Switching CO₂ Electroreduction Selectivity Between C₁ and C₂ Hydrocarbons on Cu Gas-Diffusion Electrodes

2.1 Synthesis and Characterization of Cu GDEs

The Cu GDEs (Cu-X GDEs, X denotes the deposition voltage) were prepared by the potentiostatic electrodeposition method, as illustrated in Figure 1a. During the synthesis process, Cu foam at the anode was first anodized to form Cu²⁺ ions, while the hydrogen evolution reaction (HER) occurred on the cathode of gas-diffusion layer (GDL, Sigracet 35 BC) (stage I). Most of Cu²⁺ ions were released to the KOH solution, followed by coordination with OH⁻ ions to form [Cu(OH)₄]²⁻ complexes. The [Cu(OH)₄]²⁻ complexes diffused to the cathode where they were reduced and electrodeposited onto the GDL to form Cu GDEs (stage II and III). Due to the accumulating formation of Cu(OH)₂ on the surface of Cu foam, the predominant reaction on the anode was switched from Cu anodization to oxygen evolution reaction (OER) on Cu(OH)₂ (stage III). The corresponding time-dependent current density curves revealed these three stages (Figure 1b). Stage I only lasted for about 35 s under the different deposition voltages, indicating the same reactions occurred in stage I. The period of stage II became shorter as the electrodeposition voltage increased. That is because the formation rate of Cu(OH)₂ layer was increased by increasing the electrodeposition voltage. At the transition from stage II to stage III, the current density abruptly decreased because the sluggish kinetics OER started to restrict the current density. The electrodeposition on the cathode continued at stage III. However, the electrodeposition rate would decline as the residual [Cu(OH)₄]²⁻ was gradually consumed. We propose the possible reaction mechanism of the electrodeposition process, as demonstrated in the following Equations 1-7:

The real-time electrodeposition process at 1.5 V was captured by the photo images (Figure S1). The color of Cu foam changed from brick-red to dark green as the electrodeposition proceeded (Figure S2), illustrating that the generated Cu²⁺ ions nearby the Cu foam were turned into Cu(OH)₂ which was coated onto the surface of Cu foam. We observed hydrogen gas bubbles resulting from HER on the cathode, while no oxygen gas bubbles emerged on the anode in the stages I and II of the electrodeposition process (Figure S2). However, after the formation of a thick Cu(OH)₂ layer on the Cu foam as the electrodeposition elapsed beyond 900 s, a significant number of oxygen gas bubbles appeared due to OER on the anode. These results support the proposed three-stage mechanism of the electrodeposition process.

The electrodeposition voltage plays a significant role in regulating the morphology and bulk composition of the as-prepared Cu GDEs. The scan electron microscopy (SEM) images show that Cu GDEs possess distinct morphologies by varying the electrodeposition voltage (Figure 1c and Figure S3). Of note, the different electrodeposition voltage would cause the different electrodeposition time between stages II and III, especially for the stage II. The hydrogen bubbles generated in stage II from the hydrogen evolution reaction would affect the morphology structure and size of the as-prepared Cu GDEs. The Cu-1.5 GDE exhibited a nano-mulberry structure composed of some small tetrahedron nanoparticles. When increasing the voltage to 1.6 V, tetrahedron nanoparticles were also presented in nano-mulberry structure, but the size of the nano-mulberry became larger due to the decreasing amount of the released hydrogen bubbles on the cathode. The size of the nano-mulberry increased continuously as the voltage increased up to 1.8 V. However, the morphology of nanoparticles in nano-mulberry changed from tetrahedrons to cubes as the voltage increased to 1.7 V, and evolved into irregular spheres when the voltage rose to 1.8 V. When the voltage exceeded 1.8 V, the morphology of the Cu nanoparticles became quasi-polyhedral structure with a decreased size. The TEM images show a tetrahedron shape of the Cu-1.5 GDE (Figure 2a), a cube shape of Cu-1.7 (Figure 2b), and a quasi-polyhedral structure of Cu-2.0 (Figure 2c), which are consistent with the SEM results. The HRTEM image, combined with the corresponding fast Fourier transform (FFT) diffraction pattern in Figure 2d, reveals the Cu₂O (111) plane with the lattice distance of 0.243 nm in Cu-1.5, and Figure 2e,f show the Cu (200) and Cu (111) planes with the lattice distance of 0.181 and 0.208 nm in Cu-1.7 and Cu-2.0, respectively.

The XRD results show that all the Cu GDEs present a dominant phase of metallic Cu⁰ with the diffraction peaks at 2θ = 43.3°, 50.4°, and 74.1° corresponding to (111), (200), and (220) plane, respectively (Figure 2g). The diffraction peaks of Cu₂O phase are discernable for Cu-1.5 and Cu-1.6, indicating the coexistence of Cu and Cu₂O phases in the Cu GDEs prepared by low electrodeposition voltages. The intensities are weakened with the increase in electrodeposition voltage, and Cu₂O diffraction peaks even disappear at high electrodeposition voltages of 1.9 and 2.0 V. To further obtain structural information of Cu GDEs, the XAS was performed for Cu-1.5 and Cu-2.0 GDEs. The Cu K-edge XANES of Cu-1.5 and Cu-2.0 agree well with that for the Cu foil (Figure 2h). The corresponding EXAFS results show that the Cu-1.5 and Cu-2.0 GDEs have the same coordination structure as the Cu foil (Figure 2i). Thus, the XAS results confirm that metallic Cu⁰ prevails in both Cu-1.5 and Cu-2.0 GDEs.

The electrodeposition voltage gradually regulates the Cu surface composition of the as-prepared Cu GDEs as well. The peaks of X-ray photoelectron spectroscopy (XPS) spectra for Cu 2p_3/2 and Cu 2p_1/2 can be deconvoluted into two subpeaks: a dominant subpeak at lower binding energy for Cu⁰/Cu⁺ and a secondary subpeak at higher binding energy for Cu²⁺ (Figure 3a). The formation of minor Cu²⁺ species is attributed to the oxidation of Cu⁰/Cu⁺ species, especially Cu⁺, in the alkaline solution. The ratio of Cu²⁺/(Cu⁺+Cu⁰) (or the percent of Cu²⁺ species) progressively declines with the increase in the electrodeposition voltage. For example, the Cu²⁺/(Cu⁺+Cu⁰) ratio decreases from 1.49 for Cu-1.5 GDE to 0.66 for Cu-2.0 GDE (Figure 3a and Table S1). The Cu LMM spectra distinguish Cu⁺ (kinetic energy of 916.4 eV) from Cu⁰ (kinetic energy of 918.2 eV) as shown in Figure 3b. The Cu LMM spectra show that the ratio of Cu⁺/Cu⁰ (or the fraction of Cu⁺) also decreases with the increase in electrodeposition voltage (Tables S1 and S2), corroborating the XRD results. Note that the ratio of Cu²⁺/Cu⁺ also decreases as the electrodeposition voltage rises, as derived from Cu 2p and Cu LMM spectra and confirmed by O 1s spectroscopy (Figure 3c and Table S1). Specifically, the content of Cu²⁺ is almost twofold of Cu⁺ for Cu-1.5 GDE and decreases to 1.33 times for Cu-2.0 GDE.

2.2 Electrocatalytic Performance for CO₂ Reduction

The electrochemical performance of the Cu GDEs for electrocatalytic CO₂ reduction reaction (ECO₂RR) was evaluated in a flow cell (Figure S4) operated under the potentiostatic mode. The selectivity toward C₁ hydrocarbon and C₂₊ products, including C₂H₄ and C₂H₅OH, is switchable on electrodeposited Cu GDEs (Figure 4, Figures S5 and S6). Interestingly, the Cu-1.5 GDE achieved the highest selectivity of CH₄ production among all GDEs, with the Faradaic efficiency (FE) of CH₄ reaching 65.4% at a cathodic potential of −0.83 V (Figure 4a). In contrast, the FE of C₂H₄ was less than 3.5% at the same potential (Figure 4c). However, the selectivity was gradually shifted from CH₄ to C₂₊ products as the electrodeposition voltage increased from 1.5 to 1.9 V. The FE for C₂₊ products achieved a maximum of 82.2% for the Cu-1.9 GDE at −0.73 V. Further increasing the electrodeposition voltage to 2.0 V, the top FE of total C₂₊ products slightly declined to 80.1% and that of C₂H₅OH decreased from 32.0% to 26.3% at −0.75 V (Figure 4e and Figure S5d). However, the FE of C₂H₄ kept increasing, which reached a maximum of 46.4% at −0.75 V (Figure 4c). Meanwhile, the CH₄ selectivity was suppressed to 1.6% at −0.75 V for the Cu-2.0 GDE (Figure 4a). Accompanied with the shift of selectivity to C₂₊ products, the total current densities of ECO₂RR for Cu GDEs monotonically increased as the electrodeposition voltage increased, surging from 350 mA cm⁻² for the Cu-1.5 GDE to 750 mA cm⁻² for the Cu-2.0 GDE at around −0.83 V (Figure S7). Corresponding to the selectivity shift, the partial current density of CH₄ ($j_{{\text{CH}}_4}$) decreased from Cu-1.5 to Cu-2.0 GDEs, whereas both partial current densities of C₂H₄ ($j_{{\mathrm{C}}_2{\mathrm{H}}_4}$) and C₂₊ products ($j_{{\mathrm{C}}_{2+}}$) increased (Figure 4b,d,f). The Cu-1.5 GDE exhibited $j_{{\text{CH}}_4}$ as high as 228 mA cm⁻² and $j_{{\mathrm{C}}_{2+}}$ as low as 11 mA cm⁻² at around −0.83 V. In comparison, the Cu-2.0 GDE achieved the maximum $j_{{\mathrm{C}}_2{\mathrm{H}}_4}$ of 279 mA cm⁻² and $j_{{\mathrm{C}}_{2+}}$ of 482 mA cm⁻² at the identical potential.

2.3 Identification of the Origin of Catalytic Selectivity and Activity

The switchable electrocatalytic activity and selectivity toward CO₂ electroreduction on these electrodeposited Cu GDEs are hypothesized to be resulted from the change of morphology and surface Cu species composition at different electrodeposition voltages. The effect of surface roughness or electrochemical active surface area (ECSA) in Cu GEDs with different morphologies was studied first. The ECSA, which reflects the number of accessible active sites in the Cu GDEs, can be estimated by measuring the double-layer capacitance (C_dl) of the electrode–electrolyte interface via the cyclic voltammetry method (Figures S8 and S9, Table S3). In general, the ECSA-normalized $j_{{\mathrm{C}}_2{\mathrm{H}}_4}$ increases while ECSA-normalized $j_{{\text{CH}}_4}$ decreases as the ECSA rises from Cu-1.5 GDE to Cu-2.0 GDE. However, the ratio of ECSA-normalized $j_{{\mathrm{C}}_2{\mathrm{H}}_4}$ of Cu-2.0 GDE to Cu-1.5 GDE is only 7.9 at −0.83 V, which is much lower than that of geometric $j_{{\mathrm{C}}_2{\mathrm{H}}_4}$ (23.9) (Figure S10). Conversely, the ratio of ECSA-normalized $j_{{\text{CH}}_4}$ (13.5) of Cu-1.5 GDE to Cu-2.0 GDE is higher than that of geometric $j_{{\text{CH}}_4}$ (5.8) at −0.83 V (Figure S10). These results demonstrate that the surface roughness itself cannot fully account for the switchable activity and selectivity toward CO₂ electroreduction on these different Cu GEDs.

Next, we focused on the effect of surface roughness induced difference in local pH on the selectivity toward CH₄ and C₂₊ products. The ECO₂RR performance over Cu-1.5, Cu-1.7, and Cu-2.0 GDEs was first investigated and compared in electrolytes with different bulk pH values, including 1 m KHCO₃, 1 m KOH, and 5 m KOH. For these three GDEs, the ${\text{FE}}_{{\text{CH}}_4}$ slightly decreased with the increase in bulk electrolyte pH while a reverse trend was observed for ${\text{FE}}_{{\mathrm{C}}_{2+}}$ (Figure S11). Take Cu-1.5 GDEs as an example, the ${\text{FE}}_{{\text{CH}}_4}$ decreased 14% by changing the electrolyte from 1 m KHCO₃ to 5 m KOH while the corresponding ${\text{FE}}_{{\mathrm{C}}_{2+}}$ increased 24%. The results indicate that the low pH favors the ECO₂RR toward CH₄ while high pH promotes the formation of C₂₊ products, consistent with the previously reported works.^{[31, 32]} However, there were no significant changes in selectivity of both CH₄ and C₂₊ products among these three GDEs at the same electrolyte of either 1 m KHCO₃ or 5 m KOH, suggesting that the bulk electrolyte pH has little influence on the reaction pathways between CH₄ and C₂₊ products on the same GDE. The local pH near the Cu GDE surface differentiates from that in the bulk electrolyte due to the CO₂–OH⁻ neutralization reaction. The peak area ratio of ${\text{CO}}_3^{2-}/{\text{HCO}}_3^{-}$ can be used to estimate the local pH based on the Henderson–Hasselbach equation: pH = pK_a + log(${\text{CO}}_3^{2-}/{\text{HCO}}_3^{-}$), where K_a is acid dissociation constant.^[15] The local pH nearby the surface of GDEs can be calculated from the calibration curve by measuring in situ Raman spectra (Figures S12 and S13).^[39] The local pH of all three GDEs slightly increased as the cathode potential was negatively swept (Figure S14), indicating the formation rate of OH⁻ in CO₂ reduction process was faster than its consumption rate by the neutralization reaction. Interestingly, these three GDEs displayed almost equivalent local pH values at each potential, indicating the similar local reaction environment during the ECO₂RR. Taking the results of bulk pH effect and local pH measurement together, we conclude that the local pH variation is not the main factor that controls the selectivity between CH₄ and C₂₊ products.

Afterward, we investigated the influence of Cu facet on the activity and selectivity toward CO₂ electroreduction. The morphologies of Cu GDEs were well-maintained after ECO₂RR, as shown in the SEM images (Figure S15). The HRTEM images reveal the distinguished lattice plane of Cu (111), Cu (200), and Cu (111) facets with distances of 0.208, 0.181, and 0.208 nm for Cu-1.5, Cu-1.7, and Cu-2.0, respectively (Figure S16). Of note, most of the oxide layer on Cu-1.5 was reduced to metallic Cu after ECO₂RR, as evidenced by XRD and XAS results (Figures S17 and S18). The Cu-1.5 GDE with dominant exposed Cu (111) facet exhibited the main CH₄ product. When Cu (200) facet was primarily exposed on Cu-1.7 GDE, the production rate and selectivity of C₂H₄ increased, accompanied by the decrease in CH₄ product. However, the Cu-2.0 GDE with prevailing Cu (111) facet exhibited the highest selectivity toward C₂H₄ and C₂₊ products, indicating facet alone cannot be accountable for the difference in activity and selectivity.

Finally, the dependence of the reactivity toward CO₂ electroreduction on the surface composition was studied. The ratio of Cu^δ+/Cu⁰ (Cu^δ+ = Cu²⁺ + Cu⁺) declined from 12.32 for Cu-1.5 to 2.30 for Cu-2.0 (Table S1). The ratios of Cu⁺/Cu⁰ and Cu²⁺/Cu⁺ followed the same tendency. The Cu-1.5 GDE surface, preferring CH₄ formation, was primarily constituted of Cu²⁺+Cu⁺ species with a total percentage of 92.5% (Table S2). By contrast, Cu-2.0 GDE, achieving the highest production rate of C₂H₄, had 70.0% Cu^δ+ species and 30.0% Cu⁰. This distinguishable result referred to the most important role of surface Cu species in selectivity toward ECO₂RR. Of note, the ratio of Cu^δ+/Cu⁰ in Cu GDEs decreased after ECO₂RR (Figure S19, Tables S4 and S5), indicating that the Cu^δ+ species can be reduced to Cu⁰ species during ECO₂RR, which was consistent with our XRD and XAS results as well as previous results from operando analysis.^[27] However, the content of Cu⁺ and Cu²⁺species in the Cu GDEs still followed the order of Cu-1.5 > Cu-1.7 > Cu-2.0 after ECO₂RR. The surface Cu species composition under ECO₂RR process cannot be accurately reflected by XPS due to the possible re-oxidation of Cu exposed in the air, which requires more advanced analysis techniques in the future work. Nevertheless, our result shows that both ${\text{FE}}_{{\mathrm{C}}_2{\mathrm{H}}_4}/{\text{FE}}_{{\text{CH}}_4}$ and ${\text{FE}}_{{\mathrm{C}}_{2+}}/{\text{FE}}_{{\text{CH}}_4}$ ratios are inversely proportional to Cu^δ+/Cu⁰ ratio for the pristine Cu samples (Figure 5a,b). The Cu-2.0 GDE with the lowest Cu^δ+/Cu⁰ ratio (Cu⁺/Cu⁰ ratio of 0.99) achieved the highest ${\text{FE}}_{{\mathrm{C}}_2{\mathrm{H}}_4}/{\text{FE}}_{{\text{CH}}_4}$ and ${\text{FE}}_{{\mathrm{C}}_{2+}}/{\text{FE}}_{{\text{CH}}_4}$ ratios among the six GDEs. The maximum ${\text{FE}}_{{\mathrm{C}}_2{\mathrm{H}}_4}/{\text{FE}}_{{\text{CH}}_4}$ and ${\text{FE}}_{{\mathrm{C}}_{2+}}/{\text{FE}}_{{\text{CH}}_4}$ reached 48.6 and 89.5, respectively, for Cu-2.0 GDE at −0.66 V. These results suggest that the high Cu^δ+/Cu⁰ ratio (Cu⁺/Cu⁰ ratio of 2.54) on the Cu surface favors the C₁ hydrocarbon formation, while the low Cu^δ+/Cu⁰ ratio (Cu⁺/Cu⁰ ratio of 0.99) facilitates the C–C coupling for C₂₊ products. This trend is consistent with the prior findings that demonstrated the oxide-derived Cu surface composed of Cu⁰ and Cu⁺ with equal amount enhances the selectivity toward C₂₊ products.^[40] Combining all the above results, we may safely reach a conclusion that the selectivity toward CO₂ reduction is principally governed by the Cu^δ+/Cu⁰ ratio over the surface roughness, local pH, and facet in different Cu GDEs.

2.4 Mechanism of CO₂ Reduction Reaction

Several previous studies proposed that ${}^{\ast }\text{CO}$ is the key intermediate for CO₂ conversion.^{[21, 41-43]} CH₄ pathway is proceeded by hydrogenation of ${}^{\ast }\text{CO}$ to ${}^{\ast }\text{CHO}$ intermediate, while the pathway toward C₂₊ products mainly lies in CO dimerization (C–C coupling) step, especially at the low overpotential region.^{[21, 44]} We posited that such different selectivity of CH₄ and C₂₊ products on our Cu GDEs is attributed to the difference in ${}^{\ast }\text{CO}$ surface coverage. We analyzed the CO generation rate and dimerization rate over six Cu GDEs aforementioned to unravel the underlying reaction mechanism toward CO₂ conversion. The CO generation rate is the sum of the production rates of C₂₊, CH₄, and CO products normalized by the electrons transferred per mole of product assuming CO is the reactant. The CO dimerization rate is referred to normalized production rates of C₂₊ products (See detailed calculation in Supporting Information). The CO generation rate increases in the order from Cu-1.5 to Cu-2.0 GDEs (Figure 5c), indicating that the coverage of ${}^{\ast }\text{CO}$ on the electrode is greatly enhanced from Cu-1.5 to Cu-2.0 GDEs, especially at more negative potentials (<−0.75 V). Generally, the high ${}^{\ast }\text{CO}$ surface coverage promotes the C–C coupling reaction kinetics on the Cu surface, while the low ${}^{\ast }\text{CO}$ surface coverage favors the hydrogenation of ${}^{\ast }\text{CO}$ to ${}^{\ast }\text{CHO}$, the critical step to CH₄ formation.^{[16, 45-47]} Accordingly, an analogous increasing trend of CO dimerization rate can be observed from Cu-1.5 to Cu-2.0 GDEs (Figure 5d). These results validate that the Cu-2.0 GDE generates sufficient CO to increase the ${}^{\ast }\text{CO}$ surface coverage, leading to the acceleration of CO-to-C₂₊ products conversion. To further study the reaction kinetics in terms of CO intermediate reactant, the Tafel plots of CO generation and dimerization were analyzed (Figure 5e,f). The Cu-2.0 GDE displays the smallest Tafel slopes of both CO generation and CO dimerization reactions, illustrating the fastest production rate of C₂₊ products.

In situ Raman spectroscopy was performed in a flow cell to further investigate the adsorption property of reaction intermediates during ECO₂RR (Figure 6a–c). Two weak Raman signals of ${}^{\ast }{\text{CO}}_2^{·-}$ at 735 and 1556 cm⁻¹ could be observed in Cu-1.5 GDE at potentials larger than −0.3 V and then the signals disappeared at more negative potentials. The Raman signals of ${}^{\ast }{\text{CO}}_2^{·-}$ peaks increased for Cu-1.7 GDE and were further enhanced for Cu-2.0 GDE. The higher intensity of ${}^{\ast }{\text{CO}}_2^{·-}$ peaks reflected a strong adsorption of ${}^{\ast }{\text{CO}}_2^{·-}$ intermediate on the surface of Cu-2.0 GDE, indicating more stabilized ${}^{\ast }{\text{CO}}_2^{·-}$ intermediate on Cu-2.0 GDE. The ${}^{\ast }{\text{CO}}_2^{·-}$ radical anion is regarded as the first reaction intermediate for ECO₂RR, which can be further reduced to form the second key intermediate of ${}^{\ast }\text{CO}$ for hydrocarbons.^[48] Therefore, the strong adsorption/stabilization of ${}^{\ast }{\text{CO}}_2^{·-}$ intermediate promotes the conversion of CO₂ to ${}^{\ast }\text{CO}$ intermediate.^{[48, 49]} A similar peak intensity tendency of ${}^{\ast }\text{CO}$ intermediate adsorption behavior could be revealed among Cu-1.5, Cu-1.7, and Cu-2.0 GDEs. The Cu-2.0 GDE showed the strongest Raman peaks corresponding to ${}^{\ast }\text{CO}$ at 301, 398, and 2060 cm⁻¹, which are assigned to the Cu–CO frustrated rotation, Cu–CO stretch, and C≡O stretch, respectively.^{[50, 51]} This result demonstrates that the Cu-2.0 GDE provides abundant active sites for efficient conversion of ${}^{\ast }{\text{CO}}_2^{·-}$ to ${}^{\ast }\text{CO}$ intermediate, leading to high ${}^{\ast }\text{CO}$ surface coverage on Cu-2.0 GDE, which agrees well with the calculation results of CO generation rate above.

Previous studies implied that the surface components of Cu^δ+ and Cu⁰ in Cu-based catalysts governed product selectivity by tailoring the adsorption of ${}^{\ast }\text{CO}$ intermediates.^{[52, 53]} Our in situ Raman results also suggest that Cu^δ+/Cu⁰ ratio might play an essential role in tuning ${}^{\ast }\text{CO}$ adsorption. As shown in Figure 6a–c, the peak at 536 cm⁻¹ appears, which is assigned to the CuO_x/(OH)_y species.^[54] Cu-1.5 and Cu-1.7 GDEs displayed high peak intensity of CuO_x/(OH)_y, while Cu-2.0 GDE showed a relatively low peak intensity, suggesting the higher content of Cu^δ+ species can be remained in Cu-1.5 and Cu-1.7 GDEs. Moreover, to probe the stability of the surface Cu^δ+ species in the Cu GDEs during CO₂ reduction, we further investigate the in situ Raman spectra of Cu-1.5 and Cu-2.0 GDEs under continuous CO₂ reduction at −0.45 V for 140 min, as shown in Figure S20. Both Cu-1.5 and Cu-2.0 GDEs exhibit a peak of CuO_x/(OH)_y species in the long-term test, demonstrating the stability of surface Cu^δ+ species during CO₂ reduction. Although the in situ Raman test cannot precisely quantify the amount of Cu^δ+ species in the Cu GDEs, it can be well reflected the dynamic change of Cu species during CO₂ reduction. The possible reaction mechanism of the formation of CH₄ and C₂H₄ on Cu surfaces was proposed and depicted in Figure 6d. In the case of Cu-1.5 GDE with dominant Cu⁺, the low ${}^{\ast }\text{CO}$ surface coverage results in a high energy barrier for C–C coupling that suppresses C₂₊ generation while leading to hydrogenation step for CH₄ production. When the Cu⁰ species fraction increases in the Cu GDE as the electrodeposition voltage increases, more ${}^{\ast }\text{CO}$ intermediate emerges on the Cu surface, especially at more negative overpotentials. The nearly equal content of Cu⁺ and Cu⁰ in Cu-2.0 GDE will generate abundant amount of ${}^{\ast }\text{CO}$ intermediate, providing high ${}^{\ast }\text{CO}$ surface coverage, which assists in dimerizing CO to form C₂₊ products.^{[17, 35]}