Unraveling the effect of alkali cations on Fe single atom catalysts with high coordination numbers

Abstract

Fe single atom catalysts (SACs) with high coordination numbers have emerged as high-performance catalysts for the conversion of CO₂ to CO. However, the influence of alkali cations at the catalyst–electrolyte interface has not yet been understood clearly. Here, we investigate the role of various alkali metal cations (Na⁺, K⁺, Rb⁺) in catalytic CO₂ reduction reaction (CO₂RR) behavior on high coordination number Fe SACs (FeN₅ and FeN₆) obtained from a facile hard template method. We find that larger cations can greatly promote the CO₂RR and such effects are enhanced with increasing cation concentration. Nevertheless, the hydrogen evolution side reaction (HER) on co-existing N heteroatom sites will be worsened. This trade-off highlights the importance of manipulating the reactive sites for SACs. From theoretical simulation and in situ spectroscopy results, we confirm that the functioning mechanism of cations on Fe SACs lies in the enhancement of the adsorption of key intermediates through direct coordination and indirect hydrogen bonding effects. With the rationally designed Fe SACs (FeN₅) and the electrolyte conditions (1 M KOH), our flow cell test demonstrates a maximum Faraday efficiency of CO (FE_CO) of approximately 100% at 100 mA cm⁻². This research provides significant insights for future SACs and electrolyte design.

---

1. Introduction

Carbon dioxide (CO₂) is the primary greenhouse gas responsible for global warming, and there is an urgent need to develop technologies that can utilize CO₂. Electrochemical reduction of CO₂ is a promising approach to convert CO₂ into value-added chemicals such as CO,^1,2 HCOOH,^3,4 CH₄,^5,6 C₂H₄,^7–9 and C₂H₅OH.^10,11 Among these products, CO is an important industrial raw material for the Fischer–Tropsch reaction. However, the chemical inertness of CO₂ makes the CO₂-to-CO reaction challenging.¹² To address this issue, researchers have focused on developing various catalyst design strategies to promote the CO₂ reduction reaction (CO₂RR) to improve the reaction activity and selectivity.¹³

Recently, increasing attention has been paid to the electrolyte effect at the catalyst–electrolyte interface. In most cases, the CO₂RR takes place at the gas–liquid–solid three-phase boundary, and thus the electrolyte environment is a pivotal factor affecting the CO₂RR process.¹⁴ Previously, researchers have revealed the effect of different electrolytes on metal catalysts for the CO₂RR. Significant impacts of ionic species and their concentrations in the electrolyte on the faradaic efficiency (FE) and current density (J) have been observed on Au,¹⁵ Ag,¹⁶ and Cu catalysts.¹⁷ However, to our knowledge, there is very scarce research concerning the interaction between single atom catalyst (SACs) and electrolytes. Take Fe SACs as an example, which is a promising catalyst for the CO₂-to-CO conversion with maximum atom utilization and high catalytic activity.^18–21 Ana et al. found that pH has no apparent effect on the J_CO of Fe–N–C, but significantly enhances J_H2 at lower pH.²² Li et al. observed a volcano trend for J_CO at Fe-NS-C with increasing electrolyte concentration in the high overpotential region.²³ However, the underlying mechanism of the electrolyte effect on the performance of Fe SACs remains elusive. Recently, enhanced CO₂RR performance of Fe SACs with high coordination numbers has drawn great interest as a result of the changes of the active site symmetric and electronic structure.^24,25 This raises an open question regarding how the electrolyte would interact with such SACs with high coordination numbers during the CO₂RR. As a result, it is imperative to understand the functioning mechanism of the electrolyte, especially the cation, to guide future system design and industrial application of SACs.

Herein, we develop a facile hard template method to synthesize two different Fe SACs with high coordination numbers, FeN₆ and FeN₅. For both Fe SACs, increasing J_CO was observed with a higher cation radius. However, in contrast to traditional metal catalysts, HER side reactions also aggravated when the cation radius increased, which could explain the irregular change of the FE for CO conversion with different cations. Apart from the cation, increasing the ionic concentration enhances the FE_CO at Fe SACs at low overpotentials but shows the opposite effect at high overpotentials. However, the J_CO of FeN₆ shows a volcano trend at high overpotentials, which is attributed to the significantly enhanced HER due to the abundance of heteroatom sites with high HER activity and the relatively low CO₂RR activity of the Fe–N₆ site, as confirmed by the thiocyanate poisoning test. In contrast, FeN₅ can achieve higher J_CO at high overpotentials with the selective removal of HER sites and the promoted formation of Fe–N₅ sites with high CO₂RR activity. The density functional theory in classical explicit solvent (DFT-CES) simulation and in situ synchrotron radiation Fourier transform infrared (SR-FTIR) results indicate that cations can coordinate with *CO₂ or *COOH intermediates and enhance their hydrogen bonding with H₂O, thus lowering the reaction barrier and facilitating the subsequent charge transfer rate. Further tests of the FeN₅ SACs in the flow cell system demonstrate high selectivity (FE_CO ≈ 100% at 100 mA cm⁻²) in 1 M KOH. This work provides fundamental understandings of the interaction between SACs and electrolytes and guidance for the future design of SACs.

2. Experimental

2.1 Materials

Silica fumed powder (SiO₂, S5130) was purchased from Sigma-Aldrich Co. Ltd. 3,8-Dibromo-1,10-phenanthroline (C₁₂H₆Br₂N₂, 97%) was purchased from Bidepharm Co. Ltd. Sodium bicarbonate (NaHCO₃, ≥99.5%), tetrahydrofuran (THF, 99.5%), sulfuric acid (H₂SO₄, AR) and ethanol (C₂H₅OH, ≥99.7%) were bought from Sinopharm Chemical Reagent Co. Ltd. Potassium bicarbonate (KHCO₃, 99.5%), rubidium bicarbonate (RbHCO₃, 99%), rubidium sulfate (Rb₂SO₄, 99%), potassium sulfate (K₂SO₄, 99%), sodium hydroxide (NaOH, 99%) and ferric nitrate nonahydrate (Fe(NO₃)₃·9H₂O, 99.9%) were obtained from Macklin Co. Ltd. Sodium perchlorate (NaClO₄, 99%) was purchased from Aladdin Co. Ltd. All chemicals were used as received without further purification.

2.2 Synthesis of the FeN₆ SAC

The single-atom catalyst was synthesized by a simple hard template method. In a typical synthesis, 1 g SiO₂, 1 g C₁₂H₆Br₂N₂ and 0.7 g Fe(NO₃)₃·9H₂O were first dissolved in 60 mL THF and stirred for 4 h. Then, the solvent was removed by rotary evaporation. Subsequently, the obtained brown powder was pyrolyzed under flowing N₂ for 2 h at 800 °C. After that, the product was immersed in 2 M NaOH for 2–3 days to remove the template and then was leached in 0.5 M H₂SO₄ at 90 °C for 4 h to remove unstable metallic species. The catalyst was washed with DI water until pH = 7 and collected by centrifugation. Finally, by vacuum drying at 70 °C overnight, the FeN₆ SAC was obtained.

2.3 Synthesis of the FeN₅ SAC

The FeN₅ was obtained by H₂ treatment of FeN₆. The FeN₆ SAC was further pyrolyzed under flowing H₂/Ar = 5% for 1 hour at 800 °C to remove undesired pyrrolic N and pyridinic N. After H₂ treatment, the FeN₅ SAC was successfully obtained.

2.4 Characterization

The X-ray diffraction pattern (XRD) was obtained through a Japan Rigaku Miniflex 600 using Cu Kα radiation (1.54 Å). Raman scattering spectra were recorded with a Renishaw System 2000 spectrometer. N₂ adsorption/desorption analysis were carried out at 77 K using the Micromeritics ASAP 2020 system. Scanning electron microscopy (SEM) observations were carried out on a Gemini SEM 450. Transmission electron microscopy (TEM), Energy dispersive X-ray (EDX) mapping, high-angle annular dark-field scanning transmission electron microscopy (HADDF-STEM) and Aberration-corrected (AC) HAADF-STEM images were acquired on a ThermoFisher Themis. An inductively coupled plasma atomic emission spectrometer (ICP-AES) was employed to quantify the content of Fe using a Thermo scientific iCAP 7400 series instrument. X-ray photoelectron spectroscopy (XPS) measurements were performed using an ESCALAB 250Xi high-performance electron spectrometer. The X-ray absorption spectra (XAS) including X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) of the samples were collected at the Singapore Synchrotron Light Source (SSLS) center, where a pair of channel-cut Si (111) crystals was used in the monochromator. The Fe K-edge XANES data were recorded in a transmission mode. Fe foil was used as the reference. The storage ring was working at the energy of 2.5 GeV with an average electron current of below 200 mA. The acquired EXAFS data were extracted and processed according to the standard procedures using the Athena module implemented in the IFEFFIT software packages.

The in situ Specular Reflection Fourier Transform Infrared (SR-FTIR) measurements were recorded between 600 and 4000 cm⁻¹ at the Infrared Spectroscopy and Microspectroscopy Endstation (BL01B) in the National Synchrotron Radiation Laboratory (NSRL) in Hefei, China (Fig. S29†). This end station was equipped with an FTIR spectrometer (Bruker 66 v/s) with a KBr beam-splitter and various detectors (here a liquid nitrogen cooled mercury cadmium telluride detector was used) coupled with an infrared microscope (Bruker Hyperion 3000) with a ×16 objective, and can provide infrared spectroscopy measurement with a high spectral resolution of 0.25 cm⁻¹. Each infrared absorption spectrum was acquired by averaging 128 scans at a resolution of 4 cm⁻¹. The CO₂ saturated electrolyte was selected as the flowing electrolyte and pumped into an electrochemical in situ IR cell (Beijing Scistar Technology Co., Ltd) with a flow rate of 2 mL min⁻¹. To guarantee CO₂ saturated the electrolyte, we purged CO₂ in each electrolyte over 10 minutes before the test. The conductive tape was used to gum the catalyst on the glassy carbon electrode to fabricate the working electrode. There was less electrolyte between the electrode and ZnSe window, containing active species adsorbed on the catalyst. All infrared spectral acquisitions were carried out after a constant potential was applied to the working electrode for 10 min. The background spectrum of the working electrode was acquired at an open-circuit voltage before each systemic CO₂RR measurement (Fig. S30a†), and the measured potential ranges of the CO₂RR were −1.0 to −1.6 V vs. Ag/AgCl.

2.5 Electrochemical CO₂ reduction experiment

Different electrolytes (see pH in Table S6†) were prepared by bubbling CO₂ into different carbonate and bicarbonate aqueous solutions over 30 minutes. The H-type cell (Fig. S10†) was used as the equipment for CO₂ electrochemical reduction. A Nafion 117 membrane was inserted between the cathodic chamber and anodic chamber. A gas mass flow controller was used to set the CO₂ flow rate at 30 sccm. The Pt sheet and Ag/AgCl were used as the counter electrode and reference electrode, respectively. The working electrode was prepared by the following method: first, 10 mg of the as-synthesized catalyst, 950 μL ethanol, and 50 μL Nafion 117 solution were mixed and sonicated for above 30 minutes to form a homogeneous ink. Next, 100 μL ink was coated on each side of carbon paper (the total area of paper is 1 × 1 cm²) to ensure a loading of 1 mg cm⁻². Last, this electrode was dried in an oven overnight for next use. Before all electrochemical measurement, the working electrode was activated by 100 cycles of cyclic voltammetry (CV) test from −0.6 V to −2.0 V (vs. Ag/AgCl). The LSV curves were conducted with the CHI760E electrochemical workstation with a scan rate of 5 mV s⁻¹. The potentials vs. the reversible hydrogen electrode (RHE) were measured by transforming the recorded potentials against Ag/AgCl in this work using the following equation:

E_RHE = E_Ag/AgCl + 0.204 + 0.059 × pH

The chronoamperometry tests were conducted at each potential for 10 min (Fig. S14†). The gas products of electrolysis were detected on the FULI GC9702 plus gas chromatograph. High purity helium (99.9999%) was used as the carrier gas for the chromatography. The liquid products of electrolysis were analyzed using a Bruker Avance III HD 400 NMR. The Faraday efficiency (FE) of gas products was calculated by the equation:

V (mL min⁻¹) = gas flow rate measured by a flow meter at the exit of the cell at room temperature and under ambient pressure. n = total number of electrons transferred in the reaction, CO and H₂ are 2. A = volume concentration of CO and H₂ in the exhaust gas from the electrolyzer (GC data). C = the total electricity consumed by the reaction.

Both CO₂RR and HER make contribution for the current in the aqueous electrolyte. The reduction products are detected by gas chromatography and nuclear magnetic resonance for the calculation of FE to accurately evaluate the CO₂RR activity. No liquid products are detected after electrolysis (Fig. S11†). CO is detected from the flame ionization detector (FID) and H₂ is detected from the thermal conductivity detector (TCD). In order to distinguish the contribution of the CO₂RR and HER in current, the partial current density was calculated by the equation:

J_CO = J_total × FE_CO

J_H2 = J_total × FE_H2

J_total = current density measured by chronoamperometry at different potentials.

For experiments in flow cells, the electrode was prepared by hand-painting the catalyst ink onto the gas diffusion layer (GDL). 10 mg of the FeN₅ catalyst, 50 μL of Nafion, and 950 μL of ethanol were mixed with ultrasound. The 400 μL ink was airbrushed on the GDL (the total area of paper is 2 × 2 cm²) to ensure a loading of 1 mg cm⁻². The prepared gas diffusion electrode (GDE) and Pt foam were used as the cathode and anode, respectively. Their chambers were filled with 1 M KOH. A cation exchange membrane (Nafion 117 membrane) was used to separate the cathode and anode. During the measurements, KOH electrolyte flow was kept constant at 20 mL min⁻¹, and CO₂ gas with a flow rate of 50 sccm was directly fed to the cathode GDL. In order to ensure the experiments to be more accurate, a mass flowmeter was used to monitor the export flow rate.

2.6 DFT calculation

Density functional theory (DFT) computational simulation was performed with the Vienna ab initio simulation package (VASP). The generalized gradient approximation (GGA) within the Perdew–Burke–Ernzerhof (PBE) exchange-correlation functional was used in the calculations.^26,27 A vacuum layer of 15 Å was used to avoid periodic image interaction in the z direction, and a kinetic energy cut-off of 500 eV was chosen. A 5 × 5 graphene supercell was used for simulation, and the surface Brillouin zone was sampled using a 3 × 3 × 1 Monkhorst–Pack k point grid. The force convergence criteria and energy convergence criteria were set at 10⁻⁵ eV, and the convergence threshold was set at 0.02 eV Å⁻¹, respectively. In addition, we also considered adding 25 H₂O molecules (density of ∼1 g cm⁻³) to the subsequent liquid environment for simulation.^28,29

The formation energy (E_for) can be used to characterize the stability of the substrate, the formula is: E_for = E_total − E_slab − E_FeN5/6, where E_total, E_slab, and E_FeN5/6 are the total energy of the catalyst, graphene slab and FeN_5/6 molecules respectively. Meanwhile, for the strength of molecular adsorption, adsorption energy (E_a) is defined as: E_a = E_X/sub − E_sub − E_X, where E_total, E_sub, and E_X are the total energy of the adsorbent, substrate and molecules respectively. According to this definition, the more negative the adsorption energy, the stronger the adsorption.

The reaction processes of reducing CO₂ to CO are calculated as follows:

* + CO₂ → *CO₂

*CO₂ + H⁺ + e⁻ → *COOH

*COOH + H⁺ + e⁻ → *CO + H₂O

*CO → CO + *

The Gibbs free energy for the CO₂RR process is calculated as: ΔG = ΔE − TΔS + ΔZPE, where ΔE is the total optimal energy of DFT, ΔS is the entropy difference between the gas phase and adsorption state, ΔZPE corresponds to the zero-point energy, and T is = 298.15 K.

3. Results and discussion

3.1 Synthesis and characterization of Fe single atom catalysts

Because the asymmetric electronic structure of the active site in the high coordination number Fe SACs is beneficial for the CO₂RR compared with FeN₄, we synthesized mesoporous FeN₆ using a simple hard template method (Fig. 1a, see details in the ESI†). Since the o-phenanthroline can form a six-coordinate structure with Fe³⁺, 3,8-dibromo-1,10-phenanthroline was used as the nitrogen source. A silica template was used to obtain a mesoporous structure, which in turn facilitates the contact between the catalyst and the electrolyte. In order to systematically study the electrolyte effect on high-coordination Fe SACs, FeN₅ was also synthesized from FeN₆ by hydrogen treatment. Their structures could be confirmed by various characterization methods. The X-ray diffraction (XRD) pattern (Fig. S2a and S4a†) shows two main peaks at 25° and 42°, corresponding to the (002) and (100) planes of the graphitic carbon, which are also confirmed by the Raman result (Fig. S2b and S4b†). No other obvious peaks can be observed, indicating no iron particles exist after the acid treatment. Fig. S2c and S4c† show the nitrogen adsorption/desorption curves representing the IUPAC IV type, and the pore size distribution map (Fig. S2d, S4d and Table S1†) indicates that that the mean pore sizes of the two samples are around 8 nm. Scanning and transmission electron microscopy (SEM and TEM) images further confirm the mesoporous structure (Fig. 1b, S1a, b and S3a, b†) of the materials. It can be concluded from energy dispersive X-ray (EDX) mapping (Fig. S1d and S3e†) that C, N, O and Fe are evenly distributed in the catalysts. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) shows no bright dots until using aberration-corrected (AC) HAADF-STEM (Fig. 1c and S3c–f†), further demonstrating that there are only Fe single atom sites in the two catalysts. The valence and coordination environment of Fe atoms in the prepared catalysts were analyzed using synchrotron X-ray absorption spectroscopy (XAS). Fig. 1d shows the Fe K-edge X-ray absorption near edge structure (XANES) spectra of FeN₅ and FeN₆. They exhibited different energy absorption edge profiles, demonstrating that the valence of Fe in FeN₆ is higher than that in FeN₅. The Fourier transformed extended X-ray absorption fine structure (FT-EXAFS) (Fig. 1e) indicates that only one dominate peak around 1.5 Å exists in these materials, which can be attributed to the first Fe–N/O shell.¹⁸ The absence of the Fe–Fe peak around 2.2 Å in the two catalysts further confirms the atomic dispersion of Fe. Furthermore, the fitting results (Fig. S5 and Table S3†) show that the coordination numbers of the two samples are 5.5 and 6, respectively. Moreover, the wavelet transform (WT) of χ(k) spectra (Fig. S6†) reveals that the centers of the [k, R] intensity maximum for FeN₅ and FeN₆ are similar, while significantly different from that of Fe foil, indicating that only Fe–N/O bonds are present in our samples. X-ray photoelectron spectroscopy (XPS) was used to analyze the non-metallic and metallic elements in detail in the catalyst to exclude oxygen coordination (Table S2†). The spectral profile of N 1s (Fig. 1f and g) exhibited five peaks corresponding to pyridinic N (398.4 eV), Fe–N (399.5 eV), pyrrolic N (400.1 eV), graphitic N (401.0 eV), and oxidized N (403.2 eV).^19,30 The contents of pyrrolic N and pyridinic N significantly decreased in FeN₅ due to hydrogen treatment (Table S4†). In addition, no M–O bond (530 eV) was observed in the O 1s spectrum (Fig. S8†).^25,31 These results confirmed that Fe–N₅ and Fe–N₆ are the actual Fe sites of the two catalysts. The Fe spectral profile (Fig. S9†) indicated that the oxidation state was between +2 and +3, and the content of Fe³⁺ in FeN₆ was higher than that in FeN₅, which are consistent with XANES results (Fig. S7 and Table S5†) and further confirm the higher valence of Fe in FeN₆. In conclusion, mesoporous Fe SACs with high coordination numbers were successfully synthesized. Next, these SACs are used to explore the difference in catalytic performance under different electrolyte conditions.

Fig. 1 (a) Schematic illustration of the hard template method for preparing FeN₆ and FeN₅. (b) SEM image of FeN₆. (c) AC-HADDF-STEM image of FeN₆ (red circles are the Fe single atom sites). (d) The Fe K-edge XANES spectra. (e) The Fe K-edge FT-EXAFS spectra. (f) The N spectrum of FeN₆. (g) The N spectrum of FeN₅.

3.2 CO₂ electrochemical reduction performance

The electrochemical CO₂RR performance was tested in a three-electrode setup using the H-type cell with CO₂ saturated solution (Fig. S10 and Table S6†). Although different electrolyte conditions will lead to pH changes, this change is very small, and the effect of pH on the properties of Fe SACs has been explored by previous researchers,²¹ so the effect of pH is no longer considered in this paper. The gas products were quantified by gas chromatography (Fig. S12 and S13†). No other products except CO and H₂ can be detected (Fig. S11†). It is surprising that no clear trends of cation activity on FeN₆ can be observed (Fig. 2a and S15a, b†). Similarly, the FE_CO values are close for FeN₅ for different cations, only the case of Na⁺ is slightly lower (Fig. S20b†). These contrast with metal catalysts like Ag¹⁶ and Cu,¹⁷ for which the selectivity of the CO₂RR typically increases with increasing cation radius. However, when increasing the cation concentration (Fig. 2b, S15c, d and S19b†), it can be found for both FeN₅ and FeN₆ that the FE_CO significantly improved at low overpotentials, but decreased at high overpotentials. These observations align with prior studies on SACs,^23,32 indicating that SACs share strong interactions with electrolyte species. As demonstrated in the linear sweep voltammetry (LSV) test (Fig. S16, S19a and S20a†), the activity of the catalysts increases noticeably as the electrolyte concentration increases, and the order of activity follows the sequence: Rb⁺ > K⁺ > Na⁺ under all conditions.

Fig. 2 The FE_CO of FeN₆ under the (a) same concentration conditions and (b) same cation conditions. The (c₁) J_CO and (c₂) J_H2 of FeN₆ at low overpotential (about −0.6 V vs. RHE) and high overpotential (about −1.0 V vs. RHE) under different electrolyte conditions. The (d₁) LSV and (d₂) FE_CO at a constant concentration of HCO₃⁻. The (e₁) LSV, (e₂) FE_CO at a constant concentration of Rb⁺.

To gain a better understanding of the changes in selectivity and activity, the partial CO and H₂ current densities (J_CO and J_H2) are examined, which could more accurately reflect the CO₂RR and HER activity. As seen in Fig. 2c₁ and c₂, J_CO and J_H2 follow the same order as Rb⁺ > K⁺ > Na⁺ (Fig. S17a–c, S18a–c and S20c, d†). This result may explain the reason that no clear trend of FE_CO was observed on FeN₆ and FeN₅ with different cations, as the HER is aggravated simultaneously. In addition to cations, it is noteworthy that J_CO and J_H2 of both SACs (Fig. S17d–f, S18d–f and S19c, d†) were also boosted with increasing cation concentration, except that the J_CO of FeN₆ shows a volcano trend at high overpotentials. The enhanced HER is related to the plateau or even the declining trend of J_CO at the 0.5 M concentration electrolyte. Therefore, it can be concluded that higher concentration and larger cations can simultaneously promote the CO₂RR and HER on Fe SACs. Such phenomena differ significantly from traditional metal catalysts where only the CO₂RR can be promoted but the HER did not change obviously or even be hindered under the same conditions,^33,34 and thus further investigation is required to explain these unique findings.

To determine whether the alkali metal cation or the bicarbonate is the major factor for the electrolyte effect, we conducted experiments with constant alkali metal cations or bicarbonate concentrations for FeN₆ first. As shown in Fig. 2d₁ and d₂, changing the Rb⁺ concentration significantly affects selectivity and activity. Increasing the Rb⁺ concentration promotes J_CO at low overpotentials but it reaches a plateau at high overpotentials (Fig. S26a₁ and a₂†), while J_H2 increases at the whole potential range. This is very similar to the change in RbHCO₃ concentration discussed earlier. However, when the concentration of Rb⁺ is constant, the catalytic performance has no obvious change (Fig. 2e₁, e₂ and S26b₁, b₂†). Similar results were obtained with different alkali metal cations (Na⁺ and K⁺) and FeN₅ (Fig. S25, S27, S28 and Tables S7–S12†). These experiments reveal that the concentration of alkali metal cations is the primary factor, illustrating the essential role of cations present in the electrolyte in the SAC activity. Moreover, bicarbonate may not affect the performance of SACs because HCO₃⁻ is negatively charged, making it difficult to access the interface layer and participate in the proton–electron transfer process, which is different from previous results on metal catalysts.¹⁵

3.3 Identifying the reaction site to uncover the cation influence

In traditional metal catalysts, both the CO₂RR and HER occur at the same site, which means that the competition between the two reactions is essentially a competition for the adsorbed species.³⁵ However, for SACs with a variety of active sites, the CO₂RR and HER could occur at different sites whiling competing for charges (Fig. 3d). Due to the diversity of catalytically active sites on SACs, we first distinguish the HER and CO₂RR sites through poisoning the Fe single atom sites with SCN⁻via specific adsorption (Fig. 3a and b).³⁶ It should be noted that the overall cation concentration is controlled to be constant, this is to prevent the performance change from being caused by cations. The addition of KSCN to the electrolyte indeed inhibited the CO₂RR in both catalysts. However, in an Ar atmosphere where no CO₂RR product was detected (Fig. S24†), the LSV curves corresponding to the HER did not show any significant changes (Fig. 3c). Based on our XPS and XAS results (Fig. 1d–g), it can be concluded that the Fe–N_x sites are the actual CO₂RR site, while the pyridine N and pyrrolic N sites are the HER sites, consistent with previous studies.^18,37 Therefore, the decreased HER activity of FeN₅ compared to FeN₆ under Ar conditions (Fig. 3c and S23†) results from the removal of pyridine N and pyrrolic N by hydrogen treatment. Besides, FeN₅ has higher FE_CO and J_{CO ECSA} than FeN₆ (Fig. S22†), and previous research³⁰ reveals that this is due to the more active site FeN₅ than FeN₆, however, in this work, we confirm that the N site can also influence the performance. Therefore, we propose a “trade-off effect” to describe the electrolyte effect on CO₂RR activity for Fe SACs, where the two competing active sites are responsive to cations in the electrolyte. Since previous studies^38,39 have examined the role of alkali metal cations in the HER and revealed their effect on accelerating the Volmer step (H₂O → OH⁻ + *H), our focus here is to explore the cation effect on FeN_x sites during the CO₂RR.

Fig. 3 Catalytic performance of (a) FeN₆ and (b) FeN₅ with (0.4 M KHCO₃ + 0.1 M KSCN) and without 0.1 M KSCN (0.5 M KHCO₃). (c) The LSV curves of FeN₅ and FeN₆ in an Ar atmosphere with and without 0.1 M KSCN. (d) Schematic illustration of reaction sites in Fe SACs.

3.4 Revealing the bonding interaction between cations and intermediates

To gain mechanistic understanding of the effect of different cations on the CO₂RR catalytic process, DFT simulations were respectively performed on FeN₅ and FeN₆ catalysts. First, we construct two different FeN₅ structural models based on the fine structure analysis shown in Fig. S31a and b.† Table S13† indicates that using pyrrolic N as the coordinating ligand to build FeN₅ based on the planar FeN₄ is more stable than the pyridinic N ligand. In light of this result, we continue to simulate FeN₆ using dual pyrrolic N ligands (Fig. S31c†). The FeN₆ using dual pyridinic N is also created, but this structure cannot exist stably (Fig. S31d†), so it can be confirmed that the two additional N in FeN₆ are pyrrole N. The two pyrrole N are transformed from the pyridine N in the o-phenanthroline ring at high temperature. Previous studies have shown that it is unfavourable for CO₂ adsorption when the two pyrrole nitrogen ligands are on opposite sides.³⁰ Here, we chose to investigate the CO₂RR using pyrrole N ligands positioned on the same side. Additionally, the pyrrolic N and pyridinic N are also created (Fig. S32†) as HER sites. Fig. 4b shows that FeN₅ and FeN₆ have lower *CO₂ free energy than *H while the pyrrolic N and pyridinic N are favorable for *H adsorption, supporting that the CO₂RR occurs at the FeN_5/6 and the pyrrolic N/pyridinic N is more prone to the HER. The free energy comparison further confirms the trade-off effect, where two different functional sites undergo different reactions under negative bias.

Fig. 4 (a) Schematic illustration of the effect of different cations on reaction sites. The Gibbs free energy of (b) different reactions on different sites and (c) different intermediates under various electrolyte conditions. The charge density difference (CDD) of (d₁) FeN₆ with adsorbed *COOH under Rb⁺ conditions and (d₂) FeN₅ with adsorbed *CO₂ under Rb⁺ conditions. The in situ SR-FTIR spectrum of FeN₆ in (e₁) 0.1 M NaHCO₃, (e₂) 0.5 M NaHCO₃, and (e₃) 0.5 M RbHCO₃. (f) The FE_CO and J_CO of FeN₅ under flow cell conditions (1 M KOH).

Based on the knowledge of previous research on the impact of cations on the CO₂RR, we speculate that the main reason for the influence of cation concentration and species on FeN_x sites is the change in the charge distribution, which causes the adsorption strength of intermediates to change. Because in recent years, researchers have discovered that alkali metal cations can bind with intermediates and stabilize them.⁴⁰ We first identify the critical intermediates (i.e., the adsorbed species corresponding to the rate-determining step (RDS)) on different catalysts to accurately describe the influence of alkali metal cations. Fig. S33 and S34† reveal that the adsorption of CO₂ and the formation of *COOH correspond to the RDS of FeN₅ and FeN₆, respectively. Based on these findings, *CO₂ and *COOH are designated as the key adsorbed species of FeN₅ and FeN₆ for the study of the cation effect. Additionally, the RDS energy barrier of FeN₅ is lower than that of FeN₆, which further demonstrates its superior catalytic performance (Fig. S34†). To investigate the effect of cations on the FeN₅ and FeN₆ sites and confirm our speculation, DFT-CES simulation was employed, as it allows for precise examination of atomic details at the catalyst–electrolyte interface.⁴¹ Besides, this method can also help us understand how cations affect the adsorption strength of intermediates. The DFT-CES calculation (Fig. 4c and Table S14†) indicates that the adsorption of both *CO₂ (on FeN₅) and *COOH (on FeN₆) is enhanced in the presence of cations compared to pure water. This finding suggests that higher cation concentrations can result in better performance, as more cations can reach the inner Helmholtz layer and act on more FeN_x sites to facilitate the CO₂RR.⁴² Moreover, cations with larger radii (e.g., Rb⁺) can further promote the adsorption of intermediates. The examination of the charge distribution (Fig. 4d and S35–S37†) reveals that cations can enhance the charge transfer by coordination, facilitating charge accumulation in the intermediate and promoting its adsorption. Furthermore, it is important to note that in addition to the direct coordination effect of cations, there is also an indirect effect to facilitate charge transfer. Further analysis of bonding length (Table S15†) reveals that H₂O can also interact with the intermediate. Specifically, the H atom of H₂O bonds with the O atom of *CO₂ and *COOH, forming hydrogen bonds that alter the charge distribution. Based on our analysis, the conclusion can be drawn that the cation has two kinds of effects on charge accumulation in the intermediate. First, the cation directly coordinates with the intermediate to facilitate charge transfer. Second, the coordination between the O atom in H₂O and the cation alters the charge distribution in the water molecule. This, in turn, strengthens the hydrogen bond effect between the H atom in H₂O and the O atom in the intermediate. We suspect that the direct bonding of cations has a stronger effect on the promotion of charge transfer than that of the hydrogen bond, which is relatively weak intermolecular interaction. Therefore, the presence of Rb⁺ enhances adsorption. This is due to the fact that larger cations have a thinner solvation shell, which enables them to approach the intermediates more closely.³⁵ As a result, the direct coordination is enhanced. In contrast, Na⁺ with a larger solvation shell is predominantly influenced by indirect hydrogen bonding through H₂O, resulting in weaker adsorption than Rb⁺. In conclusion, higher concentration and larger cations can facilitate the stabilization of key intermediates at more sites, thus accelerating the CO₂RR (Fig. 4a). To validate the above conclusion, in situ SR-FTIR was employed to detect the intermediate. As presented in Fig. 4e₁–e₃, the vibration band at 1700 cm⁻¹ can be assigned to the C=O stretching of the generated *COOH intermediate during the CO₂RR.⁴³ A greater extent of the critical *COOH intermediate on FeN₆ could be observed with larger cations and higher concentration, which is consistent with the DFT-CET calculation results that an increase in charge transfer promotes the adsorption of *COOH. The in situ spectroscopy tests confirm that changes in cation species and their concentration could indeed alter the distribution of intermediates on the FeN_x active sites. Based on these findings, we select FeN₅ for CO₂RR tests under flow cell conditions with a high cation concentration (1 M KOH), as it exhibits better catalytic ability and strong alkaline conditions that can inhibit the HER.²² The reason why RbOH is not selected as the electrolyte is that its price is too expensive, so the cost of configuring it into a high-concentration electrolyte is not appropriate. As shown in Fig. 4f, S38 and S39,† the maximal FE_CO can reach approximately 100% at 100 mA cm⁻². The impressive CO₂-to-CO performance of FeN₅ highlights its potential for practical applications.

4. Conclusion

In summary, FeN₅ and FeN₆ were first synthesized using a facile hard template method. These catalysts exhibit enhanced performance for both the CO₂RR and HER under higher concentration and larger cation radius conditions, owing to their unique structure. Through rational experiments and DFT simulations, we propose the trade-off effect to explain the unconventional phenomenon, where cations can act on different functional sites of the catalyst. In addition, it was also found that the promotion effect of cations on FeN₅ and FeN₆ is mainly due to the promotion of charge accumulation in *CO₂ and *COOH, thereby enhancing the adsorption strength. Cations promote charge transfer through direct coordination and indirect hydrogen bonding. Larger cations can further promote the stabilization of intermediates through direct coordination. This work provides a novel concept for understanding the interaction between alkaline metal cations and Fe SACs and offers valuable insights for future SACs and electrolyte design.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Yecheng Li: conceptualization, formal analysis, investigation, methodology, visualization, writing – original draft; Songjie Meng: methodology, formal analysis, writing – review & editing; Zihong Wang: investigation, formal analysis, writing – review & editing, supervision; Hehe Zhang: investigation; Xin Zhao: investigation, writing – review & editing, supervision; Qingshun Nian: investigation, writing – review & editing, supervision; Digen Ruan: investigation, supervision; Zhansheng Lu: methodology, formal analysis; Xiaodi Ren: conceptualization, funding acquisition, methodology, supervision, writing – review & editing.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (Grant No. 22179124, 21905265, and 12274118), the Fundamental Research Funds for the Central Universities (WK3430000007), and Henan Center for Outstanding Overseas Scientists (No. GZS2023007). The SR-FTIR was performed at the Infrared Spectroscopy and Microspectroscopy Endstation (BL01B) in the National Synchrotron Radiation Laboratory (NSRL) in Hefei, China; the DFT simulation was supported by the High Performance Computing Center of Henan Normal University.

References

1. J. Li, H. Zeng, X. Dong, Y. Ding, S. Hu, R. Zhang, Y. Dai, P. Cui, Z. Xiao, D. Zhao, L. Zhou, T. Zheng, J. Xiao, J. Zeng and C. Xia, Nat. Commun., 2023, 14, 340.
2. H. Guo, D.-H. Si, H.-J. Zhu, Q.-X. Li, Y.-B. Huang and R. Cao, eScience, 2022, 2, 295–303.
3. L. Li, Z. Liu, X. Yu and M. Zhong, Angew Chem. Int. Ed. Engl., 2023, 135, e202300226.
4. Z. Tao, Z. Wu, X. Yuan, Y. Wu and H. Wang, ACS Catal., 2019, 9, 10894–10898.
5. J. Zhao, P. Zhang, T. Yuan, D. Cheng, S. Zhen, H. Gao, T. Wang, Z. J. Zhao and J. Gong, J. Am. Chem. Soc., 2023, 145, 6622–6627.
6. X. Tan, K. Sun, Z. Zhuang, B. Hu, Y. Zhang, Q. Liu, C. He, Z. Xu, C. Chen, H. Xiao and C. Chen, J. Am. Chem. Soc., 2023, 145, 8656–8664.
7. 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.
8. D. Xiao, X. Bao, M. Zhang, Z. Li, Z. Wang, Y. Gao, Z. Zheng, P. Wang, H. Cheng, Y. Liu, Y. Dai and B. Huang, Chem. Eng. J., 2023, 452, 139358.
9. W. Li, Z. Yin, Z. Gao, G. Wang, Z. Li, F. Wei, X. Wei, H. Peng, X. Hu, L. Xiao, J. Lu and L. Zhuang, Nat. Energy, 2022, 7, 835–843.
10. A. Xu, S.-F. Hung, A. Cao, Z. Wang, N. Karmodak, J. E. Huang, Y. Yan, A. Sedighian Rasouli, A. Ozden, F.-Y. Wu, Z.-Y. Lin, H.-J. Tsai, T.-J. Lee, F. Li, M. Luo, Y. Wang, X. Wang, J. Abed, Z. Wang, D.-H. Nam, Y. C. Li, A. H. Ip, D. Sinton, C. Dong and E. H. Sargent, Nat. Catal., 2022, 5, 1081–1088.
11. S. Kuang, Y. Su, M. Li, H. Liu, H. Chuai, X. Chen, E. J. M. Hensen, T. J. Meyer, S. Zhang and X. Ma, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2214175120.
12. Y. N. Gong, C. Y. Cao, W. J. Shi, J. H. Zhang, J. H. Deng, T. B. Lu and D. C. Zhong, Angew Chem. Int. Ed. Engl., 2022, 61, e202215187.
13. J. Yin, J. Jin, Z. Yin, L. Zhu, X. Du, Y. Peng, P. Xi, C. H. Yan and S. Sun, Nat. Commun., 2023, 14, 1724.
14. Z.-M. Zhang, T. Wang, Y.-C. Cai, X.-Y. Li, J.-Y. Ye, Y. Zhou, N. Tian, Z.-Y. Zhou and S.-G. Sun, Nat. Catal., 2024, 7, 807–817.
15. G. Marcandalli, A. Goyal and M. T. M. Koper, ACS Catal., 2021, 11, 4936–4945.
16. M. R. Singh, Y. Kwon, Y. Lum, J. W. Ager and A. T. Bell, J. Am. Chem. Soc., 2016, 138, 13006–13012.
17. J. Resasco, L. D. Chen, E. Clark, C. Tsai, C. Hahn, T. F. Jaramillo, K. Chan and A. T. Bell, J. Am. Chem. Soc., 2017, 139, 11277–11287.
18. C. Liu, Y. Wu, K. Sun, J. Fang, A. Huang, Y. Pan, W.-C. Cheong, Z. Zhuang, Z. Zhuang, Q. Yuan, H. L. Xin, C. Zhang, J. Zhang, H. Xiao, C. Chen and Y. Li, Chem, 2021, 7, 1297–1307.
19. Y. Wang, B. J. Park, V. K. Paidi, R. Huang, Y. Lee, K.-J. Noh, K.-S. Lee and J. W. Han, ACS Energy Lett., 2022, 7, 640–649.
20. S. Chen, X. Li, C. W. Kao, T. Luo, K. Chen, J. Fu, C. Ma, H. Li, M. Li, T. S. Chan and M. Liu, Angew Chem. Int. Ed. Engl., 2022, 61, e202206233.
21. J. Gu, C.-S. Hsu, L. Bai, H. M. Chen and X. Hu, Science, 2019, 364, 1091–1094.
22. A. S. Varela, M. Kroschel, N. D. Leonard, W. Ju, J. Steinberg, A. Bagger, J. Rossmeisl and P. Strasser, ACS Energy Lett., 2018, 3, 812–817.
23. F. Pan, B. Li, E. Sarnello, S. Hwang, Y. Gang, X. Feng, X. Xiang, N. M. Adli, T. Li, D. Su, G. Wu, G. Wang and Y. Li, Nano Energy, 2020, 68, 104384.
24. H. Zhang, J. Li, S. Xi, Y. Du, X. Hai, J. Wang, H. Xu, G. Wu, J. Zhang, J. Lu and J. Wang, Angew Chem. Int. Ed. Engl., 2019, 58, 14871–14876.
25. T. Zhang, X. Han, H. Liu, M. Biset-Peiró, J. Li, X. Zhang, P. Tang, B. Yang, L. Zheng, J. R. Morante and J. Arbiol, Adv. Funct. Mater., 2022, 32, 2111446.
26. D. R. Hamann, M. Schlüter and C. Chiang, Phys. Rev. Lett., 1979, 43, 1494–1497.
27. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
28. T. Sheng and S.-G. Sun, Chem. Phys. Lett., 2017, 688, 37–42.
29. X. Qin, T. Vegge and H. A. Hansen, J. Am. Chem. Soc., 2023, 145, 1897–1905.
30. H. Chen, X. Guo, X. Kong, Y. Xing, Y. Liu, B. Yu, Q.-X. Li, Z. Geng, R. Si and J. Zeng, Green Chem., 2020, 22, 7529–7536.
31. S. Huang, B. Hu, S. Zhao, S. Zhang, M. Wang, Q. Jia, L. He, Z. Zhang and M. Du, Chem. Eng. J., 2022, 430, 132933.
32. 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.
33. W. Ren, A. Xu, K. Chan and X. Hu, Angew Chem. Int. Ed. Engl., 2022, 61, e202214173.
34. A. S. Malkani, J. Anibal and B. Xu, ACS Catal., 2020, 10, 14871–14876.
35. B. Pan, Y. Wang and Y. Li, Chem Catal., 2022, 2, 1267–1276.
36. Y. Chen, L. Zou, H. Liu, C. Chen, Q. Wang, M. Gu, B. Yang, Z. Zou, J. Fang and H. Yang, J. Phys. Chem. C, 2019, 123, 16651–16659.
37. C. Wang, Y. Liu, H. Ren, Q. Guan, S. Chou and W. Li, ACS Catal., 2022, 12, 2513–2521.
38. M. C. O. Monteiro, A. Goyal, P. Moerland and M. T. M. Koper, ACS Catal., 2021, 11, 14328–14335.
39. J. T. Bender, A. S. Petersen, F. C. Østergaard, M. A. Wood, S. M. J. Heffernan, D. J. Milliron, J. Rossmeisl and J. Resasco, ACS Energy Lett., 2022, 8, 657–665.
40. M. C. O. Monteiro, F. Dattila, B. Hagedoorn, R. García-Muelas, N. López and M. T. M. Koper, Nat. Catal., 2021, 4, 654–662.
41. S. J. Shin, H. Choi, S. Ringe, D. H. Won, H. S. Oh, D. H. Kim, T. Lee, D. H. Nam, H. Kim and C. H. Choi, Nat. Commun., 2022, 13, 5482.
42. J. Yu, J. Yin, R. Li, Y. Ma and Z. Fan, Chem Catal., 2022, 2, 2229–2252.
43. J. Pei, T. Wang, R. Sui, X. Zhang, D. Zhou, F. Qin, X. Zhao, Q. Liu, W. Yan, J. Dong, L. Zheng, A. Li, J. Mao, W. Zhu, W. Chen and Z. Zhuang, Energy Environ. Sci., 2021, 14, 3019–3028.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5sc00581g

‡ Y. L. and S. M. contributed equally.

---