CO₂-affinitive surface of metal/Nafion attributable to selective polarization for superior CO₂RR

Abstract

The CO₂ reduction reaction (CO₂RR) is considered a promising process that can compensate for CO₂ emissions by converting them into value-added chemicals. In the preparation of the CO₂RR catalytic electrode, Nafion is a common binder, but its effect in the catalytic process is usually ignored. Herein, we proposed the bifunctionality of Nafion in the CO₂RR based on its molecular structure and successfully developed a CO₂-affinitive surface of micro-Ag/Nafion by controlling the content of Nafion, which exhibited a superior CO faradaic efficiency (FE_CO) of 95.07% at −0.9 V (vs. RHE). This excellent FE_CO and a partial current density (j_CO) twice that of CO in the flow-cell proved its industrial prospects. Its splendid selectivity for the CO₂RR was attributed to the selective polarization of CO₂ molecules on Nafion, which resulted in a 6.8 times higher adsorption energy (E_ads) of CO₂ than that of H₂O. Therefore, the CO₂RR gained an advantage in the competitive reaction. Nafion was recognized as a favorable transporter for CO₂ on the surface of the metal catalyst. Our work describes the bifunctional role of Nafion in the CO₂RR electrode, and it will contribute to optimize the design of the metal/Nafion catalytic material for the CO₂RR.

---

Green foundation

1. We proposed the selective polarization behavior of CO₂ molecules at the surface of metal/Nafion. Its novelty lies in the presentation of the bifunctionality of Nafion in CO₂ reduction, which not only bonded with the catalyst but also enhanced the CO₂ adsorption. Based on this mechanism, Nafion can be effectively regulated to enhance CO₂ reduction and achieve the green chemical goal of reducing carbon emissions and developing renewable feedstocks.

2. By adjusting the content of Nafion, hydrogen evolution was restrained, and the surface of micro-Ag/Nafion achieved a superior CO Faraday efficiency of 95.07% at −0.9 V (vs. RHE) in both the H-cell and flow-cell.

3. Future research includes achieving a high current density and high stability on the basis of high selectivity, realizing the module integration of CO₂ reduction systems, reducing the costs, and promoting industrial and commercial applications.

Introduction

Owing to the excessive utilization of fossil fuels, the greenhouse effect caused by CO₂ emissions has become a thorny issue.^1–4 In this context, the electrocatalytic CO₂RR has become a research hotspot as a long-term solution for compensating CO₂ emissions.^2,5–8 However, the theoretical redox potential of the CO₂RR (−0.11 V vs. RHE for CO),^9–11 as well as its thermodynamic and kinetic dependence on available protons,¹² suggests that the hydrogen evolution reaction (HER) in the water splitting process is inevitable. As a competitive reaction against the CO₂RR, the HER is obviously deleterious as it causes energy waste.^13,14 Additionally, the impure products caused by side reactions require complex downstream separation processes, which dramatically restrict their practical application.¹⁵

In the electrocatalytic reduction process, the rational designing of the structures and components of the catalytic surface can tune its electroactive sites and induce ideal products.^16–21 It has been demonstrated that a hydrophobic surface can successfully inhibit the HER and raise the competitive position of the CO₂RR.^22–24 Xing et al. dispersed polytetrafluoroethylene (PTFE) nanoparticles in commercial copper nanoparticles, which achieved a greatly improved activity and faradaic efficiency for CO₂ reduction, with an inspiring partial current density (>250 mA cm⁻², around twice that without PTFE).²⁵ Wakerley et al. treated hierarchically structured Cu dendrites with 1-octadecanethiol to increase the concentration of CO₂ at the electrode-solution interface and consequently increase the CO₂ reduction selectivity.¹² Therefore, slight organic doping is proved as a feasible strategy to construct a hydrophobic and gas-affinitive catalytic surface so as to restrain the HER and promote the CO₂RR.

Perfluorosulfonic acid (PFSA, trade name Nafion) is a popular membrane material and adhesion promoter in electrochemical engineering because of its intrinsic ion transport properties.^26–28 We note that the Nafion molecule has the same main chain (MC) structure as PTFE (Fig. S1), making it exhibit similar hydrophobic and gas-affinitive properties as those of PTFE. Besides, as the functional group of the Nafion molecule, the sulfonic acid group (SAG) on the side chain (SC) shows a particular polarity.^29,30 This inequality in molecule polarity at different positions of Nafion inspired us to think that it might cause an inequality in the adsorption of CO₂ and H₂O and thus lead to an unfair competition between the CO₂RR and HER. However, in most research on the CO₂RR, which takes Nafion as a binder on the electrode, its influence is not taken in account in either the experiments or calculations.

Herein, on the basis of the excellent CO selectivity of the Ag catalyst,^31–35 we combined micro-Ag (MAg) with Nafion and constructed a CO₂-affinitive metal/organic surface. In this context, the Nafion molecule played a bifunctional role with both adhesion and CO₂ affinity. Consequently, the MAg/Nafion surface achieved a superior FE_CO of 95.07% at −0.9 V (vs. RHE) in the H-cell and an equivalent level in the flow cell. Further analysis and density functional theory (DFT) calculations demonstrated that the Nafion molecules contribute to a kinetic effect rather than a thermodynamic effect. The selective CO₂ polarization with Nafion molecules resulted in a local high concentration of CO₂ at the surface of the catalyst. This effect suppressed the HER and enhanced the CO₂RR. Therefore, this work proposes the selective polarization function of Nafion and explains its bifunctional role in the complex CO₂RR process, which will contribute to the exploration of high-efficiency CO₂RR.

Experimental

Materials

MAg powder was purchased from Shanghai Aladdin Biochemical Technology Co., LTD. Nafion solution was obtained from DuPont Engineering Polymers. All chemical reagents used in this study were analytical grade and were used as received without further purification. All aqueous solutions were prepared with deionized water.

Electrode preparation

The precursor solution was obtained by ultrasonic oscillation. First, a 0.5 mL solution was prepared with anhydrous ethanol, Nafion solution and 20 mg MAg powder. Then, the solution was placed in an ultrasonic cleaning machine at a frequency of 40 kHz for high-frequency shock for 1 h to fully disperse the precursor and obtain a precursor suspension. Hydrophobic carbon paper (1 × 3 cm) was used as the electrode substrate. Polyimide material was used to bond and seal the back and around the substrate to control the exposed area to prevent the non-catalytic sites from participating in the reaction and interfering with the product. The substrate was preheated to 90 °C to reduce its surface energy to promote the spread of the precursor dispersion, and the precursor dispersion was applied to the exposed carbon paper in drops. The electrode substrate was volatilized for 30 min until the surface was completely dried to obtain a uniformly dispersed MAg/Nafion electrode surface.

CO₂ electroreduction experiments

The electrochemical tests were carried out in both the H-cell and the flow-cell. In the H-cell, to prevent the mass dissolving of CO₂ in alkaline solution, a neutral system was applied with 0.5 M KHCO₃ (pH = 7.2) as the electrolyte, Pt as the counter electrode, and Ag/AgCl as the reference electrode. In contrast, to reduce the concentration of H⁺ and restrict the HER, the flow-cell employed an alkaline system with 1 M KOH (pH = 13.6) as the electrolyte, Pt as the counter electrode, and Hg/HgO as the reference electrode. The conversion between the reference electrode and the reversible hydrogen electrode is as follows:

E(vs. RHE) = E(vs. Ag/AgCl) + 0.0592pH + 0.197

E(vs. RHE) = E(vs. Hg/HgO) + 0.0592pH + 0.098

In situ surface-enhanced Raman spectroscopy (SERS) was conducted for the CO₂RR using a 532 nm laser under the same electrochemical conditions as the H-cell. The in situ SERS spectra were recorded from −0.8 V to −1.2 V (vs. RHE).

At the same time, the gas product measured by gas chromatography (GC) was used to compare the FE in the range of −0.8 V to −1.2 V (vs. RHE). The FE of the gas product was calculated as follows:

In this formula, n represents the number of electrons transferred for one product molecule; u represents the gas flow rate (L s⁻¹), whose value is 30 mL min⁻¹ in this experiment; V represents the product gas concentration (%); F represents the Faraday constant (C mol⁻¹), whose value is 96500 C mol⁻¹; I represents the average current under the 2000 s test (A); V_m represents the molar volume of the gas (L mol⁻¹), whose value is 22.4 L mol⁻¹.

After electroreduction, the composition of the liquid product in the cathode electrolyte was measured by nuclear magnetic resonance (NMR) with D₂O as the solvent and dimethyl sulfoxide (DMSO) as the internal standard. The FE of the liquid product was calculated as follows:

In this formula, n represents the number of electrons transferred for one product molecule; c represents the concentration of liquid product in the cathode electrolyte (mol L⁻¹); Vol represents the volume of the cathode electrolyte (L), whose value is 0.02 L in this experiment; F represents the Faraday constant (C mol⁻¹), whose value is 96 500 C mol⁻¹; I represents the average current under the 2000 s test (A); t represents the reaction time (s), whose value is 2000 s in this experiment.

All product data were obtained in three independent measurements.

Computation

The density functional theory (DFT) computation was executed using the DMoL3 module of Materials Studio. The Perdew–Burke–Ernzerhof (PBE) functional within the Generalized-Gradient-Approximation (GGA) was used to describe the exchange–correlation energy. In the geometry optimization, the convergence tolerances for energy, force, and displacement were set to 10⁻⁵ Ha, 0.002 Ha Å⁻¹, and 0.005 Å, respectively. To accelerate the optimization process, the smearing was set to 0.005 Ha and the global orbital cut-off radius in real space was set to 4.1 Å. When sampling in the Brillouin zone, the k-point grid was set to 3 × 3 × 1 in the geometric optimization for speeding up the conclusion. The electron density, deformation electron density, Mulliken charge, and some bond orders were calculated to analyze the interactions caused by electron distribution and transfer.

The change in the Gibbs free energy (ΔG) for each state in the reaction was calculated based on the standard hydrogen electrode (SHE) model proposed by Nørskov. The ΔG of the intermediates was calculated as follows:

In this formula, ΔE represents the difference in the intrinsic energy between two intermediates, ΔZPE represents the zero-point vibration energy difference of adsorbed atoms, T represents the temperature (298.15 K), ΔS represents the entropy change, represents the enthalpic temperature correction, k_B represents the Boltzmann constant (1.38 × 1023 J K⁻¹), pH is set to 0 according to SHE model, e represents the elementary charge quantity (1.6 × 10⁻¹⁹ C) and U represents the work electrode potential.

As shown in Table S1, the free energies of gas molecules (CO₂, H₂) were calculated at a standard pressure of 100 kPa, and the free energy of the liquid molecule H₂O was calculated at a pressure of 3169 Pa (saturated vapor pressure of H₂O). The entropy and formation energy data for small molecules (H₂O, CO₂, CO, H₂) were obtained from the CRC Handbook of Chemistry and Physics, 97th Edition.³⁶

Results and discussion

Synthesis and characterization

The brief synthesis process of the electrode is illustrated in Fig. 1a. The precursor, consisting of a mixture of MAg and Nafion solution (volume content x% from 5% to 50%, denoted as MAg-Nfx), was ultrasonically treated until fully dispersed. The carbon paper substrate was sealed around the polyimide material in case of marginal effects. The precursor was then dropped onto the substrate and dried, forming the working electrode.

Fig. 1 (a) Schematic of the CO₂RR electrode fabrication. (b) SEM images and EDS elemental mapping of the MAg/Nafion electrode. (c) TEM image of MAg. (d) HRTEM image of MAg. (e) XRD patterns of MAg and the MAg electrode. (f) XRD patterns of the MAg electrode with different Nafion concentrations. (g) XPS spectrum of the MAg-Nf20 electrode.

Since the CO₂RR occurs at the active sites on the catalyst surface, the component and structure of MAg are essential. The surface scanning electronic microscopy (SEM) and Energy Dispersive Spectroscopy (EDS) analysis of the prepared MAg electrode are shown in Fig. 1b. The MAg particles (largest diameter about 2 μm) were uniformly distributed on the electrode surface. There was no obvious rich or poor area, which improved the stability of the electrode. Gaps among large particles might provide channels for gas diffusion so that the MAg inside had the chance to adsorb more CO₂ molecules and contribute to the increase in the number of CO₂RR active sites. As the main component elements of Nafion molecules, C and F were not obvious in EDS, indicating that the Nafion content was much lower when compared to MAg. As a result, it played an auxiliary modification role rather than a catalyst. TEM was used to further explore the crystal structure of MAg, as shown in Fig. 1c. The smallest size of MAg was approximately 0.5 μm without agglomeration. This micro-level size effect enriched the surface area and reaction sites, providing kinetic conditions for the CO₂RR. High-resolution TEM (HRTEM) analysis further indicated that the interplanar lattice fringe spacing of MAg was 0.236 nm, which matched the crystal face orientation of the face-centered cubic Ag (1 1 1) plane (Fig. 1d).³⁷ At the same time, distorted lattice fringes (red curve) indicate the presence of abundant Ag crystal edge sites, which are commonly known as highly active sites for the CO₂RR.³⁸ This crystal structure can contribute to strong CO₂ activation and fast CO desorption.

X-ray diffraction (XRD) patterns (Fig. 1e and f) showed that the main diffraction peak of the MAg electrode at 38.1° matched the Ag (1 1 1) plane, consistent with the result in Fig. 1d. Compared with the diffraction peak of the MAg powder, only the C diffraction peak was observed in the XRD of the MAg electrode, which was attributed to the carbon paper substrate. The absence of oxide peaks suggests the pure phase structure; this was also proven by XPS survey spectra (Fig. 1g). Two characteristic peaks with the same full width at half maximum and suitable area ratio were attributed to Ag 3d_3/2 (374.7 eV) and Ag 3d_5/2 (368.6 eV).^39–42 No electron transfer indicated that no extra reduction reaction would occur in the subsequent CO₂RR, which would retain a stable electrochemical activity.^43,44 The high ratio of F atoms confirmed the stable presence of Nafion on the electrode surface (Table S2).

CO₂RR performance

The CO₂RR performance of the MAg/Nafion electrode was first tested in a traditional H-cell (Fig. S4); commercial Ag foil was also tested for comparison. The product composition is shown in Fig. 2a, which consists mainly of CO and H₂. Fig. 2b shows the FE_CO at different potentials, indicating the CO₂RR performance. The FE_CO of Ag foil was far lower than that of the MAg/Nafion electrode. This is easy to explain by its smaller catalytic active area and lack of Nafion. Besides, a clear optimization trend of Nafion content was observed. With the increase in the Nafion content, FE_CO first increased to a maximum of 95.07% at MAg-Nf20 and then decreased with the further addition of Nafion, with MAg-Nf50 performing most poorly. This behavior underscored a critical trade-off: an optimal Nafion content was beneficial for its role in ion exchange and microenvironment management, but surpassing this threshold was detrimental. We inferred that excessive Nafion molecules form a dense and non-porous layer that caused the encapsulation of the MAg catalyst. This encapsulation blocked the active sites, severely hindered the mass transport of CO₂ and CO, and ultimately stifled the catalytic reaction, leading to the observed decline in FE_CO.

Fig. 2 (a) Product composition of MAg/Nf20 in the CO₂RR. (b) FE_CO and (c) j_CO plots of the MAg electrodes with different Nafion contents. (d) Stability of MAg-Nf20. (e ) Flow-cell CO₂RR test and the performance of MAg-Nf20.

For all MAg/Nafion samples, noteworthily, the FE_CO remained at a high level for all MAg/Nafion samples from −0.8 to −1.0 V (vs. RHE, the same below). As for the MAg-Nf20, its FE_CO was up to 95.07% at the potential of −0.9 V, which was 3.4 times that of Ag foil at the same reduction potential. However, not only the foil, but for all MAg/Nafion samples, when the potential was higher than −1.0 V, the FE_CO showed a continuous decline. This phenomenon is explained in Fig. 2c. Considering the j_CO, it was found that when the potential was lower than −1.0 V, j_CO increased with the increase in potential. However, when the reduction potential increased above −1.0 V, j_CO began to hold constant, indicating that it had reached the limit of the system. This indicates that when the potential was above −1.0 V, the original competitive situation between CO₂RR and HER was broken, and the HER could continue to expand while the CO₂RR could not. This can be attributed to the concentration limit of CO₂ in the aqueous solution system, which kinetically hindered the further expansion of the CO₂RR.

In order to verify the stability of the MAg/Nafion electrode, the above Ag-Nf20 sample was tested for 12 h, and the test results are shown in Fig. 2d. In the stability test period, the current density gradually became stable, and the FE_CO decreased slightly, by 10.19% within 12 h. The SEM and XPS analyses in Fig. S5 proved the stability of its surface structure and chemical state.

To verify its feasibility in operando conditions, a flow-cell was assembled and tested in further experiments, as shown in Fig. 2e. A similar FE_CO was obtained in the flow-cell for MAg-Nf20, demonstrating its industrial application prospects. Moreover, at the same potential as the H-cell, the j_CO in the flow-cell was almost 2 times that in the H-cell. This suggests that a flow-cell structure can efficiently reduce the solution resistance and enhance CO₂ diffusion to increase its concentration at the surface of the catalyst.

The electrochemical resistance (EIS) test in Fig. 3a revealed that the resistance of MAg/Nafion was significantly lower than that of Ag foil, and proved that our electrode achieved rapid charge transfer at the interface between the electrode and electrolyte. Moreover, the short, low-frequency region (arc region) revealed that most electrons were donated to mass transfer, rather than electron transfer at the interface capacitance. This also indicated the CO₂-affinity of the surface. LSV curves in Fig. 3b showed that the current density increased with an increase in the cathodic potential. Under the same potential, Ag-Nf20 had the lowest current density. A lower current density usually indicates lower electrocatalytic activity. However, this electrocatalytic activity was the result of the accumulation of all catalytic reactions, including the CO₂RR and HER. Compared with Fig. 2b, the high current density at the same reduction potential meant low selectivity of CO. Compared with the strong HER, the j_CO progressively approached a plateau despite further increases in cathodic potential. This saturation signified that the thermodynamic driving force, dictated by electronic energy, no longer served as the rate-determining factor for the CO₂RR, and the reaction kinetics became increasingly governed by the supply of CO₂ or diffusion of CO. Consequently, the mass transfer process emerged as the most plausible bottleneck, restricting further enhancement of the CO₂RR performance. Further insight was gained by comparing the j_CO recorded in the H-cell and the flow cell. The pronounced enhancement of j_CO in the flow cell can be ascribed to its excellent delivery of CO₂ to the catalytic sites and CO from it. In contrast, in the H-cell, the low solubility and slow diffusion of CO₂ in the aqueous electrolyte led to severe mass transfer limitations. Therefore, under elevated reduction potentials, the overall performance of the CO₂RR was primarily constrained by mass transfer rather than thermodynamic or intrinsic kinetic barriers.

Fig. 3 (a) EIS, (b) LSV, (c) fitted Tafel, (d) fitted C_dl, (e) normalized LSV, and (f) normalized j_CO plots of the MAg electrode with different Nafion contents. (g ) In situ SERS spectra of the CO₂RR on MAg-Nf20.

To further verify the CO₂RR process, Tafel slopes were calculated, as shown in Fig. 3c. The Tafel slope of the MAg-Nafion electrode was significantly lower than that of Ag foil (838.3 mV dec⁻¹), which demonstrated the lower activation energy and initial kinetic barrier of the CO₂RR. It reduced the dependence of j_CO on the overpotential, which was more conducive to the CO₂RR dominating the competitive reactions. At low overpotentials, the reaction rate is highly sensitive to this initial kinetic barrier. The optimized kinetics of the MAg-Nafion electrode ensured that at any given moment, especially with low overpotential, FE is skewed towards the CO₂RR rather than the HER.⁴⁵ Consequently, this kinetic superiority directly translates to a higher selectivity for the CO₂RR and improved overall energy efficiency as the system can operate effectively at a lower whole voltage.

The double-layer capacitances (C_dl) were determined to characterize the electrochemically active area (ECSA). Fig. 3d shows that the C_dl values of the MAg/Nafion electrodes were similar. This may be explained by the insufficiency of the binding strength due to less Nafion. Meanwhile, the low C_dl of silver foil (39.6 F cm⁻²) explained its low FE_CO. To eliminate the effect of the electrochemical area on the catalytic activity, the normalized current density and partial current density were determined as shown in Fig. 3e and f. The general trend of the curve is consistent with that before normalization, which indicates that the ESCA made almost no contribution in the CO₂RR.

The in situ SERS spectra in Fig. 3g revealed the variation of the reaction intermediates of MAg-Nafion20 in the CO₂RR. The Raman peaks below 1000 cm⁻¹ and at 1520 cm⁻¹ (black dashed lines) were attributed to the intrinsic MAg. The intensities of these peaks, along with the HCO₃⁻ signal at 1068 cm⁻¹ (orange dashed line), remained constant as the reduction potential increased. In contrast, the Raman peaks at 1360 and 1590 cm⁻¹ (green dashed lines), which were assigned to the adsorbed *CO₂ intermediates on the catalyst surface, exhibited a pronounced potential dependence.^46,47 These signals were relatively weak at lower reduction potentials but underwent significant enhancement when the potential exceeded −0.95 V. Under these conditions, the peaks showed increased intensity and broader half-width, even surpassing the intensity of the adjacent Ag peak. This phenomenon indicated a substantial enhancement in CO₂ adsorption on the Ag surface at this specific potential, which is aligned with the observed maximum CO Faraday efficiency (FE_CO) near this value in Fig. 2b. The broad feature peaks in the 3000 to 3800 cm⁻¹ range are typically associated with the O–H stretching vibrations of interfacial water molecules.⁴⁸ The intensity variation within this region (blue dashed line) suggests a change in the coordination environment of water molecules at the catalyst surface, thereby confirming the occurrence of the HER.

Noteworthily, the Nafion ionomer also played a pivotal role in modulating the local microenvironment at the electrode–electrolyte interface, which significantly influenced the selectivity of the CO₂RR.⁴⁹ In this study, the observed peak in the FE_CO of 95.07% at approximately −0.9 V (vs. RHE) is likely attributable to an optimal local microenvironment established at this potential, mediated by Nafion. This microenvironment provided sufficient local alkalinity to promote the CO₂RR while avoiding issues such as carbonate precipitation or mass transfer limitations that can arise from excessively high pH.^8,50,51 Concurrently, the mass transfer resistance of Nafion towards reactants (CO₂, H₂O) and the product (CO) may also contribute to the observed potential-dependent selectivity.^52–54 Therefore, the catalytic performance reported by us was not solely a manifestation of the intrinsic activity of the Ag-based catalyst, but also a result of the specific local microenvironment created by the composite catalytic layer comprising both the catalyst and the Nafion ionomer.

DFT calculation

To further explore the enhancement mechanism of Nafion, DFT calculations were conducted to determine the CO₂RR process. Fig. 4a shows the difference in E_ads of the Nafion molecule to H₂O and CO₂ molecules. The E_ads at the SAG position was at least 4 times greater than those at the MC or SC positions, which suggests that compared with nonpolar MC, SAG on SC tends to adsorb small molecules such as H₂O and CO₂. Noteworthily, even at the SAG position, the E_ads of the CO₂ molecule reached −0.8 eV, about 6.8 times that of H₂O, which proved the CO₂-affinity of Nafion, which could be demonstrated experimentally in contact angle measurements (Fig. S7). From the optimized geometry structure, it was found that the distance between the CO₂ molecule and SAG was dramatically close, which implied that there may be chemical bonds between them.

Fig. 4 (a) E_ads of CO₂ and H₂O at different positions of Nafion. (b ) Mulliken population. (c) Electron density location. (d) Charge density difference of CO₂ and H₂O adsorbed to Nafion.

Mulliken population analysis indicated that there was little charge transfer or bond order change for H₂O after adsorption to SAG (Fig. 4b and S9). The distance between H in H₂O and O in SAG was approximately 2.6 Å, longer than most hydrogen bonds.^55–57 Therefore, the E_ads of the H₂O molecule can be attributed to this weak van der Waals interaction. However, after adsorption, the bond angle of O=C=O changed to 128.9° and this polarization of the CO₂ molecule broke its intrinsic symmetry. This was inspiring because the electronic symmetry of the CO₂ molecule is an important reason for the high bond energy of C=O (750 kJ mol⁻¹).^58,59 Since one of the O atoms in CO₂ was adsorbed to SAG, its Mulliken charge significantly decreased to −0.146e, compared with −0.242e of the farther O atom. The decrease in its C=O bond order also proved the C=O bond to be more delicate. Different from H₂O to Nafion, this selective CO₂ polarization behavior may provide an advantage for the formation of the *COOH intermediate during the CO₂RR process.

Noteworthily, the distance between the CO₂ molecule and SAG was less than 2 Å when adsorbing, which was close enough to form covalent interactions. The location of electron density in Fig. 4c illustrates that the presence of heteroatoms led to the rearrangement of electrons in CO₂. It was observed that there were more delocalized electrons between O and O, C and S. In contrast, almost no electron cloud was distributed between H₂O and SAG. This difference in electron distribution proved the stronger interaction of Nafion with CO₂ than with H₂O and explained the difference in E_ads. However, the charge density difference (Fig. 4d) indicated that the charge accumulation region (red region) was still distributed around the O atom as a ring. Depleted electrons suggest that this covalent interaction is not as strong as a covalent bond like O=O in O₂, allowing CO₂ to desorb from Nafion and adsorb to Ag.

To further research the selective polarization function of Nafion, according to the characterization result in Fig. 1d, we built the Ag (111) surface with and without Nafion (Fig. 5a and b) under RHE conditions to compare the difference in Gibbs free energy (ΔG) via DFT calculations in catalytic steps. Fig. 5c reveals that it was the most endothermic step from CO₂ to *COOH for its high energy barrier. When adsorbing to the Ag surface, the interaction between CO₂ and SAG was broken. There was a similar adsorption structure of *COOH with and without Nafion, which explained why they had a similar energy barrier. Moreover, the formation energy of the *COOH intermediate (about 1.6 eV) was dramatically higher than that of the *H intermediate (about 0.4 eV), indicative that the HER intrinsically (in the case at open circuit potentials) proceeds more easily than the CO₂RR, leading to a decrease in FE_CO. However, the experimental results proved that CO₂RR took the dominant position. This indicates that under a potential higher than that of the intrinsic CO₂RR, it was kinetics and not thermodynamics that drove the reaction. This may be due to the selective polarization effect, which indicates that it was easier for SAG to capture a CO₂ molecule rather than H₂O (Fig. 5d).

Fig. 5 Reaction intermediate formation process of MAg (a) without Nafion and (b) with Nafion. (c) Gibbs free-energy diagram of the HER and CO₂RR. (d) Schematic of selective polarization at the metal/Nafion surface.

Considering the flexibility of SC in high-humidity conditions,^60–62 the captured CO₂ molecules should be easily transferred to the surface of Ag by swinging SC. A high concentration of CO₂ adsorbed by SAG and a low concentration of H₂O excluded by MC led to the kinetic preponderant position of the CO₂RR. Consequently, the formation of *H was kinetically limited, and the limited numbers of *H were likely to combine with rich *CO₂ nearby to form the *COOH intermediate, rather than couple into H₂. Therefore, this behavior of suppressing the HER and enhancing the CO₂RR was legitimately attributed to the selective polarization function of Nafion molecules.

Conclusions

We have successfully developed a CO₂-affinitive surface of MAg/Nafion for the CO₂RR. Nafion played a bifunctional role as the binder of the electrode and as a selective sifter for CO₂. The gaps between the MAg particles provided gas diffusion channels for the CO₂RR. As a result, with a moderate Nafion content, the prepared MAg/Nafion electrode showed a superior FE_CO of 95.07% at −0.9 V (vs. RHE). Analogously, a remarkable FE_CO and 2 times the j_CO were observed in the flow-cell device, which proved its industrial prospects. Experiments indicated that the concentration of CO₂ at the catalytic surface had an important influence on the CO₂RR. DFT calculations revealed that the polarization of the CO₂ molecule at the SAG of Nafion was the main reason for its selective adsorption. Electrons of the O atom in CO₂ were covalently transferred to the O atom in SAG, resulting in the 6.8 times higher E_ads of CO₂ as compared to H₂O. In contrast, Nafion had little influence on the thermodynamic process of the CO₂RR. In view of this situation, Nafion was recognized as not only a binder component, but also a transporter for CO₂ on the surface of the metal catalyst. This work explains the mechanism of the CO₂-affinitive metal/Nafion surface that may contribute to the rational design of CO₂RR catalytic material.

Author contributions

J. Q. conceived the idea and guided the research. X. Z. carried out the experiment and analyses. X. Z. and Y. L. performed the DFT calculations. Y. Y. and X. Z. co-wrote the manuscript. C. L. and J. C. discussed the results together and assisted in the manuscript preparation.

Conflicts of interest

There are no conflicts to declare.

Data availability

The supporting data of this article has been included in the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5gc04074d.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 52175303 and 52505362), the National Science Fund for Distinguished Young Scholars (Grant No. 52125502), the Fundamental Research Funds for the Central Universities (Grant No. 2025CXPT03), Hainan Province Science and Technology Special Fund (Grant No. ZDYF2024SHFZ082), Postdoctoral Fellowship Program of CPSF (Grant No. GZB20240949), Key Research Program of Heilongjiang Province (Grant No. 2024ZXDXA19), Natural Science Foundation of Heilongjiang Province (Grant No. LH2024E031) and Heilongjiang Postdoctoral Fund (Grant No. LBH-Z24174).

References

1. A. Stechemesser, N. Koch, E. Mark, E. Dilger, P. Klösel, L. Menicacci, D. Nachtigall, F. Pretis, N. Ritter, M. Schwarz, H. Vossen and A. Wenzel, Science, 2024, 385, 884–892.
2. L. Li, X. Li, Y. Sun and Y. Xie, Chem. Soc. Rev., 2022, 51, 1234–1252.
3. M. Meinshausen, J. Lewis, C. McGlade, J. Gütschow, Z. Nicholls, R. Burdon, L. Cozzi and B. Hackmann, Nature, 2022, 604, 304–309.
4. A. Rode, T. Carleton, M. Delgado, M. Greenstone, T. Houser, S. Hsiang, A. Hultgren, A. Jina, R. E. Kopp, K. E. McCusker, I. Nath, J. Rising and J. Yuan, Nature, 2021, 598, 308–314.
5. G. Wang, J. Chen, Y. Ding, P. Cai, L. Yi, Y. Li, C. Tu, Y. Hou, Z. Wen and L. Dai, Chem. Soc. Rev., 2021, 50, 4993–5061.
6. M. Jiang, H. Wang, M. Zhu, X. Luo, Y. He, M. Wang, C. Wu, L. Zhang, X. Li, X. Liao, Z. Jiang and Z. Jin, Chem. Soc. Rev., 2024, 53, 5149–5189.
7. Z. Gu, H. Shen, Z. Chen, Y. Yang, C. Yang, Y. Ji, Y. Wang, C. Zhu, J. Liu, J. Li, T.-K. Sham, X. Xu and G. Zheng, Joule, 2021, 5, 429–440.
8. J. E. Huang, F. Li, A. Ozden, A. S. Rasouli, F. P. García de Arquer, S. Liu, S. Zhang, M. Luo, X. Wang, Y. Lum, Y. Xu, K. Bertens, R. K. Miao, C.-T. Dinh, D. Sinton and E. H. Sargent, Science, 2021, 372, 1074–1078.
9. K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard and T. F. Jaramillo, J. Am. Chem. Soc., 2014, 136, 14107–14113.
10. A. Xu, N. Govindarajan, G. Kastlunger, S. Vijay and K. Chan, Acc. Chem. Res., 2022, 55, 495–503.
11. A. S. Varela, W. Ju, A. Bagger, P. Franco, J. Rossmeisl and P. Strasser, ACS Catal., 2019, 9, 7270–7284.
12. D. Wakerley, S. Lamaison, F. Ozanam, N. Menguy, D. Mercier, P. Marcus, M. Fontecave and V. Mougel, Nat. Mater., 2019, 18, 1222–1227.
13. X. Tan, C. Yu, Y. Ren, S. Cui, W. Li and J. Qiu, Energy Environ. Sci., 2021, 14, 765–780.
14. T. Möller, F. Scholten, T. N. Thanh, I. Sinev, J. Timoshenko, X. Wang, Z. Jovanov, M. Gliech, B. Roldan Cuenya, A. S. Varela and P. Strasser, Angew. Chem., Int. Ed., 2020, 132, 18130–18139.
15. M. A. Khoshooei, X. Wang, G. Vitale, F. Formalik, K. O. Kirlikovali, R. Q. Snurr, P. Pereira-Almao and O. K. Farha, Science, 2024, 384, 540–546.
16. Z.-Z. Niu, F.-Y. Gao, X.-L. Zhang, P.-P. Yang, R. Liu, L.-P. Chi, Z.-Z. Wu, S. Qin, X. Yu and M.-R. Gao, J. Am. Chem. Soc., 2021, 143, 8011–8021.
17. K. Lv, C. Teng, M. Shi, Y. Yuan, Y. Zhu, J. Wang, Z. Kong, X. Lu and Y. Zhu, Adv. Funct. Mater., 2018, 28, 1802339.
18. G. An, K. Wang, M. Yang, J. Zhang, H. Zhong, L. Wang and H. Guo, Molecules, 2025, 30, 953.
19. C. Liang, Z. Zhang, Z. Liu, L. Wang and Y. M. Lam, Chem. Eng. J., 2025, 520, 165927.
20. P. Wang, W. Shi, L. Gu, H. Ran, Z. Pan, Y. Su, X. Song, J. Cao, H. Chen and W. Li, J. Mater. Sci. Technol., 2026, 244, 46–59.
21. Y. Yan, B. Wang, P. Li, B. Qin, Z. Ye, Y. Xia, Z. Zhang, Z. Wang, Y. Li, X. Zheng, H. A. Butt, D. V. Krasnikov, A. G. Nasibulin, J. Cao and J. Qi, Trans. Mater. Res., 2025, 1, 100038.
22. P. Yue, Q. Fu, J. Li, L. Zhang, L. Xing, Z. Kang, Q. Liao and X. Zhu, Chem. Eng. J., 2021, 405, 126975.
23. D. Zeng, C. Li, W. Wang, L. Zhang, Y. Zhang, J. Wang, L. Zhang, X. Zhou and W. Wang, Chem. Eng. J., 2023, 461, 142133.
24. K. U. Hansen and F. Jiao, Joule, 2021, 5, 754–757.
25. Z. Xing, L. Hu, D. S. Ripatti, X. Hu and X. Feng, Nat. Commun., 2021, 12, 136.
26. P. C. Okonkwo, I. Ben Belgacem, W. Emori and P. C. Uzoma, Int. J. Hydrogen Energy, 2021, 46, 27956–27973.
27. S. Chang, Y. Xuan, J. Duan and K. Zhang, Appl. Catal., B, 2022, 306, 121135.
28. K. Vezzù, G. Nawn, E. Negro, G. Crivellaro, J. W. Park, R. Wycisk, P. N. Pintauro and V. Di Noto, J. Am. Chem. Soc., 2020, 142, 801–814.
29. M. Phonyiem, S. Chaiwongwattana, C. Lao-ngam and K. Sagarik, Phys. Chem. Chem. Phys., 2011, 13, 10923–10939.
30. K. Sagarik, M. Phonyiem, C. Lao-ngam and S. Chaiwongwattana, Phys. Chem. Chem. Phys., 2008, 10, 2098–2112.
31. D. Sun, X. Xu, Y. Qin, S. P. Jiang and Z. Shao, ChemSusChem, 2020, 13, 39–58.
32. Q. Lu, J. Rosen, Y. Zhou, G. S. Hutchings, Y. C. Kimmel, J. G. Chen and F. Jiao, Nat. Commun., 2014, 5, 3242.
33. S. Liu, C. Sun, J. Xiao and J.-L. Luo, ACS Catal., 2020, 10, 3158–3163.
34. Z. Yang, M. Wan, Z. Gu and F. Che, J. Phys. Chem. C, 2023, 127, 17685–17693.
35. A. Thevenon, A. Rosas-Hernández, A. M. F. Herreros, T. Agapie and J. C. Peters, ACS Catal., 2021, 11, 4530–4537.
36. W. M. Haynes, D. R. Lide and T. J. Bruno, Thermochemistry, Electrochemistry, and Solution Chemistry, CRC Handbook of Chemistry and Physics 97th Edition, 2017.
37. S.-Q. Liu, S.-W. Wu, M.-R. Gao, M.-S. Li, X.-Z. Fu and J.-L. Luo, ACS Sustainable Chem. Eng., 2019, 7, 14443–14450.
38. D. Wang, Y. Zhu, W. Yu, Z. He, F. Dong, Y. Shen, T. Zeng, X. Lu, J. Ma, L. Wang and S. Song, Electrochim. Acta, 2022, 403, 139652.
39. K. Han, B. C. Rowley, M. P. Schellekens, S. Brugman, M. P. de Heer, L. P. S. Keyzer and P. J. Corbett, ACS Energy Lett., 2024, 9, 2800–2806.
40. T. Deka, A. Das, S. John and R. G. Nair, Int. J. Hydrogen Energy, 2024, 72, 1169–1183.
41. J. R. Perutil, S. Prashanth, C. V. Yelamaggad, P. Nukala and N. S. John, ACS Appl. Nano Mater., 2025, 8, 7998–8009.
42. Z. Bu, J. Li, S. Fang, Y. Zhan, W. He and Y. Huang, Colloids Surf., A, 2025, 718, 136791.
43. X.-Z. Wang, Q. Hu, M. Gao, S. Liu and J.-L. Luo, Electrochim. Acta, 2022, 403, 139656.
44. R. Yang, Q. Wen, Y. Yang, Y. Liu, Y. Yang, M. Wu, Y. Wei, B. Mei, Y. Liu, H. Li and T. Zhai, Adv. Mater., 2025, 37, 2414642.
45. H. Cheng, X. Wu, X. Li, X. Nie, S. Fan, M. Feng, Z. Fan, M. Tan, Y. Chen and G. He, Chem. Eng. J., 2021, 407, 126842.
46. J. Chen, B. Ma, Z. Xie, W. Li, Y. Yang, M. Mu, X. Zou, B. Zhao and W. Song, Appl. Catal., B, 2023, 325, 122350.
47. K. Wang, K. Huang, Z. Wang, G. An, M. Zhang, W. Liu, S. Fu, H. Guo, B. Zhang, C. Lian, J. Wu and L. Wang, Small, 2025, 21, 2502733.
48. Z. Li, X. Zhou, Y. Liu, X. Li, Y. Shen, M. Wen, P. Lei and L. Qian, ACS Catal., 2025, 15, 15019–15032.
49. Y. Yan, M. Wu, L. Zhou, W. Chen, L. Han, G. Gao, Y. Cui, Z. Sun and A. Cabot, Adv. Funct. Mater., 2025, 35, 2419328.
50. Y. Li, Z. Pei, D. Luan and X. W. Lou, Angew. Chem., Int. Ed., 2023, 62, e202302128.
51. J. Feng, L. Wu, X. Song, L. Zhang, S. Jia, X. Ma, X. Tan, X. Kang, Q. Zhu, X. Sun and B. Han, Nat. Commun., 2024, 15, 4821.
52. O. Christensen, S. Zhao, Z. Sun, A. Bagger, J. V. Lauritsen, S. U. Pedersen, K. Daasbjerg and J. Rossmeisl, ACS Catal., 2022, 12, 15737–15749.
53. Z. Wang, Y. Li, X. Zhao, S. Chen, Q. Nian, X. Luo, J. Fan, D. Ruan, B.-Q. Xiong and X. Ren, J. Am. Chem. Soc., 2023, 145, 6339–6348.
54. Z. Xu, X. Zhang and Q. Wang, Chem. Eng. J., 2024, 492, 152164.
55. R. Steudel and Y. Steudel, Eur. J. Inorg. Chem., 2009, 2009, 1393–1405.
56. J. Karo, A. Aabloo, J. O. Thomas and D. Brandell, J. Phys. Chem. B, 2010, 114, 6056–6064.
57. J. Kabrane and A. J. A. Aquino, J. Phys. Chem. A, 2015, 119, 1754–1764.
58. H. Li, J. Zhao, L. Luo, J. Du and J. Zeng, Acc. Chem. Res., 2021, 54, 1454–1464.
59. J. Dong, Y. Liu, J. Pei, H. Li, S. Ji, L. Shi, Y. Zhang, C. Li, C. Tang, J. Liao, S. Xu, H. Zhang, Q. Li and S. Zhao, Nat. Commun., 2023, 14, 6849.
60. K. A. Mauritz and R. B. Moore, Chem. Rev., 2004, 104, 4535–4586.
61. K. Kodama, K. Motobayashi, A. Shinohara, N. Hasegawa, K. Kudo, R. Jinnouchi, M. Osawa and Y. Morimoto, ACS Catal., 2018, 8, 694–700.
62. Z. B. Yan, D. H. Brouwer and G. R. Goward, Macromolecules, 2016, 49, 7331–7339.

---