We synthesized the Ni-N-C catalyst through a high-temperature pyrolysis route using carbon black as the support material. Typically, nickel(II) acetate (3.6 mg, Macklin) and 1,10-phenanthroline monohydrate (8.5 mg, Macklin) were dissolved in anhydrous ethanol (2 mL), then ultrasonicated to obtain a homogeneous dispersion. Subsequently, Carbon black (10 mg) was introduced into the mixture and stirred at 80 °C for 8 h in a constant-temperature water bath. The resulting mixture was then placed in an oven and dried at 60 °C overnight to obtain a powder sample. Urea (Macklin) was subsequently introduced at a urea-to-sample mass ratio of 10:1, and the combined mixture was ground mechanically. The obtained powder was sequentially heat-treated at 600 °C for 2 h and then at 800 °C for 1 h. Thereafter, the product was washed with diluted hydrochloric acid (10 vol%) over 24 h, then collected by suction filtration and dried at ambient temperature for another 24 h to yield the final Ni-N-C catalyst.

Ni-N-C gas diffusion electrodes (GDEs) containing different ionomers were fabricated via ultrasonic spray deposition. The catalyst ink consisted of Ni-N-C catalyst (40 mg), D79 ionomer dispersion (35 µL, 25 wt%, Syensqo), and anhydrous ethanol (18 mL). This mixture was ultrasonicated for 30 min to ensure uniform dispersion, then sprayed onto a carbon paper substrate (3 × 3 cm², H15C13, Freudenberg) using an ultrasonic spray coater. The resulting GDE exhibited an ionomer content of 20 wt% and a catalyst loading of 2 mg cm^-2. GDEs containing different types of ionomers (D72, 25 wt%, Syensqo; D520, 5 wt%, DuPont; D521, 5 wt%, DuPont) and varying ionomer contents (5, 10, 20, 30 and 40 wt%) were fabricated following the same procedure. The anode consisted of a 2.5 × 2.5 cm² titanium felt with a thickness of 0.25 mm (coated with iridium, Fuershun Electronic Material Co., Jiangsu, China).

Scanning electron microscopy (SEM) imaging was conducted using a Hitachi S-4800 microscope at an accelerating voltage of 3.0 kV, and energy-dispersive X-ray spectroscopy (EDS) maps were collected using the same instrument. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging was performed on an FEI Tecnai G2 F20 microscope at an accelerating voltage of 200 kV. To investigate the microporous structure below 2 nm and CO₂ adsorption performance of the ionomers, the ionomer materials were characterized by CO₂ adsorption-desorption measurements. CO₂ adsorption-desorption measurements were performed using an ASAP 2460 analyzer (Micromeritics) at 0 °C. Prior to measurement, approximately 100 mg of ionomer was degassed at 120 °C under vacuum for 12 h. The CO₂ adsorption isotherms were recorded in the relative pressure (P/P₀) range of 10^-5 to 0.99. The specific surface area was determined using the Brunauer-Emmett-Teller (BET) method. Micropore size distribution was calculated using the density functional theory (DFT) model. CO₂ uptake capacity was determined from the adsorption amount at P/P₀ ≈ 0.95. Water uptake of the ionomer materials was measured gravimetrically. Ionomer membranes (approximately 2 × 2 cm²) were dried at 80 °C under vacuum for 24 h to obtain the dry weight (W_dry). The membranes were then immersed in deionized water at room temperature for 3 h to ensure complete hydration. The wet membranes were gently blotted with filter paper to remove surface water and immediately weighed (W_wet). The water uptake was calculated as: Water uptake (%) = (W_wet - W_dry)/W_dry × 100%. Each measurement was repeated three times, and the average values are reported.

The gaseous products (CO and H₂) collected after electrolysis were analyzed by a gas chromatograph (GC) equipped with a flame ionization detector (FID) for CO and a thermal conductivity detector (TCD) for H₂.

The FE values for CO and H₂ were determined using the following equation^[24]:

$$ \begin{equation} \begin{aligned} \mathrm{FE} = \mathrm{n_i}\mathrm{Z_i}\mathrm{F/Q}×100\% \end{aligned} \end{equation} $$

where Z_i represents the number of electrons transferred per mole of product i (CO or H₂), F represents the Faradaic constant (96,485 C mol^-1), n_i (mol) denotes the molar quantity of product i produced, and Q (C) refers to the total charge quantity.

The partial current density (j_i) for each product was determined using the following equation^[24]:

$$ \begin{equation} \begin{aligned} j_i = j_{total} × \mathrm{FE_i} \end{aligned} \end{equation} $$

where j_total denotes the overall current density applied to the system, and FE_i is the Faradaic efficiency of product i.

The following instruments were used in this study: scanning electron microscopy (S-4800, Hitachi, Japan); high-angle annular dark-field scanning transmission electron microscopy (Tecnai G2 F20, FEI, USA); CO₂ adsorption-desorption measurements (ASAP 2460, Micromeritics, USA); gas chromatography (8890, Agilent Technologies, USA); electrochemical workstation (PARSTAT MC, Princeton Applied Research, USA); in situ ATR-SEIRAS (INVENIO, Bruker, Germany); ultrasonic spray coater (SimCoat Coating System, Sono-Tek, USA). All instruments were operated according to the manufacturers' specifications.

This study investigates the influence of ionomers on CO₂RR in an MEA-type electrolyzer. In the MEA, humidified CO₂ from the cathode flow channel is transported through the gas diffusion layer (GDL) to the catalyst, while anodic water concurrently crosses the membrane to reach the reaction interface. At the three-phase interface, CO₂, protons, and electrons are consumed to generate CO and H₂, rendering CO₂ transport and interfacial water structure critical determinants of MEA performance. Ni-N-C is selected as the catalyst due to its low cost as a non-precious metal catalyst and its excellent CO selectivity, making it widely adopted in this field. During the preparation of GDE, the catalyst ink is uniformly sprayed onto the GDL, forming a thin film of ionomer that encapsulates the carbon support and catalyst surface [Figure 1A]. The resulting cathode and anode are closely assembled with the membrane to form an MEA structure. As illustrated in Figure 1B, the ionomer consists of a hydrophobic polytetrafluoroethylene (PTFE) main chain and pendant side chains carrying terminal hydrophilic sulfonic acid groups (-SO₃H), which establish CO₂ and water transport channels, respectively, at the three-phase interface. The SSC ionomer features a SSC structure, namely -O-CF₂-CF₂-SO₃H, whereas the LSC ionomer possesses a longer side chain, -O-CF₂-CF(CF₃)-O-CF₂-CF₂-SO₃H. The shorter side chain of the SSC ionomer enables a higher density of sulfonic acid groups per unit polymer mass, which is corroborated by its higher ion-exchange capacity. The SSC ionomer, with its hydrophilic groups positioned closer to the backbone, exhibits lower ionomer-H₂O adsorption energy, enhanced acid resistance, improved water uptake, and higher proton conductivity, thereby promoting the formation of a unique interfacial water structure under acidic conditions^[25]. To systematically investigate the influence of ionomer side chain structure, we selected two SSC ionomers (D72 and D79, both Aquivion-type) and two LSC ionomers (D520 and D521, both Nafion-type) for comparative study. As shown in Figure 1C, the as-prepared catalyst exhibits a spherical morphology typical of carbon black supports, with no obvious agglomeration and a uniform porous structure. Elemental mapping reveals distinct Ni and F signals, confirming the successful loading of Ni and ionomer onto the GDE surface. As shown in Figure 1D, HAADF-STEM images reveal numerous isolated bright spots on the carbon black, which correspond to uniformly dispersed single Ni atoms^[26,27]. TEM images further demonstrate a distinct thin ionomer film, 2-3 nm thick, coating the catalyst surface [Figure 1E]. It is noteworthy that the SEM, TEM, and contact angle measurements of the D79- and D520-modified GDEs exhibited comparable results [Figure 1 and Supplementary Figures 4-6], indicating that the side chain structure exerts no significant influence on the electrode morphology or macroscopic wettability. Consequently, the subsequent differences in electrochemical performance cannot be attributed to these two factors.

CO₂ absorption-desorption tests were conducted on SSC-D79 and LSC-D520 to probe their microporous structures. As shown in Figure 2A-C, D79 exhibits a cumulative pore area of 3.36 m² g^-1, significantly higher than that of D520 (0.053 m² g^-1), with its micropores predominantly centered at 0.56 and 0.75 nm. The shorter side chains of D79 result in less dense packing of the polymer backbones during the formation of ion transport channels compared to their long side chain ionomers, thereby generating a greater number of smaller micropores at the nanoscale. This abundant microporous structure establishes a denser and more efficient ion transport network. Upon water uptake, these micropores form interconnected nanochannels, providing additional pathways for ion transport and enabling ions to migrate more smoothly without being isolated or blocked. These micropores not only provide additional pathways for ion transport but also serve as confinement sites for CO₂ adsorption, explaining the higher CO₂ uptake observed for D79. Studies have shown that the ionic conductivity of D79 is 2.5 times that of D520^[28]. D79 exhibits a water uptake of 56.1%, markedly higher than the 14.2% observed for D520 [Figure 2D]. This advantage is attributable to its SSC architecture, which allows a greater number of hydrophilic sulfonic acid groups, key to water absorption, to be incorporated within the same polymer chain length, thereby facilitating the formation of a more favorable ionic network on the catalyst surface. It is worth noting that while D79 exhibits substantially higher water uptake than D520, the contact angles of the corresponding GDEs are similar [Supplementary Figure 4]. This apparent discrepancy arises because water uptake reflects the bulk absorption capacity of ionomer governed by the total sulfonic acid group density, whereas the contact angle probes the surface wettability of GDE, which is dominated by the hydrophobic PTFE backbone and carbon support rather than the sulfonic acid groups at the outermost surface layer.

We first evaluated the performance of GDEs modified with SSC ionomers (D72, D79) and LSC ionomers (D520, D521) in the current density range of 100-400 mA cm^-2, which covers the typical operating window for acidic CO₂RR MEA systems reported in the literature. The results are presented in Figure 3, where both D72 and D79 maintained a CO selectivity of nearly 90% over a current density range of 100-400 mA cm^-2 in 0.5 M K₂SO₄ electrolyte (pH 1, adjusted with H₂SO₄). At 400 mA cm^-2, D79 still achieved a FE_CO of 88.6%, while the LSC ionomers D520 and D521 exhibited FE_CO of 69.6% and 59.6%, respectively. The difference in FE_CO between D521 and D79 approached 30% [Supplementary Figure 7]. Meanwhile, D521 showed an FE_H2 of 38.3% at 400 mA cm^-2, indicating that the interfacial water structure formed by LSC ionomers fails to suppress water transport near the catalyst under high current densities, leading to a marked enhancement of the competing HER. In addition, the cell voltages of D520 and D521 were generally 0.1-0.2 V higher than those of D72 and D79, which can be attributed to the higher ionic conductivity per unit mass of SSC ionomers. A control GDE containing only PTFE (without any ionomer) was also tested, delivering an FE_CO of merely 12.9% at 300 mA cm^-2. In the absence of hydrophilic sulfonic acid groups to regulate the hydrogen-bonding network near the catalyst, the high proton concentration under acidic conditions directly aggravated HER. Beyond the intrinsic properties of the ionomers, their interaction with the Nafion membrane also plays a critical role in determining the overall MEA performance. Both D79 and D520 are PFSA ionomers based on the same PTFE backbone, which ensures good compatibility at the catalyst layer-membrane interface and minimizes interfacial resistance. However, different side-chain structures result in distinct water uptake and proton conductivity, which can affect the hydration state at the interface. The higher water uptake and proton conductivity of D79 facilitate proton transport from the membrane to the catalyst layer, maintaining a sufficient proton supply for CO₂RR under high-current-density operation. At the same time, the dense sulfonic acid groups in D79 create a strongly hydrogen-bonded interfacial water network that suppresses HER from free water, achieving a desirable balance between proton availability and HER suppression. Based on the FE and cell voltage according to Faraday's law, we calculated the CO production rate and energy consumption for both D79-modified and D520-modified GDEs [Supplementary Table 2]. For the D79-modified GDE, the CO production rate increases from 1.81 mmol h^-1 cm^-2 at 100 mA cm^-2 to 6.61 mmol h^-1 cm^-2 at 400 mA cm^-2, while the corresponding energy consumption ranges from 6.04 to 8.39 kWh kg^-1 CO. In contrast, the D520-modified GDE achieves a maximum CO production rate of only 5.20 mmol h^-1 cm^-2 at 400 mA cm^-2, with a significantly higher energy consumption of 11.04 kWh kg^-1 CO. The superior performance of D79 is attributed to its higher FE_CO and lower cell voltage, both of which originate from effective regulation of the interfacial water structure and enhanced ionic conductivity of the SSC ionomer. These results substantiate the superior energy efficiency of the SSC ionomer strategy, thereby corroborating its viability for practical acidic CO₂ electrolysis. The above results revealed that the performance gap between SSC and LSC ionomers widened substantially at current densities above 200 mA cm^-2, indicating that the influence of the interfacial water hydrogen-bonding network on electrochemical performance was more pronounced at high-current-density operation. This observation motivated us to extend testing to 500 mA cm^-2 to further verify the advantages of the SSC ionomer under high-current-density conditions. Accordingly, we further evaluated the FE_CO of D79 and D520 at 500 mA cm^-2, with D79 maintaining a high FE_CO of 91.1% while D520 exhibited only 61.8%. Collectively, these results demonstrate that ionomers can modulate the interfacial water structure on the catalyst surface, thereby significantly influencing product selectivity. To evaluate the operational stability of the D79-modified GDE, long-term continuous electrolysis was performed at a constant current density of 100 mA cm^-2 for 12 h in 0.05 M K₂SO₄ electrolyte (pH 1, adjusted with H₂SO₄) [Supplementary Figure 8], where a lower K⁺ concentration was adopted to mitigate salt precipitation. As depicted, the FE_CO remained at approximately 83% for the first 3 h, and then gradually diminished to 79% by the end of the 12-h test. Meanwhile, the cell voltage exhibited a negligible increment from 3.08 to 3.11 V throughout the electrolysis period. These results collectively corroborate the good stability of the D79-modified GDE under sustained acidic CO₂ electrolysis. Finally, to verify the universality of the SSC ionomer strategy, we extended the investigation to Ag-based catalysts [Supplementary Figure 9]. Under identical testing conditions, the D79-modified Ag GDE delivered a FE_CO of 90.6% at a current density of 300 mA cm^-2, markedly surpassing that of the D520-modified Ag GDE (76.9%). This consistent performance enhancement across different catalytic systems substantiates the broad applicability of the SSC ionomer approach for acidic CO₂ electroreduction.

To elucidate the interfacial water structures of SSC and LSC ionomers, in situ ATR-SEIRAS measurements were conducted, with the results presented in Figure 4. The broad feature appearing between 3,000 and 3,800 cm^-1 is attributed to the O-H stretching vibration mode, reflecting the unique hydrogen-bonding network formed by interfacial water^[29]. Spectral deconvolution resolves this broad band into three distinct peaks centered at approximately 3,600, 3,380, and 3,250 cm^-1, which are characteristic of free water, weakly hydrogen-bonded water, and strongly hydrogen-bonded water, respectively^[30]. In the spectra of GDEs modified with different side chain ionomers, absorption peaks corresponding to weakly hydrogen-bonded water (3,380 cm^-1) and strongly hydrogen-bonded water (3,250 cm^-1) were observed. Owing to the more densely packed sulfonate groups in SSC ionomers, they can form a more extensive and stronger hydrogen-bonding network. As shown in Figure 4A and C, with increasing applied potential, the content of strongly hydrogen-bonded interfacial water on the D79-modified GDE (SSC) gradually increased from 46.6% at -0.7 V vs. RHE to 60.5% at -1.7 V vs. RHE. This enhanced intermolecular hydrogen bonding is likely attributed to the intensified interfacial electric field upon potential application, which induces a more ordered dipole alignment of water molecules^[31]. In contrast, for GDEs modified with LSC ionomers, the proportion of weakly hydrogen-bonded water consistently exceeded 60%, and the corresponding absorption peak gradually intensified with increasing potential [Figure 4B and D]. It is noteworthy that the absolute peak intensity for D79 showed minimal change with increasing potential, whereas for D520 it increased substantially. This difference arises because the SSC ionomer already forms a significant strongly hydrogen-bonded network at low potential due to its high sulfonic acid group density. The applied potential primarily drives water reorientation rather than increasing total water population. In contrast, the LSC ionomer interface is dominated by weakly hydrogen-bonded water, and the applied potential gradually induces both water accumulation and reorientation, resulting in more pronounced absolute intensity changes. When SSC ionomers induce the formation of a strongly hydrogen-bonded network, water molecules become anchored to specific lattice sites. This strong confinement increases the energy barrier for breaking hydrogen bonds, potentially hindering proton migration via the Grotthuss (hopping) mechanism. Consequently, proton transport may be forced to transition toward a vehicle mechanism, in which protons diffuse together with their hydration shells [Figure 4E and F]. This proposed transition would ultimately lead to a substantial reduction in the interfacial proton flux^[32], a scenario that is consistent with the observed increase in strongly hydrogen-bonded water and the concomitant suppression of the HER. Moreover, it is well established that water molecules exhibit a reorientation from H-up to H-down configurations when a negative potential is applied^[33]. The distance between the catalyst and hydrogen atoms is governed by differences in the orientational reconfiguration behavior of water molecules. As shown in Figure 4G, a strong hydrogen-bonding network not only increases the catalyst-H distance but also more effectively immobilizes water molecules and H⁺, preventing their approach to the vicinity of active sites and reducing water dissociation^[34]. Consequently, the HER is significantly suppressed, which explains why the D79-modified GDE maintains a high CO selectivity even at high current densities.

The catalyst/ionomer ratio within the catalyst layer significantly influences mass transport processes. An excessively low ionomer ratio exposes a larger active area to the acidic electrolyte, thereby promoting the HER. Conversely, an excessively high ionomer ratio impedes gaseous CO₂ transport [Figure 5A]. Moreover, elevated ionomer concentrations facilitate proton transport, thereby reducing CO selectivity. To investigate the impact of local CO₂ and proton transport modulated by the ionomer on electrochemical performance, we systematically varied the catalyst/ionomer ratio. The CO₂RR performance was evaluated across five different D79 loadings (5, 10, 20, 30, and 40 wt%). Supplementary Figure 10 reveals that the F content rises with increasing ionomer content, confirming that the prepared GDEs possess substantially different ionomer loadings. As shown in Figure 5B and Supplementary Figure 11, CO selectivity and cell voltage exhibited a volcano-type trend as the D79 ratio increased. The optimal performance was achieved at 20 wt% D79, yielding a FE_CO of 97.1% at 300 mA cm^-2. In contrast, at the same current density, the FE_CO was only 11.8% for the GDE with 5 wt% D79 and decreased from 97.1% to 62.5% when the D79 ratio was increased from 20 wt% to 40 wt%. This is because a high ionomer ratio impedes CO₂ diffusion, whereas a low ionomer ratio exacerbates HER by exposing active sites to the acidic environment. To investigate the underlying reasons for this phenomenon, the ECSA of GDEs with varying ionomer ratio was determined using the double-layer capacitance method [Supplementary Figure 12]^[35,36]. As shown in Figure 5C, the C_dl increased from 5.69 mF cm^-2 at 5 wt% D79 to a maximum of 10.5 mF cm^-2 at 20 wt% D79, then decreased to 6.48 mF cm^-2 at 40 wt% D79. This indicates that the GDE with 20 wt% D79 possesses the optimal ECSA, which facilitates the formation of more three-phase reaction interfaces near the active sites, thereby promoting CO₂RR. EIS is a non-destructive analytical technique capable of real-time monitoring of charge transfer and mass transport processes in electrochemical systems^[37,38]. As shown in Figure 5D, the GDE with 20 wt% D79 exhibited the lowest ohmic resistance of 16.9 Ω and the smallest charge transfer resistance of 13.3 Ω, suggesting that this ionomer ratio achieves an optimal balance between CO₂ transport and ion transport, establishing a stable three-phase reaction interface conducive to efficient CO₂RR. In summary, the experimental results demonstrate that precisely controlling the catalyst/ionomer ratio within the catalyst layer is critically important, as it profoundly influences local mass transport and overall electrochemical performance. Insufficient ionomer ratio fails to suppress the HER via the hydrogen-bonding network formed by the interfacial water structure, while excessive ionomer ratio hinders CO₂ transport. Optimizing this ratio is particularly essential in acidic electrolysis with high H⁺ concentration to balance ionic transport and CO₂ diffusion at the interface, thereby achieving maximal CO selectivity.

By converting EIS data from the frequency domain to the time domain, the DRT can effectively resolve multiple processes with close time constants, thereby simplifying the interpretation of EIS results^[39,40]. This method has been successfully applied to the diagnostic assessment of various electrochemical systems, including lithium-ion batteries, fuel cells, and water electrolyzers^[41,42]. To elucidate the influence of ionomer content on CO₂ transport, DRT analysis was conducted^[43,44]. On the basis of the characteristic frequency ranges and their responses to variations in ionomer loading, current density, and CO₂ partial pressure, the DRT peaks are systematically deconvoluted and assigned as follows: the high-frequency feature (1-10 kHz) is ascribed to ohmic and ionic transport resistance; the mid-frequency feature (10-200 Hz) is attributed to the charge transfer resistance associated with the CO₂RR; and the low-frequency feature (1-10 Hz) originates from CO₂ mass transport limitations. Figure 5E demonstrates that the FE_CO declines as the CO₂ concentration is reduced, suggesting that CO₂ mass transport plays an important role in determining the electrochemical performance. As shown in Figure 5F, as the current density increases from 100 to 500 mA cm^-2, the CO₂ transport peak in the DRT spectra shifts to higher frequencies with a concurrent decrease in peak area. This behavior reflects a significant enhancement in mass transport within the MEA electrolyzer, where the substantial increase in CO₂ consumption at high current densities leads to a greater demand for CO₂ supply^[45]. The effect of ionomer ratio on CO₂ transport was further investigated. As presented in Figure 5G, the CO₂ transport peak progressively increases with higher ionomer ratio. Reducing the CO₂ flow rate enlarges the CO₂ transport peak [Figure 5H], indicating a marked rise in transport resistance, which would consequently lead to a decline in FE_CO with decreasing CO₂ concentration. A direct comparison between the 5 and 40 wt% GDEs [Figure 5I] reveals that increasing the ionomer ratio not only enlarges the CO₂ transport peak but also reduces the CO₂ reaction peak. This suggests that although a higher ionomer ratio increases CO₂ transport resistance, it simultaneously enhances CO₂RR kinetics by promoting the formation of a favorable interfacial water structure and the establishment of efficient ion transport pathways. Therefore, optimizing the ionomer ratio is essential to balance CO₂ transport and CO₂RR kinetics, which accounts for the highest FE_CO observed for the GDE with 20 wt% D79 in the performance tests shown in Figure 5B.