Synergistic electrode design for efficient CO₂ electrolysis to multicarbon products at elevated temperatures

Abstract

Electrocatalytic CO₂ reduction reaction (CO₂RR) technology holds significant industrial potential. However, when faced with elevated-temperature environments caused by industrial-scale operations, the fundamental understanding of temperature-dependent CO₂RR behavior in flow cells remains elusive. This study points out that elevated-temperature operation (>333 K) presents both challenges and opportunities for multi-carbon (C₂₊) production. While elevated temperature enhances reaction kinetics and reduces thermodynamic energy barriers, it simultaneously induces reconstruction of Cu-based catalyst, accelerates gas diffusion electrode flooding, and promotes *CO desorption together with hydrogen evolution reaction, collectively suppressing C₂₊ product formation and compromising long-term reactor stability. Through rational design of hydrophobic-enhanced Pd-Cu₂O/polytetrafluoroethylene (PTFE)/Ag tandem electrodes, we overcome these challenges. Leveraging thermal reduced C-C coupling barriers, the optimized electrode achieves >70% Faradaic efficiency of C₂₊ across industrially relevant current densities (200–1000 mA cm⁻²) at 348 K. This strategy converts elevated temperature adversity into a kinetic and thermodynamic advantage, boosting C₂₊ cathodic energy efficiency by 1.3 times compared to ambient operation, establishing a promising paradigm for industrially viable CO₂ electrolysis.

Introduction

Electrocatalytic CO₂ reduction to C₂₊ products offers a promising route for sustainable chemical synthesis and carbon emission mitigation^1,2,3. Advances in electrocatalyst design, microenvironment engineering, and reactor optimization have enhanced C₂₊ production efficiency in ambient-temperature systems^4,5,6. Nevertheless, when electrolyzers are scaled up and operated under industrial conditions, heat from Joule heating due to ohmic resistance, and limited heat dissipation collectively elevate electrode surface temperatures (Fig. 1a)^7,8. Strikingly, scaling electrode areas from 4 cm² to 25 cm² rapidly elevates cell temperature to 353 K or induces violent boiling (Fig. 1b, Supplementary Fig. 1 and Supplementary Note 1), exceeding thermal limits of ion-exchange membranes and causing system failure. This demonstrates that elevated-temperature operation is unavoidable in industrial scaling. Understanding temperature impacts on CO₂RR systems, including reactor integrity and reaction pathways, is essential for industrial implementation^9,10.

Fig. 1: The effect of elevated temperature on the CO₂RR.

a Schematic illustration of the source pathways of Ohm heat in a flow cell for industrial CO₂RR applications. b Electrolyte temperature curve of large-area electrodes operating at high currents for 30 min in a flow cell (Room temperature: 298 K). c Schematic illustration of the effect of elevated temperature on CO₂RR products. The black represents Cu catalyst, the brown represents Pd atom, and the combination of red, gray, and blue represents an OCCOH intermediate formed by C-C coupling. d A comparison of temperature and the FE of C₂₊ for CO₂RR in previous reports (The highest FE and the FE at the highest temperature from these studies). e Breakdown of the plant-gate levelized cost per tonne of C₂₊ products with and without cooling, as calculated from a TEA. Source data are provided as a Source Data file.

Current understanding of temperature effects originates from H-cell studies, which provide preliminary insights into product selectivity and intermediate adsorption/desorption^11,12,13. Koper et al. proposed that a slight increase in temperature promotes the formation of C₂₊ products, while excessive elevation accelerates the formation of the competing HER¹⁴. However, elevated temperature alters CO₂ solubility, water vapor pressure, and electrolyte conductivity, thereby fundamentally reconstructing the reaction microenvironments^15,16. Consequently, investigating temperature effects in H-cells is substantially hindered by the coupled influences of kinetic and mass transport changes, making it challenging to infer purely thermodynamic or intrinsic kinetic effects. Gas diffusion electrodes (GDEs) with stable gas–liquid–solid interfaces effectively mitigate mass transport limitations. Compared to membrane electrode assembly (MEA) electrolyzers, the flow-cell configuration offers a key advantage for fundamental research by enabling independent and precise thermal management and potential measurement of the cathode. Moreover, it avoids complications such as membrane degradation and water evaporation at elevated temperature, which are commonly encountered in MEA electrolyzers, thereby enabling a rigorous assessment of thermal impacts on reaction thermodynamics and kinetics^17,18,19. Previous studies in flow cells demonstrate temperature-enhanced kinetics for C₁ products, exemplified by reduced overpotentials for CO and H₂ evolution on silver catalysts from 291 to 343 K²⁰, while mechanistic work reveals that temperature-dependent *COOH/*H adsorption competition dictates CO/H₂ selectivity²¹.

Despite these advances for C₁ products, premature *CO desorption and accelerated hydrogen evolution reaction (HER) kinetics severely impede C-C coupling at elevated temperatures, limiting C₂₊ formation on Cu catalysts (Fig. 1c). Takanabe et al. demonstrate that temperatures exceeding 333 K significantly reduce the selectivity of C₂₊ products²². Han et al. Further confirms that lower operating temperature (270 K, −3 °C) favors C-C coupling and hydrogenation pathways of *CH-COH intermediates on Cu-based catalysts²³, suggesting incompatibility between C₂₊ production conditions and industrial elevated-temperature environments. Additionally, elevated temperature may compromise the stability of both catalysts and GDEs in flow cells. Therefore, prior strategies have focused on circumventing, rather than addressing, the elevated-temperature challenge, for example, by lowering operating temperature via heat exchange and electrolyzers' cooling. Thus far, maintaining a high Faradaic efficiency (FE) for C₂₊ products in elevated-temperature flow cells remains a critical scientific challenge and a technological bottleneck urgently demanding resolution (Fig. 1d)^{13,14,16,23,24,25,26,27,28,29,30,31}. Consequently, a comprehensive understanding of the adverse effects of elevated temperature on CO₂RR in flow cells, together with the development of corresponding mitigation strategies, is essential for industrial implementation.

To bridge these gaps, we systematically investigate the effect of temperature (298–348 K) on CO₂RR in flow cells. Experimental results reveal that elevated temperature induced dynamic restructuring of Cu-based catalysts, accelerated GDEs flooding due to increased water vapor pressure, and enhanced *CO desorption and HER kinetics. Therefore, elevated temperature inhibits the formation of C₂₊ products and indirectly promotes product loss in flow cells. To address these challenges, we engineered hydrophobic-enhanced Pd-Cu₂O/PTFE/Ag tandem electrodes that achieve >70% FE of C₂₊ across 200–1000 mA cm⁻² at 348 K and exhibit stable operation for extended periods (Fig. 1d). Operando Raman spectroscopy and density functional theory (DFT) calculations confirm that strengthened *CO adsorption and reduced C-C coupling energy barriers synergistically promote the formation of C₂₊ at elevated temperature. These strategies convert elevated-temperature challenges into opportunities, increasing cathodic energy efficiency (CEE) by 1.3 times compared to room-temperature operation. Preliminary techno-economic analysis (TEA) indicates that this strategy not only significantly reduces electricity consumption but also eliminates approximately 14.9% of cooling-related operating costs (Fig. 1e, Supplementary Fig. 2, Supplementary Note 2 and Supplementary Table 1 to 2). This work provides essential guidance for the scale-up and practical application of CO₂ electrolysis technology.

Results

We systematically investigated the effect of temperature on CO₂RR in a flow cell by simultaneously examining catalyst structure, three-phase interface stability, and key reaction intermediates and pathways, all of which collectively govern CO₂RR activity. Given that ion exchange membranes are typically operated below 353 K, we investigated the impact of temperature on CO₂RR within a safe range of 298-348 K^32,33. To maintain a stable CO₂RR interfacial temperature, both the flow cell and catholyte were heated to simulate the elevated temperature environment encountered in flow cell (Supplementary Fig. 3 and Supplementary Note 3). The flow cell temperature was monitored using two K-type thermocouples and infrared thermal imaging technology. Measurements commenced only after the temperature had reached the set-point and remained stable for 5 min.

The impact of temperature toward the catalyst structure

Cu is widely employed for C₂₊ production owing to its efficient C-C coupling capability, yet concerns persist regarding its structural vulnerability under industrial elevated-temperature industrial operation^34,35. X-ray diffraction (XRD) analysis reveals distinct thermal oxidation pathways in magnetron-sputtered Cu electrodes (Fig. 2a). The as-synthesized electrodes exhibit characteristic Cu(111) and (200) diffraction peaks^36,37. After immersion in 348 K 1.0 M KOH electrolyte, a new peak emerges at 36.5° corresponding to Cu₂O(111), confirming surface oxidation^38,39. Subsequent CO₂RR operation restores the original Cu phase, demonstrating electrochemical reduction reversibility. X-ray photoelectron spectroscopy (XPS) verified this phenomenon (Supplementary Fig. 4a), i.e., Cu was oxidized in 348 K electrolyte, as evidenced by a distinct Cu(II) signal at 935 eV⁴⁰. Cu LMM Auger electron spectroscopy (AES) confirms increased relative ratios of Cu(I) and Cu(II) signals at elevated temperature (Supplementary Fig. 4b)⁴¹. To eliminate environmental interference and probe dynamic redox behavior of the Cu electrode, electrochemical in situ X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) measurements were conducted (Supplementary Fig. 5). Upon electrochemical reduction, the Cu K-edge XANES spectra (Fig. 2b and Supplementary Fig. 6a) confirm dominant Cu(0) states, matching the Cu foil references. However, once electrochemical control is removed, rapid surface oxidation occurs, as evidenced by a distinct Cu-O coordination peak at 1.6 Å in the EXAFS spectra (Fig. 2c, Supplementary Fig. 6b, and Supplementary Table 3)⁴². In addition to changes in chemical state, scanning electron microscope (SEM) images reveal that operating at the elevated temperature (348 K) caused slight morphological changes in Cu nanoparticles (Fig. 2d).

Fig. 2: The impact of temperature toward the structure of Cu and Cu₂O.

a XRD patterns of Cu electrode (2.5 cm×1.0 cm) under different conditions (I: Initial sample; II: After immersion in 298 K 1.0 M KOH electrolyte, pH = 13.7 ± 0.1; III: After immersion in 348 K 1.0 M KOH electrolyte; IV: After CO₂RR in 348 K 1.0 M KOH electrolyte, pH = 13.7 ± 0.1. b Normalized Cu K-edge in situ XANES spectra and c Fourier transform k³-weighted EXAFS of Cu under different conditions in a 0.5 M KHCO₃ electrolyte (pH = 6.7 ± 0.1, 6.8 ± 0.3 Ω). d SEM images of Cu under different conditions. e SEM images of Cu₂O under different conditions. f Normalized Cu K-edge in situ XANES spectra and g Fourier transform k³-weignted EXAFS of Cu under different conditions in a 0.5 M KHCO₃ electrolyte (pH = 6.7 ± 0.1, 6.8 ± 0.3 Ω). All reported potentials were not IR corrected. Source data are provided as a Source Data file.

Therefore, elevated-temperature (348 K) electrolyte immediately affects the structure of Cu catalysts, but continuous electrochemical stimulation masks the effects of elevated-temperature oxidation, so this dynamic oxidation-reduction process is easily overlooked during the CO₂RR. Critically, elevated-temperature-induced oxidation alters CO₂RR pathways. As the temperature increased from 298 K to 348 K, the FE of CH₄ increased from 3.4% to 20.8%, while the FE of C₂H₄ decreased from 30.2% to 17.6% at −200 mA cm⁻², concomitant with accelerated HER (Supplementary Fig. 7). During the 6-hour stability test (Supplementary Fig. 8), the FE of C₂H₄ and CH₄ remained nearly constant at 298 K. Conversely, at 348 K, Cu oxidation and surface restructuring shifted selectivity toward CH₄.

In contrast, the elevated-temperature-induced dynamic oxidation-reduction process shows a different pattern on Cu₂O. Cu₂O undergoes no further oxidation in 348 K electrolyte, and its morphology remains unchanged (Fig. 2e). XRD analysis confirms the absence of elevated-temperature-induced crystal structure alteration (Supplementary Fig. 9); the weak diffraction peaks at 44° and 51° are ascribed to electrochemically reduced Cu(0) (111) and (200) facets⁴³. XPS shows a slight decrease in the Cu(I)/Cu(0) ratio after electrochemical reduction at 348 K (Supplementary Fig. 10). Notably, XANES spectra reveal that a metallic Cu phase forms during the prolonged electrochemical reduction reaction (Fig. 2f and Supplementary Fig. 11). When electrochemical control is removed, the metallic Cu is re-oxidized to Cu₂O (Fig. 2g). This result confirms that the Cu₂O crystal structure is unaffected by the elevated temperature, yet the metallic Cu generated during CO₂RR is still oxidized, just like the Cu electrode mentioned above. In contrast, as the temperature increased from 298 K to 348 K, the Cu₂O electrode exhibited no appreciable changes in the distribution of carbon products (Supplementary Fig. 12), apart from a drastic increase in the FE of H₂. Operando surface-enhanced Raman spectroscopy (SERS) revealed the temperature dependence of product selectivity on Cu and Cu₂O electrodes (Supplementary Figs. 13–17 and Supplementary Notes 4–7), that elevated‑temperature‑induced structural changes in Cu significantly promote the activation of interfacial water. Under the combined effects of low *CO coverage and excessive protonation, the product selectivity shifts from C₂₊ to CH₄. In contrast to Cu, Cu₂O maintains its structural integrity at 348 K and, crucially, preserves its favorable interfacial water structure. Therefore, in subsequent research, selecting Cu₂O as the subject of study can effectively avoid the adverse effects arising from structural changes in the catalyst induced by elevated temperatures. While Cu₂O nanocubes demonstrate superior structural stability relative to metallic Cu at elevated temperature, their CO₂RR performance at 348 K remains severely compromised. This performance degradation points to additional elevated-temperature barriers beyond structural preservation.

The impact of temperature toward the GDE

Beyond catalyst instability, elevated temperatures compromise electrode integrity by disrupting the gas-liquid equilibrium. Product distribution analysis on the Cu₂O electrodes shows a sharp rise in the FE of H₂ and a simultaneous drop in the FE of CO and C₂₊ when the temperature increases from 298 K to 348 K (Supplementary Fig. 12). This anomalous enhancement in HER does not stem from catalyst degradation but from catastrophic flooding of the GDE, a direct consequence of elevated water vapor pressure that destabilizes the three-phase boundary (Fig. 3a). In the flow cell configuration, the three-phase interface separates the CO₂ feed from the electrolyte via the catalyst layer. The short gas diffusion path greatly enhances mass transfer efficiency from the CO₂ feed to the catalyst-electrolyte interface, enabling high current densities. However, this design also renders the cathode susceptible to flooding⁴⁴. When the hydrophobic GDE structure is compromised during electroreduction, electrolyte infiltrates the catalyst layer, the GDE, and the gas channel, blocking CO₂ access to active sites. At elevated temperature, the electrolyte’s saturated vapor pressure rises; water vapor formation accelerates flooding, further diminishing the population of active sites for CO₂RR. The remaining sites in contact with the electrolyte continue to drive the hydrogen evolution reaction.

Fig. 3: The impact of temperature toward the GDE.

a Schematic representation of water flooding of cathode structures in a flow cell at elevated temperature. b FEs of H₂ and c FEs of C₂₊ at different temperatures and current densities on Cu₂O. d Double-layer charging current plotted against the CV scan rate at different temperatures on Cu₂O and Cu₂O/PTFE. e EIS of Cu₂O and Cu₂O/PTFE at different temperatures. f FEs of H₂ and C₂₊ on Cu₂O/PTFE at different current densities at 348 K. All error bars represent the standard deviation calculated from at least 3 data points. Source data are provided as a Source Data file.

The Cu layer formed by magnetron sputtering is highly dense and therefore hydrophobic, effectively preventing water vapor intrusion at elevated-temperature conditions (Supplementary Fig. 18a and Supplementary Note 8). In contrast, most Cu-based composite catalysts must be sprayed onto the GDE in the powder form; their high porosity compromises hydrophobicity and renders them more susceptible to flooding at elevated temperature (Supplementary Fig. 18b). The electrochemical in situ reconstruction of Cu₂O is largely independent of H₂ evolution; at 298 K the FE of H₂ actually decreases when the current density is raised to −600 mA cm⁻² (Fig. 3b). However, as the temperature rises further, the FE of H₂ increases sharply, indicating that elevated temperatures disrupts CO₂ transport and suppresses C₂₊ products formation (Fig. 3c and Supplementary Fig. 19). Damage to the three-phase interface also leads to product loss; the total FE falls to only 88.5%, of which H₂ accounts for 74.9%. Thus, the electrode structure is more vulnerable at elevated temperatures and high current densities. Cyclic voltammetry (CV) in non-Faradaic regions shows that the C_dl value of Cu₂O increases with temperature, giving a larger electrochemical active surface area (ECSA) at 348 K (Fig. 3d and Supplementary Table 4)⁴⁵. In the flow cell, however, the ECSA encompasses both CO₂RR and HER active sites; higher temperature therefore intensifies flooding and enlarges the HER-active area (Supplementary Fig. 20). After a 6-h stability test, the gas-solid-liquid interface is almost completely destroyed and H₂ becomes the predominated reduction product (Supplementary Fig. 21).

Hydrophobic polytetrafluoroethylene (PTFE) particles were introduced into the catalyst layer to suppress water vapor intrusion (Supplementary Fig. 22)⁴⁶. The C_dl value of Cu₂O/PTFE remains nearly constant with increasing temperature, indicating that PTFE effectively stabilizes the three-phase interface at elevated temperature (Supplementary Fig. 23). Electrochemical impedance spectroscopy (EIS) shows that while elevated temperature lowers the interfacial charge transfer resistance, the insulating PTFE induces a slight increase in this value (Fig. 3e). The thickness (δ) of the CO₂ diffusion layer on the catalyst was fitted at different temperatures (Supplementary Table 5). For Cu₂O, δ decreased from 4.51 to 3.54 as temperature rose, consistent with water vapor instructions shortening reducing the CO₂ diffusion path (Supplementary Fig. 24). In contrast, the almost unchanged δ for Cu₂O/PTFE confirms that a stable three-phase interface that is insensitive to temperature. Consequently, PTFE effectively inhibits water flooding and limits the FE of H₂ to only 13.0% at 348 K, even at −600 mA cm⁻² (Fig. 3f). Thus, the shift of CO₂RR products toward H₂ on Cu₂O at elevated temperature is mainly attributable to thermally accelerated water flooding that impedes CO₂ diffusion. The PTFE not only mitigates flooding but also ensures robust GDE operation at elevated temperatures. Nevertheless, the FE of H₂ on Cu₂O/PTFE remains higher at 348 K than at 298 K (Supplementary Fig. 25), indicating that intrinsic kinetic limitations, beyond flooding, also contribute to suppressed C₂₊ formation.

The impact of temperature toward the C₂₊ product

The efficiency of C-C coupling is a key determinant of C₂₊ yield. Even with PTFE-stabilized GDE operation, elevated temperatures suppress C-C coupling, as evidenced by reduced C₂₊ FE and concomitant promotion of H₂ and CO generation (Supplementary Fig. 25 and Fig. 4a). Investigating the temperature effect on C-C coupling therefore requires suppressing competitive hydrogen evolution and preventing *CO desorption. Although the Cu₂O/PTFE electrode remains structurally stable at elevated temperatures, temperature rise still enhances its HER. The same trend was observed on magnetron-sputtered Cu electrodes. To further suppress H₂ reduction, we coated an Ag layer onto the Cu₂O/PTFE electrode, creating a spatially separated tandem architecture (Supplementary Fig. 26 and Supplementary Note 9)⁴⁷. The Ag catalyst layer suppresses water dissociation and effectively reduces CO₂ to CO, increasing the local *CO concentration while inhibiting *H adsorption on Cu₂O surface (Fig. 4a and Supplementary Fig. 27)^31,48 The CO₂RR activity tests show that the Cu₂O/PTFE/Ag tandem structure lowers the FE of H₂ from 22.5% to 7.2% at 348 K and −200 mA cm⁻² (Fig. 4b). In previous studies, Ag serves as a *CO generator, and higher *CO coverage can lower the C-C coupling barrier. However, Cu₂O/PTFE/Ag exhibits a higher FE of CO than Cu₂O/PTFE at both 298 K and 348 K (Supplementary Fig. 28 and Fig. 4b), indicating that the abundant *CO produced by Ag is not efficiently spilled over to Cu₂O for coupling but is instead desorbed and vented. This suggests that the *CO-adsorption capability of Cu₂O is equally critical; its capacity as a *CO receiver dictates the amount of *CO available for C-C coupling⁴⁹. Because local *CO concentration and pH scale with current density, the *CO coverage at −200 mA cm⁻² is lower than that at higher current densities, and the higher C-C coupling barriers yield lower FE of C₂₊. Moreover, Cu₂O/PTFE/Ag shows a higher FE of CO at 348 K than at 298 K, confirming that elevated temperature not only fails to promote C-C coupling but also accelerates *CO desorption.

Fig. 4: The impact of temperature toward the C₂₊ product.

a Schematic diagram of elevated temperature promotion of CO₂RR on the surface of Cu₂O/PTFE, Cu₂O/PTFE/Ag, and Pd-Cu₂O/PTFE. b Corresponding FEs at different current densities on Cu₂O/PTFE/Ag at 298 K and 348 K in 1.0 M KOH electrolyte (pH = 13.7 ± 0.1, 0.62 ± 0.03 Ω). c Product distributions and corresponding FEs at different current densities on Pd-Cu₂O/PTFE at 348 K in 1.0 M KOH electrolyte. d Product distributions and corresponding FEs at different current densities on Pd-Cu₂O/PTFE/Ag at 348 K in 1.0 M KOH electrolyte. e FEs and C₂₊/C₁ ratio at −200 mA cm^-2 on CP (Cu₂O/PTFE), CPA (Cu₂O/PTFE/Ag), PCP (Pd-Cu₂O/PTFE), and PCPA (Pd-Cu₂O/PTFE/Ag) at 298 K and 348 K. f Corresponding potentials and CEEs at different current densities on Pd-Cu₂O/PTFE/Ag at 298 K and 348 K in 1.0 M KOH electrolyte. g Comparison of some typical results at −200 mA cm^-2 at 348 K. All error bars represent the standard deviation calculated from at least 3 data points. All reported potentials were not IR corrected. Source data are provided as a Source Data file.

Recently, Xu et al. showed that alkali metal cations modulate both the activation enthalpy of reaction and the enthalpy of CO adsorption⁵⁰. The CO₂RR tests in electrolytes containing different alkali metal cations reveal that larger cations suppress water dissociation more effectively, with the FE of H₂ following the order Li⁺> Na⁺> K⁺> Cs⁺ (Supplementary Fig. 29). Conversely, larger cation weakens *CO adsorption, so the FE of CO increases in the order Li⁺ <Na⁺ <K⁺ <Cs⁺. As the temperature rises, the FEs of H₂ and CO increase in all alkali metal cation-containing electrolytes, confirming that elevated temperatures accelerate *CO desorption. Cs⁺, which exhibits the weakest *CO binding, gives the highest FE of CO. These results demonstrate that the interfacial *CO adsorption capacity governs C-C coupling efficiency at elevated temperature (Supplementary Note 10).

Therefore, converting the inherently adverse elevated temperature environment into a beneficial one hinges on suppressing *CO desorption. Because Pd possesses a markedly stronger *CO binding energy than Cu, highly dispersed Pd was deposited onto Cu₂O via chemical reduction to stabilize *CO intermediates at elevated temperatures (Fig. 4a)^51,52. Elemental maps confirm the uniform distribution of Pd across the Cu₂O surface, and aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM) further verifies its atomic-level dispersion (Supplementary Fig. 30). At 298 K, the CO₂RR tests show that Pd incorporation lowered the FE of CO and promoted the C-C coupling, yielding 53.1% FE of C₂₊ at −200 mA cm⁻² (Supplementary Fig. 31). However, Pd also strongly binds *H, enhancing water dissociation and increasing the FE of H₂⁵³. When the temperature is raised to 348 K, H₂ becomes the dominant side product (Fig. 4c). To balance these competing effects, Pd and Ag were co-incorporated into the Cu₂O/PTFE electrode. In the resulting Pd-Cu₂O/PTFE/Ag architecture, Pd selectively suppresses *CO desorption, while Ag inhibits HER and continuously supplies *CO. The presence of Pd and Ag is confirmed by XRD and XPS spectra (Supplementary Fig. 32), and SEM imaging reveals in the Pd-Cu₂O/PTFE/Ag electrode (Supplementary Fig. 33). Leveraging this synergy, the Pd-Cu₂O/PTFE/Ag electrode delivers a C₂₊ FE to 60.2% at −200 mA cm⁻² and 298 K (Supplementary Fig. 34). Pd strengthens *CO adsorption on Cu₂O sites fed by Ag, and the additional thermal energy at elevated temperature further increases the probability of C-C coupling. Consequently, the Pd-Cu₂O/PTFE/Ag electrode achieves a C₂₊ FE of 72.6% at −200 mA cm⁻² upon raising the electrolyte temperature to 348 K (Fig. 4d and Supplementary Fig. 35). Optimization of Pd loading reveals that 0.7 wt% Pd provided optimal *CO and *H adsorption, maximizing FE of C₂₊ products while minimizing FE of H₂ (Supplementary Fig. S36). The low loading and high dispersion of Pd prevent its aggregation during prolonged elevated-temperature electroreduction, thereby maintaining the structural function (Supplementary Figs. 37 and 38). Notably, elevated temperature simultaneously suppresses CO/H₂ generation and enhances C-C coupling at low current densities (Fig. 4e), overcoming the CO concentration limitation inherent to low current density operation. Benefiting from robust interfacial water management, the Pd-Cu₂O/PTFE/Ag electrode remains stable during a 48-h test at 348 K (Supplementary Fig. 39). Furthermore, the elevated temperature accelerates CO₂RR kinetics and lowers the required overpotential (Supplementary Fig. 40), thereby improving cathodic energy efficiency (CEE). Assisted by the elevated temperature, the Pd-Cu₂O/PTFE/Ag electrode delivers a CEE_C2+ of 48.9% at −200 mA cm⁻² (Fig. 4f), 1.3-fold higher than at room temperature (37.7%) and 5.5-fold higher than the initial Cu₂O/PTFE (8.9%)⁵⁴. These results confirm that elevated temperature can promote C-C coupling under controlled conditions. Thus, by synergistically optimizing interfacial water management, strengthening *CO adsorption, and facilitating C-C coupling, the hydrophobic-enhanced Pd-Cu₂O/PTFE/Ag electrode surmounts elevated-temperature limitations and achieves optimal C₂₊ production in flow cells at 348 K (Fig. 4g and Supplementary Table 6).

Mechanistic studies

Complementing our experimental observations of *CO desorption, Operando SERS provides direct evidence for interfacial intermediate dynamics in flow cells. Because the Ag tandem layer functions solely as a CO generator while C-C coupling occurs exclusively on Cu₂O or Pd-Cu₂O surfaces, the electrodes for Operando SERS measurements were fabricated without Ag layers to avoid interference from its strong surface-enhanced Raman scattering effect (Supplementary Fig. 41 and Supplementary Note 11). Raman spectra (Fig. 5a) reveal that at 298 K, when the Cu₂O/PTFE electrode potential is lowered to -0.6 V (vs. RHE), characteristic peaks emerge at 2030-2090 cm⁻¹, assigned to *CO atop stretching vibrations (*CO_atop)⁵⁵. These signals indicate that *CO intermediate adsorbs predominantly atop Cu atoms and participates in C-C coupling. The concurrent appearance of a peak at 360 cm⁻¹ (assigned to Cu-CO stretching) confirms strong potential-dependent Cu-CO interactions (Supplementary Fig. 42a). At 348 K, the disappearance of the *CO_atop peak signals (Fig. 5b) signifies suppressed C-C coupling⁵⁶. The peak at 360 cm⁻¹ now emerges only at higher potentials (Supplementary Fig. 42b), demonstrating that elevated temperature weakens Cu-CO binding and promotes *CO desorption. Collectively, these results underscore that catalyst-specific *CO adsorption strength is critical for C₂₊ formation. Depositing Pd onto Cu₂O strengthens Cu-CO interactions; both the Cu-CO (360 cm⁻¹) and *CO_atop peaks intensify at lower potentials at 298 K (Supplementary Fig. 42c and Fig. 5c), confirming Pd-enhanced *CO binding⁵⁷. Importantly, Pd mitigates the elevated-temperature limitation; *CO_atop signals intensify at low potentials at 348 K (Fig. 5d), suggesting that elevated temperature accelerates localized *CO conversion, thereby facilitating low-potential C-C coupling. Additionally, the Pd-induced enhanced interfacial electric field strengthens the adsorption and activation of CO₂⁻, thereby indirectly facilitating the protonation process of CO₂⁻, evidenced by an enhanced CO₂⁻ peaks at 700 cm⁻¹ and 1530–1560 cm⁻¹ at low potentials (Supplementary Fig. 42d and Supplementary Note 12).

Fig. 5: Mechanism of elevated temperature-enhanced formation of C₂₊ products.

Operando SERS for CO₂RR over the Cu₂O/PTFE catalyst at different applied potentials in 1.0 M KOH electrolyte (pH = 13.7 ± 0.1, 1.2 ± 0.1 Ω) at a 298 K and b 348 K. Operando SERS for CO₂RR over the Pd-Cu₂O/PTFE catalyst at different applied potentials at c 298 K and d 348 K. e The *CO adsorption free energy and *H adsorption free energy on Cu₂O and Cu₂O/Pd at 298 K and 348 K. All reported potentials were not IR corrected. f Energy profiles of asymmetric coupling of *CO with *COH on Cu₂O and Cu₂O/Pd surfaces with different CO coverages. g Energy profiles of asymmetric CO-COH coupling on Cu₂O and Cu₂O/Pd surfaces at 298 K and 348 K. Source data are provided as a Source Data file.

To quantify the temperature-dependent adsorption energetics inferred from Raman spectra, DFT calculations were employed to model *CO binding on Cu₂O and Pd-Cu₂O. The Cu₂O (100) surface and atomically dispersed Pd on Cu₂O (100) (denoted Pd-Cu₂O (100)) served as models (Supplementary Data 1). The adsorption energy calculations reveal that both *CO and *H binding weaken monotonically with temperature (Supplementary Fig. 43 and Fig. 5e), promoting *CO desorption and impeding subsequent C-C coupling, consistent with experimental results. Although *CO binding on Pd-Cu₂O remains temperature-dependent, Pd-Cu₂O exhibits stronger CO absorption than Cu₂O across the entire temperature range. This enhancement arises from the increased density of states near the Fermi level in Pd 4 d orbitals, which upshifts the d-band center of Cu₂O. Projected crystal orbital Hamiltonian population (pCOHP) analysis further reveals stronger Pd-C bonding, confirming the higher *CO adsorption energy (Supplementary Fig. 44). Given the critical role of surface *CO coverage in C-C coupling, we further examined the asymmetric *CO-COH coupling pathway under varying coverages^58,59,60. Increasing *CO coverage on Cu₂O from 1/18 ML to 5/18 ML (Supplementary Fig. 45 and Fig. 5f) progressively lowers the coupling barrier from 1.715 eV to 1.011 eV. This aligns with electrochemical observations. Higher current densities raise *CO coverage, reduce the C-C coupling barrier, and enhance the FE of C₂₊ products. At identical *CO coverage (5/18 ML), Pd-Cu₂O exhibits a lower C-C coupling barrier than Cu₂O (1.011 eV vs. 0.947 eV), rationalizing its superior C₂₊ FE at low current densities. Crucially, although elevated temperature promotes intermediate desorption, it simultaneously reduces C-C coupling barriers (Fig. 5g)⁶¹. Raising the temperature from 298 K to 348 K decreases the *OCCOH formation barrier by 0.124 eV for Cu₂O and by 0.136 eV for Pd-Cu₂O, underscoring the thermodynamic advantage of elevated temperature C-C coupling.

The influence of elevated temperature on C₂₊ production can be summarized as follows: (i) Elevated temperature accelerates *H and *CO desorption from the Cu₂O surface, leading to substantial H₂ and CO formation; (ii) Introducing an Ag tandem layer suppresses *H participation and continuously generates CO intermediates that migrate to the Cu₂O surface. Although the locally high CO concentration facilitates the C-C coupling, weak *CO adsorption on Cu₂O limits its ability to effectively capture and convert the migrated CO at elevated temperature; (iii) Incorporating Pd significantly strengthens the *CO adsorption on Cu₂O, enabling localized accumulation of *CO and promoting the *OCCOH generation and subsequent C₂₊ formation. Additionally, Pd site mitigates *CO desorption at elevated temperature, leveraging the reduced C-C coupling barrier to further enhanced C₂₊ yields. Consequently, the Pd-Cu₂O/PTFE/Ag electrode demonstrates significantly enhanced CO₂RR activity at elevated temperature (Supplementary Note 13).

Discussion

In summary, this work fundamentally reframes the paradigm for elevated-temperature CO₂ electrolysis. Rather than regarding elevated temperatures solely as a liability for C₂₊ production, we demonstrate that its intrinsic kinetic and thermodynamic advantages can be leveraged through rational electrode design. We first delineated the three primary challenges encountered under industrially relevant elevated-temperature (>333 K) operation in flow cells: (i) dynamic reconstruction of Cu catalysts, (ii) accelerated GDE flooding caused by increased vapor pressure, and (iii) detrimental *CO desorption coupled with enhanced HER. By transforming these challenges into opportunities, we engineered a hydrophobically stabilized tandem electrode (Pd-Cu₂O/PTFE/Ag). This integrated architecture operates synergistically: Hydrophobic PTFE mitigates flooding to preserve the vital gas-liquid-solid interface; the Ag layer suppresses water dissociation while generating abundant *CO intermediates; atomically dispersed Pd significantly strengthens *CO adsorption on Cu₂O, counteracting thermal desorption and enabling efficient C-C coupling by exploiting thermally lowered activation barriers. As a result, the optimized system achieves exceptional performance under industrially demanding conditions, achieving >70% FE of C₂₊ products across a wide current density range (200–1000 mA cm⁻²) at 348 K. Most notably, this strategy of converting thermal energy into catalytic advantage boosts the CEE of C₂₊ by 1.3 times compared to ambient operation. This study provides a transformative perspective for CO₂ electrolysis scale-up. By strategically managing interfacial processes at elevated temperatures, we unlock the dual benefits of accelerated reaction kinetics and enhanced thermodynamic favorability for C₂₊ production. TEA confirms that elevated-temperature C₂₊ production not only reduces electrolysis costs but also eliminates additional cooling energy consumption. This approach enables waste-heat utilization and represents a critical advance toward economically viable carbon-negative chemical manufacturing. Furthermore, the electrode design strategy proposed in this work may also provide valuable insights for other electrochemical synthesis fields, such as amine synthesis, biomass conversion, and fuel cells.

Methods

Chemicals and materials

Copper nitrate (Cu(NO₃)₂·3H₂O, 98%), Sodium hydroxide (NaOH, 96%), L(+)-Ascorbic acid (99%), L-cysteine (98%), Sodium borohydride (NaBH₄, 96%), and isopropanol (IPA) (C₃H₈O, 99%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Potassium chloropalladite (K₂PdCl₄, 98%) was purchased from Shanghai Acmec Biochemical Technology Co., Ltd. Ag nanoparticles were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Polytetrafluoroethylene (PTFE, particle size: 200 nm) was purchased from Shanghai Macklin Biochemical Technology Co., Ltd. GDL (YLS-30T) and Nafion (5.0 wt%, DuPont 520) were purchased from Suzhou Sinero Technology Co., Ltd. All chemicals were analytical grade and used without additional purification. The ultrapure water used in the experiments was deionized water (DI) with a resistivity of 18.2 MΩ·cm.

Fabrication of the Cu₂O catalyst

Cu₂O nanocube particles were prepared by a wet chemical method. First, 120.0 mg Cu(NO₃)₂·3H₂O was added to 90.0 mL DI and stirred for 5 min. Subsequently, 200.0 mg NaOH, which had been dissolved in 5.0 ml DI, was added under stirring conditions to form a blue precursor solution. Finally, L(+)-Ascorbic acid dissolved in 5.0 mL DI was slowly dripped into the precursor solution and stirred vigorously for 60 min until the solution turned orange. The product was collected, washed with DI water and ethanol three times by centrifugation, and dried in a vacuum oven at 60 °C for 12 h. The final product was Cu₂O nanocube particles with a size of about 200 nm.

Fabrication of the Pd-Cu₂O catalyst

The above-prepared 70.0 mg Cu₂O was added to 70.0 mL ethanol and dispersed by sonication for 1  h. Subsequently, 1.2 mg of L-cysteine was added and sonication was continued for 30 min. A certain volume (2.0 mL, 4.0 mL, and 6.0 mL) of K₂PdCl₄ precursor solution (0.01 mol·L⁻¹) was added to the mixed solution under high-speed stirring conditions. The amount of Pd added was varied by changing the volume. After 60 min of stirring, 1.42 mg NaBH₄ dissolved in 1.25 mL DI was added to the mixed solution and stirred continuously for 60 min. The products were collected and washed by centrifugation with DI water and ethanol. The Pd-Cu₂O composites were obtained after drying in a vacuum oven for 12 h. All samples labeled Pd-Cu₂O represent samples of 4.0 mL of K₂PdCl₄ precursor solution added, and 0.7 wt% Pd was obtained by inductively coupled plasma optical emission spectrometer (ICP-OES) testing.

Preparation of Cu₂O electrodes

All electrodes were prepared by the spraying method. 20.0 mg Cu₂O was dispersed in 2.0 mL of isopropanol, and 50 μL of Nafion solution was added as a binder. Dispersing was carried out under ultrasonic conditions for 1 h to form a homogeneous ink. The GDE was placed on a heater at 70 °C and sprayed with a spray gun fitted with ink. The quality of the catalyst was determined by weighing the change in mass of the GDE before and after spraying. The loading of Cu₂O was controlled to be 1.0 mg·cm⁻² by spraying several times. Finally, the Cu₂O electrode was cut to 2.5 cm × 1.0 cm for testing.

Preparation of Cu₂O/PTFE electrodes

Cu₂O/PTFE electrodes were prepared in a similar manner to that described above. 20.0mg of Cu₂O and 50μL Nafion were added to 2.0 mL IPA to make solution A. 10.0 mg of polytetrafluoroethylene (PTFE) powder was added to 2.0 mL IPA to make solution B. Solution A and solution B, after ultrasonication for 60 min, were mixed, and ultrasonication was continued for another 30min to form ink. The steps for spraying were the same as described above.

Preparation of Cu₂O/PTFE/Ag electrodes

A total of 10.0 mg of Ag and 50 μL Nafion were added to 2.0 mL IPA and sonicated for 30 min to form inks. Spraying of Ag was continued on Cu₂O/PTFE electrodes to obtain Cu₂O/PTFE/Ag electrodes. The loading of Ag was controlled to be 1.0 mg cm⁻².

Preparation of Pd-Cu₂O/PTFE/Ag electrodes

The Pd-Cu₂O/PTFE/Ag electrode was prepared in the same way as the Cu₂O/PTFE/Ag electrodes, but the Cu₂O was replaced with Pd-Cu₂O.

Preparation of Cu electrodes

Cu electrodes were prepared using a magnetron sputtering system (MSP-300B). A Cu layer with a thickness of about 400 nm was sputtered on carbon paper at a rate of 1 Å/sec.

Product analysis and performance evaluation for CO₂RR

All electrocatalytic CO₂ reduction reaction (CO₂RR) measurements were tested on a CHI 760e electrochemical workstation with three electrodes. The flow cell reactor consisted of a cathodic gas runner plate, a cathodic flow chamber, and an anodic flow chamber. The volume of the cathode flow chamber is 0.7 cm³, and the reaction area is 1.0 cm². The components are separated by PTFE gaskets. The cathode was a prepared gas diffusion electrode (2.5 cm × 1.0 cm), and the anode was IrO_x/Ti (2.0 cm × 2.0 cm). Fresh 1.0 M KOH (pH = 13.7 ± 0.1) was used as the cathodic and anodic electrolyte. The electrolytes were all transported to the flow cell by a peristaltic pump at a flow rate of 13.0 mL min⁻¹ and separated by an anion exchange membrane (3.0 cm × 3.0 cm, 130 μm). The anion exchange membrane was repeatedly flushed with pure water prior to use. CO₂ (Messer gas, 99.99%) controlled by a mass flow controller (MFC) was delivered to the cathode gas flow channel plate. The gas flow rate at the outlet during the reaction was controlled to 30.0 mL·min⁻¹ and monitored using a digital flowmeter. The product was collected after operating for 30 min in constant current mode. The gas phase products of CO₂RR were analyzed using an online gas chromatograph. The chromatograph (PerkinElmer Clarus 680) was equipped with a thermal conductivity detector (TCD) and a flame ionization detector (FID). Ar (Messer gas, 99.999%) was used as carrier gas. The liquid phase products were analyzed using ¹H nuclear magnetic resonance (¹HNMR) spectroscopy on an Advance III 500-MHz Unity plus spectrometer (Bruker). A mixture of 500 μL of liquid phase product with 100 μL of DMSO was diluted to 100 ppm with D₂O for testing. A saturated Ag/AgCl electrode was used as the reference electrode and was calibrated in H₂‑saturated 0.5 M H₂SO₄ (pH = 0) prior to testing. A Pt electrode and a graphite rod were employed as the working electrode and counter electrode, respectively. Cyclic voltammetry (CV) was conducted at a scan rate of 1 mV s⁻¹, and the potential at zero current was taken as the calibrated E_Ag/AgCl. No IR compensation was employed for all electrochemical tests. All the potentials used were based on the reversible hydrogen electrode (RHE) using the following equation (Eq. 1):

$${E}_{{RHE}}={E}_{{Ag}/{AgC}l}+0.059\times {pH}+0.198\,V$$

(1)

Characterization

Scanning electron microscope (SEM) was performed on a HITACHI S-4800. X-ray photoelectron spectroscopy (XPS) measurements were conducted on a Thermo Scientific ESCALAB 250Xi instrument. Contact angle images were obtained by Bruker Theta Flex/DSA25. X-ray diffraction (XRD) patterns were obtained on a Rigaku Ultima IV diffractometer using Cu Kα radiation. High-resolution transmission electron microscope (HRTEM) measurements were carried out on a FEI Talos F200s electron microscope. Aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM) was captured using a JEOL GrandARM300F with double Cs correctors at an acceleration voltage of 300 kV. Inductively coupled plasma optical emission spectroscopy (ICP-OES) measurements were carried out by ThermoFisher iCap 7000. The electrochemical surface area (ECSA) of the electrode was quantified by measuring its double-layer capacitance (C_dl). CV was performed at various scan rates within a potential window where no Faradaic processes occur. The value of C_dl was then obtained by fitting the slope of the plot of the charging current against the CV scan rate. Electrochemical impedance spectra (EIS) measurements were carried out in the frequency range from 0.1 Hz to 100 kHz at 0 V (vs. RHE) with an amplitude of 10 mV. The resistance of the flow cell was measured using EIS under open circuit potentials (0.62 ± 0.03 Ω). The room temperature in the experiment was consistently maintained at 298 K (25 °C).

In situ X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectroscopies were carried out using the Rapid XAFS HE (Anhui Absorption Spectroscopy Analysis Instrument Co., Ltd.) by transmission mode at 20 kV and 20 mA. A Si (553) spherically bent crystal analyzer (SBCA) with a radius of curvature of 500 mm served as the monochromator, thereby ensuring a diffraction geometry approaching a 90-degree backscatter angle at the absorption edge. After monochromatization, the X-rays pass through the sample and were collected using a high-energy-resolution silicon drift detector (SDD) to obtain the X-ray intensity. In measurements, a custom-made cell was utilized (The volume of the chamber is approximately 15.0 cm³), the samples were uniformly dispersed on carbon paper and immersed in a 0.5 M KHCO₃ electrolyte (pH = 6.7 ± 0.1) saturated with CO₂ gas. A saturated Ag/AgCl electrode and a Graphite rod electrode were used as the reference and counter electrodes, respectively. The XAFS data were acquired in transmission mode. During the XAFS measurements, the position of the absorption edge (E₀) was calibrated using a standard Cu foil, and all data were collected within a single time period.

Operando surface-enhanced Raman spectroscopy (SERS) was carried out using a Horiba LabRAM HR Evolution Raman microscope in a modified flow cell (The configuration was identical to that of the flow cells used in electrochemical testing). The volume of the cathode flow chamber is 8.0 cm³, and the reaction area is 0.5 cm². The electrolytes were all transported to the flow cell by a peristaltic pump at a flow rate of 13.0 mL min⁻¹ and separated by an anion exchange membrane (3.0 cm × 1.5 cm, 130 μm). A saturated Ag/AgCl electrode and a IrO_X/Ti electrode (2.0 cm ×1.0 cm) were used as the reference and counter electrodes, respectively. 1.0 M KOH (pH = 13.7 ± 0.1) was used as the electrolyte, and CO₂ gas is continuously passed into the cathode. A 632.8 nm excitation laser was used. The signals were recorded at different applied potentials, using a 60-second integration and by averaging two scans.

Performance evaluation for CO₂RR

The thickness of the diffusion layer δ can be determined as follows:

$$\delta=\sqrt{3{R}_{d}{D}_{0}{C}_{d}}$$

(2)

where R_d and C_d are the equivalent resistance and capacitance of the diffusion layer, and the values are obtained from the equivalent circuit fitting of EIS. D₀ is the diffusion coefficient of CO₂ in the electrolyte, which is affected by temperature changes^62,63. D₀ at different temperatures is calculated by the following equation:

$${D}_{0}=\frac{435.7{T}^{\frac{3}{2}}}{p{({{V}_{A}}^{\frac{1}{3}}+{{V}_{B}}^{\frac{1}{3}})}^{2}}\times \sqrt{\frac{1}{{M}_{A}}+\frac{1}{{M}_{B}}}$$

(3)

where T is the thermodynamic temperature (298 K, 323 K, and 348 K), p is the pressure (Pa), M_A and M_B are the molecular weight of H₂O and CO₂, and V_A and V_B are the molar volumes of H₂O and CO₂. The D₀ values at 298 K, 323 K, and 348 K are 0.1574, 0.1776, and 0.1986 cm²·s⁻¹, respectively⁶².

FE for each product was calculated based on the following equation:

$$F{E}_{i}=\frac{{z}_{i}\times {x}_{i}\times F}{Q}\times 100\%$$

(4)

where z_i is the number of electrons transferred for product, x_i is the number of moles of the product, F is Faraday’s constant and Q is the total charge passed during electrolysis.

CEE for C₂₊ product was calculated based on the following equation:

$${CE}{E}_{C2+}=\frac{\sum {{FE}}_{i}\times ({1.23-E}_{i})}{1.23-{E}_{{cathode}}}$$

(5)

where FE_i is the FE for each C₂₊ product. E_i is the thermodynamic reduction potential of each C₂₊ product, E_ethylene = 0.06 V vs RHE, E_ethanol = 0.08 V vs. RHE, E_n-propanol = 0.10 V vs. RHE, and E_acetate = 0.12 V vs. RHE for the CO₂RR^64,65. E_cathode is the measured potential values in the experiment.

DFT calculations

In this work, all density functional theory (DFT) calculations were performed using the plane-wave code QUANTUM ESPRESSO^66,67,68. The exchange-correlation potential was approximated with the Perdew−Burke−Ernzerhof functional^69,70. The interaction between the cores and the valence electrons was represented with the projector augmented wave approach⁷¹, kinetic energy cutoff for wavefunctions and charge density were 50 and 500 Ry, respectively. The convergence criteria for electronic structure iterations and force on each atom during structure relaxation were set to 10⁻⁸Ry and 5×10⁻⁴Ry/Bohr, respectively. The 3×3 Cu₂O (100) were modeled by two-layer slab with a 15 Å vacuum region to avoid interactions between periodic images. We sample the Brillouin zone with a 3×3×1 Γ-centered Monkhorst-Pack k-points for geometry optimizations. Gaussian smearing with a width of 10⁻³Ry was used. For all the calculations, the van der Waals (vdW) interactions are described with the DFT-D3 method in Grimme’s scheme⁷². The transition states were determined using the climbing image nudged elastic band (CI-NEB) approach. The TS structure was optimized using quasi-Newton algorithm, and the TS were converged to a residual force smaller than 0.05 eV/Å. Additional, on-site Coulomb repulsion between the Cu 3 d orbitals were chosen as U = 6.0 eV was adopted. One of the surface Cu atom in the unit cell was substituted by Pd atom, which was used as the model of Pd₁/Cu₂O catalyst. Atomic coordinates of computational models for Cu₂O (100) and Pd-Cu₂O (100) are shown in Supplementary Data 1.