Tailoring Microenvironments and In Situ Transformations of Cu Catalysts for Selective and Stable Electrosynthesis of Multicarbon Products

Electrochemical CO2 reduction is of tremendous interest for storing chemical energy from renewable sources while reducing CO2 emissions. While copper is one of the most effective catalysts, it suffers from low selectivity and limited long‐term durability. Here, these limitations are overcome by engineering Nafion coatings on CuO nanoparticle‐based catalysts supported on glassy carbon. By tuning the Nafion thickness and internal structure, it is shown that both the selectivity to multicarbon (C2+) products and long‐term stability can be dramatically enhanced. Optimized catalyst layers reach Faradaic efficiencies for C2+ products of 86% during long‐term testing for 200 h, with no evidence for performance degradation. Indeed, the C2+ Faradaic efficiency increases during testing, which is attributed to favorable in situ electrochemical fragmentation of catalytic nanoparticles. Finally, the optimized Nafion/Cu catalytic coatings are utilized to create scalable membrane electrode assemblies for CO2 electrolysis, yielding significantly enhanced C2H4 selectivity (≈58%) and activity at technologically‐relevant currents of 1–2 A. These results highlight the potential for creating multi‐functional Nafion coatings on CO2 reduction catalysts to favorably tune the reaction environment, while also promoting in situ transformations to active and selective nanoscale structures and morphologies, not just on model surfaces but also in state‐of‐the‐art gas diffusion electrodes.


Introduction
Electrochemical CO 2 reduction (CO 2 R) to value-added feedstock chemicals using electricity from renewable energy is of DOI: 10.1002/aenm.202303936considerable interest in the current transition to a carbon-neutral society. [1] number of metal-based electrocatalysts have been developed to catalyze CO 2 reduction reactions (CO 2 RR) in aqueous solutions.Among them, copper (Cu)based materials have received the most attention due to their unique ability to produce a range of multicarbon (C 2+ ) products, including ethylene (C 2 H 4 ), ethanol (C 2 H 5 OH), acetate (CH 3 COO − ), and n-propanol (C 3 H 7 OH). [2]Despite intensive efforts to optimize Cu-catalyzed CO 2 RR over the past decades, the low selectivity toward desired value-added chemicals and their poor long-term durability remain critical challenges. [3,4]vercoming these challenges is particularly difficult due to the complex reaction network associated with CO 2 electrolysis, in which subtle modifications of the catalytic system during continuous operation can lead to substantial changes in the product distribution, often resulting in increasing production of hydrogen at the cost of valuable C 2+ products.Indeed, both the structure of Cu-based catalysts and the local reaction environment near the catalyst surface have significant impacts on selective electrosynthesis of C 2+ products in CO 2 RR. [2]Flat Cu foils tend to favor generation of single carbon products, especially methane (CH 4 ), rather than multicarbon products. [5]In contrast, rough Cu structures, often formed via the reduction of copper oxides or the electrochemical reconstruction of initially smooth surfaces, display particularly high selectivities toward C 2+ products.Such surfaces are characterized by an abundance of undercoordinated sites that can strongly bind important intermediates, such as *CO, thereby enhancing C-C coupling. [6,7]Moreover, the C 2+ selectivity can be tuned by varying the size and shape/facet of Cu nanostructures.For example, a high Faradaic efficiency (FE) of 41% for C 2 H 4 has been demonstrated on Cu nanocubes with an optimal size of 44 nm. [8]n addition to the engineering of catalyst structures and morphologies, C 2+ selectivity can be further enhanced by modifying the reaction microenvironment, including the electrolyte cation, pH, and molar ratio of CO 2 /H 2 O in the vicinity of the Cu surface.For example, compared to larger hydrated cations of Li + , Na + , and K + , smaller hydrated ions of Cs + and Rb + induce a larger surface charge density and stronger interfacial electric field, which enhances adsorption of CO 2 and C 2+ intermediates, thereby facilitating C 2+ production. [9]Furthermore, recent studies have shown that adjusting the local pH and CO 2 /H 2 O availability by tailoring organic coatings (binders or ion conductive polymers, ionomers) on Cu also influences the activity and selectivity of CO 2 RR.12] In another example, a double layer of both anion-and cationconducting ionomers was introduced on the Cu surface to regulate the water concentration and to increase local pH by trapping hydroxide ions (OH − ), which resulted in significant suppression of the hydrogen evolution reaction (HER) and a record high 90% Faradaic efficiency of C 2+ products. [13]longside product selectivity, the poor long-term durability of Cu-based catalysts introduces another significant barrier to sustainable generation of C 2+ products from CO 2 R. Indeed, bare Cubased catalysts have been shown to be unstable when subjected to continuous CO 2 electrolysis.Severe structural and morphological changes of catalyst layers, including their delamination from substrates, the agglomeration of catalyst particles, and facet reconstruction, lead to degradation of catalytic performance. [14]herefore, it is of crucial importance to develop strategies to improve the durability of Cu-based catalysts for selective synthesis of C 2+ products.
[18][19][20][21][22][23] Recently, we reported that the activity, selectivity, and stability of CO 2 R can be drastically impacted by engineering the internal structure and properties of Nafion ionomer coatings on oxidederived Cu catalysts supported on carbon paper. [24]Importantly, thin layers of Nafion can beneficially stabilize the catalyst and create favorable microenvironments for CO 2 R, but can also lead to the blocking of active sites or introduce undesired mass transport limitations.In that previous study, we found that increasing the solvent dielectric constant by increasing the water content leads to thinner Nafion coatings with improved network connectivity, which results in an increased molar ratio [CO 2 ]/[H 2 O] and, thus, increased CO 2 availability and suppressed HER. [24]In addition, the reduced water content within the Nafion films may have a significant impact on the local pH, yielding more alkaline microenvironments that promote C-C coupling.While that study provided important insights into the function of Nafion binders for CO 2 R and demonstrated a generalizable approach to optimizing different reaction outcomes via modification of catalyst ink formulations, the complex microporous structure of the carbon paper support may have contributed to a broadened distribution of reaction microenvironments near its surface and deeper within its pores, which could limit the Faradaic efficiency to selected high value products.Therefore, in the present work, we first apply the Nafion/solvent engineering strategy to well-defined and planar glassy carbon substrates in H-type cells, showing that the resulting Nafion/CuO nanoparticle coatings yield significantly enhanced C 2+ product selectivity with exceptional long-term stability during sustained durability testing.In addition, we show that the optimized Nafion/Cu catalytic coatings can be translated to scalable gas-fed membrane electrode assemblies (MEAs) featuring enhanced production of C 2+ products under industryrelevant operating conditions.For the H-cell configuration, optimized catalyst layers on glassy carbon are shown to yield stable and highly selective electrosynthesis of C 2+ products, with maximum C 2 H 4 FE of 65% and total C 2+ FEs of 86% during long-term CO 2 electrolysis for at least 200 h.Remarkably, the selectivity of C 2+ products is found to increase during the 200 h performance test, which is attributed to in situ electrochemical fragmentation of catalyst particles and the favorable reaction microenvironment provided by Nafion coatings prepared from inks containing optimal Nafion contents and solvent compositions.This behavior stands in contrast to most previous reports of Cu-based CO 2 R catalysts, in which long-term operation frequently leads to increased hydrogen evolution at the cost of reduced CO 2 RR selectivity.Importantly, integration of these Nafion/Cu nanoparticle coatings into scalable gas-fed MEAs via spray coating of the optimized catalyst inks also results in highly selective and active production of C 2+ products, with the C 2 H 4 FE and partial current density peaking at 58% and 136 mA cm −2 , respectively.Together, these results reveal the importance of engineering multi-functional Nafion coatings that not only provide favorable microenvironments and stabilize electrochemically active interfaces against degradation, but also promote in situ transformations to desirable nanoscale structures and morphologies, thereby enabling highly selective and stable CO 2 R in H-cell and integrated gas diffusion electrode (GDE) configurations alike.

Effect of Catalyst Ink Composition on CO 2 Electrolysis
As a starting point for understanding and controlling the activity of Cu-based catalysts, we first compare their morphologies and CO 2 electrolysis characteristics in the presence and absence of Nafion ionomer.Using commercial CuO nanoparticles (NPs) as the model pre-catalyst, we utilized various ink compositions containing four different Nafion weight ratios (0, 0.005, 0.05, and 0.5 wt%) and two water volume fractions (0 and 50 vol%, with isopropanol as balance of solvent).Following preparation of the inks, working electrodes were fabricated by drop casting catalyst layers on glassy carbon substrates with a constant mass loading of 0.255 mg cm −2 .Throughout this work, the prepared Nafion coatings and catalyst layers are denoted by their volumetric water content in solvent-Nafion weight ratio in the form of, for example, 50 vol%-0.5wt%.
X-ray diffraction (XRD) data and scanning electron microscopy (SEM) images demonstrate that the pre-catalytic nanoparticles comprise pure monoclinic phase CuO and have an average size of ≈50 nm, respectively (Figures S1 and S2, Supporting Information).In the absence of Nafion, large aggregates of CuO NPs and extended regions of bare glassy carbon are observed, regardless of the water content in the solvent.In contrast, addition of Nafion into the catalyst inks leads to significantly suppressed particle aggregation and more uniform CuO NP coatings (Figure S2, Supporting Information).Notably, the homogeneity of the catalyst layers increases with increasing Nafion content in the ink, yielding the smoothest and most uniform films for those prepared with 0.5 wt% Nafion.This result is consistent with our prior findings, indicating the importance of Nafion in the catalyst ink for producing stable dispersions and suppressing NP agglomeration upon catalyst layer formation. [24]In addition, magnified SEM images for a fixed Nafion content of 0.25 wt% indicate the presence of thicker Nafion coatings around CuO NPs with decreasing water content in the solvent (Figure S3, Supporting Information).Complementary measurements of the double layer capacitances of these electrodes reveal a reduction of the electrochemically active surface area (ECSA) with decreasing water content (Figure S4, Supporting Information), suggesting the blocking of catalyst surfaces as the Nafion thickness increases.These observations are consistent with the known dependence of the internal Nafion structure on solvent polarity.In particular, the increasing solvent dielectric constant with increasing water fraction promotes strong interactions between the hydrophobic perfluorocarbon backbones of Nafion, resulting in thinner films possessing improved inner network connectivity. [14,18]In this regard, the presence of Nafion in the solvent ink is necessary to suppress CuO NP agglomeration and obtain uniform catalyst layers, whereas increasing water content in the solvent is beneficial for maintaining thin Nafion coatings with internal structures that permit facile mass transport without blocking surface active sites.
We next examine the CO 2 R activities and selectivities of working electrodes prepared from catalyst inks with different solvent and Nafion contents in CO 2 -saturated 0.1 m KHCO 3 aqueous solutions using an H-type cell.Gaseous products were analyzed by online gas chromatography during 1 h of chronoamperometric testing at each fixed potential.Remarkably, electrodes prepared with Nafion-free inks exhibited negligible activity for CO 2 R, with H 2 being the only major product detected.As shown in Figure 1a, the measured average FEs of H 2 were ≈100% for both Nafion-free electrodes, regardless of solvent composition, as measured at -1.0 V vs. the reversible hydrogen electrode (V RHE ).To understand this surprising behavior, we performed post-mortem SEM imaging, which revealed the complete absence of catalyst layers on the glassy carbon surfaces after sustained operation (Figure 1b,c).Thus, we conclude that poor adhesion of Nafion-free Cu NP films results in exfoliation of catalyst layers from substrates and, consequently, complete loss of CO 2 R activity.
In sharp contrast, electrodes prepared with Nafion-containing formulations exhibited much lower FEs (i.e., <45%) for H 2 production across a wide potential range, suggesting the possibility for appreciable generation of CO 2 R products (Figure 1d).Indeed, the minimum H 2 FE of <20% near -1.0 V RHE indicates significant suppression of HER upon addition of Nafion.Consistent with this observation, post-mortem analysis reveals that catalyst layers were preserved with no significant exposure of the underlying glassy carbon substrates after the complete set of chronoamperometric tests on all Nafion-containing electrodes (Figure 1e,f; Figure S5, Supporting Information).Such considerable stability variations between Nafion-free and Nafion-containing electrodes underscore the necessity of incorporating Nafion in catalyst inks to ensure robust electrodes with strong interaction between catalyst layers and substrates.Furthermore, we note that this behavior differs from previous experiments with Nafion-free CuO NPs on porous carbon paper substrates, [24] where significant catalyst restructuring was observed but catalytic CO 2 R activity was retained even in the absence of Nafion.This difference indicates the challenge of stabilizing catalytic NPs on planar surfaces, as well as the ability of porous supports to inhibit catalyst detachment due to physical confinement.Thus, we conclude that Nafion plays an especially important role in ensuring the durability of catalyst layers on planar supports, such as the glassy carbon substrates used here.
To quantify the dependence of CO 2 R performance on the catalyst ink composition, potential-dependent total current densities (TCDs), FEs, and the partial current density for individual products were analyzed during chronoamperometric testing of 5 h, as shown in Figures 2 and S6 (Supporting Information).Over-all, the TCDs increased with decreasing Nafion content in the catalyst ink (Figure 2a).The maximum TCDs of ≈-20 mA cm −2 near -1.05V RHE were achieved at the lowest Nafion content of 0.005 wt%, regardless of water content in solvent.However, for the highest Nafion content of 0.5 wt%, the electrode prepared with 50 vol% water exhibited significantly larger TCDs across the entire potential range than the analogous one prepared without water.We ascribe this enhanced TCD at higher water content to the larger electrochemically active surface areas of catalyst layers produced from high water content inks, as previously reported and discussed above. [24]he potential-dependent FEs of gaseous products from the set of cathodes prepared from different solvent ink compositions are plotted in Figure 2b-e and Figure S6 (Supporting Information).For all electrodes, C 2 H 4 and H 2 were detected as two major products.While HER was significantly suppressed near -1.0 V RHE , reaching a minimum H 2 FE of ≈16% for most electrodes, a high C 2 H 4 FE was observed near this working potential.In general, higher C 2 H 4 FEs and larger partial current densities of C 2 H 4 were observed on electrodes prepared from high water content solvents relative to those prepared with water-free formulations.Indeed, a maximum C 2 H 4 FE of ≈60% at -1.05 V RHE was observed from the electrode prepared from the 50 vol% H 2 O-0.5 wt% Nafion catalyst ink.However, this beneficial improvement of selectivity appears to come at the cost of reduced activity, with the partial current densities of C 2 H 4 decreasing with increasing Nafion content (Figure 2f).Nevertheless, the magnitude of the activity loss is reduced with increasing water content, thus motivating further exploration of higher-water-content formulations to improve the activity of producing C 2 H 4 or other C 2+ products, as described in the next section below.
Alongside C 2 H 4 and H 2 , CO and CH 4 were also detected as minor products, with increasing Nafion content leading to somewhat increasing CO FEs and decreasing CH 4 FEs across the investigated potential range (Figure 2c,d).In particular, at -1.05 V RHE , where the C 2 H 4 FE was maximized, the CO FEs were less than 1% on all electrodes, while CH 4 FEs were <1% on electrodes prepared from inks with the highest Nafion content (0.5 wt%).We note that the CH 4 FE has been reported to be an indicator for local pH in CO 2 electrolysis, as the production of one CH 4 molecule from CO 2 requires the consumption of 8 protons. [25]he extremely low CH 4 FE measured on electrodes prepared with the highest Nafion content implies that thicker Nafion coatings result in higher local pH, thus suppressing HER and promoting the selectivity of C 2+ products. [13,25,26]In addition, the overall increase of C 2 H 4 FE for films prepared with 50 vol% water compared to those prepared without water is consistent with the higher inner network connectivity afforded by the higher dielectric constant of the solvent, which has been shown to enhance the local CO 2 /H 2 O concentration and pH to favor C-C coupling reactions.
Overall, the catalyst ink composition-dependent CO 2 R performance characteristics on the glassy carbon substrate are consistent with prior findings regarding the role of Nafion structure and thickness on the local reaction environment. [24]However, important differences between CO 2 R on planar glassy carbon supports compared to previously reported findings on carbon paper supports are apparent. [24]In particular, use of glassy carbon leads to significantly greater suppression of H 2 and single carbon (C 1 ) product generation, along with a pronounced enhancement of C 2+ products, particularly C 2 H 4 .While this could indicate that the planar system provides more homogeneous reaction microenvironments that promote higher C 2+ selectivity, these results motivate a broader exploration of the substrate-specific trends of CO 2 R performance on catalyst ink composition, with the aim of identifying optimal formulations to achieve necessary combinations of selectivity and stability, as described below.

Optimized Nafion/Solvent Compositions for Enhanced Generation of C 2+ Products
As a next step, we systematically assessed the effect of catalyst ink composition on the selectivity of CO 2 R during longer CO 2 electrolysis experiments, with the aim of identifying the conditions needed to maximize selective generation of C 2+ products.For this purpose, a series of catalyst inks containing six Nafion weight fractions (0.005, 0.025, 0.05, 0.125, 0.25, and 0.5 wt%) and four solvent compositions (0, 25, 50, and 75 vol% H 2 O, with isopropanol as the balance) were employed to produce working electrodes.All electrodes were subjected to 5 h chronopotentiometry tests at a constant current density of -12.8 mA cm −2 and both gas and liquid products, including H 2 , CO, CH 4 , C 2 H 4 , formate, ethanol, acetate, and n-propanol, were quantified.The current density was fixed at -12.8 mA cm −2 , which corresponds to an operating condition that yields lowest HER activity and highest FEs for C 2 H 4 for the majority of electrodes described above (Figure 2).The resulting product distribution obtained from each of the 24 independent electrodes is shown in Figure S7 (Supporting Information).For all cases, the three major products were ethylene, ethanol, and hydrogen, with FEs that varied as a function of catalyst ink composition.All other analyzed compounds (CO, CH 4 , formate, acetate, n-propanol) represented minor contributions (<10%) to the overall product distribution.The total FE for all products was ≈100% in all cases, indicating the lack of significant quantities of undetected products or corrosion currents.
To further analyze the influence of catalyst ink composition on the selectivity of each product, the average FEs were individually compared in heatmaps (Figure 3a-c; Figure S8, Supporting Information).For most catalyst ink compositions, the H 2 FE was maintained at a low level of ≈25%.However, for the case of high Nafion contents of 0.25 and 0.5 wt%, HER was highly dependent on water content.For example, vigorous HER with a H 2 FE of ≈40% was measured for the 0 vol%-0.5 wt% sample, but was significantly reduced with increasing water content.Overall, the lowest H 2 FEs of 16% were recorded with the combination of relatively high Nafion content of 0.25 wt% and high water content of 50-75 vol% (Figure 3a).Accordingly, the selectivity toward CO 2 R was significantly enhanced in this catalyst ink composition range, with the FEs of C 2 H 4 increasing to ≈58% (Figure 3b).Similarly, the production of ethanol, another major C 2 product, was also promoted in the region of both high Nafion (0.25-0.5 wt%) and water content (25-75 vol%), with FEs reaching above 20% (Figure 3c).The FEs for other C 2+ products, such as acetate and n-propanol (denoted as PrOH), were below 5% on all electrodes and represented a minor contribution to the total CO 2 R product distribution (Figure S8f,h, Supporting Information).Thus, the production of C 2+ products was maximized with a FE of ≈84% and a partial current density of approximately -10 mA cm −2 on electrodes prepared with inks containing 0.25 wt% Nafion and >25 vol% water (Figure 3d,e).
In contrast to the case of C 2+ products, the selectivities of single carbon (C 1 ) products such as CO and CH 4 were very low on all electrodes, with FEs of <1.8% for CO and <3.7% for CH 4 (Figure S8b,c, Supporting Information).Formate was the only comparatively major C 1 product and its FE exhibited a strong dependence on Nafion content.Specifically, increasing Nafion content resulted in decreasing formate production regardless of water content, and the corresponding FE was reduced from a peak value of 7.6% at 0.005 wt% Nafion to 1% at 0.5 wt% Nafion (Figure S8d, Supporting Information).Considering the minimal FEs of both CO and CH 4 , the overall production of C 1 products was therefore predominantly influenced by Nafion content (Figure S9b, Supporting Information).Consequently, as shown in Figure S9d (Supporting Information), the ratio of C 2+ FE to C 1 FE (C 2+ /C 1 ) also displayed a strong dependence on Nafion content.
As discussed above, increased Nafion content leads to thicker Nafion coatings, suggesting that the competition between C 2+ and C 1 reaction pathways can be modulated by tuning the thickness of Nafion coatings.Indeed, it has been shown that thick Nafion films can raise local pH values by trapping hydroxide ions (OH − ) generated during CO 2 electrolysis and preventing buffering bicarbonate species from entering the local microenvironment, thereby enhancing C 2+ selectivity. [13]However, thicker ionomer coatings can also block active sites, as evidenced by electric double layer capacitance measurements (Figure S4, Supporting Information) and may also leads to inefficient mass transport at the catalyst/electrolyte interface.To better understand these competing effects on C 2+ FE, Figure 3f shows a heatmap of Nafion thickness as a function of catalyst ink composition.As expected, the thickness increases with increasing Nafion and decreasing water content.Comparison of the thickness heatmap to the product distribution heatmaps reveals a clear trend of increasing H 2 and decreasing C 2+ product generation with increasing Nafion thickness, indicating that thicknesses of >10 nm are detrimental.In contrast, while the yield of C 2+ products is enhanced for catalyst inks containing both high water and Nafion contents, resulting in Nafion films of ≈2.5-7.0 nm (Table S1, Supporting Information), no clear trend between thickness and electrocatalytic properties is observed.However, as previously reported, the increasing solvent dielectric constant with increasing water content also leads to improved network connectivity and reduced water uptake within the Nafion films, which is consistent with atomic force microscopy (AFM) analysis (Figure S10, Supporting Information).Thus, the C 2+ product partial current density peaks at -10.8 mA cm −2 , realized with an ink composition of 75 vol% H 2 O-0.25 wt% Nafion (Figure 3e), which yields a favorable balance of thickness and internal Nafion structure.
Finally, we note that the heatmap profiles for both C 2 H 4 and ethanol exhibit a striking similarity, which is reasonable considering that they share the same critical intermediate, *CO, in their respective reaction pathways.This aligns well with the observed trends seen in heatmaps for all C 2+ products (Figure S9, Supporting Information), where the hot spots (with high FEs) for each C 2+ product are consistently located in the regions with low CO FEs, suggesting efficient C-C coupling rather than release of CO.Overall, we find that the optimal catalyst ink compositions for not only selective CO 2 R but also enhanced production of C 2+ products involve a combination of high water content (>50 vol%) and relatively high Nafion content (i.e., 0.25 wt%).

Stabilized C 2+ Selectivity During Long-Term CO 2 Electrolysis
Having obtained a maximum C 2+ FE of 84% during 5 h of continuous CO 2 electrolysis using the working electrode prepared from a catalyst ink containing 75 vol% H 2 O-0.25 wt% Nafion, we subsequently assessed its durability during 200 h of chronopotentiometric CO 2 electrolysis at -12.8 mA cm −2 .As described in the Experimental Section, this long-term stability test was performed in five sequential experiments of ≈40 h duration, immediately following one another, in order to allow the electrolyte to be periodically refreshed.For comparison, a second electrode was prepared from a catalyst ink containing 0 vol% H 2 O-0.25 wt% Nafion and was tested under identical conditions.Immediately prior to the long-term electrolysis experiments, the potential-dependent performance characteristics of both electrodes were confirmed by performing 5 h chronoamperometric tests (Figure 4a-c).Compared to the 0 vol%-0.25 wt% electrode, the 75 vol%-0.25wt% electrode delivered higher TCDs over the potential range from -0.9 to -1.1 V RHE , as well as higher C 2 H 4 FEs and much lower H 2 FEs across the complete range of investigated potentials.These findings are consistent with the conclusions discussed above that high water content catalyst inks contribute to suppressed HER and enhanced generation of C 2+ products.
Throughout the entire course of the 200 h long-term durability test, the 75 vol% H 2 O-0.25 wt% Nafion electrode exhibited high FEs for CO 2 R (Figure 4e,f).The corresponding operating potential was also stable, near -1.10 V RHE , decreasing slightly to approx.-1.05 V RHE with time (Figure 4d).Surprisingly, the C 2 H 4 FE increased from ≈55% at the beginning of the chronopotentiometric experiment to ≈60% after 40 h of operation, eventually reaching a peak average FE of 65% during the last 30 h of testing.This represents an overall relative increase of the C 2 H 4 FE of ≈19% (≈10% absolute increase) during 200 h of sustained operation.The increased C 2 H 4 FE was accompanied by progressively decreased FEs for the production of both H 2 and ethanol, with the averaged FEs of H 2 decreasing from 17% to 13% and of ethanol decreasing from 20% to 17%.
Previous mechanistic studies have revealed that the selective production of C 2 H 4 and ethanol is determined by the C-C coupling process and the subsequent hydrodeoxygenation of C 2 intermediates, respectively. [27]][34][35] Given the pure CuO-derived catalysts used here, as well as the comparatively minor decrease of ethanol FE with operational time, we infer that the constant rise of C 2 H 4 selectivity during the 200 h chronopotentiometric test is primarily driven by enhanced C-C coupling rather than an increase of the hydrodeoxygenation efficiency.In addition, we note that the production of all other minor products, including CO, CH 4 , formate, acetate, and n-propanol, was maintained at minimal levels, accounting for FEs in the range of just 0.5% to 3%.The FE of C 2+ products hence exhibited a steady increase over the course of 200 h and reached a remarkable value of 86% by the conclusion of the experiment.As summarized in Table 1, this represents one of the highest C 2+ FEs and longest stabilities for long-term liquid phase aqueous CO 2 R reported to date.Importantly, the majority of studies report durability tests of Cuassisted CO 2 electrolysis for operational periods of less than 10 h due to the inevitable degradation of catalytic activity.In contrast, we find no evidence for degradation during 200 h of testing and, on the contrary, observe an increase of selectivity and decrease of operating potential during extended testing.
To better understand the remarkable combination of selectivity and stability achieved by these Nafion-coated Cu nanoparticle catalyst layers, we compare the long-term performance characteristics of the optimal 75 vol%-0.25wt% electrode with that of the 0 vol%-0.25 wt% electrode (Figure S11, Supporting Information).While the cathode formed from the 0 vol%-0.25 wt% ink composition also demonstrated reasonable stability, the FEs  S11b,c, Supporting Information).Furthermore, this electrode was characterized by a nearly 2× larger H 2 FE, which gradually increased from 28% to 31% during the long-term stability test (Figure S11b, Supporting Information).The FE of C 2+ products was ≈65%, which is more than 20% lower than that found for the 75 vol%-0.25wt% electrode as described above (Figure 4f; Figure S11c, Supporting Information).Importantly, the sole difference between these two electrodes is the water content of the solvent used in the catalyst ink, which is known to have a profound impact on the internal structure and thickness of the resulting Nafion coating.In particular, electrodes prepared from high water content solvents possess thinner Nafion coatings with improved inner network connectivity, resulting in significantly enhanced CO 2 /H 2 O local concentration ratios and greater OH − accumulation that favors CO 2 R over HER and promotes C-C coupling, respectively.In contrast, those prepared without water in the solvent tend to be thicker and are characterized by greater water uptake, resulting in lower CO 2 /H 2 O concentration ratios and increased HER.While detection of Nafion by transmission electron microscopy (TEM) after stability testing was challenging (Figure 5a,b) due to the ultrathin and amorphous nature of the coatings, complementary detection of fluorine by energy dispersive X-ray (EDX) spectroscopy verifies that Nafion is retained on both electrodes following sustained CO 2 electrolysis for 200 h (Figure S12, Supporting Information).
Although the favorable reaction microenvironment accounts for higher C 2+ selectivity on the electrode prepared with 75 vol% water compared to that prepared with 0 vol% water, the reason for the progressive enhancement of the FE of C 2+ products, especially C 2 H 4 , during the course of 200 h chronopotentiometric testing is not obvious.Here, we note that most previous studies on the stability of Cu-based catalysts report increasing HER and decreasing CO 2 RR selectivity during sustained operation, typically on the order of 10 h (Table 1).A notable exception to this can be found in the work of Jung et al., who reported the in situ electrochemical fragmentation of Cu 2 O NPs, with decreased particle size leading to a doubling of the FE of C 2 H 4 , from 27% to 57%. [41]Importantly, TEM analysis of catalysts from the two different electrodes indicates the presence of small particles (< 20 nm) in both cases (Figure 5a,b), suggesting that electrochemical fragmentation may proceed to some extent on both electrodes.To determine whether such a mechanism could be responsible for the increase of C 2+ FE with time for our Nafion-coated CuO nanoparticles, we collected SEM images of the 75 vol%-0.25wt% and the 0 vol%-0.25 wt% electrodes before and after 200 h of operation (Figure 5c,d).In both cases, the starting CuO NPs are well dispersed and are characterized by an average diameter of ≈50 nm (Figure S13, Supporting Information).While SEM indicates that this morphology and average size is largely preserved for the case of the 0 vol%-0.25 wt% electrode (Figure 5c), which exhibited increasing HER activity during sustained operation, a notable decrease of the average particle diameter is observed for the 75 vol%-0.25wt% cathode (Figure 5d), with a large fraction of particles possessing diameters below 20 nm.This finding suggests that a similar electrochemical fragmentation process may be responsible for the improved performance of the 75 vol%-0.25wt% electrode during 200 h of electrolysis.Furthermore, the comparatively low degree of electrochemical fragmentation for the 0 vol%-0.25 wt% electrode is consistent with its more common behavior, with increasing H 2 FE over a similar operational time.
Here, it is important to note that electrochemical fragmentation comprises two distinct processes: the breaking of Cu NPs into smaller particles and the growth of small Cu NPs through the re-deposition of dissolved Cu species.These processes can collectively contribute to the overall fragmentation behavior.To better understand these phenomena, we analyzed the dissolved Cu 2+ content in electrolytes as a function of electrolysis time using inductively coupled plasma mass spectrometry (ICP-MS).After the first cycle of ≈40 h, we detected a high Cu content for the bestperforming electrode (75 vol%-0.25 wt%), with a Cu 2+ concentration of 19.1 μg L −1 (Figure 5e).This concentration was nearly twice as large as the value of 9.8 μg L −1 obtained for the electrode prepared with the water-free formulation.Nevertheless, following continued operation, the dissolved Cu ion concentration dropped to lower levels of ≈5 μg L −1 Cu 2+ for both electrodes, indicating redeposition on the cathode surfaces.These results indicate that catalyst particles with Nafion coatings fabricated from high water content ink formulations undergo a greater degree of dissolution and redeposition at the beginning of electrolysis, resulting in the generation of smaller nanoparticles.It has been demonstrated that such fragmentation or reconstruction of Cu catalysts can produce abundant undercoordinated sites near the surface, functioning as active sites to promote C-C coupling and, thus, enhancing the selectivity C 2+ products, particularly C 2 H 4. [6,7,28,41,46] In the present case, such a mechanism can also explain the observed increase of CO 2 R performance during longterm operation for at least 200 h.It has been previously reported that such electrochemical fragmentation of Cu NPs typically occurs in the initial 5-6 h of operation and is closely related to three main factors, including the negative applied potential, the interaction between surface Cu atoms and intermediates in CO 2 RR, and the local reaction environment (e.g., electrolyte). [41]Although the specific mechanism by which the catalyst ink composition and associated Nafion properties impact the electrochemical fragmentation process is not yet known, it is possible that the thicker Nafion films with poor inner network connectivity generated from water-free inks inhibit fragmentation by reducing the interaction between surface Cu atoms and CO 2 R intermediates.While this hypothesis is consistent with the decreased electrochemically active surface area with reduced water content, additional in situ studies will be required to further clarify the complex interactions that govern fragmentation of hybrid Nafion/Cu catalyst layers.

Translation to Gas Diffusion Electrodes for Scalable CO 2 Electrolysis
Due to the low solubility of CO 2 in aqueous environments, CO 2 electrolysis is increasingly performed in flow cells or gas-fed MEAs incorporating GDEs, which enable significantly larger operational currents in scaled devices that are relevant for industrial applications.With this in mind, we next examined the general applicability of using optimized Nafion/solvent formulations to spray-coat Nafion/Cu nanoparticle catalyst layers onto GDEs and assessed their performance for CO 2 electrolysis in gas-fed MEAs.Based on the findings reported above, we chose to comparatively assess the performance characteristics of two such MEAs, derived from catalyst inks containing 75 vol% H 2 O-0.25 wt% Nafion and containing 0 vol% H 2 O-0.25 wt% Nafion.As shown in Figure 6a, the full cell voltages at the specified current densities (100, 150, 200, 300, and 400 mA cm −2 ) were similar in both cases, to within the experimental uncertainty.However, the high water content formulation yielded significantly higher partial current densities for C 2 H 4 (Figure 6b), which were associated with dramatically suppressed FEs for H 2 , CO, and CH 4 compared to the water-free ink formulation (Figure 6c,d).Quantitatively, the MEA produced with the optimal ink formulation reached a maximum C 2 H 4 FE of 57.5% at a total current of 1 A (200 mA cm −2 ), while the corresponding partial current density peaked at 136 mA cm −2 , measured at a total current of 1.5 A (300 mA cm −2 ).This represents a 1.8-fold enhancement in the C 2 H 4 yield rate compared to the maximum value (74 mA cm −2 ) obtained on the electrode produced from the water-free solvent formulation (Figure 6b-d).Thus, we conclude that Nafion in high water content catalyst inks can positively influence C 2+ selectivity, particularly C 2 H 4 , at industry-relevant current densities.As indicated by Table S2 (Supporting Information), the measured CO 2 R to C 2 H 4 performance characteristics reported here represent notable progress in scaling-up CO 2 electrolysis from proof-of-concept level (1 cm −2 ) to larger area MEA CO 2 electrolyzers (5 cm −2 ).While full optimization of such MEAs is outside the scope of the present work and is likely to also depend on a complex set of cell parameters within the complete cell, these results suggest that solvent-engineered Nafion/Cu nanoparticle catalyst layers can be applied to scalable CO 2 electrolysis systems operating at high current densities, including gas-fed MEAs.

Conclusion
In conclusion, we find that the exceptional CO 2 R performance of Nafion-coated Cu-based catalysts arises from a complex interplay of several beneficial and synergistic factors that can be tuned via the catalyst ink composition.In particular, appropriately formed Nafion coatings can strike a careful balance between providing enhanced particle adhesion and reduced agglomeration to stabilize long-term performance, while also enabling in situ morphological reconstructions to achieve more active fragmented particle nanostructures.At the same time, the Nafion coatings provide favorable reaction microenvironments for suppressing HER and promoting C-C coupling.Here, the combination of high Nafion content and high water fraction in the solvent of the catalyst ink yields Nafion coatings possessing high inner network connectivity without significant blocking of surface sites, which is known to enable enhanced CO 2 /H 2 O concentration ratios and pH values in the local reaction microenvironment.Compared to our previously reported carbon paper supports, planar glassy carbon supports yield significantly higher FEs for C 2+ products, which may be a consequence of the more homogeneous nanoscale structure.Overall, optimization of the Nafion content and the solvent composition of the catalyst ink yielded glassy carbon-supported CuO NP catalysts with Nafion coatings exhibiting remarkable FEs of 65% for C 2 H 4 and 84% for C 2+ products, along with excellent durability during 200 h of operation.Notably, the CO 2 R performance, especially for the production of C 2+ products, showed continuous improvement over the 200 h testing period, reaching a total C 2+ FE of 86% with no indications of performance degradation.This combination of high selectivity and durability is attributed to an increased loading of active sites for C-C coupling resulting from the electrochemical fragmentation of catalyst particles.
Despite the remarkable combination of selectivity and stability enabled by tailored Nafion coatings, we note that use of planar glassy carbon electrodes in an H-cell configuration introduces fundamental limitations to the maximum achievable current densities during CO 2 electrolysis.To overcome these restrictions, we extended this approach to MEA-based CO 2 electrolysis at industry-relevant current densities.In particular, we show that optimized catalyst ink formulations can be used to spray coat MEAs, resulting in scalable electrochemical CO 2 reduction systems featuring significantly enhanced C 2 H 4 selectivity and activity.Thus, our findings present a straightforward and effective approach to engineer catalyst layers for highly selective and stable CO 2 electrolysis, including in gas diffusion electrode-based reactors.

Experimental Section
Preparation of Nafion Coatings and Working Electrodes: Catalyst inks were prepared using commercial Nafion ionomer dispersions (5 wt% in alcohol-water mixture containing 45% water + 55% propanol and methanol, Sigma-Aldrich), which were first diluted into 100% isopropanol or isopropanol-water solutions to achieve Nafion dispersions with different defined Nafion weight ratios.For example, 2 μL of the commercial Nafion dispersion was added to a solvent comprising 999 μL deionized water and 999 μL isopropanol, followed by 30 min of ultrasonic treatment, to make a specific Nafion dispersion with 50 vol% H 2 O and 0.005 wt% Nafion.To complete the catalyst ink preparation, 10 mg of CuO powder (50 nm average diameter, Sigma-Aldrich) was added to each 2 mL Nafion dispersion.
To prepare cathodes, catalyst inks were deposited onto glassy carbon substrates (diameter of 5 mm, Gauss Union) with a working area of 0.196 cm 2 .The mass loading of CuO catalyst was fixed at 0.255 mg cm −2 , which was obtained by drop-casting a total of 10 μL (2 μL, 5 times) of the catalyst ink onto the substrate.After drop casting, the electrodes were baked at 80 °C in air for 10 min on a hot plate.The Nafion weight ratios studied in this work were 0 wt%, 0.005 wt%, 0.025 wt%, 0.05 wt%, 0.125 wt%, 0.25 wt% and 0.5 wt%, and the investigated water contents in the solvent were 0 vol%, 25 vol%, 50 vol%, and 75 vol% (with the remaining fraction being isopropanol).Throughout this work, the prepared Nafion coatings and catalyst layers are denoted by their volumetric water content-Nafion weight ratio in the form of, for example, 75 vol%-0.25wt%.
Physical and Chemical Characterization: Scanning electron microscopy (SEM, Zeiss NVision40 FIB-SEM) and transmission electron microscopy (TEM, FEI Titan ThemisG2 60-300, 300 kV) were used to analyze the morphologies of Nafion coatings and catalyst layers.For TEM analysis, the catalyst layers on the glassy carbon support were gently scratched away after 200 h of CO 2 electrolysis and transferred to a TEM grid (Cu grids from Plano, mesh 400 μm, 3 nm carbon film on lacey carbon).In a typical process, the catalyst coating was suspended in 10 μL ethanol, and 1 μL of the dispersed solution was dropped on the grid, followed by drying under an infrared lamp for ≈1 min.The collected grids were stored in a vacuum desiccator until TEM characterization.In addition, energy-dispersive X-ray (EDX) spectra were recorded using a four-quadrant Bruker SuperX chemiSTEM system to assess the presence of Nafion.
A Cu anode-equipped X-ray diffractometer (SmartLab, Rigaku) was utilized for crystallographic characterization in grazing incidence.The Cu content in electrolytes after long term CO 2 electrolysis was measured by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7900 ICP-MS) in the He-collision mode.The ICP-MS samples were prepared by diluting the electrolyte 5 times with a solution mixture of 2% HNO 3 and 0.5% HCl prepared from 70% HNO 3 and 35% HCl stock solutions (both CarlRoth, ROTIPURAN Supra, < 1ppb trace metals) and 18.2 MΩcm DIwater.The Cu calibration series was prepared from a certified multielement standard (TraceCERT, Transition Metal Mix 1) and the same mixture of 2% HNO 3 and 0.5% HCl.
Electrochemical Characterization: For CO 2 electrolysis in an H-type cell, an air-tight two-compartment H-type cell (C023-2, Gauss Union) with a Nafion 115 membrane separating the catholyte (45 mL) and anolyte (35 mL) was employed.All electrochemical measurements were performed using a potentiostat (Biologic VSP-300) connected in a threeelectrode configuration, with a Ag/AgCl (filled with 3 m KCl) electrode (Gauss Union) and a Pt foil (99.99%,Gauss Union) electrode as the reference electrode and counter electrode, respectively.The working electrode and the reference electrode were placed at a defined constant distance from one another for all measurements.
The series resistance (R), measured by electrochemical impedance spectroscopy and the current interrupt method, was compensated by 85%.All potentials were converted to the RHE scale using the equation below, in which i was the current recorded during CO 2 electrolysis: Prior to CO 2 electrolysis measurements, the 0.1 m KHCO 3 (99.95%,Sigma-Aldrich) aqueous electrolyte was continuously purged with CO 2 gas (99.999%,Linde) for 1 h, ensuring CO 2 saturation.Continuous CO 2 bubbling was maintained during all CO 2 electrolysis experiments at a constant flow rate of 40 sccm, controlled by a mass flow controller (MKS).The electrochemical activation of working electrodes was carried out by 20 cycles of cyclic voltammetry (CV) scans between -1.0 and -1.8 V versus the Ag/AgCl reference electrode.Chronopotentiometry tests were conducted for 5 h or 200 h, as indicated in the text, at a constant current density (-12.8 mA cm −2 ) and chronoamperometry tests were conducted for 1 h at each of the fixed potentials.For long-term chronopotentiometric CO 2 electrolysis over a period of 200 h, the electrolytes in both the cathodic and anodic compartments were refreshed after each cycle of ≈40 hours.Notably, to minimize the influence of oxygen in air on the catalyst layers, the working electrode was quickly transferred into degassed electrolytes when refreshing electrolytes.Afterwards, both compartments were purged with CO 2 gas for 1 h before starting the next cycle of durability testing.
For GDE-based CO 2 electrolysis, the MEA reactor (Sci-Materials Hub) consists of a titanium anode plate and a cathode plate with flow fields, along with insulating gaskets, integrated into a compression cell.The geometric area of each flow field is 5 cm 2 .An anion exchange membrane (PiperION, A40-HCO3, Versogen) was used to separate the anode and the cathode.The anode was made of iridium oxide coated on titanium mesh (IrO x /Ti mesh, Suzhou Borui Anodes Industry Inc.).The cathode was prepared by spray-coating of CuO-based catalyst inks (prepared with two different formulations containing 5 mg mL −1 CuO powder: 0 vol% H 2 O-0.25 wt% Nafion and 75 vol% H 2 O-0.25 wt% Nafion) on gas diffusion layers (2.3 cm × 2.3 cm carbon paper with micro-structures, JNT20A6L, JNTG), followed by baking at 80 °C for 15 min to achieve a constant loading of ≈1 mg cm −2 .The 0.1 m KHCO 3 aqueous anolyte was circulated at a flow rate of 30 mL min −1 using a peristaltic pump.Humidified CO 2 gas, realized by flowing CO 2 gas through a homemade humidifier at room temperature, was fed into the cathode flow field at a rate of 40 mL min −1 using a mass flow controller (MC-Series, Alicat Scientific).A potentiostat (PTC-05100EW, 5 V-100 A, IPS China Ltd.) was used to apply constant currents to the MEA and record the cell voltages accordingly.No iR compensation was performed for the electrochemical measurements in of the MEAs.Gas products were analyzed by online gas chromatography (GC2060, Shenzhen Zhentai Tech Ltd.) during CO 2 electrolysis at each current setpoint (0.5, 0.75, 1, 1.5, 2 A) for 1 h.The gas flow rate at the outlet of the gas chromatographer was measured for calculation of Faradaic efficiencies of gas products.
CO 2 Reduction Product Analysis: Gaseous products (hydrogen, carbon monoxide, methane, ethylene, and ethane) were quantified using an online gas chromatograph (GC, SRI 8610C) outfitted with both a thermal conductivity detector and a flame ionization detector.Ar (99.9999%,Linde) was used as the carrier gas.After CO 2 electrolysis, cathodic electrolytes were analyzed using 1 H nuclear magnetic resonance (NMR) spectroscopy (Bruker AV 500 HD) to identify the yield of liquid products, including methanol, formate, acetate, ethanol, and n-propanol.For this purpose, 400 μL of electrolyte and 100 μL of D 2 O with 200 ppm dimethyl sulfoxide as the internal standard were mixed, after which the concentration of each liquid product was measured and quantified using predefined calibration curves.The corresponding Faradaic efficiency (FE) was determined according to: where n is the number of electrons required to produce one molecule of the specified product, F is the Faraday constant of 96,485 C mol −1 , C is the measured quantity of product molecules in mols, and Q is the total number of electrons consumed by electrolysis.

Figure 1 .
Figure 1.Electrochemical and structural characterization of electrodes prepared from inks with different Nafion/solvent compositions.a-c) The average Faradaic efficiencies (FEs) of H 2 during 1 h chronoamperometric CO 2 electrolysis at -1.0 V vs. RHE on electrodes produced in the absence of Nafion, with ink compositions of 0 vol% H 2 O-0 wt% Nafion and 50 vol% H 2 O-0 wt% Nafion (a) and corresponding SEM images of the electrodes before and after CO 2 electrolysis (b, c).d-f) The potential-dependent H 2 FEs during chronoamperometric CO 2 electrolysis on electrodes produced from inks possessing two H 2 O volume fractions (0 and 50 vol%) and three Nafion weight fractions (0.005, 0.05, and 0.5 wt%) of Nafion (d), along with the corresponding SEM images of electrodes obtained before and after CO 2 electrolysis (e,f).

Figure 2 .
Figure 2. Effect of the composition of Nafion/solvent inks on potential-dependent CO 2 electrolysis.The catalytic performance of electrodes prepared from inks containing different Nafion (0.005, 0.05, and 0.5 wt%) and H 2 O/isopropanol (0 and 50 vol% H 2 O) contents, evaluated in terms of total current density a), as well as Faradaic efficiencies for H 2 b), CO c), CH 4 d), and C 2 H 4 e), as well as the partial current density of C 2 H 4 f).

Figure 3 .
Figure 3. Analysis of the impact of catalyst ink composition on selectivity, activity, and thickness of Nafion coatings.Heatmaps of Faradaic efficiency for H 2 a), C 2 H 4 b), ethanol c), and C 2+ products d), as well as partial current density for C 2+ products e), during chronopotentiometric CO 2 electrolysis at -12.8 mA cm −2 on glassy carbon compared to the thickness of Nafion coatings f), prepared from dispersions with varied Nafion contents and solvent compositions.

Figure 4 .
Figure 4. a-c) Potential-dependent performance of two electrodes prepared with 0.25 wt% Nafion and different water contents (0 and 75 vol%) in terms of total current densities (a) and Faradaic efficiencies (FEs) of gas products (b, c).d,e) Evaluation of durability in 200 h chronopotentiometric CO 2 electrolysis at -12.8 mA cm −2 on the electrode prepared from the ink containing 75 vol% H 2 O and 0.25 wt% Nafion.The dynamic change of cathodic potential (d) and FEs of gas products measured every 20 min over 200 h (e).f) The average FEs of all products in each cycle of ≈40 h.

Figure 5 .
Figure 5. Characterization of catalyst layers and electrolytes after 200 h of CO 2 electrolysis.TEM a,b) and SEM c,d) images of catalyst layers prepared from inks with 0 vol% H 2 O-0.25 wt% Nafion (a, c) and 75 vol% H 2 O-0.25 wt% Nafion (b, d).e) Concentrations of dissolved Cu ions in electrolytes collected before electrolysis and after each cycle of CO 2 electrolysis measured by ICP-MS.

Figure 6 .
Figure 6.Current-dependent performance characteristics of gas-fed MEAs (active area 5 cm 2 ) using 0.1 m KHCO 3 aqueous solution as the anolyte.The cathodes were prepared with inks of 0.25 wt% Nafion and different water contents (0 and 75 vol%).The resulting cells were evaluated in terms of full cell voltages a), the partial current density of C 2 H 4 b), and FEs of gas products c,d) examined in the galvanostatic mode at five different current densities, of 100, 150, 200, 300, and 400 mA cm −2 .

Table 1 .
Comparison of optimized production of C 2+ products on various Cu-based catalysts in neutral-pH aqueous electrolyte.