermediate (*CO) and the process of C-C coupling. Due to the low coverage of *CO, the process of C-C coupling has always been restricted for Cu catalysts. Through regulating the molar ratios of CO2 to CO in co-feeding gases, we found that the moderate surface coverage of *CO was bene�cial for the electroreduction of CO2 into ethylene (C2H4). Meanwhile, the cross-coupling process between CO2 and CO was proved to be responsible for the enhanced activity of C2H4 according to isotopic labeling experiments. We constructed a tandem model with cobalt phthalocyanine (CoPc) as CO-generating component on Cu to further verify the tandem mechanism. With the introduction of CoPc onto the surface of Cu, the partial current density for C2H4 reached as high as 313 mA cm-2 at applied current density of 480 mA cm-2, which was one-fold higher than that (165 mA cm-2) of Cu �lm. In-situ Raman measurements further revealed that CO generated by CoPc increased the coverage of *CO on the surface of Cu, facilitating the process of C-C coupling.


Introduction
The excessive utilization of fossil fuels emits a large amount of CO 2 , giving rise to energy shortage and environmental crisis [1][2][3][4][5] .To meet the ever-increasing energy demands and to mitigate ever-fragile environmental status, researchers have exploited a series of strategies such as chemical reforming 6,7 , photoreduction 8,9 , biological conversion 10,11 , and electroreduction 12-16 via transforming CO 2 into useful chemicals and feedstocks.Among these strategies, CO 2 electroreduction, which operates under ambient pressure at room temperature, serves as one of the promising ways to promote CO 2 utilization and global carbon recycling [17][18][19][20][21] .During the process of CO 2 electroreduction, a variety of carbonaceous products, including carbon monoxide (CO), formate (HCOO -), methane (CH 4 ), ethylene (C 2 H 4 ), propylene (C 3 H 6 ), ethanol (C 2 H 5 OH), and acetate (CH 3 COO -) have been reported [22][23][24] .Especially, multi-carbon (C 2+ ) products have been widely used as key feedstocks for industrial manufacture and liquid fuels due to their unique chemical functional groups and high energy density 25 .As such, it is of pivotal signi cance to construct the highly e cient electrocatalysts for C 2+ products towards CO 2 electroreduction.Among all catalysts, Cu-based nanocrystals have been regarded as the most e cient catalysts to generate C 2+ products towards CO 2 electroreduction.Great efforts have been devoted to improve the selectivity for C 2+ products via engineering active phases [26][27][28] and constructing compositions [29][30][31][32] .In addition, the strategy for designing tandem catalysts where Cu-based nanocrystals coupled with COgenerating catalysts have been considered as the most effective method to improve catalytic performance for C 2+ products.Up till now, a series of Cu-based tandem catalysts including Cu 4 Zn 33 , Ag-Cu 2 O PB 34 , Au/Cu bimetallic 35 , Cu dots on Ag substrate 36 , Cu-Ag bimetallic 25 , CuO x /NiNC 37 , and Cu/Zn bimetallic 38 have been developed for CO 2 electroreduction into C 2+ products.Despite all those achievements, the underlying reaction mechanism for tandem catalytic system remains ambiguity.Therefore, the understanding of the intrinsic reaction mechanism for tandem catalysts is of great importance to design e cient catalysts for the electroreduction of CO 2 into C 2+ products.
Herein, we unraveled the relationship between the behavior of adsorbed CO intermediate ( * CO) and the process of C-C coupling.Due to the low coverage of * CO, the process of C-C coupling has always been restricted for Cu catalysts.Through regulating the molar ratios of CO 2 to CO in co-feeding gases, we found that the moderate surface coverage of * CO was bene cial for the electroreduction of CO 2 into C 2 H 4 .
Meanwhile, the cross-coupling process between CO 2 and CO was proved to be responsible for the reached as high as 313 mA cm -2 at applied current density (j) of 480 mA cm -2 , which was one-fold higher than that (165 mA cm -2 ) of Cu lm.In-situ Raman measurements further revealed that CO generated by CoPc increased the coverage of * CO on the surface of Cu, facilitating the process of C-C coupling.

Results
Synthesis and structural characterizations of gas diffusion electrode (GDE) of Cu.The GDE of Cu was prepared via sputtering Cu on the porous polytetra uoroethylen (PTFE) frames.As shown in scanning electron microscopy (SEM) images, Cu lm was uniformly coated on the PTFE (Fig. 1a and b).1e, a characteristic peak at 918.8 eV was observed in the Cu LMM Auger spectrum, demonstrating the metallic Cu species in the GDE of Cu.Fig. 1f shows the K-edge X-ray absorption near-edge spectroscopy (XANES) spectra of Cu foil, Cu 2 O, and GDE of Cu.The GDE of Cu exhibited a similar energy absorption edge pro le with Cu foil in the range between 8970.0 and 9000.0 eV.Moreover, the rst-order derivative peak of XANES pro le for the GDE of Cu was located at 8979.0 eV, further revealing the metallic state of Cu in the GDE of Cu (Supplementary Fig. 1).
To improve structural stability for the GDE of Cu, we constructed a hydrophobic layer via spraying carbon black (Vulcan, XC-72) on Cu lm [39][40][41] .After spraying a layer of carbon black onto Cu lm, the GDE of Cu was regulated from hydrophilic to hydrophobic with the contact angles increased from 65.5 o to 121.5 o (Fig. 1g).As shown in Fig. 1h, the average thickness of Cu lm and carbon black in the GDE of Cu were about 0.6 and 1.5 μm, respectively.The structure for the GDE of Cu was further proved by energy dispersive X-ray (EDX) elemental mapping with clear contrasts among carbon black, Cu lm, and PTFE frames (Fig. 1i).
Catalytic performance for the GDE of Cu towards CO 2 electroreduction.A ow-cell system with 1 M KOH as electrolyte was used to evaluate the catalytic performance for the GDE of Cu (Supplementary Fig. 2).
During the process of chronopotentiometric electrolysis, the gaseous products including hydrogen (H 2 ), CO, CH 4 , C 2 H 4 , and C 3 H 6 were quanti ed via online gas chromatography (GC) (Supplementary Fig. 3).
Meanwhile, the liquid products containing HCOO -, C 2 H 5 OH, and CH 3 COO -were quantitatively analyzed via 1 H nuclear magnetic resonance ( 1 H NMR) (Supplementary Figs. 4 and 5).As shown in Fig. 2a, a  (Supplementary Fig. 6).Apart from the C 2+ products, the GDE of Cu also exhibited a relatively steady FE for HCOO -(FE HCOO-) of about 15% and FE for CH 4 (FE CH4 ) of around 2% at all applied j (Supplementary Fig. 7).Fig. 2b shows the FE for CO (FE CO ).FE CO decreased with the increase of applied j.When the applied j was higher than 360 mA cm -2 , FE CO was limited within 3%.As shown in Fig. 2c, the highest j for C 2 H 4 (j C2H4 ) was 165 mA cm -2 at applied j of 480 mA cm -2 .Nevertheless, j for CO (j CO ) decreased with the increased tendency of j C2H4 (Fig. 2d).A reasonable explanation for this phenomenon is that * CO is the essential intermediate for the formation of C 2 H 4 .
The investigation of tandem mechanism for the GDE of Cu.To gain insight into the C-C coupling mechanism for the formation of C 2 H 4 , we monitored the FE C2H4 for the GDE of Cu at applied j of 300 mA cm -2 with various partial pressures of CO 2 and CO as co-feeding gases.As shown in Fig. 3a, the GDE of Cu exhibited a volcano-type relationship between FE C2H4 and partial pressure of gaseous CO.When the partial pressure of CO reached 8%, the GDE of Cu displayed the highest FE C2H4 of 59%, which was 1.4 times as high as the GDE of Cu with pure CO 2 as feeding gas (43%).We deduced that the enhanced FE C2H4 was attributed to the increased surface coverage of * CO which originated from the increased partial pressure of gaseous CO.Further increasing the partial pressure of CO to 16% leaded to the decrease of FE C2H4 .To unravel the role of gaseous CO in the formation of C 2 H 4 , we adopted isotopic labeling experiments using 13 CO 2 and CO as co-feeding gas ( 13 CO 2 :CO=92:8).At applied j of 300 mA cm - According to the quantitative equations and the peak areas of fragments with m/z of 30, 27, and 26, the ratios for 13 C 2 H 4 , 13 CH 2 CH 2 , and C 2 H 4 were calculated to be 57.1%,42.5%, and 0.4%, respectively (Fig. 3b).The abundant product of 13 CH 2 CH 2 proved that the cross-coupling process proceeded between CO 2 and CO.
To investigate the interferences of the local pH of electrolyte around the surface of catalysts and partial pressure of gaseous CO 2 , we conducted CO 2 electroreduction over the GDE of Cu using CO 2 and Ar as cofeeding gases at applied j of 300 mA cm -2 .As shown in Fig. 3c, a relatively steady FE C2H4 of 40% was acquired for the GDE of Cu as the partial pressure of CO 2 ranged from 100% to 84%.This result indicated that the enhanced FE C2H4 in the co-feeding gases of CO 2 and CO was not ascribed to the increased local pH.In addition, the FE C2H4 gradually decreased as the partial pressure of CO 2 decreased from 70% to 10% using CO 2 and Ar as co-feeding gases (Supplementary Fig. 9).The decreased FE C2H4 could be attributed to the low partial pressure of gaseous CO 2 which hindered the kinetic process for the formation of * CO and thus the coupling process of * CO into * OCCO.To verify whether the enhanced FE C2H4 was originated from the coupling process of CO, we conducted CO electroreduction in various co-feeding gases of Ar and CO at applied j of 300 mA cm -2 .When partial pressure of CO ranged from 0% to 16%, FE C2H4 gradually increased from 0% to 4% (Fig. 3d).Nevertheless, when the partial pressure of CO was further increased from 30% to 90%, all of the FE C2H4 for the GDE of Cu remained around 4% (Supplementary Fig. 10).
Especially, when pure CO served as feeding gas, FE C2H4 was still less than 5% as applied j was higher than 300 mA cm -2 (Supplementary Fig. 11).The poor catalytic activity for CO electroreduction into C 2 H 4 could be ascribed to the overloaded coverage of * CO on Cu surface.The overloaded coverage of * CO restricts the coverage of * H, depressing the hydrogenation kinetics of * OCCO to * OCCOH 36,42 .As a consequence, we concluded that a moderate coverage of * CO was bene cial for the formation of C 2 H 4 .
To validate the tandem process between CO 2 and CO, we constructed a type of tandem catalyst via coupling CoPc with the GDE of Cu (denoted as Cu-CoPc).In these tandem catalysts, the electronic structure of Cu maintained original state due to the van der Waals contact between CoPc and Cu, in which CoPc served as highly e cient catalyst for the formation of CO. 43 Notably, CoPc exhibited remarkable catalytic performance in ow-cell system with FE CO higher than 90% in a wide region of applied j from 60 to 360 mA cm -2 (Supplementary Fig. 12).To explore the matched degree between CoPc and Cu, we prepared a series of tandem catalysts by spraying different amount of CoPc onto Cu lm.The compositions of Cu-CoPc were investigated by Raman spectra.As shown in the Supplementary Fig. 13, all of the Cu-CoPc catalysts exhibited the obvious peaks at 681.9, 1337.2, and 1537.9 cm -1 , which were attributed to the characteristic peaks for CoPc.In addition, the intensity of the characteristic peaks for the four Cu-CoPc catalysts gradually increased associated with the increased contents of CoPc.Determined by the inductively coupled plasma atomic emission spectroscopy (ICP-AES), the molar ratios of CoPc to Cu were 3.1%, 6.2%, 10.4%, and 16.6%, denoting these samples as Cu-CoPc-1, Cu-CoPc-2, Cu-CoPc-3, and Cu-CoPc-4, respectively (Supplementary Table 2).Supplementary Fig. 14 shows the EDX elemental mapping of the four Cu-CoPc catalysts, where Co and N uniformly distributed on the GDE of Cu.
To evaluate the catalytic performance of these tandem catalysts, we conducted chronopotentiometric electrolysis.As shown in Fig. 4a and Supplementary Fig. 15, Cu-CoPc-2 achieved the highest FE C2+ of 82% at applied j of 480 mA cm -2 , which was 1.4, 1.2, and 2.0 times as high as Cu-CoPc-1 (57%), Cu-CoPc-3 (70%), and Cu-CoPc-4 (41%) under the similar applied potentials, respectively.Fig. 4b shows the FE C2H4 for the four Cu-CoPc catalysts.As the applied j ranging from 240 to 480 mA cm -2 , Cu-CoPc-2 always exhibited the highest FE C2H4 among all of the Cu-CoPc catalysts.When applied j was set at 420 mA cm -2 , the FE C2H4 of Cu-CoPc-2 was up to 67%, which was 1.3, 1.1, and 1.6 times as high as those of Cu-CoPc-1 (50%), Cu-CoPc-3 (60%), and Cu-CoPc-4 (42%), respectively.Fig. 4c shows the j C2H4 for the four Cu-CoPc catalysts.Cu-CoPc-2 also exhibited the highest j C2H4 with respect to the other counterparts at the applied j ranging from 240 to 480 mA cm -2 .Especially, at the applied j of 480 mA cm -2 , the j C2H4 of Cu-CoPc-2 reached 313 mA cm -2 , whereas the j C2H4 of Cu-CoPc-1, Cu-CoPc-3, and Cu-CoPc-4 were 220, 282, and 184 mA cm -2 , respectively.Notably, the catalytic performance of Cu-CoPc-2 was comparable to those of the state-of-the-art tandem catalysts up to date (Supplementary Table 3).The durability test for Cu-CoPc-2 was conducted at applied j of 480 mA cm -2 for 20 h.Cu-CoPc-2 exhibited around 16% decay for FE C2H4 and 5% ascent for cathodic potential (Fig. 4d).The degradation of the catalytic performance was mainly attributed to the destruction of the carbon black layers and the transformation of hydrophobic surface into hydrophilic surface (Supplementary Fig. 16).
To explore the effect of CoPc on the intrinsic activity for the formation of C 2 H 4 , we evaluate the ECSAs via Pb under potential deposition (UPD) measurements.As shown in Supplementary Fig. 17 and Supplementary Table 4, the four Cu-CoPc catalysts exhibited similar electrochemical surface areas (ECSAs) with the GDE of Cu.Nevertheless, all of the tandem catalysts exhibited higher FE C2H4 and j C2H4 than the GDE of Cu (35% for FE C2H4 , 165 mA cm -2 for j C2H4 ) at applied j of 480 mA cm -2 .As such, the enhancement of FE C2H4 was attributed to the tandem process between CoPc and Cu.As shown in Supplementary Fig. 18, the FE CO for all of Cu-CoPc was higher than that for the GDE of Cu at the same applied j.Meanwhile, the FE CO increased with the increased molar ratios of CoPc to Cu, indicating that CoPc served as CO-generating component.The FE HCOO-decreased with the increased molar ratios of CoPc to Cu, further revealing that the introduction of CoPc into Cu weakened the competing 2e process on Cu (Supplementary Fig. 19).Hence, the introduction of CoPc modulated the local coverage of * CO on the surface of Cu, regulating the catalytic activity for C 2+ products towards CO 2 electroreduction.
To gain deeper insight into the tandem mechanism between Cu and CoPc, we carried out in-situ Raman spectroscopy measurements.Fig. 5, a and b show the in-situ Raman spectra towards CO 2 electroreduction over the GDE of Cu and Cu-CoPc-2.The peaks at 2037 and 2085 cm -1 were observed for both the GDE of Cu and Cu-CoPc-2, which were assigned to the intramolecular stretching of the sparse and local-enrichment * CO, respectively. [29,44]As shown in Supplementary Fig. 20, no obvious peaks associated with the intramolecular stretching of * CO was observed for CoPc, indicating that CoPc served as CO-generating component with weak * CO adsorption.As such, the peaks at 2037 and 2085 cm -1 for both the GDE of Cu and Cu-CoPc-2 were attributed to the intramolecular stretching of * CO on Cu sites.At each applied potential, Cu-CoPc-2 always exhibited a higher peak intensity for the intramolecular stretching of * CO in comparison with the GDE of Cu, indicating that the introduction of CoPc increased the coverage of * CO on the surface of Cu.Moreover, the peak-area ratios for the local-enrichment * CO to the sparse * CO increased from 0.66 to 2.01 when the applied potential increased from -0.4 to -0.8 V vs RHE, further revealing that the surface coverage of * CO around Cu increased at high overpotential over Cu-CoPc-2.Hence, the increased surface coverage of * CO around Cu facilitated C-C coupling, improving the catalytic activity for C 2+ products at high overpotential over Cu-CoPc-2.These results were in good agreement with the tendency of catalytic performance.Accordingly, the tandem process between CoPc and Cu was responsible for the enhanced activity for C 2+ products.

Discussion
In summary, we unraveled that the moderate surface coverage of * CO was bene cial for the electroreduction of CO 2 into C 2 H 4 using the co-feeding gases of CO 2 and CO on the GDE of Cu.The crosscoupling process between CO 2 and CO was proved to be responsible for the enhanced activity of Sigma-Aldrich.Cu disk (99.995%, φ 50.8 × 3, mm) was purchased from Zhongnuo Advanced Material. 13CO 2 (99.99%) was purchased from Wuhan Newradar special gas Co.Ltd..The deionized water was produced using a Millipore Milli-Q grade with a resistivity of 18.2 MΩ cm.All the chemicals were used as received without any further puri cation.

Preparation for the GDE of Cu
In the typical preparation for the GDE of Cu, Cu disk and PTFE frame were used as target material and substrates, respectively.The sputtering was conducted via direct current magneton-sputtering methods with sputtering current of 60 mA at room temperature under Ar atmosphere (2.4 × 10 -3 mbar, 20 SCCM) for 1 h.
Treatment of the GDE with Carbon black 4.0 mg of carbon black was dispersed in the mixed solution containing 1.5 mL of EtOH, 0.4 mL of H 2 O, and 0.1 mL of Na on solution via sonication for 30 min.Typically, 0.2 mL of dispersion was uniformly sprayed on the GDE with geometric area of 2.0 × 2.0 cm 2 .The prepared GDE were used as working electrodes.

Electrochemical measurements
The catalytic performances for the GDE of Cu and four Cu-CoPc tandem catalysts were carried out in a ow-cell system, in which anodic and cathodic electrolyte were wriggled by peristaltic pump with the ow rate of 5.0 mL min -1 and separated by Na on 115 membrane.The nickel foam and Ag/AgCl electrodes were used as the counter electrode and reference electrode, respectively.The chronopotentiometric electrolysis was controlled by an Autolab potentiostat/galvanostat (CHI1140E).All potentials were measured versus the Ag/AgCl reference electrode (vs Ag/AgCl) and converted to the reversible hydrogen electrode reference scale (vs RHE) on account of the equation: E (vs RHE) = E (vs Ag/AgCl) + 0.21 V + 0.0591 × pH.The applied potential was iR-corrected using voltage drop from the resistance of electrolyte.
During CO 2 electroreduction, the gaseous products were monitored by an online GC equipped with a ame ionization detector (FID) and thermal conductivity detector (TCD) once every ve minutes.The liquid products were detected by 1 H NMR. Speci cally, 0.4 mL of KOH solution was mixed with 0.1 mL of DMSO-d 6 .0.1 mL of 6.0 mM DSS solution was added as an internal standard.The 1 H NMR spectra were measured on a Bruker 400 MHz NMR spectrometer.
Pb UPD was conducted in a mixed solution containing 100 mM of HClO 4 , 0.5 mM of PbCl 2 , and 50 mM of KCl.The ve electrodes with geometric area of 1.0 × 1.0 cm 2 were held at -0.375 V vs Ag/AgCl for 10 min in the mixed solution.The stripping of Pb was conducted via cyclic voltammetry (CV) sweeping from -0.25 to 0.15 V vs RHE with scan rate of 10 mV s -1 .The calculations of ECSAs assume the coverage of adsorbed Pb monolayer over Cu catalysts with 2e transfer for the oxidation of a Pb atom.According to previously reported literature, the conversion factor was 310 μC cm -2 for Pb UPD 25 .
Instrumentations SEM images were obtained with a scanning electron microscope (SEM, JSM-6700F) operated at 5 kV.HAADF-STEM images were carried out on a JEOL ARM-200F eld-emission transmission electron microscope operating at an accelerating voltage of 200 kV using Mo-based TEM grids.Cu LMM Auger spectroscopy were performed on a VG ESCALAB MK II X-ray photoelectron spectrometer with an exciting source of Mg Kα = 1253.6eV.Contact angles were conducted on Attension Theta (Biolin Scienti c) with 1 M KOH as droplet.The gaseous products of CO 2 electroreduction were monitored by an online GC (SHIMADZU, GC-2014).Liquid products were examined on a Varian 400 MHz NMR spectrometer (Bruker AVANCE AV III 400).Isotopic gases were detected by GC-MS (Agilent Technologies, GC7890A and MS5973C).The X-ray absorption ne structure (XAFS) spectra of Cu K (E 0 = 8979 eV) edge was performed at BL14W1 beamline of Shanghai Synchrotron Radiation Facility (SSRF) operated at 3.5 GeV under "top-up" mode with a constant current of 220 mA.The XAFS data were recorded under transmission mode.The energy was calibrated accordingly to the absorption edge of pure Cu foil.Athena code was used to extract the data and t the pro les.For XANES, the experimental absorption coe cients as function of energies μ(E) were processed by background subtraction and normalization procedures, and reported as "normalized absorption" to compare with Cu foil (Cu 0 ) and Cu 2 O (Cu 1+ ).The Raman spectra of the samples were obtained using a LABRAMHR Raman spectrometer with an excitation wavelength of 532 nm.The in-situ Raman spectra towards CO 2 electroreduction were obtained using a LABRAMHR Raman spectrometer with an excitation wavelength of 785 nm.
Tandem mechanism for the GDE of Cu. a, FEC2H4 for the GDE of Cu using various co-feeding gases of CO2 with CO. b, The mass spectra of the C2H4-correlated fragments with m/z of 30, 27, and 26 during electrolyzing co-feeding gases of 13CO2 and CO with molar ratio of 92:8 for the GDE of Cu at applied j of 300 mA cm-2.c,d, FEC2H4 for the GDE of Cu using various co-feeding gases of CO2 with Ar (c) and CO with Ar (d) at applied j of 300 mA cm-2.Scale bar represents the standard deviations for the three separated tests.The applied potential was iR-corrected using voltage drop from the resistance of electrolyte.
Tandem mechanism for the GDE of Cu. a, FEC2H4 for the GDE of Cu using various co-feeding gases of CO2 with CO. b, The mass spectra of the C2H4-correlated fragments with m/z of 30, 27, and 26 during electrolyzing co-feeding gases of 13CO2 and CO with molar ratio of 92:8 for the GDE of Cu at applied j of 300 mA cm-2.c,d, FEC2H4 for the GDE of Cu using various co-feeding gases of CO2 with Ar (c) and CO with Ar (d) at applied j of 300 mA cm-2.Scale bar represents the standard deviations for the three separated tests.The applied potential was iR-corrected using voltage drop from the resistance of electrolyte.
Fig.1cshows the high-angle annular dark eld scanning transmission electron microscopy (HAADF-STEM) image of Cu nanocrystal.The interplanar spacings of 0.205 and 0.181 nm were assigned to (111) and (200) facets of face-centered-cubic (fcc) Cu, respectively.The single-crystal structure of Cu nanocrystal was identi ed by the selected area electron diffraction (SAED) pattern (Fig.1c, inset).Fig. 1d shows X-ray diffraction (XRD) patterns of PTFE and the GDE of Cu.Except for the characteristic peaks for PTFE, other peaks located at 43.3 o , 50.4 o , and 74.1 o for the GDE of Cu were indexed to the (111), (200), and (220) facets of fcc Cu (JCPDS No. 04-0836), respectively.We further investigated the valence state of Cu in the GDE of Cu.As shown in Fig.

C 2 H 4 according
to isotopic labeling experiments.In addition, the tandem mechanism was further veri ed via constructing tandem catalysts of CoPc on Cu, in which Cu-CoPc-2 exhibited one-fold enhancement for the activity of C 2 H 4 relative to Cu. In-situ Raman measurements further revealed that CO generated by CoPc increased the coverage of * CO on the surface of Cu, facilitating the process of C-C coupling.This work provided an insight into the tandem mechanism to design highly e cient catalysts towards CO 2 electroreduction into C 2+ products.MethodsChemicals and materialsPotassium hydroxide (KOH, >85%), Absolute ethanol (EtOH, 99.7%), Absolute methanol (MeOH, 99.5%), Perchloric acid (HClO 4 , 70%), Lead chloride (PbCl 2 , 99.9%), Potassium chloride (KCl, 99.8%), and Carbon black (Vulcan XC-72) were purchased from Sinopharm Chemical Reagent Co. Ltd.. Cobalt(II) phthalocyanine (CoPc, 97%), 1-Propanesulfonic acid 3-(trimethylsilyl) sodium salt (DSS), (Methyl sulfoxide)-d 6 (DMSO-d 6 ), Na on solution (5 wt%), and Na on 115 membrane were purchased from characterization for the GDE of Cu. a,b, SEM images for the GDE of Cu. c, HAADF-STEM image of Cu nanocrystal.The inset shows the corresponding SAED pattern.d, XRD patterns for PTFE and GDE of Cu. e, Cu LMM spectrum for the GDE of Cu. f, XANES spectra of Cu foil, Cu2O, and GDE of Cu. g, Contact angles of PTFE, GDE of Cu, and GDE of Cu with hydrophobic treatment.h,i, The SEM image (h) and EDX elemental mapping (i) of cross-sectional GDE of Cu with hydrophobic treatment.

Figure 5 In
Figure 5

Figure 5 In
Figure 5 enhanced activity of C 2 H 4 according to isotopic labeling experiments.We constructed a tandem model with cobalt phthalocyanine (CoPc) as CO-generating component on Cu to further verify the tandem mechanism.With the introduction of CoPc onto the surface of Cu, the partial current density for C 2 H 4

Table 1 .
To exclude the in uence of COcorrelated fragments ( 13 CO with m/z of 29 and CO with m/z of 28), we chose the mass spectrometry signals with m/z of 30, 27, and 26 to quantitatively analyze the ratios for 13 C 2 H 4 , 13 CH 2 CH 2 , and C 2 H 4 .
2, C 2 H 4 -correlated fragments were quantitatively analyzed by gas chromatography-mass spectrometer (GC-MS) measurements over the GDE of Cu.The calibrations and ratios of C 2 H 4 -correlated fragments were shown in Supplementary Fig.8and Supplementary