Integrating CO₂ electroreduction with phenol hydrogenation on an oxygen-affinity tailored catalyst

Abstract

Electrocatalytic CO₂ reduction (ECR) to formic acid faces challenges in separating and purifying a formate-electrolyte mixture. In situ utilization of this mixture presents a promising yet underexplored solution. Here, we report the synthesis of Bi_xPd_1−xTe nanocrystals (NCs) via a microwave-assisted cation topological exchange approach, enabling the precise tuning of surface oxygen affinities to simultaneously optimize the ECR and catalytic transfer hydrogenation (CTH) of phenol. Optimized Bi_0.1Pd_0.9Te NCs achieve a 92% formate Faradaic efficiency at −0.9 volts versus reversible hydrogen electrode and a production rate of 860 millimoles per hour per gram of catalyst at 100 milliamperes per square centimeter. This formate-electrolyte mixture serves as an effective hydrogen donor, enabling 98% selectivity toward cyclohexanone in phenol hydrogenation. Mechanistic studies show uniformly dispersed Bi sites create an oxygen affinity gradient, enhancing *OCHO adsorption for formate production and promoting noncoplanar phenol adsorption for selective cyclohexanone formation. This work pioneers synergistic ECR-CTH integration, establishing an innovative CO₂ valorization and biomass upgrading strategy.

An oxygen-affinity tailored catalyst enables integrated CO₂-to-formate electroreduction and formate-driven phenol hydrogenation.

INTRODUCTION

The conversion of carbon dioxide (CO₂) into value-added products through electrocatalytic CO₂ reduction (ECR) processes, powered by renewable electricity, offers an appealing approach for the sustainable production of chemical commodities (1–7). Among the potential products, formic acid stands out for its economic potential and the efficiency of its two-electron transfer process. Recent advancements have highlighted the efficacy of p-block metal catalysts—such as Sn, Bi, Pb, In, and their metal oxides—in achieving high selectivity toward formic acid production (8–15). Despite the considerable strides made in the ECR to formic acid, practical challenges persist that extend beyond the development of high-performance electrocatalysts. A critical limitation stems from conventional aqueous electrolyte systems, which typically produce a dilute formate-electrolyte mixture rather than pure formic acid (13–15), necessitating energy-intensive and costly downstream purification and concentration procedures to fulfill the requirements of various industrial applications.

A potential solution for producing pure formic acid involves the use of solid electrolytes, an approach whose feasibility has been validated by several studies (16–21). However, this method remains costly and requires stringent equipment and operational conditions, currently restricting its application to laboratory settings on a short-term basis. Therefore, developing direct applications for the formate-electrolyte mixture holds a substantial practical value to extend the scope of ECR. Formate compounds have gained recognition as efficient hydrogen donors in catalytic transfer hydrogenation (CTH) reactions, offering a safe, convenient, and cost-effective alternative to the molecular hydrogen source (22–24). Despite this potential, the integration of CTH with the formate-electrolyte mixture derived from ECR remains underexplored, representing a promising yet largely untapped avenue for downstream applications.

In ECR, the electrolyte is typically an aqueous solution, requiring the high water solubility of the hydrogenated substrate. Phenol, which is miscible with water at any ratio when heated above 65°C, primarily yields cyclohexanone and cyclohexanol as hydrogenation products (25–29). Cyclohexanone, a vital industrial raw material, is essential for nylon production and the synthesis of fine chemicals, including pesticides, paints, and dyes. The intrinsic hydrophobicity of cyclohexanone facilitates spontaneous phase separation from the aqueous formate-electrolyte mixture, enabling efficient product recovery. This approach offers substantial advantages over conventional cyclohexane oxidation methods by streamlining product isolation and enhancing cost efficiency through improved atomic economy. However, the thermodynamic preference for cyclohexanol formation during phenol hydrogenation highlights the critical need for achieving high cyclohexanone selectivity to enable broader industrial adoption. Consequently, the development of a catalyst capable of efficiently integrating ECR and CTH processes represents a compelling yet challenging endeavor.

Herein, we report the synthesis of a series of Bi_xPd_1−xTe NCs by using a microwave-assisted cationic topological exchange (CE) approach, using CuTe NCs as a sacrificial template to enable simultaneous substitution of Pd and Bi cations. This method yields a finely tuned surface oxygen affinity gradient in Bi_xPd_1−xTe NCs, endowing them with exceptional catalytic performance in both ECR and CTH, thereby establishing an effective tandem catalyst system (Fig. 1). In ECR, Bi_0.1Pd_0.9Te NCs outperform PdTe NCs, delivering a higher formate Faradaic efficiency (FE_formate) and production rate, enabled by Bi incorporation that optimizes HOCO* adsorption energy. Using the ECR-generated formate-electrolyte mixture for in situ phenol hydrogenation, Bi_0.1Pd_0.9Te NCs also demonstrate high cyclohexanone selectivity because of its uniformly distributed Bi sites, which effectively mitigates σ-π adsorption while preserving the hydrogenation efficiency of Pd sites, promoting the noncoplanar phenol adsorption configuration conducive to cyclohexanone formation. This work sets a precedent for integrating ECR and CTH into a unified strategy for CO₂ valorization and the upgrading of biomass-derived materials.

Fig. 1. Schematic illustration of synergistic integration of ECR and CTH.

Schematic illustration of the in situ hydrogenation of phenol to cyclohexanone catalyzed by Bi_xPd_1−xTe NCs using the formate-electrolyte mixture derived from ECR as the hydrogen source.

RESULTS

Preparation and characterizations of Bi_xPd_1−xTe NCs

In a typical synthesis, Bi_xPd_1−xTe NCs were prepared through a microwave-assisted cationic topological exchange approach. CuTe NCs were initially synthesized using a facile wet-chemical method (fig. S1) (30), followed by the simultaneous exchange of Bi and Pd cations under microwave-assisted conditions, resulting in the formation of a series of Bi_xPd_1−xTe NCs (Fig. 2A and see Materials and Methods for detailed procedures). The crystal structure of Bi_xPd_1−xTe NCs was analyzed by x-ray diffraction (XRD) (Fig. 2B). In the absence of Bi, only the PdTe crystalline phase (JCPDS no. 29-0971) is observed. However, with increasing Bi precursor, the diffraction peaks at 29.6°, 40.6°, and 43.7° conspicuously shift toward lower angles (Fig. 2, C and D), indicating an expansion in the unit cell volume and an escalation in lattice constants as a result of the incorporation of larger Bi atoms. The atomic ratio of Bi_xPd_1−xTe NCs and PdTe NCs was further corroborated by the energy-dispersive x-ray spectroscopy and inductively coupled plasma-optical emission spectrometry analyses (fig. S2), being consistent with XRD results. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images reveal that both Bi_xPd_1−xTe NCs and PdTe NCs exhibit a relatively uniform rectangular shape with an average diagonal length of ~200 nm (Fig. 2E and fig. S3A). Elemental mappings confirm the uniform distribution of Bi, Pd, and Te elements throughout Bi_xPd_1−xTe NCs (Fig. 2F and figs. S4A and 5A). The selected area electron diffraction pattern demonstrates that the hexagonal close-packed structure (P63/mmc) of PdTe is retained upon Bi incorporation (Fig. 2G). The chemical states of Bi_xPd_1−xTe NCs were further probed using x-ray photoelectron spectroscopy (XPS). The metallic Pd 3d_5/2 core levels for Bi_xPd_1−xTe NCs are observed at 334.49, 334.45, and 334.38 eV for Bi_0.05Pd_0.95Te NCs, Bi_0.1Pd_0.9Te NCs, and Bi_0.2Pd_0.8Te NCs, respectively, showing a shift to lower binding energies compared to PdTe NCs at 334.52 eV (fig. S6A). This shift intensifies with increasing of Bi content, implying charge redistribution between Pd and Bi atoms and resulting in an electron-enriched state at Pd sites. Moreover, contrary to the oxide predominance in Bi powder because of its oxophilic character, most Bi species in Bi_xPd_1−xTe NCs exist in a metallic state, affirming a robust interaction between Bi and Pd/Te atoms. Upon Bi incorporation into PdTe NCs, the metallic Bi 4f_7/2 core levels shift to higher binding energies of 156.42, 156.46, and 156.55 eV for Bi_0.05Pd_0.95Te NCs, Bi_0.1Pd_0.9Te NCs, and Bi_0.2Pd_0.8Te NCs, respectively, compared to Bi powder at 156.39 eV (fig. S6B). This shift toward higher binding energies becomes more pronounced with decreasing Bi content, suggesting a greater surrounding of Bi atoms by Pd atoms. In addition, the relative valence state proportions of Te remain constant despite changes in Bi site distribution, confirming the stability of the Te anionic framework (fig. S6C).

Fig. 2. Synthesis and characterizations of Bi_xPd_1−xTe NCs.

(A) Schematic illustration of the fabrication process for a series of Bi_xPd_1−xTe NCs using a microwave-assisted cationic topological exchange (CE) approach. (B) XRD patterns and (C and D) enlarged patterns from the selected regions in (B) for Bi_xPd_1−xTe NCs and PdTe NCs. (E) HAADF-STEM image, (F) elemental mapping images, and (G) corresponding selected area electron diffraction pattern of Bi_0.1Pd_0.9Te NCs.

To determine the distribution and incorporation of Bi within PdTe crystals following cationic topological exchange, advanced surface atomic-level characterization techniques were used. The HAADF-STEM image and the enlarged colored region reveal that Bi_0.1Pd_0.9Te NCs are abundant in voids and cracks, a consequence of the cation-exchange process, consistent with previously reported PdTe NCs (Fig. 3, A and B) (30). The aberration-corrected HAADF-STEM image further resolves the atomic arrangement (Fig. 3C and figs. S4B and 5B). Intensity profiles across the orange rectangular regions in the corresponding HAADF-STEM images provide clear evidence for Bi incorporation into the PdTe crystal lattice. Notably, Bi sites are uniformly distributed in Bi_0.05Pd_0.95Te NCs and Bi_0.1Pd_0.9Te NCs (Fig. 3, D to F, and fig. S4, C and D), while more pronounced clustering of Bi sites is observed in Bi_0.2Pd_0.8Te NCs (fig. S5, C and D), indicating subtle structural modifications as the Bi content exceeds a critical threshold. X-ray absorption spectra at the Bi L₃-edge were further analyzed to probe local structural changes induced by varying Bi content. The x-ray absorption near-edge fine structure at the Bi L₃-edge reveals that the electronic structure and spectral features of Bi_xPd_1−xTe NCs closely resemble those of metallic Bi foil rather than Bi₂O₃ or Bi₂Te₃, confirming the predominance of the metallic Bi state in all Bi_xPd_1−xTe NCs (fig. S7A). Specifically, the Bi L₃-edge spectra of Bi_xPd_1−xTe NCs exhibit a high-energy shift with increasing Bi content, indicating a transition to a higher valence state, consistent with XPS results. Analysis of the extended x-ray absorption fine structure (EXAFS) spectra in the k-space reveals substantial differences compared to Bi foil, Bi₂O₃, and Bi₂Te₃ (fig. S7B), suggesting that Bi atoms occupy Pd sites during the cationic topological exchange process while maintaining the hexagonal close-packed crystal structure of PdTe. Noticeably, the Fourier transform (FT)-EXAFS results of Bi_0.05Pd_0.95Te NCs and Bi_0.1Pd_0.9Te NCs differ markedly from Bi foil, as the prominent peak assigned to the Bi─Te/Pd path at the Bi L₃-edge is shorter than the Bi─Bi path at 3.0 Å for Bi foil. In contrast, Bi_0.2Pd_0.8Te NCs exhibit a distinct Bi─Bi path at 3.0 Å, indicating that Bi atoms form continuous sites as their content increases (fig. S7C). In addition, wavelet transform contour plots of Bi_xPd_1−xTe NCs as well as references are presented in Fig. 3G. The intensity maxima of Bi_0.05Pd_0.95Te NCs (k = 4.52 Å⁻¹, R = 2.66 Å) and Bi_0.1Pd_0.9Te NCs (k = 4.86 Å⁻¹, R = 2.66 Å) align closely with the Bi─Te scattering model (k = 4.42 Å⁻¹, R = 2.70 Å) rather than the Bi─Bi scattering path (k = 8.38 Å⁻¹, R = 2.95 Å). Given that Bi is surrounded by Te and Pd in the Bi_xPd_1−xTe structure, these intensity maxima are attributed to the Bi─Te/Pd scattering path. However, the intensity maximum of Bi_0.2Pd_0.8Te NCs (k = 6.32 Å⁻¹, R = 2.68 Å) shifts to higher R and k values with increasing Bi content, suggesting the emergence of a Bi─Bi scattering path. This trend reflects lattice expansion upon Bi incorporation, consistent with XRD results. These findings demonstrate that topological substitution at lower Bi content levels results in a uniform and isolated dispersion of Bi sites within the PdTe lattice, whereas higher Bi content promotes Bi─Bi bond formation (Fig. 3H). In addition, O₂ temperature-programmed desorption analysis reveals that the main oxygen desorption peaks of Bi_xPd_1−xTe NCs shift to higher temperatures with increasing Bi content compared to PdTe NCs (fig. S8). This trend suggests that incorporating Bi into the PdTe lattice alters the surface oxygen affinity, thereby affecting catalytic behaviors.

Fig. 3. Advanced characterizations of Bi_xPd_1−xTe NCs.

(A) HAADF-STEM image and (B) corresponding colored image highlighting the orange marked region in (A) for Bi_0.10Pd_0.90Te NCs. (C) Aberration-corrected HAADF-STEM image depicting the surface atomic arrangement of Bi_0.10Pd_0.90Te NCs. (D to F) Intensity profiles extracted from the orange-marked regions in (C) and atomic model overlays of the corresponding regions. a.u., arbitrary units. (G) Wavelet transform plots of the Bi L₃-edge k³-weighted EXAFS signals for Bi_xPd_1−xTe NCs, Bi₂O₃, Bi₂Te₃, and Bi foil. (H) Schematic diagram of the crystal structure model showing isolated Bi sites and Bi─Bi bond formation. Blue, red, and green spheres represent Te, Bi, and Pd atoms, respectively.

ECR performance of Bi_xPd_1−xTe NCs

The ECR performance of Bi_xPd_1−xTe NCs was initially evaluated in a three-electrode H-type electrolytic cell using controlled potential electrolysis to analyze the reduction products, with PdTe NCs as a benchmark. Gas and liquid products were analyzed separately using gas chromatography (GC) and nuclear magnetic resonance (NMR) spectroscopy (fig. S9). No liquid products are detected for PdTe NCs, which exhibit a high Faradaic efficiency for CO of 86% at −0.8 V versus reversible hydrogen electrode (RHE; fig. S10A). In contrast, formate is detectable on Bi_xPd_1−xTe NCs, with FE_formate increasing substantially and the hydrogen evolution reaction being markedly suppressed as the Bi content increased. Bi_0.1Pd_0.9Te NCs and Bi_0.2Pd_0.8Te NCs demonstrate high FE_formate across the applied voltage range, with Bi_0.1Pd_0.9Te NCs achieving a maximum FE_formate of 92% at −0.9 V versus RHE (Fig. 4A and fig. S10, B to D). To enhance the formate concentration by accelerating CO₂ diffusion and mass transfer in the electrolyte, ECR experiments were further conducted in a flow-through electrochemical cell equipped with a gas diffusion electrode (fig. S11). As depicted in Fig. 4B and fig. S12, Bi_xPd_1−xTe NCs present a volcano-shaped trend in formate production rate with the current density from 50 to 200 mA cm⁻². In particular, Bi_0.1Pd_0.9Te NCs exhibit the highest formate production rate of 1.85 mmol hour⁻¹ cm⁻² at 150 mA cm⁻², outperforming Bi_0.05Pd_0.95Te NCs and Bi_0.2Pd_0.8Te NCs across the entire current density range. Furthermore, the formate production rate was normalized by catalyst loading (Fig. 4C and fig. S13). Balancing catalyst utilization and catalytic efficiency, Bi_0.1Pd_0.9Te NCs demonstrate optimal performance at 100 mA cm⁻², achieving a formate production rate of 860 mmol hour⁻¹ g_cat⁻¹ with an FE_formate of 91%, surpassing those of Bi_0.05Pd_0.95Te NCs (580 mmol hour⁻¹ g_cat⁻¹, 62% FE_formate) and Bi_0.2Pd_0.8Te NCs (820 mmol hour⁻¹ g_cat⁻¹, 87% FE_formate). To optimize formate production, key ECR parameters were meticulously evaluated for Bi_0.1Pd_0.9Te NCs at 100 mA cm⁻². Catalyst utilization efficiency peaked at a catalyst loading of 2 mg cm⁻², with higher loadings leading to reduced formate production rates and FE_formate, potentially due to catalyst shedding during the flow-through reaction (Fig. 4D). The formate production rate increased from 750 to 860 mmol hour⁻¹ g_cat⁻¹ as the electrolyte concentration rose from 0.25 to 0.5 M, with no notable improvement observed at higher concentrations (Fig. 4E). The initial lower production rate and FE_formate at 0.25 M are primarily attributed to the competitive hydrogen evolution reaction, which is effectively suppressed in more concentrated electrolytes. In addition, electrolyte flow rates had a minimal impact on formate production, with optimal efficiency achieved at 15 ml min⁻¹ (Fig. 4F). Under these optimized conditions, the cumulative formate concentration for Bi_0.1Pd_0.9Te NCs at 100 mA cm⁻² increased linearly over the first 8 hours of reaction (Fig. 4G). Beyond this period, device efficiency declined, potentially due to formate crossover or carbon loss associated with the prolonged use of anion exchange membranes. In addition, the crystal structure of the catalyst remains largely intact after prolonged ECR testing (fig. S14), while XPS analysis indicates a slight increase in high-valence species on the catalyst surface (fig. S15).

Fig. 4. ECR performance and mechanism investigation over Bi_xPd_1−xTe NCs.

(A) FE_formate over Bi_xPd_1−xTe NCs in a three-electrode H-type electrolytic cell. (B) Formate production rate of Bi_xPd_1−xTe NCs at various current densities from 50 to 200 mA cm⁻² in a flow cell. (C to F) FE_formate and formate production rate normalized on Bi_0.1Pd_0.9Te NCs under different conditions: (C) current densities, (D) catalyst dosages, (E) KHCO₃ concentrations, and (F) electrolyte flow rates. (G) Cumulative formate concentration and FE_formate of Bi_0.1Pd_0.9Te NCs after a long-term chronoamperometric test at 100 mA cm⁻². In situ ATR-FTIR spectra depicting intermediate adsorption on (H) PdTe NCs and (I) Bi_0.1Pd_0.9Te NCs at various potentials. Gibbs free energy profiles for formate and CO production on (J) PdTe NCs and (K) Bi_0.1Pd_0.9Te NCs.

To probe the catalytic intermediates and reaction pathways, we performed in situ attenuated total reflectance-FT infrared (ATR-FTIR) spectroscopy over the potential range of −0.4 to −1.2 V versus RHE. As depicted in Fig. 4H, a distinctive peak at 1350 cm⁻¹, assigned to the O─C─O vibrations in the *OCHO species—a key intermediate in formate formation—emerges at −0.6 V for Bi_0.1Pd_0.9Te NCs and intensifies progressively with increasing potentials from −0.6 to −1.2 V. In stark contrast, *OCHO species are not clearly visible at 1350 cm⁻¹ for PdTe NCs, where prominent peaks around 1930 cm⁻¹, corresponding to *CO species, are observed (Fig. 4I). These results align with the ECR performance of Bi_0.1Pd_0.9Te NCs and PdTe NCs, suggesting that Bi incorporation redirects the ECR pathway. Density functional theory (DFT) calculations were further used to analyze the Gibbs free energy changes for reaction steps on Bi_0.1Pd_0.9Te (101) and PdTe (101) surfaces. The adsorption behaviors of *OCHO and *COOH are critical in determining the reaction pathways and product distribution. *OCHO, with its characteristic oxyphilic bidentate adsorption configuration, is identified as a key intermediate for formate production, whereas *COOH, with a carbophilic unidentate configuration, is the primary intermediate for CO formation. The formation of *COOH (0.83 eV) on PdTe (101) is energetically more favorable than that of *OCHO (1.09 eV) (Fig. 4J), representing the potential determining step for CO formation. In contrast, the Gibbs free energy for *OCHO (0.43 eV) on Bi_0.1Pd_0.9Te (101) is substantially lower than that for *COOH (1.48 eV) (Fig. 4K). This suggests that Bi incorporation substantially alters the ECR pathway, steering it away from the dominance of CO production in PdTe to a more formate-selective route, aligning with the experimentally observed structure-dependent selectivity. Therefore, it can be concluded that PdTe NCs electrochemically reduce CO₂ to CO through the *COOH intermediate. On the contrary, the introduction and increase in Bi content in Bi_xPd_1−xTe NCs enhance surface oxygen affinity, altering adsorption configurations to favor *OCHO and thereby promoting the selective production of formate during ECR.

CTH performance of Bi_xPd_1−xTe NCs

Using the formate-electrolyte mixture generated from ECR as the hydrogen source, the performance of Bi_xPd_1−xTe NCs in phenol hydrogenation was evaluated, with products validated and quantified using NMR and GC (see Materials and Methods for detailed procedures and fig. S16). To optimize phenol conversion and cyclohexanone selectivity, we systematically investigated the effects of reaction temperature and formate concentration. With Bi_0.1Pd_0.9Te NCs as the catalyst and a formate-electrolyte mixture (1.37 wt %, 10-hour electrolysis) as the hydrogen source, phenol conversion increases with temperature, reaching complete conversion at 90°C with 92% cyclohexanone selectivity after 15 hours. Further temperature increases slightly reduced cyclohexanone selectivity, possibly due to the reaction temperature approaching the boiling point of the solvent, promoting the readsorption and subsequent hydrogenation of cyclohexanone to cyclohexanol (fig. S17A). In addition, the stoichiometric ratio of formate to phenol substantially affects phenol conversion. By adjusting the electrolysis time, various formate concentrations were obtained, with full phenol conversion and 92% cyclohexanone selectivity achieved at ~4 equiv of formate after 15 hours (fig. S17B). Under these optimized conditions, the effect of reaction time on phenol hydrogenation was examined for Bi_xPd_1−xTe NCs and PdTe NCs. As shown in Fig. 5A, Bi_0.1Pd_0.9Te NCs achieve 100% phenol conversion within 10 hours, with 98% selectivity toward cyclohexanone. This selectivity remained high at 92% even when the reaction time was extended to 15 hours. In contrast, PdTe NCs exhibit a notable increase in phenol conversion but a sharp decline in cyclohexanone selectivity (Fig. 5B). To further elucidate the influence of Bi content, the hydrogenation activity of Bi_0.05Pd_0.95Te NCs and Bi_0.2Pd_0.8Te NCs was assessed. For Bi_0.05Pd_0.95Te NCs, which feature lower Bi content, the phenol conversion rate exceeds that of Bi_0.1Pd_0.9Te NCs; however, this is accompanied by reduced cyclohexanone selectivity (Fig. 5C). Conversely, Bi_0.2Pd_0.8Te NCs, with higher Bi content, exhibit a substantially lower phenol conversion rate, failing to achieve complete conversion even after 12 hours, yet maintain high cyclohexanone selectivity (Fig. 5D). Moreover, Pd/C achieves full phenol conversion within 6 hours, but cyclohexanone selectivity decreases substantially with prolonged reaction time (fig. S18). These results demonstrate that the strategic incorporation of Bi into the PdTe lattice enhances cyclohexanone selectivity without compromising phenol conversion efficiency. To assess its potential for industrial applications, the reusability of Bi_0.1Pd_0.9Te NCs was evaluated in phenol hydrogenation. The catalyst maintained high phenol conversion and cyclohexanone selectivity over six consecutive cycles (Fig. 5E). Postrecycling structural characterizations of Bi_0.1Pd_0.9Te NCs reveal no notable alterations in the crystal structure and surface valence state (figs. S19 and S20), highlighting the robust stability of Bi_0.1Pd_0.9Te NCs in phenol hydrogenation processes.

Fig. 5. CTH performance and mechanism investigation over Bi_xPd_1−xTe NCs.

(A to D) Phenol conversion and product selectivity over time using the formate mixed electrolyte as the hydrogen source for (A) Bi_0.1Pd_0.9Te NCs, (B) PdTe NCs, (C) Bi_0.05Pd_0.95Te NCs, and (D) Bi_0.2Pd_0.8Te NCs. (E) Reusability of Bi_0.1Pd_0.9Te NCs on consecutive phenol hydrogenation reactions. (F) Evaluation of the external diffusion effect on Bi_0.1Pd_0.9Te NCs over a 2-hour period. (G) Comparison of bicarbonate and formate concentrations before and after CTH. (H) Time-conversion plots for the hydrogenation of phenol and cyclohexanone over Bi_0.1Pd_0.9Te NCs at 90°C. (I) Arrhenius plots for the hydrogenation of phenol. (J) Different adsorption configurations of phenol on BiPdTe (101) and PdTe (101) surfaces. (K) Schematic representation of the proposed CTH mechanisms on PdTe NCs and Bi_0.1Pd_0.9Te NCs surfaces. Abbreviations: Ph, phenol; CHE, cyclohexanone; CHL, cyclohexanol.

To understand the high cyclohexanone selectivity observed on Bi_0.1Pd_0.9Te NCs, we examined the external mass transfer effect by correlating stirring rates with catalyst activity. Figure 5F illustrates that phenol conversion and cyclohexanone selectivity remained constant over the 2-hour reaction period across stirring speeds ranging from 100 to 600 rpm, indicating negligible external mass transfer limitations. This constancy is likely attributed to formate molecules adsorbed on the catalyst surface, which generate adsorbed hydrogen (H*) that directly engages in phenol hydrogenation, bypassing the diffusion barrier associated with gaseous hydrogen. This hypothesis is further corroborated by the substantial decrease in cyclohexanone selectivity over Bi_0.1Pd_0.9Te NCs when the formate solution was replaced with gaseous hydrogen (fig. S21). To further clarify the transformation of formate after its involvement in CTH, the concentrations of bicarbonate and formate were quantified by using acid-base titration and ion chromatography, respectively (fig. S22). The results reveal an obvious decrease in formate concentration and a corresponding increase in bicarbonate concentration after the reaction (Fig. 5G), indicating that CO₂ is initially sequestered in formate during the ECR process and subsequently converted into bicarbonate during the CTH process (fig. S23).

The reaction kinetics of phenol hydrogenation to cyclohexanone and cyclohexanone hydrogenation to cyclohexanol over Bi_0.1Pd_0.9Te NCs were systematically investigated (Fig. 5H). The rate constant k₁ for phenol hydrogenation to cyclohexanone is determined to be 0.289 hour⁻¹, while k₂ for cyclohexanone hydrogenation to cyclohexanol is 0.017 hour⁻¹. Clearly, the value of k₂ is substantially smaller than k₁, indicating that the hydrogenation rate of cyclohexanone is much slower than that of phenol, kinetically favoring cyclohexanone selectivity. In addition, the apparent activation energy (E_a) for phenol hydrogenation over PdTe NCs and Bi_xPd_1−xTe NCs is determined using Arrhenius plots (Fig. 5I and fig. S24). The E_a for PdTe NCs is calculated to be 33.2 kJ/mol, whereas Bi_xPd_1−xTe NCs exhibit a progressive increase in E_a with higher Bi content, reaching 53.2 kJ/mol for Bi_0.2Pd_0.8Te NCs. This trend implies that excessive Bi content can detrimentally affect the phenol hydrogenation efficiency, leading to reduced phenol conversion. DFT calculations were further performed to analyze the adsorption energies and configurations of phenol on PdTe (101) and BiPdTe (101) surfaces, providing insights into the role of Bi sites (Fig. 5J). Phenol adsorbs on the PdTe (101) surface in a partially coplanar configuration, stabilized by strong σ-π interactions between the phenyl ring and metal atoms, with an optimal adsorption energy of −0.11 eV. In contrast, on the BiPdTe (101) surface, the presence of Bi weakens the σ-π interaction because of its oxygen affinity, resulting in a noncoplanar adsorption configuration with a lower adsorption energy of −0.21 eV. These results demonstrate that incorporating an optimal amount of uniformly dispersed Bi sites substantially enhances the noncoplanar adsorption of phenol while preserving the hydrogenation efficiency of Pd sites, thereby markedly improving cyclohexanone selectivity. Furthermore, the presence of Bi sites moderately reduces the activation energy for the transfer of adsorbed hydrogen (H*) to phenol, effectively inhibiting the further hydrogenation of cyclohexanone to cyclohexanol (Fig. 5K).

DISCUSSION

In summary, we developed a microwave-assisted cation topological exchange approach to construct a series of Bi_xPd_1−xTe NCs. This strategy facilitates the precise modulation of surface oxygen affinity, fulfilling the dual requirements for ECR and CTH, thereby establishing an efficient tandem catalytic system. The incorporation of Bi into the PdTe lattice substantially enhances *OCHO adsorption, enabling Bi_0.1Pd_0.9Te NCs to achieve a high FE_formate of 92% at −0.9 V versus RHE and a formate production rate of 860 mmol hour⁻¹ g_cat⁻¹ at a current density of 100 mA cm⁻². Bi_0.1Pd_0.9Te NCs can achieve a cyclohexanone selectivity of up to 98% with complete phenol conversion when using the formate-electrolyte mixture from ECR as an in situ hydrogen source for phenol hydrogenation. The enhanced selectivity is primarily attributed to the presence of Bi sites, which weakens σ-π interactions and encourages noncoplanar phenol adsorption, facilitating the conversion of phenol to cyclohexanone. This work paves the way for integrating two sustainable processes—ECR and CTH—into a comprehensive strategy for sequestering CO₂ and valorizing biomass-derived materials.

MATERIALS AND METHODS

Chemicals

Palladium(II) chloride (PdCl₂; 99%) and telluric(VI) acid (H₆TeO₆; 98%) were purchased from Sigma-Aldrich. Copper(II) formate hydrate [Cu(COOH)₂·4H₂O; 99%], ascorbic acid (C₆H₈O₆, AA; 99.9%), bismuth(III) acetate (BiAc₃; 99.8%), and polyvinylpyrrolidone (molecular weight, 58,000) were purchased from J&K. Dimethyl sulfoxide (C₂H₆OS, DMSO, analytical reagent) was purchased from Aladdin. Phenol (C₆H₆O₆, analytical reagent), potassium bicarbonate (KHCO₃, analytical reagent), and benzyl alcohol (C₇H₈O, analytical reagent) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All the chemicals were used as received without further purification. The water (18 megohms/cm) used in all experiments was prepared using an ultrapure water purification system (Aqua Solutions).

Synthesis of CuTe NCs

In a typical preparation of CuTe NCs, Cu(COOH)₂ (5.6 mg), H₆TeO₆ (5.8 mg), AA (18 mg), polyvinylpyrrolidone (50 mg), and benzyl alcohol (10 ml) were added to a glass pressure vessel (volume: 30 ml). The mixture was ultrasonicated at room temperature for around 0.5 hours. The homogeneous solution was then heated to 160°C and maintained at 160°C for 3 hours in an oil bath. After cooling to room temperature, the products were collected by centrifugation (12,000 rpm, 3 min) and washed three times with ethanol and acetone.

Synthesis of PdTe NCs and Bi_xPd_1−xTe NCs

PdTe NCs and Bi_xPd_1−xTe NCs were synthesized via microwave-assisted cationic topological exchange transformation from CuTe NCs with Pd and Bi cations. For PdTe NC synthesis, CuTe NCs were dispersed in DMSO (5 ml) by ultrasonication, followed by the addition of PdCl₂ (4.4 mg). The mixture was sealed under nitrogen in a 20-ml glass microwave tube and heated to 100°C for 15 min using an XH-100A microwave synthesizer. After cooling to room temperature, the products were collected by centrifugation (12,000 rpm, 3 min) and washed three times with ethanol and acetone. For the synthesis of Bi_0.05Pd_0.95Te NCs, Bi_0.1Pd_0.9Te NCs, and Bi_0.2Pd_0.8Te NCs, an additional 0.5, 1.0, and 2.0 mg of BiCl₃ were added concurrently with PdCl₂ while keeping all other reaction conditions identical to those used for PdTe NC synthesis.

Characterizations

The transmission electron microscopy (TEM) samples were prepared by depositing an ethanol dispersion of the product onto carbon-coated copper TEM grids using a pipette, followed by drying under ambient conditions. The TEM, HAADF-STEM, and high-resolution TEM images were acquired using an FEI Tecnai F30 transmission electron microscope operated at an accelerating voltage of 300 kV. Atomic-resolution HAADF-STEM images and energy-dispersive x-ray spectroscopy mappings were obtained using a JEM-ARM300F scanning transmission electron microscope/transmission electron microscope operated at 300 kV. XRD patterns was collected on an X’Pert-Pro MPD diffractometer (Netherlands PANalytical) with a Cu Kα x-ray source (λ = 1.542 Å). Catalyst concentrations were determined using inductively coupled plasma-optical emission spectroscopy (Varian 710-ES). XPS was performed using an SSI S-Probe XPS Spectrometer, with the carbon peak at 284.6 eV serving as a reference to correct for charging effects. In situ ATR-FTIR experiments were conducted in an internal reflection configuration using a gold (Au) thin film–deposited silicon (Si) prism (Au/Si) on a Nicolet iS50 FT-IR spectrometer. X-ray absorption fine structure measurements at the Bi L₃-edge were conducted on the BL14W1 beamlines and BL11B beamlines with Si (111) crystal monochromators at the Shanghai Synchrotron Radiation Facility (Shanghai, China). Data were processed according to standard procedures using the Demeter software package (version 0.9.24) (31).

Electrochemical measurements

The catalyst ink was prepared by ultrasonically dispersing 10 mg of catalyst, 1 mg of XC-72 C, 0.50 ml of ethanol, 0.45 ml of ultrapure water, and 0.05 ml of Nafion solution (5 wt %) for 30 min. Twenty microliters of the catalyst ink was coated on a glass carbon electrode (3-mm diameter) to serve as the working electrode for linear sweep voltammetry and potentiostatic tests in an H-type cell configuration. In addition, 200 μl of the catalyst ink was coated onto carbon paper (1 cm² area) to function as the working electrode for galvanostatic tests in a flow-through reactor setup. All electrochemical measurements were performed at room temperature and ambient pressure using a CHI 760E electrochemical workstation. CO₂ electrocatalysis was evaluated in both H-cell and flow-through reactors under potentiostatic and galvanostatic conditions, respectively.

In the H-type cell setup, a three-electrode system was used, consisting of a Ag/AgCl electrode as the reference, a platinum sheet as the counter electrode, and a catalyst-coated glassy carbon electrode as the working electrode. The cathode and anode compartments were separated by a Nafion 117 membrane and filled with a 0.1 M KHCO₃ electrolyte. Before CO₂ electroreduction, the cathode chamber was purged with CO₂ for 30 min and maintained under a continuous CO₂ flow during the reaction. In the flow-through reactor, a two-electrode system was used, with both the catholyte and anolyte consisting of 0.5 M KHCO₃. The CO₂ flow rate was maintained at 30 ml min⁻¹ using a gas chamber, and the 0.5 M KHCO₃ electrolyte was circulated between the anode and the cathode at a flow rate of 15.0 ml min⁻¹ during ECR. Electrode potentials were converted to the RHE scale using the equation E(RHE) = E(Ag/AgCl) + 0.059 × pH + 0.197 V.

The concentration of liquid-phase products in the collected electrolyte was quantified using a 600-MHz ¹H NMR spectrometer (Agilent DirectDrive2), with DMSO as an internal reference for chemical shifts. Calibration curves for formate were established using pure formate at various concentrations. The Faradaic efficiency (FE) of CO₂ electroreduction products was calculated using the equation FE(%) = (Z × n × F)/Q × 100%, where Z is the number of electrons transferred, n is the number of moles of the product, and Q is the total charge transferred.

Phenol hydrogenation

In a typical reaction, phenol (140 mg, 1.5 mmol), Bi_xPd_1−xTe NCs (10 mg, 0.04 mmol), and formate-electrolyte mixtures of varying concentrations (20 ml) were added to a 48-ml round-bottom flask. The reaction mixture was heated to 90°C for an optimized duration under continuous stirring at 400 rpm. At the end of the reaction, the flask was cooled to room temperature. The supernatant was centrifuged, extracted with dichloromethane, and analyzed by GC. Products were identified by comparing their GC retention time and ¹H NMR spectrum with those of standards. Following the reaction, the catalyst was recovered by centrifugation (12,000 rpm, 3 min), washed three times with deionized water, dried in an oven, and reused under identical conditions for subsequent recycling experiments.

The conversion of reactants and the selectivity of the target product were calculated using Eqs. 1 and 2, respectively

$$\begin{array}{c}\text{Conversion}(\%)=\frac{n_{\text{reactant}}^0-n_{\text{reactant}}^t}{n_{\text{reactant}}^0}\times 100\hfill \end{array}$$ (1)

$$\begin{array}{c}\text{Selectivity}(\%)=\frac{n_{\text{product}}^t}{n_{\text{reactant}}^0-n_{\text{reactant}}^t}\times 100\hfill \end{array}$$ (2)

where $n_{\text{reactant}}^0$ is the initial amount of the reactant in moles in the reactor, and $n_{\text{product}}^t$ and $n_{\text{reactant}}^t$ are the amounts of the product and reactant in moles at reaction time t, respectively.

DFT calculation

The spin-polarized DFT calculations were performed using the Vienna Ab initio Simulation Package (32). The Perdew-Burke-Ernzerhof generalized-gradient approximation functional was used to describe electron interactions (33). An energy cutoff of 400 eV was applied, and the Monkhorst-Pack k-point grid was set to 2 by 3 by 1 for all the calculations. A vacuum region of 15 Å was established in the z direction to prevent interactions between adjacent surfaces. The convergence criteria for energy and force were set to less than 10⁻⁵ eV and 0.02 eV/Å, respectively, with the bottom layers held fixed during calculations. Adsorption energies (E_ads) were calculated using the equation E_ads = E_tot − E_surf − E_gas. The total energies of the surface with and without cyclohexanone adsorption are denoted as E_tot and E_surf, respectively, while E_gas represents the energy of the gas-phase cyclohexanone.

The reaction Gibbs free energy (ΔG) is calculated as ΔG = ΔE + ΔΕ_ZPE − TΔS, where ΔE is the reaction energy, ΔΕ_ZPE is the zero-point energy, T is the temperature (298.15 K), and ΔS is the entropy change derived from vibrational frequency calculations. The entropies of the gas-phase CO₂ and H₂ were sourced from the National Institute of Standards and Technology database under standard conditions.

Supplementary Materials

This PDF file includes:

Figs. S1 to S24