Research Article ChemSusChem doi.org/10.1002/cssc.202300344 www.chemsuschem.org Hydrothermal Fabrication of Carbon-Supported Oxide- Derived Copper Heterostructures: A Robust Catalyst System for Enhanced Electro-Reduction of CO2 to C2H4 Venkata S. R. K. Tandava,[a, b] Maria Chiara Spadaro,[c] Jordi Arbiol,[c, d] Sebastián Murcia-López,*[a] and Joan Ramón Morante[a, e] Anthropogenic CO2 can be converted to alternative fuels and at industrially relevant high current densities over 160 to value-added products by electrocatalytic routes. Copper-based 200 mA · cm 2. It is understood that the in-situ modification of catalysts are found to be the star materials for obtaining longer- CuO to Cu2O during the electrolysis is the driving force for the chain carbon compounds beyond 2e products. Herein, we highly selective conversion of CO2 to ethylene through the *CO report a facile hydrothermal fabrication of a highly robust intermediates at the onset potentials followed by C C coupling. electrocatalyst: in-situ grown heterostructures of plate-like The excellent distribution of Cu-based platelets on the carbon CuO Cu2O on carbon black. Simultaneous synthesis of copper- structure enables rapid electron transfer and enhanced catalytic carbon catalysts with varied amounts of copper was conducted efficiency. It is inferred that choosing the right composition of to determine the optimum blend. It is observed that the the catalyst by tuning the catalyst layer over the gas diffusion optimum ratio and structure have aided in achieving the state electrode can substantially affect the product selectivity and of art faradaic efficiency for ethylene > 45 % at 1.6 V vs. RHE promote reaching the potential industrial scale. Introduction demands and to curb the increasing amounts of CO2 in the atmosphere, developing functional advanced nanostructured The circular economy of CO2 is a hot issue amid global catalyst materials and systems for electrochemical conversion of warming. Nevertheless, the successful implementation of novel CO2 to alternative fuels or value-added products is the need of technologies still demands a lot of new research to orient into the hour.[5,6] The selectivity for the ECO2R is based on the reliable systems. Significant research has been carried out on applied voltages/currents and the number of electrons trans- thermal catalysis,[1] photocatalysis,[2] photoelectrocatalysis and ferred/exchanged depending on the catalyst material used. biocatalysis[3] for the effective mitigation of CO2 to various However, CO2 being a linear molecule is highly stable and its chemical feedstocks. Of all these methodologies, electrocata- low solubility in the electrolyte makes the overall process quite lytic CO2 reduction (ECO2R) is proven to be a promising strategy challenging. Issues like low product selectivity, high over- to mitigate CO2 emissions.[4] To meet the global energy potentials and the highly competitive Hydrogen Evolution reaction (HER) are to be considered to make ECO2R industrially viable. Owing to this demand, a wide variety of materials [a] V. S. R. K. Tandava, Dr. S. Murcia-López, Prof. Dr. J. R. Morante Catalonia Institute for Energy Research (IREC) including metals and metal oxides,[7] single-atom catalysts,[8] Jardins de les Dones de Negre 1, 08930, Sant Adrià de Besòs (Spain) MOF-derived catalysts,[9,10] and heteroatom-doped carbon-based E-mail: smurcia@irec.cat materials[11] have been tested for the selectivity towards various [b] V. S. R. K. Tandava Universitat Autonoma de Barcelona (UAB) value-added products like carbon monoxide (CO),[12] ethylene, Bellaterra 08193 (Spain) ethanol[13,14] and n-propanol.[15] [c] Dr. M. C. Spadaro, Prof. Dr. J. Arbiol Of all the various products, ethylene is the simplest C2 Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB product with a high market price thus remaining an attractive Bellaterra, Barcelona, Catalonia, 08193 (Spain) target for the reduction of CO2 emissions.[16] It is formed [d] Prof. Dr. J. Arbiol through multiple electron-proton transfer mechanisms (12 ICREA Pg. Lluís Companys 23, 08010, Barcelona, Catalonia (Spain) electrons) carrying a high energy density. Although several [e] Prof. Dr. J. R. Morante metallic catalysts were examined, extensive research has been Faculty of Physics done by deploying copper and copper-based materials and Universitat de Barcelona Barcelona, 08028, Catalonia (Spain) found to be exceptionally active and effective in the selective Supporting information for this article is available on the WWW under conversion of CO2 to greater than 2 electron products. Copper- https://doi.org/10.1002/cssc.202300344 based materials are found to possess adequate adsorption © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH. This is energy towards *CO intermediate[16,17] and have shown a unique an open access article under the terms of the Creative Commons Attribution ability in aiding C C coupling to two or more carbon products. Non-Commercial NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non- Significant studies have been made on various forms of copper commercial and no modifications or adaptations are made. including copper foils,[18] nanoparticles,[19] nanopores,[20] ChemSusChem 2023, 16, e202300344 (1 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 nanowires,[21] nanosheets,[22] and more recently oxide-derived composition of the catalyst layer over the GDE can fetch the species.[23] overall CO2 reduction operating at high current densities for Interestingly, oxide-derived heterostructures have shown a long hours and can make an industrially viable alternative. significant effect on CO dimerization through the C C coupling mechanism leading to C2H4 selectivity.[24] Also, the addition of oxides on Cu sites has been reported to alter the adsorption Results and Discussion behaviour of the reaction intermediates, thus tuning the reduction mechanism.[25] However, oxide-derived Cu catalysts Synthesis and structural characterisation of the catalyst are prone to reduce to metallic copper (Cu0) when working materials under reduction potentials or currents as the ones required to drive the ECO2R. Once they are reduced to metallic copper, it is We have carried out the synthesis of CuxO/C heterostructures in highly likely to notice that the catalysts tend to become inactive a two-step manner as described in the experimental section. In or to decrease their ability towards C C coupling, thereby the typical synthesis, C-Black was sonicated to widely disperse hampering the C2 product formation in a short span,[26] and following the addition of copper acetate precursor. The mixture even favouring the competitive HER. Therefore, the rational solution was heated to 60 °C before transferring to the hydro- design of highly stable oxide-derived copper-based catalysts is thermal autoclave which was maintained at 160 °C for 12 h. For required to achieve higher CO2 conversions operating at a comparative study, we have also synthesised different industrially relevant current densities and with greater stabil- combinations of Cu/C ratios by altering the copper content to ities. In the recent past, a strong focus is levied on carbon-based understand the best blend for the in-situ growth. When materials as efficient electrocatalysts for the ECO2R owing to compared with the physical addition of C-Black with the as- their intrinsic properties, such as high surface areas, excellent synthesised bare CuxO catalyst particles during the ink electrical conductivity, and high mechanical strength, largely formulations, the direct in-situ growth of oxide-derived copper facilitating the diffusion of CO2. Different types of carbon-based heterostructures showed a significant influence in the improved materials like pure carbon-derived catalysts,[27,28] hetero-atom- catalytic efficiency indicating the copious amounts of exposed doped carbon catalysts[29] and metal-carbon single-atom active sites on the C-Black facilitating the rapid electron transfer catalysts[30] have been explored to achieve high catalytic mechanism. A detailed synthesis protocol with materials and efficiencies. Moreover, carbon black (C-Black) has been used as methods of electrode fabrication is given in the experimental standard support in electrode formulations for electrolysis, methods section. batteries, and fuel cell applications.[31] Although several strat- Initial compositional and structural properties were exam- egies have been followed including direct support of electro- ined using XRD analysis of the as-prepared samples: bare CuxO, catalysts on carbon black,[32,33] the physical mixture of carbon CuxO/C1, CuxO/C2 and CuxO/C3 heterostructures, along with the and catalyst for ink formulations is the most extended Vulcan (C-Black). As shown in Figure 1a, the standard CuxO/C2 is approach[26,34] leading to non-optimal catalyst-to-carbon con- deemed to have the CuO tenorite monoclinic phase (JCPDS tact. 048-1548) at 35.5°, 38.7°, 38.9°, 48.7°, 53.3° and 66.2° corre- As a result, we propose the quasi-functionalisation of C- sponding to ( 111), (111), (200), ( 202), (020), ( 311) planes. Black through in-situ growth of oxide-derived copper hetero- The peaks located at 36.4°, 42.3°, 61.4° and 73.5° correspond to structures as alternative to having stronger adhesion, attaining (111), (200), (220) and (311) planes of Cu2O cuprite cubic phases high current densities, facilitating higher gas diffusion, and respectively (JCPDS 005-0667). On the other hand, samples of improving the catalytic efficiency with possible aid in the bare CuxO, CuxO/C1, CuxO/C2 and CuxO/C3 were analysed under electron transfer mechanism. the same conditions. The presence of metallic copper (JCPDS Herein, we report a study on Cu/C-based electrocatalysts for 004-0836) in the bare CuxO sample (prepared without any the efficient electro-reduction of CO2 to ethylene operating at addition of C-Black) along with dominant CuO and Cu2O phases high current densities. We present a facile hydrothermal route are seen. The sample CuxO/C1 revealed the sole presence of to design a robust electrocatalyst having C-Black supported CuO having low copper content, and the CuxO/C3 seemed to oxide-derived copper heterostructures. Initial chemical reduc- be very coherent with CuxO/C2 but with a higher content of tion of a copper precursor followed by in-situ growth of CuO Cu2O along with CuO. Table S1 shows the quantitative compar- and Cu2O heterostructures over C-Black was obtained. Different ison of CuO, Cu2O and metallic Cu, estimated from the XRD of nominal Cu/C ratios were synthesised to determine the all the blends. As seen, CuO content in the CuxO/C samples optimum blend and to attain the state-of-art conversion to decreases inversely with the initial ratio of Cu precursor to ethylene at industrially relevant high current densities. As- Vulcan, while the opposite occurs with the Cu2O phase. The synthesised electrocatalysts were deposited on gas diffusion presence of metallic Cu is only observed in the bare CuxO electrodes (GDEs) and evaluated in a flow electrolyser. We sample, in agreement with having more reduced species (Cu, elucidate the high conversion and selectivity towards ethylene Cu2O) in the materials with lower C-black content. This suggests as a factor of in-situ reduction of the derived CuO Cu2O that copper precursor is more easily reduced in the absence (or heterostructures to Cu2O, along with the blending with C-Black in the synthesis with lower content) of carbon. that helps to increase the conductivity and enhance the CO2 FTIR-ATR analysis confirmed the presence of CuO and Cu2O distribution. It is also understood that choosing the right in the CuxO/C2 heterostructures as shown in Figure 1b. Strong ChemSusChem 2023, 16, e202300344 (2 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Figure 1. (a) XRD of the as-prepared CuxO/C heterostructures; (b) ATR-FTIR spectrum of CuxO/C2 and comparison with Vulcan (Inset of CuxO/C2 representing Cu2O and CuO). adsorption bands appearing at 520 cm 1 are attributed to the in Cu O and Cu2O at 530.18 and 531.30 eV, along with a vibrations of Cu O from CuO structures[35] and the sharp peak prominent peak at 532.65 eV corresponding to the C O C at 612 cm 1 corresponds to the vibration of Cu O Cu originat- species on the Cu/C interface. Also, the high-resolution C 1s ing from Cu2O.[36] Strong brands at 3200–3500 cm 1 correspond spectrum (Figure 2c) shows the C C bonding at 284.76 eV from to hydroxyl groups and are observed in correlation with the C- C-Black. The presence of C O and C=O species are attributed to Black band spectrum. At 1560 cm 1, carboxyl functional groups 285.71 and 287.08 eV peaks, respectively. Similarly, we have are identified with higher intensity compared to the pure C- analysed the bare CuxO without any C-Black added. The Black. deconvoluted high-resolution spectrum of Cu 2p level in The nitrogen adsorption-desorption measurements were Figure 2d shows the presence of metallic copper in correlation performed to assess any changes in the surface area of copper with XRD. Metallic copper (Cu0) is identified, unlike the Cu2/C, heterostructures as a result of the incorporated C-Black. The along with Cu + and Cu2 + at 933.74 and 935.02 eV respectively, surface area is calculated using Brunauer-Emmett-Teller (BET) indicating Cu Cu2O CuO heterostructures in the absence of C- analysis and the resultant pore size distributions were analysed. Black. Also, considering the high-resolution deconvolution of O The BET adsorption plots and linear isotherms with the 1s, Figure 2e indicates the presence of O2 in CuO and Cu2O at corresponding pore size distributions shown in Figure S1a and 530.11 and 531.45 eV, with much lower content of C O C b clearly depict a type III isotherm. The relative surface area of species as expected, given the absence of C-Black in this CuxO/C2 (CuO Cu2O/C-Black) is around 57.3 m2 · g 1 whereas sample. Considering the C 1s spectrum Figure 2f, adventitious the one of the CuxO is only around 6.3 m2 · g 1 showing a 9-fold carbon derived from partial exposure to atmospheric conditions increase in the relative surface area. Thus, this analysis in turn during the synthesis is noticed as C C and C=O species at proves that the direct growth of supported copper on carbon 284.85 and 285.74 eV, respectively. Additional XPS analyses of black leads to heterostructures with higher surface area and CuxO/C1 and CuxO/C3 samples show similar trends to the XRD pore volumes with respect to CuxO particles. This increased area results in terms of Cu oxidation states, as depicted in the Cu 2p and porosity might facilitate the diffusion of active species (i. e., spectra in Figure S2. In the CuxO/C1 sample, Cu is mostly in fully CO2 and ions). oxidized form (Cu2 +), while in CuxO/C3, an important contribu- X-ray photoelectron spectroscopy (XPS) was used to inves- tion of Cu + is also observed. tigate the surface and compositions of the synthesised catalysts. Initial FE-SEM imaging of the as-synthesised samples in A comparative study has been done on the carbon-supported Figure 3 and Figure S3 reveal that oxide-derived Cu hetero- CuxO/C2 catalyst and as prepared bare CuxO. The obtained XPS structures were aligned over the C-Black. Interestingly, the spectrum confirmed the presence of the elements Cu, O and C samples with different compositions exhibited different mor- in correlation with the corresponding binding energies of Cu phologies while anchored over the C-Black, involving platelet to 2p, O 1s and C 1s orbitals. Figure 2a shows the high-resolution rod- and cuboid-like structures surrounded by carbon particles. spectrum of the core 2p level revealing the main characteristic The bare CuxO particles obtained in the absence of C-Black, peaks. Upon further deconvolution of the set of two peaks Figure 3a and Figure S3a and b show mixed morphologies like representing the main Cu 2p3/2 and 2p1/2 spin orbits, the rods, cubes, and rectangular structures with varied particle presence of Cu + in the Cu2O structures was confirmed by two sizes. CuxO/C2 sample (Figure 3c and Figure S3e and f) is found peaks at 933.13 and 953.01 eV. Similarly, peaks with the binding to have a more uniform morphology consisting mainly of energies 934.8 and 955.2 eV and corresponding to Cu2 + in CuO platelet-like structures along with a minor presence of rods heterostructures were found. On the other hand, the deconvo- particles, and a homogenous distribution of CuxO/C hetero- lution of O 1s, Figure 2b shows the typical peaks of lattice O2 structures. CuxO/C1 sample, with lower copper content (Fig- ChemSusChem 2023, 16, e202300344 (3 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Figure 2. High-resolution deconvoluted XPS spectra of CuxO/C2 and comparison with bare CuxO (a) Cu 2p; (b) O 1s; (c) C 1s spectra of Cu2/C; (d) Cu 2p; (e) O 1s; (f) C 1s spectra of bare Cu. and CuxO/C3 is similar and lower than the theoretical one as per the amount of Cu precursor. In fact, both samples depict similar Cu/C atomic ratios of 0.1, which ultimately implies that the differences in the performance between both materials are not associated with different Cu content. The presence of partial oxidized copper species surrounded by cloud-like carbon structures is evident from transmission electron microscopy (TEM) (as shown in Figure 4a and b) images. Low magnification TEM confirmed the presence of Cu- based heterostructures surrounded by C-based clouds. HAADF- STEM images in Figure 4c and d clearly show Cu heterostruc- tures with dominating platelet-like morphologies, which present various faceting. The power spectra (fast Fourier trans- forms) calculated on the HRTEM images obtained on the C- based “cloud-like” structures (see Figure 4e and f) show that Figure 3. FE-SEM images of as-prepared samples (a) bare CuxO (b) CuxO/C1 (c) CuxO/C2 (d) CuxO/C3. they present a periodicity corresponding to an interplanar distance of about 0.34 nm, in perfect agreement with a graph- itic-like crystal structure. The C-based structures are evidenced in red in the frequency-filtered map reported in Figure 4f. These ure 3b and Figure S3c and d) has exhibited similar morphology carbon nanostructures resemble graphitic onion-like rings. to CuxO/C2 but with lesser coverage of Cu. The CuxO/C3 blend, The HRTEM analyses performed on the Cu-based nano- with higher nominal copper content (Figure 3d and Figure S3g platelets suggest that their structure corresponds to the and h) has a drastic change in terms of morphology showing monoclinic tenorite CuO crystal phase, in complete good cubic and large octahedral crystals with heterogeneous particle agreement with our initial XRD studies. The observed CuO size. The EDX analysis of the different compositions in Fig- structures imaged in Figure 5a and b are oriented along the ure S4a–c revealed the presence of Cu, C and O with Cu [011] and [101] zone axes parallel to the electron beam. averaging around 25–40 %, O around 10 % and the rest by Electron energy loss spectroscopy (STEM-EELS) (Figure 5c and carbon. A comparison of elemental composition along with EDX d) was used to obtain the elemental composition of the sample, analysis is given in Table S3 along with the nominal Cu/C ratios. focusing the attention on C K-edge at 284 eV (green), O K-edge Interestingly, despite the higher nominal amount of Cu at 532 eV (blue) and Cu L-edge at 931 eV (red). The analysis precursor on the CuxO/C3 sample, the Cu content on CuxO/C2 performed over the platelet-like structures revealed the homo- ChemSusChem 2023, 16, e202300344 (4 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Figure 4. TEM and STEM-HAADF images of CuxO/C2: (a, b, e) low magnification TEM, (c, d) low magnification STEM-HAADF images, (f) high-resolution TEM combined with the frequency filtered map to evidence the graphitic structure (in red). Figure 5. HR-TEM and STEM-EELS images of CuxO/C2: (a, b) high-resolution TEM together with the detailed image and corresponding power spectra. (c, d) STEM-HAADF and corresponding STEM-EELS compositional analysis. genous distribution of Cu and O surrounded by carbon axis. STEM-EELS maps (Figure S5g) on C K-edge at 284 eV structures. Low magnification TEM of the bare CuxO also (green), O K-edge at 532 eV (blue) and Cu L-edge at 931 eV confirms the presence of Cu-based structures with sizes in the (red) indicate that the Cu and O are homogeneously distributed range of a few microns with varied morphologies like platelets, within the nanostructures, in good agreement with the rods, or cubes, as shown in Figure S5a and b. These large-sized structural analysis. particles are highlighted using the image contrast of HAADF- STEM as shown in Figure S5c and d. Detailed investigation using the power spectrum analysis of HRTEM (Figure S5e and f) Linear sweep voltammetry and ECSA analysis confirmed the presence of both CuO tenorite monoclinic crystal phase as on the left side, imaged along [011] and Cu2O cuprite Electrochemical double-layer capacitance measurements were cubic structure as on the right side, imaged along [11-1] zone performed to analyze the active surface area of the synthesised ChemSusChem 2023, 16, e202300344 (5 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 electrocatalysts. Figure S6 depicts the ECSA analysis along with configuration have been carried out: flow-by and flow-through double-layer capacitance CDL plots. ECSA analysis was per- operational modes. Electrolysis was performed via chronoam- formed in a standard three-electrode cell configuration. A glassy perometric measurements for at least 60 min at each applied carbon electrode with the catalyst ink deposited was used as potential ranging between 1.1 and 1.5 V vs. RHE. Gaseous the working electrode, a platinum mesh as the counter products were analysed and quantified using an online GC. The electrode and a Hg/Hg2SO4 electrode as the reference. A non- electrolysis was performed for 60 min with periodic injections faradaic region with the applied potential range of 0.4 to to quantify the gas products. 0.5 V vs. RHE was chosen, and cyclic voltammetry was The performed ECO2R fetched the potential-dependent performed at various scan rates in Ar-saturated 1 M KHCO3 selectivity of the main products as ethylene, carbon monoxide, aqueous solution. The plots of scan rates (mV · s 1) vs. current (I) hydrogen and trace amounts of methane and ethane. The are shown, and the slope represents the specific double-layer oxide-derived copper heterostructures are highly selective capacitance CDL. CuxO/C1 and CuxO/C2 are found to have CDL towards ethylene as observed during the gradual sweep values of 0.3 and 0.27 mF, respectively, higher than bare CuxO towards increasing cathodic potentials. All different composi- (0.11 mF), suggesting the presence of a higher number of active tions were evaluated and compared, and the faradaic efficien- sites. Though the sample with higher copper nominal content, cies of various catalysts for major gas products at different CuxO/C3, presents the highest capacitance (0.83 mF), the applied potentials are plotted in Figure 7. The respective gas morphology and the ratio of CuO to Cu2O seem to be the chromatograms of the analysed sample along with the limiting factors in its selectivity towards ethylene via ECO2R. calibration chromatogram are shown in Figure S8 indicating Linear sweep voltammetry studies (in Figure S7) were per- C2H4, CO and H2 as the major products while negligible amounts formed under CO2 and Ar-saturated electrolytes to study the of CH4 and C2H6 were noticed (< 0.5 % FE). The observed current behaviour of the electrocatalysts at different potentials. In densities were also represented in Figure 7d for all the samples general, all the Cu-based samples exhibit lower onset potentials measured. Figure S9 gives a detailed representation of corre- than C-Black. Although minor differences were observed in the sponding faradaic efficiencies and current densities for each LSV curves under CO2 and Ar flows, CuxO/C2 displays slightly analysed sample. Figure S9d comparatively shows the efficien- higher cathodic currents in the presence of CO2, thus confirm- cies with the different materials at the best operation ing its intrinsic higher activity towards ECO2R. On the contrary, conditions for ethylene generation along with the obtained both CuxO/C3 and bare CuxO show high currents with Ar at current densities. Comparatively, enhanced selectivity towards more cathodic potentials. More intense reduction peaks are C2H4 is seen with the CuxO/C2 sample, while the other samples observed in all the samples at less negative potentials (0 to led to much lower FE C2H4 values. At low potential values ( 1.1 0.2 VRHE) in the measurements under Ar flow, suggesting a to 1.3 V vs. RHE), CuxO/C1 exhibited only slightly lower lower tendency of CuxO species to be reduced in the presence efficiencies than CuxO/C2, although the values rapidly dropped of CO2. after further increasing the polarisation values, with a concom- itant increase in the H2 generation. This composition (CuxO/C1) resulted in an initial faradaic efficiency of 41 % towards CO, but Electrochemical reduction of CO2 in a flow electrolyser the lower number of available active sites (low Cu/C ratio) turned the catalyst layer inactive towards carbon products We have performed the electrochemical reduction of CO2 by within a short time. Similarly, the material with higher nominal deploying the synthesised CuxO/C heterostructures in an copper content, CuxO/C3, which is seen to have a completely electrochemical flow cell in a three-electrode configuration as different morphology and higher initial Cu2O compared to shown in Figure 6. We have assessed all the synthesised CuxO/C2, shows higher overpotential towards C2H4 formation, catalysts and made a comparative study to find the best blend with the FE C2H4 linearly increasing with working potential, and morphology. Two approaches to the gas-electrolyte while efficiency to CO is kept relatively constant between 20– Figure 6. Experimental setup of flow electrolyser for ECO2R. ChemSusChem 2023, 16, e202300344 (6 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Figure 7. Electrocatalytic performance of CO2 reduction of different samples. Faradaic efficiencies (FE %) of (a) C2H4, (b) CO, and (c) H2; (d) current densities at different potentials. 25 %. Therefore, despite the higher nominal Cu-content and the ences between bare CuxO and CuxO/C2 has been given in higher ECSA value in CuxO/C3, it is inferred that the morphology Table S2. and crystalline structure are not favouring the higher selectivity In general, flow-by mode has been considered to be more towards C2H4 over H2. It is understood that, even at higher suitable for industrial applications than flow-through config- applied potentials and currents, the morphology, and the ratio uration, based on certain advantages such as lower ohmic of initial oxide states of copper of the synthesised hetero- drops and higher stability attained by controlled perspiration structures play a crucial role in tuning the selectivity towards (provided flooding is avoided) and decreased salt formation.[37] the desired ethylene. We have also analysed bare CuxO along Therefore, an additional test was performed with the CuxO/C2 with the addition of C-Black (Vulcan) to understand the material by preparing an electrode on commercial carbon Toray efficiency of the synthesis and to see if the in-situ growth of with microporous layer (MPL) and using flow-by conditions. In CuO over C-Black is advantageous over the addition of physical Figure 8, we show the performance of the prepared CuxO/C2 mixture during the catalyst ink formulations. It is noted that blend along with flow-through and flow-by analyses. At 1.5 V, these catalysts are poorly selective towards C2H4 and only a slightly higher faradaic efficiency towards ethylene (FE C2H4) competitive HER is favoured, with faradaic efficiencies to H2 was achieved (36 %) at a current density of 160 mA · cm 2. When higher than 75 %. A table representing the productivity differ- a current density of 200 mA · cm 2 was continuously applied in flow-by mode, an exponential surge of FE C2H4 to 46 % was Figure 8. Electrocatalytic performance of CO2 reduction via flow-by (a) and flow-through (b) operation modes. ChemSusChem 2023, 16, e202300344 (7 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 achieved, while limiting the one of H2 to around 40 %. Although [001] zone axis parallel to the electron beam. To gain further slightly enhanced values are observed in the flow-by test, insights into the elemental distribution, STEM-EELS mapping intrinsic differences associated to the presence of MPL was carried out (Figure S15g and h) focusing the attention on C precludes exhaustive comparison. K-edge at 284 eV (green), O K-edge at 532 eV (blue) and Cu L- At the end of the electrolysis, the liquid products were edge at 931 eV (red). A good agreement with the structural analysed by taking a fraction of the catholyte and using GC analysis is observed as the Cu and O are homogeneously with a head space accessory with an auto-sampler. Trace distributed within the nanostructures. It is understood that the amounts of methanol (96 ppm), ethanol (1350 ppm) and n- C-based nanostructures are in good contact with the Cu2O propanol (390 ppm) were formed and accumulated in the nanoparticles. This summarises the whole mechanism involving catholyte, which accounts for insignificant FE values. A the reduction of the CuO heterostructures to Cu2O nano- chromatogram representing the major liquid products is shown particles, thereby acting as the catalytic active sites for the in Figure S10. An additional stability test was performed at the reduction of CO2. The post-electrolysis characterisation of the standard operating galvanostatic conditions in the flow-by catalysts indicates the presence of Cu2O (Cu +) as the final state mode with a constant applied current of 160 mA · cm 2 yielding of copper. However, in-situ minor formation of Cu0 during the a faradaic efficiency of over 38 % holding excellent stability of electrolysis is not completely excluded, with the possibility of more than 12 hours. Figure S11 shows the stability performance further re-oxidation prior to ex-situ analysis. in the aforementioned conditions. Figure S12 provides a From the obtained results and the characterisation studies, comparative analysis of the electrocatalytic performance of CO2 the derived Cu2O is found to be the main catalytic active site reduction of FE % of CO and C2H4 of different compositions for the highly selective conversion of CO2 to C2H4 at high prepared and analysed concerning the applied potentials. With current densities. Cu2O is partially formed during the synthesis CuxO/C electrocatalysts, CO is the favoured product noticed at of the materials but is also derived from the electrochemical lower polarisation values, while a selectivity shift towards C2H4 pre-reduction of CuO occurring under the initial steps. In fact, is observed with gradual cathodic sweep, indicating the unique LSV studies show slightly unfavourable Cu2 + /Cu + /Cu0 reduction ability of the synthesised electrocatalysts to promote dimeriza- steps in presence of CO2, in comparison to Ar atmosphere, tion reactions. Unlike this, the C2 selectivity with bare CuxO suggesting a certain stabilisation of oxidised forms. Besides the sample does not significantly change with the applied potential, crystallographic related aspects, mechanistic considerations and H2 remains to be the majoritarian product. already described in the literature regarding C2H4 formation and XRD post-electrolysis measurements in Figure S13a confirm its correlation with the initial generation of CO* intermediate, that CuO is reduced to Cu2O in all the compared samples. indicate that CO* needs to be formed and adsorbed on the Figure S13b depicts the change in the Cu phases of CuxO/C2 electrocatalyst surface in order to dimerise through a C C during the stages of electrolysis showing that the complete in- coupling mechanism.[17] As seen at low polarisation values, situ modification of CuO to Cu2O phase takes place rapidly even mostly CO is formed with CuxO/C1 and CuxO/C2, which seem to after preliminary polarisation at low voltages. It can be further have adequate adsorption energies to promote CO* formation understood that the as-seen reduction to Cu2O is the driving (and further desorption as CO product). This agrees with the factor for the selectivity along with the optimum Cu/C ratio of fact of initially observing mostly plate-like morphology on both active sites. Figure S14 shows the FE-SEM morphological samples. At higher polarisation, C2H4 generation is promoted comparison of the CuxO/C2 deposited on carbon paper: (a) as with CuxO/C2 at the expense of a concomitant decrease in CO deposited on the GDE, (b) after an initial pre-reduction step and formation, thus suggesting a correlation between the two (c) post-electrolysis. The morphology depicts platelet-like products. In the case of CuxO/C1, the lower Cu content seems structures on the as-deposited electrodes changing to small to be the reason for the poor selectivity towards C2H4 at high cube-like particles after the pre-reduction and in whole after polarisation values, although the comparison between CuxO/C2 the CO2 electrolysis. The post-electrolysis TEM and HR-TEM and CuxO/C3, with similar Cu/C ratios (as seen by EDX analysis), analysis of the electrocatalysts in Figure S15 (obtained from demonstrates that the differences are not only related to the Cu carefully scratching the GDE surface) revealed the morpholog- content. Moreover, even while having higher ECSA (and ical changes to cubic and rectangular-shaped particle agglom- achieving slightly higher current densities), CuxO/C3 depicts a erates as a result of the applied cathodic potentials. Modifica- different behaviour with FE to CO being constant along all the tions are seen in both structures and morphologies as polarisation conditions, and the one of C2H4 increasing linearly mentioned above. Figure S15 depicts the low magnification with the applied potential (although much lower than those of TEM images of Cu-based nanostructures possessing an average CuxO/C2). These results clearly suggest that the polymorphic size in the range of 100 mm and the HAADF-STEM studies; nature of CuxO/C3 has a detrimental effect on the ECO2R. The Figure S15c and d clearly distinguishes the C-based graphitic effect of C-black is not only in enhancing CO2 diffusion but also onion rings from the agglomerated Cu-based nanostructures. in attaining higher current densities. In fact, the evaluation of Further detailed investigation with the HR-TEM in Figure S15e bare CuxO (in physical mixture with C-black), reaching signifi- and f led to the identification of corresponding structural cantly lower currents than the Cux/C materials, shreds of analysis of the Cu-based nanostructures. Power spectrum evidence that CuxO/C interaction obtained by the in-situ analysis evaluated that all the Cu-based nanostructures possess- synthetic methods is necessary to achieve higher operating ing Cu2O Cuprite-based cubic structures are imaged with the current. ChemSusChem 2023, 16, e202300344 (8 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Related synthetic approaches have been followed in other way for developing industrially viable alternatives to produce studies for the in-situ preparation of CuO/C structures including high value-added products. carbon black and other carbonaceous materials. For instance, Yang et al. prepared CuO-supported catalysts, demonstrating improved FE to C2H4 in correlation with CO2 uptake capacity Experimental Section attained with N-doped carbon materials.[33] Although evaluated under different conditions, comparatively, the CuO material Materials supported on XC-72 Vulcan, reached FE to C2H4 of 20 % at Vulcan XC-72 from Cabot Corp, copper (II) acetate monohydrate, current densities of 140 mA · mgCu 1, while in the present work NaOH from Alfa Aesar, cetyltrimethylammonium bromide (CTAB), we achieve 46 % with estimated current of 133 mA · mgCu 1. In a NaBH4 and KHCO3, Nafion, 2-propanol were obtained from Sigma- different approach, Wang et al. have developed CuOx/C materi- Aldrich. All the chemicals were of analytical grade and used directly als with mixed oxidation states based on a chlorine induced without any further purification. Toray Carbon Paper with and without MPL was purchased from The Fuel Cell Store and used transformation, achieving FE to C2H4 of 22 % at 60 mA · cm 2, directly without any modification. and demonstrating stabilisation of mixed Cu0/Cu + valence states for long operation times in H-cell.[28] In general, this strategy has been demonstrated to be an effective alternative Synthesis of CuO Cu2O/C-Black heterostructures to obtain electrocatalysts with state-of-the-art activities for C2 + Inspired by a previous report of Z. Zhang et al.,[38] a modified products, with scalable and relatively simple synthesis methods, approach was applied to obtain oxide-derived Cu/C-Black hetero- while also deploying catalysts with low metal content with structures. The schematic illustration of the synthesis protocol is respect to other approaches with pure Cu-derived materials. shown in Figure S16. As depicted, 40 mg of commercial Vulcan XC- Differently from those studies, we have developed in the 72 (C-Black) were dispersed in 30 mL of Milli-Q water by strong ultrasonication for 2 h resulting in a uniformly well-dispersed present work systems with larger area electrodes, implemented homogenous mixture 1. In another beaker, solution 2 was prepared in flow-cell devices with optimised flow configuration, thus by adding 200 mg of Cu(CH3COO)2 into 5 mL of MilliQ water and attaining higher efficiencies and productivities. stirring vigorously for over 30 min. 20 mg of NaOH and 20 mg of CTAB were added to solution 2. Both the solutions (1 and 2) were further mixed and NaBH4 was added to the above mix. The final Conclusions suspension was further stirred for 30 min in a hot water bath maintained at 60 °C. The contents were finally transferred to a Teflon-lined stainless-steel autoclave. The oven was programmed to Here, we demonstrated an easy and scalable method to operate at 160 °C for 12 h and allowed to cool naturally to room produce in-situ growth of oxide-derived copper-based hetero- temperature. The final product suspension named Cu2/C was structures over C-Black as a highly robust electrocatalyst for the cleaned by centrifugation with water followed by ethanol and effective conversion of CO2 to ethylene. At the interface of vacuum dried at 60 °C for 24 h. A similar method is followed for bare CuxO without having mixture 1. By keeping the amount of C- CuO Cu2O/C-Black deposited over the gas diffusion electrode Black (40 mg) constant, the quantity of Cu precursor was varied in (GDE), we attained a faradaic efficiency of 46 % at a current order to obtain samples with different CuxO/C ratios: CuxO/C1 density of 200 mA · cm 2 in an aqueous flow electrolyser in (100 mg of Cu precursor), CuxO/C2 (200 mg of Cu precursor) and neutral media and room temperature electrolysis. GDE-based CuxO/C3 (300 mg of Cu precursor). cathodes exhibited a stable performance for a minimum of 12 hours when sweeping through increasing cathodic poten- Physico-chemical characterisation tials after an initial pre-reduction at 0.5 V vs. RHE. Structural and compositional characterisation revealed that nominal Cu/C Structural evaluation of the synthesised heterostructures was ratios were determinant in the final structural and morpholog- recorded on a Bruker type X-Ray D8 diffractometer (Cu Kα radiation, λ = 1.5418 Å, 40 kV, 40 mA) with a scanning range of 2 theta from ical properties of the heterostructures. In the blending with 10° to 80° and step size of 0.025°/2 s. Morphology studies were optimum performance, we have confirmed the presence of done using a Scanning electron microscope (FE-SEM, Zeiss Auriga initial dominating CuO phases with platelet-like morphologies 60) and the composition of the synthesised heterostructures was having homogenous distribution over the C-Black. The effective studied by energy-dispersive X-ray spectroscopy (EDX, Oxford Inca reduction of CuO to Cu2O indicates the initial self-evolution of Energy). Further investigations were carried out using high- the catalysts at low polarisation conditions. The presence of C- resolution transmission electron microscopy (HR-TEM) using a field emission gun FEI Tecnai F20 microscope. High-angle annular Black had a cumulative effect in not only enabling the high darkfield (HAADF) scanning TEM (STEM) and electron energy loss local CO2 diffusion but also high current densities, while having spectroscopy (EELS) were performed in the Tecnai microscope impact in the formation of more oxidized Cu phases with more using a GATAN QUANTUM filter for in-depth morphological and homogeneous morphologies during the synthetic procedure. compositional analysis. X-ray photoelectron spectroscopic (XPS) This work helps in understanding the concept of carbon- studies were performed to examine the surface structural composi- supported metal oxide heterostructures synthesised by a facile tion of the as-prepared catalyst heterostructures using a PHI instrument model 5773 Multitechnique with A1 Kα radiation hydrothermal route as robust and highly efficient electro- (1486.6 eV). FTIR studies were performed using a Bruker FTIR-ATR. catalysts achieving high conversion and selectivity towards The relative surface area was studied using N2 nitrogen adsorption- ethylene operating at high current densities thus paving the desorption measurements by Brunauer-Emmett-Teller (BET) method using TriStar II 3020-Micromeritics equipment at liquid nitrogen ChemSusChem 2023, 16, e202300344 (9 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 temperature. Initially, the samples were degassed at 90 °C for 1 h The electrolyte pH was recorded before and after the electrolysis. and then at 250 °C overnight in a FlowPrep 060-Micromeritics. The During the electrolysis, as mentioned above the product stream is BET method was used to calculate the active surface area for a directly connected to the inlet of the GC and automatically injected relative pressure (P/P0) range of 0.05–0.025. to quantify the products. The Shimadzu GC is equipped with both the flame ionization detector (FID) and thermal conductivity detectors (TCD) with an HP-PLOT column and He as the carrier gas. Electrode fabrication Catholyte and anolyte were collected at the end of electrolysis and measured using the head space GC equipped with an autosampler. Catalyst inks were prepared by adding 20 mg of catalyst to 2 mL of As the liquid products obtained were negligible (less than 0.05 % of 2-propanol and 100 μL of 5 wt.% Nafion. The mixtures were FE), we propose these Cu-based heterostructures are prime sonicated over an hour and directly deposited via drop-casting on selective towards gas products. the GDEs (Toray Carbon Paper 060 with and without MPL) covering the whole surface with uniform loading of the catalyst ink. All the The Faradaic Efficiency (FE) for each specific product is calculated electrodes prepared have a geometric area of 2×2.5 cm2 exposed to using the following Equation (2) the electrolyte and CO2 interaction. The final loading of the electrocatalyst on the electrode, Toray Carbon paper is measured � � M�v�n �F by weighing the electrode before and after the drop-casting of the FE ½%� ¼ � 100 % (2) Jtotal catalyst ink. The final loading of the catalyst (Cu/C) on the GDE is found to be around 2.5 � 0.2 mg · cm 2. A commercial DSA (dimen- sionally stable anode) was directly used as the anode. Where M is the number of moles of the specific product; v is the CO2 flow; n is the number of electrons required/exchanged to form each product (12 e for ethylene); F is the Faraday’s constant; Jtotal is CO2 electrolyser the total current at each potential. A commercial electrochemical flow filter-press cell (Micro Flow Cell, Electrocell A/S), which was previously modified and reported in our Flow-by and flow through operation mode of CO2 works by Urbain et al.[12] was deployed as the CO2 electrolyser. The electrolyser schematic representation is shown in Figure 6. The cell is accom- panied by a titanium plate holding the GDE cathode coupled with The CO2 electrolyser was operated in two different modes. Initially, a commercial DSA. A leak-free Ag/AgCl (3.5 M KCl) reference a flow-through mode was employed where the CO2 gas is forced to electrode is mounted between the cathode and membrane pass through the GDE matrix, and the outlet gas is mixed with the chambers. An anionic exchange membrane (Fumasep FAB-PK-130) catholyte (Figure S17b). Additionally, flow-by configuration was also separates the cathode and anode compartments. Titanium current evaluated (Figure S17a), with CO2 entering the cell from the top collector and DSA act as the electrical contact with the Bio-Logic inlet to the gas compartment and flowing by the GDE. In this SP50 Potentiostat. The active exposed area of the cathode is limited mode, the diffused CO2 passes through the MPL of GDE interacting to 2×2.5 cm2 and the components are separated by EPDM (ethylene at the liquid-solid (electrolyte-catalyst) interface and thus leaves the propylene diene monomer) gaskets. A constant flow of CO2, system together with the gas products at the bottom of the cell optimised to 10 mL · min 1 after studying the efficiency of the without mixing with the catholyte. system at various flow rates is supplied to the electrolyser. Aqueous 1 M KHCO3 (pH = 8.4) solution is the common electrolyte used in all the measurements. Catholyte and anolyte are continuously recircu- lated at a constant flow of 20 mL · min 1 connected through Tygon Acknowledgements tubing using a peristaltic pump. V.S.R.K.T. has received funding from the European Union’s Horizon Electrolysis and product quantification 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 754397. This work has An electrochemical filter-press cell as shown in Figure 8 was used been done in the framework of the doctorate programme in to perform the electrolysis of CO2. Bio-Logic SP-50 Potentiostat Materials Science of the Autonomous University of Barcelona. We along with a power source was used for the electrochemical measurements. For every fresh electrode, a pre-reduction at 0.5 V thank Dr. M. Biset-Peiró and Dr. N.M. Carretero for their vs. RHE was carried out for 10 minutes before the actual electrolysis. contributions to this work. IREC and ICN2 acknowledge Generalitat Electrolysis was performed over a potential ranging from 1.1 V to de Catalunya for financial support through the CERCA Program. 1.5 V for at least one hour of polarisation at each applied The authors thank the support from NANOGEN (PID2020- potential. Using Equation (1), all potentials are corrected concerning 116093RB-C43) and CERES projects (PID2020-116093RB-C42), the reversible hydrogen electrode (RHE) scale, funded by MCIN/AEI/10.13039/501100011033/ and by “ERDF A way of making Europe”, by the “European Union”. ICN2 acknowl- E ðvs: RHEÞ ¼ E ðvs: Ag=AgClÞ þ 0:209 V þ 0:059 � pH (1) edges funding from Generalitat de Catalunya 2017 SGR 327. This study was also supported by MCIN with funding from European At each potential, the electrolysis was carried out for an hour and after knowing the electrolyser reached the steady state, the product Union NextGenerationEU (PRTR-C17.I1). ICN2 is supported by the gas stream was directly connected to the in-line gas chromatog- Severo Ochoa program from Spanish MCIN/AEI (Grant No.: raphy (GC) equipment (Shimadzu) followed by micro-GC (Varian CEX2021-001214-S). 490) with TCD and FID detectors to quantify all the gas products at 30-minute intervals. Current at the given potentials are obtained by considering an average value over time. ChemSusChem 2023, 16, e202300344 (10 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH 1864564x, 2023, 19, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/cssc.202300344 by Readcube (Labtiva Inc.), Wiley Online Library on [26/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article ChemSusChem doi.org/10.1002/cssc.202300344 Conflict of Interests [17] S. Nitopi, E. Bertheussen, S. B. Scott, X. Liu, A. K. Engstfeld, S. Horch, B. Seger, I. E. L. Stephens, K. Chan, C. Hahn, J. K. Nørskov, T. F. Jaramillo, I. Chorkendorff, Chem. Rev. 2019, 119, 7610–7672. The authors declare no conflict of interest. [18] S. Asperti, R. Hendrikx, Y. Gonzalez-Garcia, R. Kortlever, ChemCatChem 2022, 14, e202200540. [19] I. Merino-Garcia, J. Albo, A. Irabien, Nanotechnology 2018, 29, 014001. [20] Y. Peng, T. Wu, L. Sun, J. M. V. Nsanzimana, A. C. Fisher, X. Wang, ACS Data Availability Statement Appl. Mater. Interfaces 2017, 9, 32782–32789. [21] M. Ma, K. Djanashvili, W. A. Smith, Angew. Chem. Int. Ed. 2016, 55, 6680– 6684. The data that support the findings of this study are available [22] B. Zhang, J. Zhang, M. Hua, Q. Wan, Z. Su, X. Tan, L. Liu, F. Zhang, G. from the corresponding author upon reasonable request. Chen, D. Tan, X. Cheng, B. Han, L. Zheng, G. Mo, J. Am. Chem. Soc. 2020, 142, 13606–13613. [23] A. H. Shah, Y. Wang, S. Hussain, M. B. Akbar, A. R. Woldu, X. Zhang, T. Keywords: electrocatalytic CO2 reduction · oxide-derived He, Phys. Chem. Chem. Phys. 2020, 22, 2046–2053. copper · gas diffusion electrodes · in-situ modification · [24] S. Y. Lee, H. Jung, N.-K. Kim, H.-S. Oh, B. K. Min, Y. J. Hwang, J. Am. Chem. Soc. 2018, 140, 8681–8689. ethylene [25] X. Wei, Z. Yin, K. Lyu, Z. Li, J. Gong, G. Wang, L. Xiao, J. Lu, L. Zhuang, ACS Catal. 2020, 10, 4103–4111. [26] J. Kim, W. Choi, J. W. Park, C. Kim, M. Kim, H. Song, J. Am. Chem. Soc. [1] M. D. Porosoff, B. Yan, J. G. Chen, Energy Environ. Sci. 2016, 9, 62–73. 2019, 141, 6986–6994. [2] G. Sahara, O. Ishitani, Inorg. Chem. 2015, 54, 5096–5104. [27] C. Jia, K. Dastafkan, W. Ren, W. Yang, C. Zhao, Sustain. Energy Fuels 2019, [3] N. S. Weliwatte, S. D. Minteer, Joule 2021, 5, 2564–2592. 3, 2890–2906. [4] M. G. Kibria, J. P. Edwards, C. M. Gabardo, C. Dinh, A. Seifitokaldani, D. [28] X. Wang, M. Miao, B. Tang, H. Duan, F. Zhu, H. Zhang, X. Zhang, W. J. Sinton, E. H. Sargent, Adv. Mater. 2019, 31, 1807166. Yin, Y. Fu, Nano Res. 2023, DOI: 10.1007/s12274-023-5554-9. [5] J. Rogelj, G. Luderer, R. C. Pietzcker, E. Kriegler, M. Schaeffer, V. Krey, K. [29] M. Zheng, P. Wang, X. Zhi, K. Yang, Y. Jiao, J. Duan, Y. Zheng, S.-Z. Qiao, Riahi, Nat. Clim. Change 2015, 5, 519–527. J. Am. Chem. Soc. 2022, 144, 14936–14944. [6] S. Fuss, J. G. Canadell, G. P. Peters, M. Tavoni, R. M. Andrew, P. Ciais, R. B. [30] J. Zhang, W. Cai, F. X. Hu, H. Yang, B. Liu, Chem. Sci. 2012, 12, 6800– Jackson, C. D. Jones, F. Kraxner, N. Nakicenovic, C. le Quéré, M. R. 6819. Raupach, A. Sharifi, P. Smith, Y. Yamagata, Nat. Clim. Change 2014, 4, [31] S. Khodabakhshi, P. F. Fulvio, E. Andreoli, Carbon 2020, 162, 604–649. 850–853. [32] E. H. Dias, G. T. S. T. Da Silva, J. C. Da Cruz, C. Ribeiro, ChemElectroChem [7] D. Ren, B. S.-H. Ang, B. S. Yeo, ACS Catal. 2016, 6, 8239–8247. 2022, 9, e202200206. [8] J. Liu, Y. Cai, R. Song, S. Ding, Z. Lyu, Y.-C. Chang, H. Tian, X. Zhang, D. [33] H. J. Yang, H. Yang, Y. H. Hong, P. Y. Zhang, T. Wang, L. N. Chen, F. Y. Du, W. Zhu, Y. Zhou, Y. Lin, Mater. Today 2021, 48, 95–114. Zhang, Q. H. Wu, N. Tian, Z. Y. Zhou, S. G. Sun, ChemSusChem 2018, 11, [9] Z. H. Xhao, H. L. Zhu, J. R. Huang, P. Q. Liao, X. M. Chen, ACS Catal. 2022, 881–887. 12, 7986–7993. [34] Y. C. Tan, K. B. Lee, H. Song, J. Oh, Joule 2020, 4, 1104–1120. [10] H. L. Zhu, H. Y. Chen, Y. X. Han, Z. H. Zhao, P. Q. Liao, X. M. Chen, J. Am. [35] L. Zheng, X. Liu, Mater. Lett. 2007, 61, 2222–2226. Chem. Soc. 2022, 144, 13319–13326. [36] S. Deng, V. Tjoa, H. M. Fan, H. R. Tan, D. C. Sayle, M. Olivo, S. Mhaisalkar, [11] H. Cui, Y. Guo, L. Guo, L. Wang, Z. Zhou, Z. Peng, J. Mater. Chem. A 2018, J. Wei, C. H. Sow, J. Am. Chem. Soc. 2012, 134, 4905–4917. 6, 18782–18793. [37] M. Duarte, B. De Mot, J. Hereijgers, T. Breugelmans, ChemElectroChem [12] F. Urbain, P. Tang, N. M. Carretero, T. Andreu, J. Arbiol, J. R. Morante, 2019, 6, 5596–5602. ACS Appl. Mater. Interfaces 2018, 10, 43650–43660. [38] Z. Zhang, L. Sun, Z. Wu, Y. Liu, S. Li, New J. Chem. 2020, 44, 6420–6427. [13] S. Ma, M. Sadakiyo, R. Luo, M. Heima, M. Yamauchi, P. J. A. Kenis, J. Power Sources 2016, 301, 219–228. [14] C. M. Gabardo, C. P. O’Brien, J. P. Edwards, C. McCallum, Y. Xu, C.-T. Dinh, J. Li, E. H. Sargent, D. Sinton, Joule 2019, 3, 2777–2791. [15] G. Wu, Y. Song, Q. Zheng, C. Long, T. Fan, Z. Yang, X. Huang, Q. Li, Y. Manuscript received: March 6, 2023 Sun, L. Zuo, S. Lei, Z. Tang, Adv. Energy Mater. 2022, 12, 2202054. Revised manuscript received: June 7, 2023 [16] Y. Hori, Electrochemical CO2 Reduction on Metal Electrodes. In: Modern Accepted manuscript online: June 12, 2023 Aspects of Electrochemistry, 42, Springer, New York, NY 2008, p. 89–189. Version of record online: August 1, 2023 ChemSusChem 2023, 16, e202300344 (11 of 11) © 2023 The Authors. ChemSusChem published by Wiley-VCH GmbH