Unveiling synergy of strain and ligand effects in metallic aerogel for electrocatalytic polyethylene terephthalate upcycling

Significance

Electrochemical upcycling of polyethylene terephthalate holds significant promise for generating value-added chemicals and achieving carbon neutrality. Nevertheless, electrocatalytic ethylene glycol oxidation reaction involves multiple possible C₁ and C₂ pathways, resulting in complex potential products that requires additional processing paths for separation and purification. Here, we demonstrate that tuning the electronic and lattice structure of metallic aerogel catalyst is a powerful approach to simultaneously achieve the high activity and high glycolic acid selectivity. Our Pd₆₇Ag₃₃ catalyst exhibits outstanding performance over most of the state-of-the-art noble metal-based catalysts. This work not only provides insights into the structure–activity relationship at the molecule level but also presents innovative attempt for the construction of electrocatalytic waste carbonaceous coreforming system.

Abstract

Recently, there has been a notable surge in interest regarding reclaiming valuable chemicals from waste plastics. However, the energy-intensive conventional thermal catalysis does not align with the concept of sustainable development. Herein, we report a sustainable electrocatalytic approach allowing the selective synthesis of glycolic acid (GA) from waste polyethylene terephthalate (PET) over a Pd₆₇Ag₃₃ alloy catalyst under ambient conditions. Notably, Pd₆₇Ag₃₃ delivers a high mass activity of 9.7 A mg_Pd⁻¹ for ethylene glycol oxidation reaction (EGOR) and GA Faradaic efficiency of 92.7 %, representing the most active catalyst for selective GA synthesis. In situ experiments and computational simulations uncover that ligand effect induced by Ag incorporation enhances the GA selectivity by facilitating carbonyl intermediates desorption, while the lattice mismatch-triggered tensile strain optimizes the adsorption of *OH species to boost reaction kinetics. This work unveils the synergistic of strain and ligand effect in alloy catalyst and provides guidance for the design of future catalysts for PET upcycling. We further investigate the versatility of Pd₆₇Ag₃₃ catalyst on CO₂ reduction reaction (CO₂RR) and assemble EGOR//CO₂RR integrated electrolyzer, presenting a pioneering demonstration for reforming waste carbon resource (i.e., PET and CO₂) into high-value chemicals.

The exponential surge in plastic consumption, combined with inadequate waste management protocols, has raised substantial global concern regarding the issue of plastic waste. The accumulation of plastic waste in natural environments has already exceeded 6,300 million tons (Mt), with projections indicating a rise to approximately 12,000 Mt by 2050 (1). Polyethylene terephthalate (PET), extensively employed in packaging and textiles, exhibits an annual global production exceeding 70 Mt (2). However, a large proportion (>80%) of PET plastics end up in landfills and environment, rather than being recycled. Catalytic PET conversion presents a sustainable strategy for utilizing this carbon-rich resource, which not only produces valuable chemicals but also minimizes potential carbon emissions (3). Nevertheless, conventional thermal catalysis relying on the high temperature (~200 °C) and high pressure (>20 atm) results in significant energy consumption, which is not in line with the concept of environment-friendliness and sustainability (4–6). Hence, developing novel catalytic approaches allowing PET upcycling under milder conditions is of necessity (7–10).

Electrocatalytic PET upcycling, driven by renewable electricity, provides an attractive alternative to selectively oxidize PET-derived ethylene glycol (EG) into valuable chemicals under ambient conditions. It is pivotal to acknowledge that electrocatalytic ethylene glycol oxidation reaction (EGOR) involves multiple possible C₁ and C₂ pathways, thus resulting in complex potential products such as carbon dioxide (CO₂), formic acid (FA), glycolic acid (GA), and oxalic acid. Given GA is a high-value chemical and a cornerstone for degradable polyglycolic acid production (11), directing efforts toward conserving the C–C bond and fostering the selective formation of high-value GA from waste PET plastics is of paramount interest. Most recently, pioneers have discovered the potential to modulate the electron structure of active sites, thereby regulating catalyst-intermediate interactions to achieve the desired selective EG-to-GA conversion (12, 13). However, it remains a challenge to construct advanced catalysts to push the catalytic activity for GA synthesis to even higher levels. Pd-based materials have been recognized as the high active catalysts for thorough oxidation of EG into CO₂ in the fuel cell (14–18). Particularly intriguing is the integration of a second element (denoted as M) into Pd nanocrystals to form a PdM alloy phase, where the ligand and strain effects have been substantiated significance in tuning the interactions between catalyst and reactants, thus boosting the activity for EGOR in fuel cell (19–21). Despite its firmly entrenched reputation for exceptional activity in EGOR, the latent potential for the selective synthesis of high-value GA from EG on Pd-based alloy catalysts has been underestimated for a long time, leading to a noticeable research gap. Therefore, the prospect of tuning crystal and electronic configurations in Pd-based alloys and uncovering their effects on intermediate adsorption for the selective GA synthesis assumes paramount significance in steering the design of highly active and selective catalysts for electrocatalytic PET valorization.

In this work, we present a PdAg alloy aerogel as an electrochemical catalyst for the synthesis of GA from waste PET plastics. Remarkably, Pd₆₇Ag₃₃ not only delivers a high Faraday efficiency (FE) for GA of 92.7% at 1.0 V (vs. reversible hydrogen electrode, RHE) with high stability but also exhibits an excellent activity of 9.7 A mg_Pd⁻¹, 12.4 times higher than commercial 20% Pd/C. Evidently, it stands as the most potent catalyst for selective EG-to-GA conversion. Depth understanding of the structure–performance relationship on Pd₆₇Ag₃₃ in EGOR is acquired by operando Fourier transform infrared spectroscopy (FTIR) and density functional theory (DFT) calculations, which indicate the high activity and GA selectivity originates from the synergistic of strain and ligand effect: 1) the ligand effect triggered by electron transfer from Ag to Pd leads to the downshift of Pd’s d-band center, which suppresses the excessive oxidation by means of facilitating the desorption of carbonyl intermediates; 2) the tensile strain can accelerate the adsorption of EG molecules and *OH species on catalyst, thus optimizes the EGOR kinetics. Additionally, the unique electron structure of Pd₆₇Ag₃₃ suppresses the poison of CO, thus favoring the electrochemical CO₂ reduction reaction (CO₂RR) with an impressive FE_FA of 91.2% at −0.35 V. Notably, we employ Pd₆₇Ag₃₃ as a bifunctional catalyst for the coupling of EGOR and CO₂RR in an integrated electrolyzer and demonstrate >80% FE on both electrodes. In summary, this investigation not only provides invaluable perspectives on the pivotal roles played by the ligand effect and tensile strain in augmenting the selectivity and activity during the EG-to-GA conversion mediated by PdAg alloy catalysts but also presents a pioneering demonstration for reclaiming high-value chemicals from waste carbon resource via electrochemical covalorization of PET and CO₂.

Results

Material Synthesis and Characterization.

The PdAg hydrogel was synthesized via a two-step salting-out and cross-linking strategy (Fig. 1A). Briefly, the presence of trisodium citrate ions (CA^–) allows the formation of stable Pd and Ag colloidal solution, while the introduction of fluorine anions (F^–) disrupts the stability of colloidal solutions and facilitates the assembly of PdAg nanoparticles (NPs) that fuse into a cross-linked framework (22). The final PdAg aerogels were obtained via freeze drying (SI Appendix, Fig. S1). Transmission electron microscopy (TEM) images (Fig. 1B) visualized the cross-linked structure of Pd₆₇Ag₃₃, which was constituted by ultrathin nanowires interconnected with each other. The integrity of the cross-linked porous structure could be preserved across varying Pd/Ag ratios (SI Appendix, Fig. S2). The fused cross-linked pure Ag aerogel with a large diameter of about 42 nm was formed, signifying the occurrence of the fusion process in the case of Ag precursor exclusively. Simultaneously, an increase in nanowire diameters was noted with increasing concentrations of the Ag precursor, which could be ascribed to the structure-directing function of Ag during the nucleation and gelation of PdAg aerogels. Moreover, N₂ adsorption-desorption isotherms and the pore size distribution curves indicated a hierarchical structure with the coexistence of mesopores and macropores in aerogels (SI Appendix, Fig. S3). The Pd₆₇Ag₃₃ exhibited a large Brunauer–Emmett–Teller (BET) surface areas (89.8 m²/g) and pore volumes (0.285 cm³ g⁻¹), which is conducive to exposure of the active sites and mass transfer. Particularly, with increasing Ag content, the specific surface area decreased, ranging from 144.8 m²/g for the pure Pd aerogel to 29.6 m²/g for the pure Ag aerogel (SI Appendix, Table S1). This phenomenon further underscored the structure-directing function of Ag on the porosity of the PdAg alloy aerogel, matching well with the TEM results.

Fig. 1.

Fabrication and characterization of Pd₆₇Ag₃₃ aerogel. (A) Schematic illustration of the preparation of PdAg aerogel. (B) TEM images (Inset: photograph) of Pd₆₇Ag₃₃ aerogel. (C) XRD patterns of Pd₆₇Ag₃₃ aerogel, pure Ag, and Pd aerogel. (D) High-resolution Ag 3d XPS spectra of Pd₆₇Ag₃₃ and pure Ag. (E) High-resolution Pd 3d XPS spectra of Pd₆₇Ag₃₃ and pure Pd. (F) Surface valence band photoemission spectra of Pd₆₇Ag₃₃ and pure Pd.

The wide-angle powder X-ray diffraction (XRD) pattern (Fig. 1C) revealed a face-centered cubic (fcc) polycrystalline structure. The characteristic peak positioning between the corresponding peaks of pure Pd and pure Ag indicated the formation of bimetallic PdAg alloys (23). In contrast to pure Pd (prepared as a contrast catalyst), the reflections of PdAg alloy exhibited noticeable shifts toward lower 2θ values as the fraction of Ag increased (SI Appendix, Fig. S4), which suggested that doping Ag into Pd lattice could increase the lattice parameter and thus produce tensile strain. X-ray photoelectron spectroscopy (XPS) illustrated negative and positive shifts in Ag 3d doublet and Pd 3d doublet of Pd₆₇Ag₃₃, respectively, compared to the spectra of pristine Ag and Pd (Fig. 1 D and E) (24). Similar trends were observed in other PdAg alloys (SI Appendix, Fig. S5). These shifts indicated the electronic donation from Ag to Pd, which could be attributed to the higher working function of Pd (23). The increasing of Pd⁰(strain)/Pd⁰ area ratios further validated the abundant tensile strains in Pd₆₇Ag₃₃ (25), which is consistent with the XRD result. Furthermore, the surface valence band photoemission spectrum revealed a 0.28 eV downshift of the Pd d-band center in Pd₆₇Ag₃₃, which can be attributed to the increased electron filling of the Pd d-band resulting from the electron transfer from Ag to Pd (Fig. 1F). This downshift of the d-band center is expected to facilitate the desorption of carbonyl intermediates by weakening the binding strength on catalyst, which is essential to suppress the C–C bond cleavage and improve the selectivity toward C₂ products (12, 26).

Aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image demonstrated the existence of abundant amorphous areas, defects, and boundaries (Fig. 2A). In contrast, Pd exhibited regularly arranged lattice fringes, indicating a lower presence of defects (SI Appendix, Fig. S6). As illustrated in Fig. 2B, the lattice spacings of 0.233, 0.210, and 0.196 nm correspond to the (111), (220), and (200) planes of Pd₆₇Ag₃₃ alloy, respectively (27). The fast Fourier transformation (FFT) pattern (Inset in Fig. 2B) further revealed the fcc polycrystalline structure. Elemental mapping images exhibited the evenly distribution of Pd and Ag elements with an atom ratio of 67.9:32.1 along the cross-linked nanowires (Fig. 2C and SI Appendix, Fig. S7). Similar atom Pd/Ag ratio of Pd₆₇Ag₃₃ (i.e., 68.3/31.7 and 65.2/34.8) was characterized by the inductively coupled plasma optical emission spectroscopy (ICP-OES) and XPS result (SI Appendix, Tables S2 and S3). As shown in Fig. 2 D and E, the lattice spacing corresponding to (111) crystal plane of Pd gauged in HRTEM is 2.244 Å, while the value detected by Pd₆₇Ag₃₃ in the identical region is 2.328 Å, indicating the introduction of tensile strain in Pd₆₇Ag₃₃ by alloying Ag into Pd lattice. Consequently, an approximate 3.7% lattice expansion (3.7% tensile strain) in the (111) lattice plane of Pd₆₇Ag₃₃ was occurred compared with pure Pd (Fig. 2F). Furthermore, we measured the specific lattice strain distribution by geometric phase analysis (GPA) to identify the strain distribution of Pd₆₇Ag₃₃. The strain mapping was obtained from the abreaction-corrected HAADF-STEM results after FFT and inverse FFT, significant chromatic aberration changes indicated a tensile strain–dominated structure (Fig. 2 G–I).

Fig. 2.

Aberration-corrected HAADF-STEM images of Pd₆₇Ag₃₃ aerogel. (A) TEM image and (B) high-resolution TEM image and the corresponding FFT pattern of Pd₆₇Ag₃₃ aerogel. (C) HAADF-STEM and the corresponding elemental mapping images of Pd₆₇Ag₃₃ aerogel. (D and E) Inverse IFFT images of Pd and Pd₆₇Ag₃₃, the Insets correspond to the lattice spacing distribution. (F) Illustration for the lattice expansion of Pd and Pd₆₇Ag₃₃. (G–I) Maps of the in-plan strain tensors xx, yy, xy, processed via GPA (the color regions ranging from green to dark blue denote the compressive strain, while the regions from red to bright yellow represent the tensile strain).

Selective Electrochemical EG-to-GA Conversion Over Pd₆₇Ag₃₃.

Electrocatalytic activity and selectivity of the pure Pd and PdAg aerogels toward EGOR were evaluated in a standard three-electrode system. The linear sweep voltammetry (LSV) curves obtained from the pure Pd and Pd₆₇Ag₃₃ aerogel in a N₂-saturated 1 M KOH solution with and without 1 M EG at a scan rate of 10 mV s⁻¹. For comparison, a lower applied anodic potential of 0.53 V (vs. RHE) was required by Pd₆₇Ag₃₃ aerogel to reach 10 mA cm⁻² in the presence of 1 M EG (Fig. 3A). The significant reduction in overpotential (1,080 mV) suggested the thermodynamically favorable feature of EGOR compared to OER. Meanwhile, the comparation of LSV curves and Tafel slope plots between Pd₆₇Ag₃₃ and Pd indicated the boosted oxidation kinetics on Pd₆₇Ag₃₃, attributed to the fine modulation in electron and lattice structure (Fig. 3B). Cyclic voltammetry CV curves displayed a significantly lower onset potential of Pd₆₇Ag₃₃, implying the optimized adsorption and activation of EG molecules on Pd₆₇Ag₃₃ (Fig. 3C). Notably, Pd₆₇Ag₃₃ delivered exceptional mass activity of 9.7 A mg_Pd⁻¹, 3.2 and 12.4 times more than that of Pd aerogel catalyst (3.0 A mg_Pd⁻¹) and commercial 20 % Pd/C (0.8 A mg_Pd⁻¹) (SI Appendix, Figs. S8–S11). Furthermore, the comparison of the forward current density (I_f) and the backward current density (I_b) of catalysts illustrated higher I_f/I_b values of Pd₆₇Ag₃₃, suggesting the Ag incorporation allows superior poisoning tolerance regarding the intermediate carbonaceous species in EGOR (SI Appendix, Fig. S12) (28). Electrochemical impedance spectroscopy (EIS) displayed lower electron transfer resistance (R_ct) of Pd₆₇Ag₃₃, suggesting the enhanced electron transfer kinetics in catalytic reactions (SI Appendix, Fig. S13). The electrochemical active surface areas (ECSAs) of samples were evaluated by the double-layer capacitance (C_dl), which implied more exposed active sites on Pd₆₇Ag₃₃ (SI Appendix, Fig. S14) (29). These active sites potentially originated from the defect-rich and amorphous structure, which endows Pd₆₇Ag₃₃ with the enriched coordination unsaturated surface sites, thereby facilitating the engagement in performance of EGOR (30, 31). Moreover, the catalytic activities in EGOR over the Pd_xAg_y (y = 13 to 67%) alloy aerogels were further investigated (SI Appendix, Fig. S15). The mass activities exhibited a distinctive volcanic trend with increasing Ag content, initially ascending and subsequently descending. As mentioned above, the incorporation of Ag into Pd lattice structure triggered tensile strains and ligand effect, thereby promoting the adsorption of *OH species and the desorption of *CO species (12, 25). Thus, the enhanced EG adsorption and boosted intermediate conversion occurred on the surface of catalyst, which both contributed to heightened the catalytic activity. The CV curve of Ag aerogel indicated the inferior intrinsic catalytic activity of Ag for EGOR, with the mass activity of Ag observed to be two orders of magnitude lower than that of Pd (SI Appendix, Fig. S16). This result further revealed that the activity improvement of PdAg catalysts originated from the synergy of ligand and strain effects by Ag incorporation. However, as the Ag fraction reached an elevated level, the catalysts may lose active surface Pd sites along the nanochain structure, thus impeding reactant adsorption and conversion. Meanwhile, an increased Ag fraction enlarges the sizes of assembly unit of aerogels, resulting in reduced specific surface area that hampers mass transfer processes, leading to unfavorable kinetics for EGOR (32). Consequently, the Pd₆₇Ag₃₃ possessing a moderate fraction of Ag delivered an optimal activity.

Fig. 3.

Electrocatalytic performance of Pd₆₇Ag₃₃ aerogel toward anodic EGOR. (A) LSV curves, (B) Tafel slope plots, and (C) CV curves of Pd and Pd₆₇Ag₃₃ catalyst in 1 M KOH/1 M EG. (D) ¹H NMR spectra of electrolyte with different charges passed at the applied potential of 1.0 V. (E) Potential dependence of FE for GA production on Pd and Pd₆₇Ag₃₃. (F) Comparison of mass activity for various EGOR catalysts reported previously, where the hollow symbols denote as the nonselective catalysts, while the solid symbols denote as the catalysts for selective GA synthesis. (G) 80-h stability test of Pd₆₇Ag₃₃ toward GA production.

To identify the product selectivity of the EG-to-GA conversion, ¹H and ¹³C NMR spectroscopy was conducted. As shown in Fig. 3D, the evolution of peak of GA (~3.77 ppm, green shadow) at different charge passed demonstrated the dynamic EG-to-GA conversion on Pd₆₇Ag₃₃. Meanwhile, neither FA (~8.3 ppm, green shadow) nor other potential EGOR products could be found during electrolysis, indicating the circumvention of C–C bond cleavage and the selective GA production. ¹³C NMR spectrum also exhibited the same result, further verifying the selective synthesis of GA on Pd₆₇Ag₃₃ (SI Appendix, Fig. S17). The quantitative analysis of the Faradic efficiencies toward GA was carried out at different applied potentials (Fig. 3E and SI Appendix, Fig. S18). At the potential of 1.0 V vs. RHE, a maximum FE_GA of 92.7% was rendered over the Pd₆₇Ag₃₃, which is much higher than that of pure Pd and is of advantage among previous catalysts for EG-to-GA conversion. The concentration variation of EG and GA within 600 C charge implied the linear correlation between the consumption of EG and the production of GA during electrolysis (SI Appendix, Fig. S19). The FE_GA of as-synthesized Pd_xAg_y aerogels with different Pd/Ag ratios were also evaluated to investigate the influence of Ag addition on EG-to-GA conversion (SI Appendix, Fig. S20). Volcanic trends of FE_GA with increasing Ag content were illustrated, presenting the Pd₆₇Ag₃₃ as the most selective catalyst for GA synthesis (SI Appendix, Fig. S21). As previous works reported, although the introduction of Ag could downshift the d-band center of Pd to preserve the C–C bond, substantial Ag incorporated into Pd lattice triggered high tensile strain contributes to upshift the d-band center (33, 34). Therefore, highly strained Pd_xAg_y catalyst could enhance the surface binding strength toward *CO intermediate, thus facilitating the C–C bond cleavage and decreasing the GA selectivity (18, 25). This finding revealed the trade-off of strain effect in PdAg alloy catalysts to optimize the selective EG-to-GA conversion. To the best of our knowledge, the Pd₆₇Ag₃₃ represented the best active electrocatalysts for selective GA synthesis in EGOR to date (Fig. 3F and SI Appendix, Tables S5 and S6) (12–18, 20, 35–37), where the mass activity of 9.7 A mg_Pd⁻¹ is much higher than that of other catalysts (0.8 to 2.2 A mg_Pd⁻¹) for EG-to-GA conversion. Stability of catalysts was evaluated through a chronoamperometric method. The chronoamperometric curve of Pd₆₇Ag₃₃ displayed a dramatically improved durability of an 83.8 % retention of current density after 2 h measurement, which is much higher than that of Pd (24.4%) (SI Appendix, Figs. S22 and S23), aligning with the superior antipoisoning ability predicted by the higher I_f  /I_b values (SI Appendix, Fig. S12). During the 80-h-cyclic chronoamperometric tests, the initial current of Pd₆₇Ag₃₃ could be recovered as the electrolyte was refreshed (SI Appendix, Fig. S24). As illustrated in Fig. 3G, the average GA selectivity of 91.4% and GA productivity of 2.56 mmol cm⁻² h⁻¹ could be well preserved on Pd₆₇Ag₃₃, indicating excellent stability in the long-term electrocatalysis. XRD, XPS, and TEM measurements further verified the excellent durability of Pd₆₇Ag₃₃ catalyst with respect to structure and composition after long-term electrolysis (SI Appendix, Figs. S25–S27). To align PET upcycling with practical applications, our endeavor involved the valorization of real-word drink bottles via base catalysis depolymerization followed by electrocatalytic upcycling (SI Appendix, Fig. S28). First, terephthalate (TPA-K) and EG constituent monomers were obtained via the hydrolysis of PET bottles in 3 M KOH. Subsequently, EG monomer was utilized as a feedstock for further electrochemical valorization into glycollate. Following this, high-purity GA products can be obtained from the electrolyte by a series of separation and purification processes. XRD characterization, ¹³C and ¹H NMR spectroscopy demonstrated the high purity of collected GA and TPA powders (SI Appendix, Figs. S29 and S30). Finally, 10 g of PET feedstock gave 8.26 g TPA (yield: 95.5%), 2.28 g GA (yield: 57.4%), demonstrating the readiness of this technique for practical applications.

Insight of the Selective Adsorption and Conversion of Intermediates on Catalysts.

To give in-depth understanding of the structure–performance relationship for the EG-to-GA conversion over Pd₆₇Ag₃₃, in situ electrochemical measurements, radical probe tests, and quenching experiments were employed. First, we evaluated the EG adsorption on catalysts by the open-circuit potential (OCP) test. By adding 0.1 M EG into the electrooxidation system, the most marked decrease in OCP over Pd₆₇Ag₃₃ electrode illustrated the optimized EG adsorption (SI Appendix, Fig. S31). The adsorption behavior of ⁻OH species on catalysts was investigated by CV measurements in 1 M KOH electrolyte. As illustrated in Fig. 4A, stronger Pd oxidation peaks (0.9 to 1.2 V) and PdO reduction peaks (0.5 to 0.9 V) of Pd₆₇Ag₃₃ suggested the sufficient defects and boundaries supplying more available active sites to participate the EGOR (38). The observed negative shift of the ⁻OH adsorption peak indicated an optimized ⁻OH adsorption process on Pd₆₇Ag₃₃, suggesting the alloying strategy could enhance the affinity of ⁻OH on catalyst. Meanwhile, the CV curve of Pd₆₇Ag₃₃ displayed a negative shifted reduction peak of Pd along with higher intensity, signifying an increased binding energy of *OH adspecies (*OH_ad). To further identify the *OH_ad on catalysts, electron paramagnetic resonance (EPR) experiment was carried out with the presence of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as spin trapping agents. A strong typical four-line characteristic DMPO−^·OH signals with an intensity ratio of 1:2:2:1 was detected in the Pd₆₇Ag₃₃/KOH system, implying the accelerated ^·OH evolution upon Pd₆₇Ag₃₃ (Fig. 4B). As the injection of EG, new signal peaks identified as alkoxy radicals (DMPO−^·OR) were detected and the DMPO−^·OH signals were slightly weakened, implying that *OH_ad played a vital role in converting EG into ^·OR (Fig. 4C). Furthermore, the productivity of *OH_ad over catalysts were quantified by fluorescence probe experiments using coumarin as a fluorescence probe, where the higher fluorescence intensity in Pd₆₇Ag₃₃/KOH system suggested the superior *OH_ad productivity (Fig. 4D) (39). In comparison with the Pd₆₇Ag₃₃/KOH system, the Pd₆₇Ag₃₃/KOH/EG system displayed a decreased fluorescence intensity (SI Appendix, Fig. S32), which corroborates with the result that *OH serves as an active species for EGOR.

Fig. 4.

Insight of the selective adsorption and conversion of intermediates on catalysts. (A) CV curves of Pd and Pd₆₇Ag₃₃ aerogel in 1 M KOH. (B and C) EPR spectra from the electrolyte using DMPO as a trapping agent: (B) OER and (C) EGOR. (D) Fluorescence spectra for the detected ^·OH radicals in electrolyte using a coumarin (0.2 mM) indicator. (E and F) In situ electrochemical FTIR spectra under EGOR operation on Pd and Pd₆₇Ag₃₃.

We further employed Cu as a substitute for Ag to alloy with Pd to uncover the pivotal role of tensile strain in accelerating the kinetics for EGOR. The substitution of Cu in the Pd lattice structure induces compressive strain due to its smaller atomic radius than Pd (40, 41). Accordingly, we prepared a Pd₆₇Cu₃₃ alloy aerogel catalyst using the same salting-out and cross-linking method to investigate the structure–performance relationship. Both XRD pattern and HR-TEM image of Pd₆₇Cu₃₃ revealed a reduction in lattice spacing of the (111) crystal plane, confirming the successful fabrication of compressive strained Pd₆₇Cu₃₃ (SI Appendix, Fig. S33). XPS spectra depicted the electron interaction between Pd and Cu (SI Appendix, Fig. S34). Intriguingly, the CV curve of Pd₆₇Cu₃₃ displayed a weak ⁻OH adsorption peak at 0.51 V, more negative than that of Pd₆₇Ag₃₃ (0.49 V), suggesting that the tensile strained lattice structure could contribute to enhancing the affinity of ⁻OH on catalyst and allowing the oxidation of ⁻OH into *OH_ad at lower potential (SI Appendix, Fig. S35A). The EPR spectra and Fluorescence spectra also indicated the accelerated *OH_ad productivity over the tensile strained Pd₆₇Ag₃₃ (SI Appendix, Fig. S35 B and C). The promotional effect of the lattice tensile strain is reflected by the lower mass activity of Pd₆₇Cu₃₃ with a lower I_f/I_b value relative to those of Pd₆₇Ag₃₃, attributed to the significant role of *OH_ad in boosting the conversion of intermediates and circumventing the poisoning effect (SI Appendix, Fig. S35D) (42, 43). Furthermore, the much lower onset potential of Pd₆₇Ag₃₃ in comparison to that of Pd₆₇Cu₃₃ further supported the accelerated EGOR kinetics by tensile strain effect (SI Appendix, Fig. S35E). Our investigation unveiled that the compressive strain on Pd₆₇Cu₃₃ resulted in the weakened adsorption of ⁻OH and hindered the subsequent activation process. This observation is aligned with previous studies suggesting the reduction in the binding strength of oxygen-containing intermediates on compressive strained catalysts (44). In summary, these findings affirmed that the tensile strain effect induced by Ag doping could signally improve the adsorption/activation behavior of ⁻OH on the catalyst, thereby optimizing the kinetics for EGOR.

To gain in-depth understanding of reaction mechanisms at the molecule level during the EG-to-GA conversion, electrochemical in situ FTIR spectroscopy was measured. In situ FTIR spectra were recorded within the potential range of 0 to 1.2 V (vs. RHE) at an interval of 100 mV in 1.0 M KOH and 1.0 M EG. The peak at 1,660 cm⁻¹ belongs to the 2-hydroxyacetyl (*OC−CH₂OH) intermediates, which were regarded as a key intermediate in the EG-to-GA conversion (26). The peak at 1,236 cm⁻¹ is assigned to C–O stretch of GA (45, 46). The bonds at 1,411 cm⁻¹ are responsible for symmetric stretch of COO⁻ in GA. The peaks at 1,326 and 1,580 cm⁻¹ belong to the symmetric and antisymmetric stretching bands of COO⁻ of GA, and the distinct vibration peak at 1,076 cm⁻¹ can be ascribed to the stretching vibration of aldehyde (−CHO) from glyoxal and GA species (18, 47). The bands at 1,307 and 1,340 cm⁻¹ are attributed to oxalate and carbonate (CO₃²⁻) (46, 48). As depicted in Fig. 4 E and F, the bands at 1,076 and 1,580 cm⁻¹ appeared at 0.5 V vs. RHE on Pd₆₇Ag₃₃, lower than that of Pd (0.6 V vs. RHE), indicating that Pd₆₇Ag₃₃ exhibits a diminished energy barrier for the EG-to-GA conversion. Meanwhile, the comparison of in situ FTIR spectra between Pd and Pd₆₇Ag₃₃ revealed the significantly enhanced signals of GA (1,076, 1,236, 1,326, 1,411, and 1,580 cm⁻¹) on Pd₆₇Ag₃₃, while the intensity of bands associated with undesirable products (1,307 and 1,340 cm⁻¹) showed no noticeable change. This result further elucidated the improved GA selectivity on Pd₆₇Ag₃₃, aligning with the NMR results. In summary, the in situ FTIR spectra uncovered underlying reason why Pd₆₇Ag₃₃ could deliver excellent EGOR performance and GA selectivity, demonstrating that alloying Pd with other metals to modulate its electron and lattice structure is a promising strategy for enhancing the EG-to-GA conversion.

DFT Calculations.

To elucidate the underlying factors responsible for the enhanced activity and GA selectivity originating from the ligand and strain effects triggered by Ag incorporation, DFT calculations were implemented. Specifically, we evaluated the adsorption energies of active species and intermediates on the model surfaces of Pd (111) and Pd₆₇Ag₃₃ (111) with 3.7% tensile strain (referred to as Pd₆₇Ag₃₃) (SI Appendix, Figs. S36–S40). Initially, the projected density of states (PDOS) analysis for the Pd d-band was calculated, revealing d-band center values of −1.65 eV for Pd (111) and −1.96 eV for Pd₆₇Ag₃₃ (111) (Fig. 5A), which were in good agreement with experimental measurements. The summarized adsorption energies in Fig. 5B demonstrate the weakened adsorption of *OCCH₂OH intermediates and CO on Pd₆₇Ag₃₃, which is in accordance with the previous studies that lowering the d-band center can weaken the adsorption of carbonyl species (12). Therefore, we confirmed that the facilitated release of *CO and *OCCH₂OH are primarily attributed to the optimized electron structure (ligand effect), where the incorporation of Ag could effectively suppress excessive oxidation and poisoning allowing the high GA selectivity and stable long-term performance. Meanwhile, Fig. 5B illustrated the stronger adsorption of ⁻OH and EG on Pd₆₇Ag₃₃, which also aligns with the experimental observations. To further investigate the contribution of the strain effect, we discussed the adsorption of these key molecules on the model surface of Pd₆₇Ag₃₃ without tensile strain (referred to as Pd₆₇Ag₃₃-0%) (SI Appendix, Fig. S41). Notably, tensile strain plays a crucial role in enhancing the adsorption of EG and ⁻OH, thus boosting the activity for EGOR. Furthermore, we investigated the free energy diagrams of the involved intermediates during the selective EG-to-GA conversion on Pd (111) and Pd₆₇Ag₃₃ (111) (SI Appendix, Figs. S42 and S43). Fig. 5C demonstrated a lower energy barrier (0.55 eV) of H-abstraction on Pd₆₇Ag₃₃. Moreover, the *OC−CH₂OH to *COOH−CH₂OH conversion, regarded as a rate-determining step (RDS) in previous works, is facilitated by a reduced energy barrier on Pd₆₇Ag₃₃ (0.41 eV), indicating the rapid conversion toward the target GA product.

Fig. 5.

DFT calculations. (A) PDOS (d-band center) of Pd and Pd₆₇Ag₃₃; the Inset shows the optimized configurations of Pd and Pd₆₇Ag₃₃. (B) Calculated adsorption energy of *EG, *OH, *CO, and *OCCH₂OH on Pd (111) and Pd₆₇Ag₃₃ (111); the Inset shows the most energetically favorable adsorption configurations of molecules on Pd₆₇Ag₃₃. (C) Free-energy profile of the selective EGOR on Pd₆₇Ag₃₃ (111) and Pd (111). (D) Scheme of the proposed EGOR mechanism on Pd and Pd₆₇Ag₃₃.

Based on the above experimental investigations and theoretical calculations, we proposed a mechanistic understanding of the electrochemical EG-to-GA conversion over Pd₆₇Ag₃₃ (Fig. 5D). First, the electron interaction between Pd and Ag atoms, known as ligand effect, decreased the d-band center of Pd, which weaken the binding strength of carbonyl species (e.g., *CO and *OC−CH₂OH) on catalyst. This feature effectively suppressed excessive oxidation and mitigated CO poisoning, thereby ensuring the C–C bond preserving pathway and superior long-term stability over the Pd₆₇Ag₃₃. Second, doping the Ag with larger atomic radius into Pd lattice introduced a 3.7% tensile strain in Pd₆₇Ag₃₃, which plays a vital role in strengthening the affinity toward reactants (e.g., EG and ⁻OH). Thus, Pd₆₇Ag₃₃ delivered the enhance catalytic activity for EGOR. As a contrast, the insufficient affinity for reactants and restrained desorption of carbonyl species on Pd should be responsible for the inferior activity and unsuppressed C₁ products pathway. In summary, these findings collectively demonstrated that the synergy of the ligand effect and strain effect in PdAg alloy for the highly active and selective GA synthesis from PET wastes, which could provide guidance for the design of future catalysts for effective PET upcycling.

Covalorization of PET-Derived EG and CO₂ in an Integrated Electrolyzer.

Electrochemical coupling of waste plastics valorization with CO₂ conversion provides a promising approach to upcycle waste carbon resources into valuable chemicals on both sides of the electrolyzer. Prior to the covalorization of PET and CO₂, we evaluated the electrochemical CO₂RR performance over Pd₆₇Ag₃₃ (SI Appendix, Figs. S44–S47), which delivered high current density of 10.8 mA cm⁻² at −0.35 V and a maximum FE of 92.5% for FA production at −0.27 V, consistent with the performance of previously reported Pd-based alloy electrocatalysts (23, 24, 49). The superior CO₂RR performance could be contributed to the ligand effect in Pd₆₇Ag₃₃, where electron interaction between Pd and Ag lowered the d-band center of Pd and thereby significantly enhanced its FA selectivity and CO tolerance. Consequently, the exceptional bifunctional properties of the Pd₆₇Ag₃₃ catalyst in electrochemical EGOR and CO₂RR hold great promise in developing a two-electrode integrated system. To this end, a two-electrode H-type electrolyzer was assembled by employing Pd₆₇Ag₃₃ catalysts on carbon paper as the anode and cathode. The anodic chamber contained 1 M KOH/1 M EG with N₂ gas flow, while the cathodic chamber was filled with 0.1 M KHCO₃ solution bubbled with CO₂, and two chambers were separated by a bipolar membrane (BPM) (SI Appendix, Fig. S48). Notably, this system first enables simultaneous anodic glycolate synthesis from PET-derived EG and cathodic CO₂-to-FA conversion, yielding attractive system-level performance. Specifically, the inclusion of EG leads to a significant reduction in overpotential of approximately 1,000 mV, at a current density of 2 mA cm⁻², indicative of the pronounced enhancement in cell performance achieved through the replacement of OER at the anode with EGOR (SI Appendix, Fig. S49). The chronoamperometric curve of the EGOR//CO₂RR integrated electrolyzer during electrolysis illustrated that Pd₆₇Ag₃₃ catalyst possesses good catalytic stability (SI Appendix, Fig. S50). In contrast to the EGOR//CO₂RR cell, bubble formation was observed on the surfaces of both electrodes in the OER//CO₂RR cell, with the cathodic bubbles primarily associated with the intensified side reaction of hydrogen evolution reaction (HER) under higher applied cell voltage (SI Appendix, Fig. S51). This finding suggests that a lowered applied potential can be achieved in the integrated cell as the substitution of OER with EGOR, which contributes to energy conservation and mitigates the undesired anodic HER.

¹H NMR spectrum of the electrolytes in each chamber revealed the product distribution, where GA and FA were identified as the primary products in the anodic and cathodic compartments, respectively (SI Appendix, Fig. S52). Quantitative analysis demonstrated that a maximum overall FE of approximately 163.5% for ideal carbonaceous products was achieved at a low cell voltage of 0.4 V, encompassing anodic FE_GA of 81.7% and cathodic FE_FA of 81.8% (SI Appendix, Fig. S53). This study provides compelling evidence for the successful production of separated FA and GA through the coelectrolysis of CO₂ and PET-derived EG over the bifunctional Pd₆₇Ag₃₃ catalyst. In comparison to previous studies that combined EGOR with cathodic HER (50, 51), the adoption of CO₂RR on cathode not only enables the utilization of carbonaceous resources for high-value chemicals production but also contributes to achieving carbon neutrality. However, it should be noted that the faradaic efficiency of cathodic FA production decreases as the cell voltage increases, primarily due to the intensified HER side reaction. To address this, future advancements in the rational design of cathodic CO₂RR catalysts to align with the working potential range of the anodic EGOR are expected to facilitate highly efficient co-upcycling of CO₂ and PET waste, opening a sustainable and economic path for the generation of value-added chemicals.

In conclusion, we designed a tensile strained Pd₆₇Ag₃₃ alloy aerogel as a highly active and selective catalyst for GA synthesis from waste PET plastics. The experiments and calculations demonstrated that the synergistic of tensile strained effect and the ligand effect of Ag doping affected the adsorption behaviors of reaction intermediates, thus contributing to the construction of the best active catalyst for EG-to-GA conversion observed thus far. Meanwhile, both the mass activity and GA selectivity of different Pd_xAg_y catalysts displayed a volcanic trend along with the increasing of Ag fraction. This finding underscored the necessity, in the design of alloy catalysts for EG-to-GA conversion, to consider not only the ligand effect but also the strain effect, size effect, and mass transfer. The modulation of electron structure of Pd sites by Ag doping also allowed the superior cathodic CO₂RR performance on Pd₆₇Ag₃₃, which endow its promising as bifunctional catalyst for the covalorization of PET plastics and CO₂. Therefore, we first assembled an integrated electrolyzer for reforming PET and CO₂ into GA and FA, respectively, which demonstrates >80% FE on both electrodes for the desirable carbonaceous molecule at a low cell voltage of 0.4 V. Our findings not only provide insights into the synergistic of train effect and ligand effect in the enhancement of performance for EG-to-GA conversion over alloy catalyst, but also contribute to the construction of advanced covalorization system that achieves economic and sustainable utilization of carbon-rich wastes into high-value chemicals.

Materials and Methods

Chemicals.

Palladium chloride (PdCl₂) and silver nitrate (AgNO₃) were purchased from Aldrich Corp. Trisodium citrate dihydrate (NaCA) and potassium hydroxide (KOH) were purchased from Shanghai Aladdin Chemical Reagent Co. Ltd. D₂O (for NMR, 99.9%) was purchased from Innochem Technology Co., Ltd. (Shanghai, China). All other reagents were purchased from Guoyao Chemical Company. All chemicals were of analytical grade and used without further purification. Deionized water (18 MΩ cm) was applied for all experiments.

Pretreatment of Plastics.

10 g of PET derived from waste plastic bottle was soaked in 100 mL 3 M KOH aqueous solution at 80 °C for 48 h with continuous stirring in a flask. The constituent unit (terephthalic acid, TPA) was separated from the solution hydrothermally treated via an acidification treatment using concentrated H₂SO₄. Residual EG was concentrated, purified, and solved in 1 M KOH for electrochemical upcycling.

Synthesis of Pd_xAg_y Aerogel.

Typically, the Pd₆₇Ag₃₃ aerogel was synthesized by a two-step strategy. First, 0.06 mmol PdCl₂, 0.12 mmol HCl, and 0.6 mmol NaCA were added into 300 mL of deionized water with stirring to form solution A. Next, 0.03 mmol AgNO₃ and 0.3 mmol NaCA were dissolved in 150 mL of deionized water to form solution B. Then, 1.2 and 0.6 mL freshly prepared NaBH₄ aqueous solution (200 mM) was rapidly injected into the solution A and solution B to form ligands stabilized Pd colloidal solution and Ag colloidal solution. After stirring for 90 s, the two colloidal solutions were uniformly mixed together. Subsequently, 4.5 mL aqueous solution containing 30 mmol NH₄F was rapidly injected into the as-obtained mix solution. After continuous stirring for 180 s, Pd₆₇Ag₃₃ hydrogel was generated at the bottom of the glassware. To remove possible residues, the hydrogel was carefully washed with water for three times in 2 d. Afterward, the resulting hydrogel was flash-frozen by liquid nitrogen and dried by a freeze dryer. As such, other Pd_xAg_y aerogels with various mole ratios were prepared by varying the mole ratios of Pd and Ag sources in the precursor solutions.

Synthesis of Pd₆₇Cu₃₃ Aerogel.

Pd₆₇Cu₃₃ aerogel was synthesized using the similar two-step strategy by simply changing the species of metal precursors.

Material Characterization.

TEM, high-resolution TEM (HRTEM) images, and elemental mapping analysis were performed on a Talos F200S. The aberration-corrected HAADF-STEM (AC- HAADF-STEM) and STEM-EDS were conducted on a Thermofisher Titan Themis Z. XRD patterns were conducted on a D8 Advance instrument (AXS-Bruker) with Cu Kα radiation. XPS measurement were done on an ESCALAB 250Xi. Elements and valence information were analyzed by the XPS peak software using setting to 20% Lorentzian and 80% Gaussian. The binding energy (B.E.) of all XPS data was calibrated vs. the standard sp² C 1 s peak at 284.6 eV. The d-band centers were calculated by following Eq. 1 (26). The surface morphologies were conducted by nitrogen sorption isotherms at 77 K on a Autosorb iQ instrument (Quantachrome), and the specific surface areas were calculated via a Brounauer–Emmett–Teller (BET) method. ¹H NMR spectroscopy was measured on a 400 MHz/AVANCE NEO (Bruker). The *OH adspecies were detected by EPR (Bruker, Germany) with DMPO as capture agents.

$$\begin{array}{c}{\int }_{-10\text{eV}}^{0\text{eV}\left({\mathrm{E}}_{\mathrm{f}}\right)}\left(\text{binding} \text{energy} \left(\mathrm{E}\right)\times \text{intensity} \left(\mathrm{E}\right) \text{dE}\right)/{\int }_{-10 \text{eV}}^{0 \text{eV}\left({\mathrm{E}}_{\mathrm{f}}\right)}\text{intensity} \left(\mathrm{E}\right) \text{dE}.\hfill \end{array}$$ [1]

Working Electrode Preparation.

The catalyst ink was prepared by ultrasonicating the suspension containing 1 mg of as-prepared catalyst, 400 isopropanol (IPA), 100 deionized water, and 2 µL of Nafion (5 wt.%) for 1 h. Then catalyst ink was dropped onto a L-type glassy-carbon electrode (GCE, 4 mm, ~0.1256 cm²) and dried under ambient air. For the EGOR and CO₂RR half-reaction tests, the catalyst loading was 0.1 mg cm⁻² and 0.5 mg cm^–2, respectively. The working electrode for mass activity evaluation was prepared by using loading 2 μg catalyst onto GCE. For the long-term stability and H-cell tests, carbon paper (CP, 0.25 cm²) with a catalyst loading of 0.5 mg cm^–2 served as the electrode.

Electrochemical In Situ FTIR Reflection Spectroscopy.

To investigate the mechanism of EG-to-GA conversion at molecule level, in situ FTIR spectroscopy was collected by a Thermo 8700 spectrometer equipped with a liquid nitrogen-cooled mercury cadmium telluride detector (MCT-A). Electrochemical in situ FTIR spectra were carried out by utilizing a home-made thin-layer infrared (IR) cell with a three-electrode configuration, in which the electrolyte consists of 1 M potassium hydroxide (KOH) and 1 M ethylene glycol (EG). The as-synthesized electrocatalyst was used as the working electrode, while an Hg/HgO (1.0 M KOH) electrode and a platinum wire served as the reference and counterelectrodes, respectively. The FTIR spectra were collected over a potential range of 0 to 1.1 V (vs. RHE) at intervals of 0.1 V. During test, an incident IR beam was transmitted through a CaF₂ optical window and a thin-layer solution prior to being reflected off the electrode surface. The resulting spectra were recorded as the relative change in reflectivity and were calculated using the following equation:

$$\Delta \mathrm{R}/\mathrm{R}=(\mathrm{R}({\mathrm{E}}_{\mathrm{S}})-\mathrm{R}({\mathrm{E}}_{\mathrm{R}}))/\mathrm{R}({\mathrm{E}}_{\mathrm{R}}),$$ [2]

where R(E_S) and R(E_R) are the single-beam spectra obtained at the catalyst potential (E_S) and reference potential (E_R), respectively.

Electrochemical Measurements.

Electrochemical EGOR measurements were equipped with a standard three-electrode single cell at room temperature with a working electrode, a platinum counterelectrode (1 cm²), and an Hg/HgO (1.0 M KOH) reference electrode. An electrolyte solution containing 1 M EG and 1 M KOH was used for all EGOR tests. As for the electrochemical measurements, the electrolyte was bubbled with N₂ (99.99%) for ~30 min to ensure N₂ saturation during tests. Chronoamperometric curves were collected at the potentials between 0.6 and 1.1 V. LSV curves were recorded at 10 mV/s unless special situation. EIS results were performed over a frequency range from 0.01 Hz to 100 kHz with a perturbation of 5 mV. A custom-made three-electrode system was used to investigate the electrochemical CO₂RR in H-cell, where a platinum electrode (1 cm²) and Ag/AgCl electrode (4.0 M KCl) served as the counterelectrode and reference electrode. The two chambers contained 15 mL of 0.1 M KHCO₃ aqueous solution (pH = 8.2) and were separated by a cation exchange membrane (Nafion 117). Before tests, the cathodic electrolyte was purged with CO₂ for 30 min at a constant rate of 20 sccm to maintain its saturation (pH = 6.8) during measurements. Chronoamperometric curves were collected at the potentials between −0.35 and 0 V (vs. RHE).

All electrochemical measurements were connected to an electrochemical workstation (VMP3, Bio-logic, France). All potentials were calibrated to the RHE based on the following equation:

$$\begin{array}{c}{\mathrm{E}}_{\text{RHE}}={\mathrm{\text{E}}}_{\text{Hg}/\text{HgO}} + 0.098 + 0.059 \text{pH},\hfill \end{array}$$ [3]

$$\begin{array}{c}{\mathrm{E}}_{\text{RHE}}={\mathrm{\text{E}}}_{\text{Ag}/\text{AgCI}} + 0.2046 + 0.059 \text{pH}.\hfill \end{array}$$ [4]

For the EGOR//CO₂RR two-electrode system, the electrochemical measurements were carried out in an H-type cell with a BPM to separate the two chambers.

Product Analysis.

The NMR samples contained 500 µL of electrolyte, 100 mL of D₂O, and 100 μL of dimethyl sulfoxide (DMSO) water solution (1% Vol.). The presaturation method was used to suppress the water peak.

The FE (%) of the GA formation can be determined by the following Eq. 5, n is the number of exchanged electrons to produce the final products (i.e., GA and FA):

$$\begin{array}{c}\text{FE}(\%) = \frac{\mathrm{n} \times \text{obtained} \text{product} \left(\text{mol}\right) \times \mathrm{96,485} \left(\mathrm{C}/\text{mol}\right)}{\text{Total} \text{charge} \text{passed} \left(\mathrm{C}\right)} \times 100\%,\hfill \end{array}$$ [5]

where n is the number of electron transfer for each product formation, n = 2 for FA, n = 4 for GA; 96,485 C mol⁻¹ is the Faraday constant.

Computational Methods.

First-principles calculations based on spin-polarization DFT have been conducted using the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) formulation (52–54). The ionic cores have been described using projected augmented wave (PAW) potentials (55, 56), while the valence electrons have been treated within a plane wave basis set with a kinetic energy cutoff of 450 eV. Van der Waals interactions have been accounted for using the DFT-D3 method proposed by Grimme (57, 58). The electronic energy convergence criterion was set at 10⁻⁵ eV, indicating self-consistency. Geometry optimization was considered converged when the energy change was below 0.02 eV Å⁻¹. A 2 × 2 × 1 gamma-centered grid was employed for Brillouin zone sampling during relaxation, and a 15 Å vacuum layer was added to eliminate spurious interactions between periodic images. Spin-polarized calculations were performed to obtain the desired results. The adsorption energy (E_ads) is calculated as

$${\mathrm{E}}_{\text{ads}}={\mathrm{E}}_{\text{total}}-{\mathrm{E}}_{\text{slab}}-{\mathrm{E}}_{\text{adsorbate}},$$ [6]

where the E_total is the total energy of an optimized slab with the adsorbate on it, E_slab is the energy of a relaxed clean slab, and E_adsorbate is the energy of an adsorbate molecule.

The Gibbs free energy diagram (ΔG) of the EGOR was determined using the computational hydrogen electrode (CHE) model proposed (59). In this model, the free energy of the proton-electron pair was equated to that of 1/2 H₂(g). The free energy change for each step was computed by

$$\Delta \mathrm{G}=\Delta \mathrm{E}+\Delta \text{ZPE}-\Delta \text{TS},$$ [7]

where ΔE was the energy difference of the reactants and the products directly obtained from DFT calculation. ΔZPE is the contribution of variation of zero-point energy (ZPE), ΔS is the entropy (S) change, T is the temperature (T = 298.15 K).

It is imperative to underscore that while we employed the most commonly proposed mechanism along as well as well-defined and widely accepted models, constructing computational models becomes notably challenging in multimetallic alloys with complex contents. Meanwhile, the inclusion of additional factors into the model may introduce potential influences on the computational results (60).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.