English

• With energy crisis increasing and pollution spreading, the need for renewable energy technologies is growing [1]. Among them, direct ethanol fuel cells have garnered significant attention due to their high theoretical energy density, low toxicity, and easy accessibility from biomass [2,3]. Previous studies have demonstrated that Pd-based materials are widely used catalysts for ethanol oxidation reaction (EOR), which is the anodic reaction of direct ethanol fuel cells [4–7]. However, during the EOR process the multiple electron pathways lead to sluggish kinetics and the formation of intermediate by-products such as CO easily deactivates the Pd catalyst. In addition, the high cost of Pd limits their large-scale application in EOR. Therefore, Pd-based electrocatalysts with outstanding catalytic ability are still urgently needed.

  In order to modulate the configurations of Pd sites, introducing p-block elements into Pd is an effective strategy. The higher energy of p-orbitals compared to d-orbitals suggests that p-d hybridization may be a more effective strategy for modulation than d-d hybridization by alloying with d-block metals in terms of orbital energy [8–12]. At present, the common strategies to introduce p-d hybridization include constructing Pd-based alloys with p-block metals or metalloids, and doping p-block elements into metallic Pd [13–15]. However, the disordered arrangement of Pd and p-block metals in these catalysts makes the p-d hybridization distribute unevenly, limiting the fine modulation of electronic structure of the electrocatalyst. In contrast, Pd-based intermetallics containing p-block metals or metalloids exhibit an ordered atomic configuration and specific atomic bonding between Pd and p-block metals or metalloids [16–18]. But the influence of p-d hybridization in Pd-based intermetallics on the catalytic performance are neglected [17,19]. In addition, the generally used thermal annealing method leads to aggregation and destroys the morphology of nanocatalysts in many cases, decreasing the exposed active sites and reducing the catalytic performance [20–22]. Therefore, preparing Pd-based intermetallics containing p-block metals through a simple wet-chemical strategy and investigating the effect of p-d hybridization are highly desired.

  In this work, we successfully synthesized intermetallic Pd₈Sb₃ nanodisks (NDs) with an average diameter of 19.5 nm and a thickness of only 3.2 nm through a straightforward one-step hydrothermal process. The ordered atomic arrangement within the intermetallic Pd₈Sb₃ NDs facilitate the strong p-d orbital hybridization of Pd and Sb atoms. The Pd₈Sb₃ NDs deliver an excellent EOR performance with a mass activity (MA) of 4.6 A/mg, approximately 6 times as high as that of commercial Pd/C. Furthermore, the unconventional p-d orbital hybridization of Pd and Sb atoms has been successfully established according to the density functional theory (DFT) calculations. This unique electronic interaction significantly enhances the catalytic performance by optimizing the adsorption energy of reaction intermediates and lowering the energy barriers for key reaction steps. Our findings elucidate the relationship between electronic interaction modulation and EOR activity in Pd-based intermetallics, and provide valuable guidance for designing highly efficient catalysts.

  The intermetallic Pd₈Sb₃ nanodomains (NDs) are fabricated through a hydrothermal synthesis approach. In this method, palladium(Ⅱ) acetylacetonate (Pd(acac)₂) and antimony trichloride (SbCl₃) served as the primary metal precursors. Poly(vinylpyrrolidone) (PVP) was employed as a stabilizing surfactant, while ethylene glycol acted as the reaction solvent. Further details of the synthesis procedure are provided in Supporting information. The X-ray diffraction (XRD) pattern (Fig. 1a) reveals the highly crystalline of the products, exhibiting an intermetallic Pd₈Sb₃ phase (JCPDS No. 04–007–0905) with regular atomic arrangement of Pd and Sb (inset in Fig. 1a), which belongs to the trigonal R3¯c space group. Transmission electron microscope (TEM) image demonstrates that uniform Pd₈Sb₃ NDs are obtained (Fig. 1b). Statistical analysis reveals that the average diameter of Pd₈Sb₃ NDs is approximately 19.5 nm (Fig. 1c). Atomic force microscopy (AFM) analysis indicates that the Pd₈Sb₃ NDs have a thickness of approximately 3.2 nm (Fig. S1 in Supporting information), which is smaller than their diameter, indicating the disk-like morphology of the obtained samples. Moreover, the thickness at the center of the NDs is greater than that at the edge. The lattice spacing of 3.8 Å observed in the high-resolution TEM (HRTEM) image is characteristic of the (300) plane of Pd₈Sb₃, thereby verifying the successful synthesis of intermetallic Pd₈Sb₃ phase (Fig. 1d). Elemental mapping images obtained via high-angle annular dark-field scanning transmission electron microscope energy-dispersive X-ray spectroscopy (HAADF-STEM-EDS) reveal uniform distribution of Pd and Sb elements across the NDs (Fig. 1e), consistent with the EDS line scanning profile (Fig. S2 in Supporting information)). The atomic ratio of Pd to Sb in the NDs is approximately 71.1:28.9 from HAADF-STEM EDS (Fig. 1f). This ratio aligns closely with the ratio obtained from inductively coupled plasma optical emission spectrometry (ICP-OES), which measured a Pd:Sb ratio of 72.3:27.7.

  Figure 1

  

  Figure 1.  (a) XRD pattern of Pd₈Sb₃ NDs. The inset illustrates the corresponding crystal structure model. (b) TEM image and (c) diameter distribution of Pd₈Sb₃ NDs. (d) HRTEM image, (e) EDS mapping and (f) EDS spectrum of Pd₈Sb₃ NDs.

  DownLoad: Full-Size Img PowerPoint

  To investigate the atomic structure of Pd₈Sb₃ NDs, aberration-corrected HAADF-STEM characterization is conducted. The ordered arrangement of atoms is clearly visible (Fig. 2a). The corresponding fast Fourier transform (FFT) pattern of Fig. 2a is displayed in Fig. 2b, in which the diffraction pattern is consistent with the simulated FFT pattern of trigonal Pd₈Sb₃ along the [001] zone axis (Fig. 2c). The (110) superlattice spot observed in Fig. 2b further demonstrates the formation of the trigonal Pd₈Sb₃ phase. Furthermore, the magnified views of regions d and e in Fig. 2a (Figs. 2d and e) reveal atomic columns with distinct contrasts. The atomic structure is consistent with the trigonal Pd₈Sb₃ phase, belonging to the R3¯c space group. The Pd and Sb atoms are located in different atomic columns, as shown in the HAADF-STEM image with atomic-resolution elemental mappings (Fig. 2f), line scanning profile (Fig. S3 in Supporting information) and the atomic models of Pd₈Sb₃ (Figs. 2g and h). Therefore, the contrast variation cannot be only attributed to the atomic number (Z) of Pd (Z = 46) and Sb (Z = 51). The variation in the number of Pd and Sb atoms within distinct atomic columns also accounts for the contrast differences observed in Figs. 2d and e.

  Figure 2

  

  Figure 2.  (a) HAADF-STEM image of a single Pd₈Sb₃ ND. (b) The corresponding FFT pattern of (a). (c) The simulated FFT pattern of Pd₈Sb₃ intermetallic along [001] direction. (d, e) The high-resolution HAADF-STEM image taken from areas marked in (a). (f) Atomic-resolution elemental mappings of a Pd₈Sb₃ ND. Crystal structure model viewed along the (g) [001] and (h) [100] directions. Red and green spheres denote Pd and Sb atoms, respectively.

  DownLoad: Full-Size Img PowerPoint

  To uncover the formation mechanism of Pd₈Sb₃ NDs, a series of products obtained at various reaction time are analyzed using XRD and TEM. As shown in the XRD patterns of the products (Fig. S4 in Supporting information), when the reaction time is 15 min, the products contain Pd₈Sb₃ intermetallic compound and face-centered cubic (fcc) Pd. This result indicates that Pd is reduced first at the initial stage. As the reaction time increases, more Sb is reduced to form the Pd₈Sb₃ intermetallic compound. As a result, the peak of Pd at 40.1° gradually disappears while the signal of Pd₈Sb₃ intermetallic compound becomes more pronounced. Additionally, an increase in reaction time from 15 min to 120 min results in an enlargement of the product particle size from 10.6 nm to 19.6 nm (Fig. S5 in Supporting information). Apart from reaction time, reaction temperature also plays a crucial role in the synthesis of intermetallic compound as a high temperature could provide sufficient energy to overcome the atomic diffusion energy barrier to induce ordered atomic arrangement. Consequently, only a Pd–Sb solid solution is obtained when the reaction temperatures are 120 ℃ or 140 ℃ (Fig. S6 in Supporting information). However, Pd₈Sb₃ intermetallic compound forms when the reaction temperature exceeds 160 ℃. These control experiments demonstrate that appropriate reaction time and temperature are essential for the synthesis of pure Pd₈Sb₃ intermetallic NDs. To unravel the chemical states of elements in the synthesized Pd₈Sb₃ NDs, X-ray photoelectron spectroscopy (XPS) characterization is conducted. Prior to the XPS analysis, Pd₈Sb₃ NDs are loaded on commercial carbon (named Pd₈Sb₃ NDs/C) to compare the binding energy with that of commercial Pd/C. Pd and Sb are present in the Pd₈Sb₃ NDs (Fig. S7a in Supporting information). The Pd 3d peaks in Fig. S7b (Supporting information) indicate that Pd is primarily in the metallic state. Notably, the Pd⁰ peaks of Pd₈Sb₃ NDs/C shift negatively by 0.59 eV compared to those of Pd/C (Fig. S7c in Supporting information), indicating the electron donation from Sb to Pd atoms due to the lower electronegativity of Sb (2.05) than that of Pd (2.20) [23,24]. Compared to Pd/C, Pd₈Sb₃ has a lower proportion of oxidized Pd, attributing to the alloying effect in Pd₈Sb₃ intermetallic compound that inhibits the oxidation of Pd. In addition, the Sb 3d peaks in Fig. S7d (Supporting information) show that Sb also mainly exists in the metallic state. The O 1s peak in the Sb 3d XPS spectrum can be deconvoluted into three characteristic peaks (i.e., O_Ⅰ at 530.7 eV, O_Ⅱ at 532.1 eV, and O_Ⅲ at 533.8 eV). These peaks are attributed to lattice oxygen in metal oxides, defect-oxide/hydroxyl-like groups, and oxide in adsorbed or residual water molecules, respectively [25].

  The electronic structures of Pd and Sb are further investigated using the X-ray absorption near-edge fine structure (XANES) measurements. Fig. 3a shows the K-edge XANES spectra of Pd₈Sb₃ NDs with Pd foil and PdO as references. The adsorption edge of Pd₈Sb₃ NDs shifts negatively compared to Pd foil and PdO, suggesting an increase in electron density at the Pd sites. Conversely, the white line intensity of the Sb K-edge in Fig. 3b is higher than that of Sb foil, indicating a decrease in electron density around Sb sites. Meanwhile, the adsorption threshold of Pd₈Sb₃ NDs is higher than those of Sb foil and Sb₂O₃ but lower than that of Sb₂O₅, illustrating that the valence state of Sb is between Sb³⁺ and Sb⁵⁺. The above results suggest the electron transfer from Sb to Pd because of the p-d orbital hybridization in Pd₈Sb₃ NDs. The coordination environment of Pd and Sb in Pd₈Sb₃ NDs is elucidated through Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) spectra. In Fig. 3c, a distinct signal centered at ≈ 2.74 Å in the FT-EXAFS spectrum of Pd K-edge in Pd₈Sb₃ NDs is observed, which is near the position of Pd–Pd bond in Pd foil. No signal of Pd–O bond exists in Pd₈Sb₃ NDs, in contrast to the obvious peak of Pd–O bond in PdO. The FT-EXAFS spectrum of Pd K-edge in Pd₈Sb₃ NDs could be fitted by the Pd–Pd path (≈ 2.70 Å) and Pd–Sb path (≈ 2.49 Å) as shown in Fig. 3d, with the fitted data provided in Table S1. Notably, the small peak at ≈ 1.90 Å could be assigned to the Pd–Sb bond and Pd–Pd bond as well. It should be noted that the presence of multiple Pd–Sb and Pd–Pd shells is expected to lead to the spectral cancellation due to destructive interference of the respective EXAFS oscillations [26]. Therefore, the coordination number of Pd is underestimated. However, the trends in relative coordination numbers and the bonding of atoms remain consistent. From the FT-EXAFS spectrum of Pd K-edge and corresponding fitting result, we can conclude that Pd–Sb bond exists in Pd₈Sb₃ NDs. In the Sb K-edge FT-EXAFS spectra of Pd₈Sb₃ NDs (Fig. 3e), a peak located at 2.68 Å is attributed to the Sb–Pd bond, which is distinct from the Sb–Sb bond in Sb foil (2.91 Å), and the Sb–O bond in Sb₂O₃ (1.50 Å) and Sb₂O₅ (1.46 Å). The least-squares EXAFS fitting results are displayed in Fig. 3f and Table S2 (Supporting information). The Pd K-edge wavelet transform EXAFS (WT-EXAFS) spectra of Pd₈Sb₃ NDs differ from those of Pd foil and PdO, indicating a different coordination environment (Fig. 3g). In the WT-EXAFS of Pd K-edge in Pd₈Sb₃ NDs, the strong signal at k = 5.0 Å^–1 could be assigned to the Pd–Pd and Pd–Sb bonds, based on the comparisons with the Pd–Pd and Pd–O–Pd signals in Pd foil and PdO. In the WT-EXAFS spectra of Sb K-edge in Pd₈Sb₃ NDs (Fig. S8a in Supporting information), only one strong signal observed at 7.6 Å^–1 is attributed to Pd–Sb bond. The position shifts negatively compared to the signal of Sb–Sb bond in Sb foil (Fig. S8b in Supporting information), and shifts positively compared to the signal of Sb–O–Sb bond in Sb₂O₃ and Sb₂O₅ (Figs. S8c and d in Supporting information). The above results confirm the presence of Pd–Sb bond in the Pd₈Sb₃ NDs and Sb atoms donate electrons to Pd atoms because of the p-d orbital hybridization.

  Figure 3

  

  Figure 3.  (a) Normalized XANES spectra at the Pd K-edge of Pd₈Sb₃ NDs, Pd foil, and PdO. (b) Normalized XANES spectra at the Sb K-edge of Pd₈Sb₃ NDs, Sb foil, Sb₂O₃, and Sb₂O₅. (c) Pd K-edge FT-EXAFS spectra of Pd₈Sb₃ NDs, Pd foil, and PdO. (d) EXAFS fitting curves at the Pd K-edge of Pd₈Sb₃ NDs. (e) Sb K-edge FT-EXAFS spectra of Pd₈Sb₃ NDs, Sb foil, Sb₂O₃, and Sb₂O₅. (f) EXAFS fitting curve at the Sb K-edge of Pd₈Sb₃ NDs. (g) WT-EXAFS contour plots of Pd₈Sb₃ NDs and the reference samples.

  DownLoad: Full-Size Img PowerPoint

  Given that Pd-based nanomaterials are widely acknowledged as highly promising anode catalysts for direct alcohol fuel cells (DAFCs) under alkaline conditions, our prepared Pd₈Sb₃ NDs are employed as electrocatalyst for EOR in alkaline conditions. Prior to the evaluation of electrocatalytic performance, Pd₈Sb₃ NDs are loaded on commercial carbon black (i.e., Ketjen Black) to prepare the carbon-supported electrocatalysts. Commercial Pd/C is also tested under the same conditions for comparison. Fig. 4a illustrates the cyclic voltammograms (CVs) of Pd₈Sb₃ NDs and commercial Pd/C in 1.0 mol/L KOH. The reduction peaks located at about −0.35 V (vs. saturated calomel electrode (SCE)) are attributed to the reduction peak of PdO. The stronger reduction peak of PdO in Pd₈Sb₃ compared to that in Pd/C suggests the more exposed Pd active sites for catalytic reaction. The integrated charge associated with the reduction peak is used to estimate the electrochemical surface areas (ECSAs) of Pd₈Sb₃ NDs and Pd/C. Pd₈Sb₃ NDs possess a larger ECSA (66.4 m²/g) than that of commercial Pd/C (29.5 m²/g), indicating much higher utilization efficiency of Pd in Pd₈Sb₃ NDs. As shown in Figs. 4b and c, CV curves are further recorded in 1.0 mol/L KOH + 1.0 mol/L ethanol to assess the catalytic activities of the catalysts. Impressively, the MA and specific activity (SA) of Pd₈Sb₃ NDs significantly increases compared to that of commercial Pd/C. Notably, Pd₈Sb₃ NDs possess a MA of 4.6 A/mg, which is about 6 times as high as that of commercial Pd/C (Fig. 4d, left). And the SAs of Pd₈Sb₃ NDs and commercial Pd/C are 6.9 and 2.6 A/cm², respectively (Fig. 4d, right). As displayed in Table S3, the MA of Pd₈Sb₃ NDs outperforms most of the reported Pd-based electrocatalysts. Meanwhile, the onset potential of EOR on Pd₈Sb₃ NDs is 70 mV lower than that on Pd/C (Fig. S9 in Supporting information), indicating that the reaction kinetics of EOR is facilitated on Pd₈Sb₃ NDs. These results indicate that Pd₈Sb₃ NDs exhibit excellent catalytic performance towards EOR.

  Figure 4

  

  Figure 4.  (a) CV curves of Pd₈Sb₃ NDs and commercial Pd/C tested in a 1.0 mol/L KOH aqueous solution. CV curves normalized to (b) the mass of Pd and (c) ECSAs tested in 1.0 mol/L KOH + 1.0 mol/L ethanol. (d) Histogram of mass activities (MAs) and specific activities (SAs) for EOR. (e) i-t-test of Pd₈Sb₃ NDs and Pd/C recorded at −0.3 V (vs. SCE). CV curves of (f) Pd₈Sb₃ NDs and (g) Pd/C before and after the i-t-test. (h) CO-stripping measurements of different catalysts in 1.0 mol/L KOH (scan rate: 50 mV/s). (i) ¹H NMR spectra of the electrolytes before and after i-t-test.

  DownLoad: Full-Size Img PowerPoint

  Apart from catalytic activity, the stability of the catalysts during the EOR process is crucial as well. To evaluate the durability of highly active Pd₈Sb₃ NDs, long-term i-t measurements are conducted at −0.3 V (vs. SCE) for 30,000 s. As shown in Fig. 4e, the MA of Pd₈Sb₃ NDs remains 0.9 A/mg after the i-t-test while the MA of commercial Pd/C decreases rapidly, nearly losing the mass activity. In addition, after 30,000 s, the current density of Pd₈Sb₃ NDs only decays by 8.7%, compared to a 51.2% decrease for commercial Pd/C (Figs. 4f and g), further demonstrating the superior stability of Pd₈Sb₃ NDs. Besides, no obvious change is found from the TEM image, the XRD pattern, and the EDS mapping of the Pd₈Sb₃ NDs after the i-t-test (Fig. S10 in Supporting information), demonstrating the high stability on the crystal structure and composition. Meanwhile, the atomic ratio of Pd and Sb is 73.1:26.9 after the i-t-test (Table S4 in Supporting information), close to the atomic ratio of the synthesized Pd₈Sb₃ NDs. This result indicates negligible Sb dissolution occurs during the long-term stability test. And the Pd and Sb are still mainly in the metallic state (Fig. S11 in Supporting information), further demonstrating the structural stability of Pd₈Sb₃ NDs. It is widely acknowledged that CO-induced poisoning of Pd active sites is unavoidable during the EOR [19,27]. The ratio of forward current density to backward current density (I_f/I_b) indicates the resistance of catalysts to poisoning by reaction intermediates. As shown in Fig. 4b, Pd₈Sb₃ NDs possess an I_f/I_b value of 1.06, larger than that of commercial Pd/C (0.63), revealing the good tolerance ability to intermediates. To further confirm the improved anti-poisoning ability of Pd₈Sb₃ NDs, CO stripping experiments are conducted. The peak potential of CO oxidation on Pd₈Sb₃ NDs (−251 mV) is 37 mV more negative than that of commercial Pd/C (−214 mV), indicating the easier removal of CO from Pd₈Sb₃ NDs and their substantially enhanced CO tolerance (Fig. 4h) [28,29]. Previous work has reported that Sb can form oxidation species at low potential, facilitating the oxidation of CO adsorbed on adjacent Pd site [25,30]. To investigate the product of EOR on Pd₈Sb₃ NDs, ¹H nuclear magnetic resonance (NMR) spectra of the electrolytes are measured before and after i-t-test. As shown in Fig. 4i, before electrolysis, the signals of H atoms in terminal hydroxymethyl (-CH₂OH) and methyl (-CH₃) groups of CH₃CH₂OH are observed clearly at 3.64 and 1.15 ppm, respectively. After the i-t-test, an obvious NMR signal appears at 1.76 ppm, indicating that acetic acid (CH₃COOH) is the only liquid product after EOR. Furthermore, the Faradaic efficiency of the acetic acid is calculated based on the calibration curve method and the standard curve is shown in Fig. S12 (Supporting information). The Faradaic efficiency of the acetic acid for Pd₈Sb₃ NDs is 80.7%, confirming that EOR on Pd₈Sb₃ NDs mainly proceed through the C2 pathway with 4 electrons transfer.

  To shed light on the superior EOR performance of Pd₈Sb₃ intermetallic NDs, DFT calculations are carried out (see Supporting Information for computational details). Two typical models, including Pd (111) surface and Pd₈Sb₃ (001) surface are constructed for the calculations (Fig. S13 in Supporting information). Based on the NMR results, the main product of EOR on Pd₈Sb₃ is CH₃COO^-. During the EOR process, multiple intermediates are generated, indicating different reaction pathways [31]. To elucidate the reaction mechanism, three possible pathways and corresponding models are shown in Fig. 5a and Fig. S14 (Supporting information) [32]. In Pathway Ⅰ, the O—H bond is first broken, leading to the formation of ^*CH₃CH₂O intermediates, and the C—H bond in -CH₂ is further broken to form ^*CH₃CHO. In Pathway Ⅱ and Pathway Ⅲ, the C—H bond, rather than O—H bond, is dissociated to form ^*CH₃CHOH. After that, the O—H bond is broken, resulting in the formation of ^*CH₃CHO in Pathway Ⅱ, while ^*CH₃COH is obtained after further dissociation of C—H bond in Pathway Ⅲ. Figs. 5b-d illustrate the free energies of intermediates adsorbed on Pd₈Sb₃ (001) and Pd (111). Obviously, in Pathway Ⅰ, the potential determining step (PDS) on Pd₈Sb₃ (001) is the oxidation of ^*CH₃CHO to ^*CH₃CO, which requires 1.25 eV to overcome the energy barrier. In both Pathway Ⅱ and Ⅲ, the PDS on Pd₈Sb₃ (001) is the dehydrogenation of ^*CH₃CH₂O to ^*CH₃COH, with a higher energy barrier of 1.49 eV. The above results demonstrate that EOR follows Pathway Ⅰ on Pd₈Sb₃ (001). As for Pd (111), the required energies for PDS in Pathway Ⅰ, Ⅱ and Ⅲ are 1.54, 1.54 and 1.70 eV, respectively, all of which are higher than those required on Pd₈Sb₃ (001). To identify the factors contributing to the enhanced electrocatalytic activity of Pd₈Sb₃ (001) in EOR, we investigate the electronic structure and the binding energies of key intermediates using theoretical calculations. As shown in the total density of states of Pd₈Sb₃ (001), high electronic states at the Fermi level are observed, suggesting the metallic characteristic of the catalyst (Fig. S15 in Supporting information) [33,34]. Therefore, Pd₈Sb₃ NDs possess high conductivity during the EOR, benefiting the catalytic activity. Moreover, the projected DOS (PDOS) of Pd₈Sb₃ in Fig. 5e illustrates that the peaks of Pd 4d orbitals closely overlap with those of Sb 5p orbitals. This, combined with the absence of Sb 4d orbitals (Fig. S16 in Supporting information), suggests the strong p-d orbital hybridization between Pd and Sb, as opposed to weaker d-d interactions [8,13,35]. From the differential charge distribution on Pd₈Sb₃ (001) (Fig. S17a in Supporting information), the redistribution of electrons of Pd induced by the p-d orbital hybridization is observed, indicating the charge transfer between Pd and Sb atoms. Through Bader charge analysis, it is shown that each Pd atom obtains an average of 0.308 |e| electrons from the Sb (Fig. S17b in Supporting information), aligning with the XPS and XANES results. The change of electronic configuration determines the adsorption energy of the key intermediates, thereby tuning the catalytic activity of catalysts. For example, the adsorption energy of CH₃CH₂OH and ^•OH on Pd₈Sb₃ (001) is −0.33 eV and 0.65 eV, respectively, which are lower than those on Pd (111) (−0.20 and 0.94 eV) as shown in Fig. S18 (Supporting information), confirming the easier adsorption of CH₃CH₂OH and ^•OH on Pd₈Sb₃. The d-band center of Pd₈Sb₃ (001) (−2.21 eV) shifts negatively compared with that of Pd (111) (−1.80 eV), making it easier to adsorb CH₃CH₂OH and ^•OH on Pd₈Sb₃ NDs (Fig. S19 in Supporting information). This result is consistent with the calculated adsorption energies of CH₃CH₂OH and ^•OH on Pd₈Sb₃ (001) and Pd (111). The easier adsorption of CH₃CH₂OH on Pd₈Sb₃ facilitates the first step of EOR. And the enhanced ^•OH adsorption benefits the removal of poisoning intermediates such as CO, which is consistent with the results of CO-stripping test. In addition, the adsorption energy of CO on Pd₈Sb₃ (001) is higher than that on Pd (111), indicating the weaker CO bonding. Furthermore, to investigate the bonding behavior of CO, the DOS of CO adsorbed on Pd (111) and Pd₈Sb₃ (001) are calculated (Fig. 5f). Clearly, the interaction of the bonding (π) and anti-bonding (π^*) orbitals in CO are different over Pd and Pd₈Sb₃. The energy levels of CO π orbitals match more well with Cu 3d orbitals of Pd in Pd (111) than in Pd₈Sb₃ (001). And the energy dispersity of d-π^* in Pd (111)/CO is broader than that in Pd₈Sb₃ (001)/CO above the Fermi level. The above results illustrate that the p-d hybridization in Pd₈Sb₃ changes the electronic structure of Pd, which can reduce the adsorption of CO and thus benefit for the anti-poisoning ability. Overall, the preparation of intermetallic Pd₈Sb₃ NDs by incorporating p-block Sb with d-block Pd leads to the transition from d-d hybridization of the fcc Pd (111) surface to p-d hybridization of the trigonal Pd₈Sb₃ (001) surface. This change allows for the reconfiguration of electrons, which optimize the intrinsic adsorption properties for key intermediates. As a result, Pd₈Sb₃ NDs exhibit superior catalytic activity than commercial Pd/C.

  Figure 5

  

  Figure 5.  (a) The scheme of reaction pathways on Pd₈Sb₃ (001). Free energy profiles of EOR on Pd₈Sb₃ (001) and Pd (111) through (b) Pathway Ⅰ, (c) Pathway Ⅱ and (d) Pathway Ⅲ. (e) The PDOS of Pd₈Sb₃ (001). (f) The interaction between CO molecule and Pd 3d orbital in Pd (111) and Pd₈Sb₃ (001).

  DownLoad: Full-Size Img PowerPoint

  In summary, intermetallic Pd₈Sb₃ NDs with a trigonal crystal structure have been synthesized using a simple hydrothermal strategy. A strong p-d orbital hybridization interaction is present in intermetallic Pd₈Sb₃ NDs. The synthesized intermetallic Pd₈Sb₃ NDs exhibit outstanding EOR performance with a MA of 4.6 A/mg, which is about 6 times as high as that of commercial Pd/C. The intermetallic Pd₈Sb₃ NDs also possess an improved stability after the i-t-test for 30,000 s with a slight decay of 8.7% in current density, as well as strong resistance to CO poisoning. DFT calculation results demonstrate that the intermetallic Pd₈Sb₃ NDs reduce the energy barrier of the PDS in EOR. Meanwhile, the p-d orbital hybridization interaction in intermetallic Pd₈Sb₃ NDs modifies the electronic structure, thereby optimizing the adsorption/desorption energies with the key intermediates.

  CRediT authorship contribution statement

  Sumei Han: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Chaoqun Ma: Visualization, Software, Data curation. Fukai Feng: Visualization, Methodology, Investigation. Xinran Jiao: Data curation. Xiao Ma: Validation. Caihong He: Formal analysis. Huaifang Zhang: Validation. Gang Lin: Visualization. Xiangmin Meng: Software, Funding acquisition. Jing Xia: Supervision, Funding acquisition. Qinbai Yun: Writing – review & editing, Supervision, Funding acquisition. Wenbin Cao: Supervision, Funding acquisition. Qipeng Lu: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was financially supported by the Beijing Natural Science Foundation (No. 2242008), the National Natural Science Foundation of China (Nos. 52471220, U2441264 and 51972026), the Fundamental Research Funds for the Central Universities (Nos. GJRC2023003 and 00007826), and the Guangdong Basic and Applied Basic Research Foundation (Nos. 2023A1515110010 and 2024A1515012653). This work was also supported by the National Key Research and Development Program of China (No. 2021YFF0704705), Major Research Plan of National Natural Science Foundation of China (No. 92263205), Scientific Equipment Development Project of Chinese Academy of Sciences, Youth Innovation Promotion Association Project of Chinese Academy of Sciences (No. 2020026), and Technical Support Talent Project of Chinese Academy of Sciences.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111974.

  
  
• 1. [1]

     J. Chow, R.J. Kopp, P.R. Portney, Science 302 (2003) 1528–1531. doi: 10.1126/science.1091939

  2. [2]

     J. Chang, G. Wang, X. Chang, et al., Nat. Commun. 14 (2023) 1346. doi: 10.1038/s41467-023-37011-z

  3. [3]

     A. Kowal, M. Li, M. Shao, Nat. Mater. 8 (2009) 325–330. doi: 10.1038/nmat2359

  4. [4]

     Q. Yun, Q. Lu, C. Li, et al., ACS Nano 13 (2019) 14329–14336. doi: 10.1021/acsnano.9b07775

  5. [5]

     X. Zhou, Y. Ma, Y. Ge, et al., J. Am. Chem. Soc. 144 (2022) 547–555. doi: 10.1021/jacs.1c11313

  6. [6]

     G. Liu, W. Zhou, Y. Ji, et al., J. Am. Chem. Soc. 143 (2021) 11262–11270. doi: 10.1021/jacs.1c05856

  7. [7]

     Y. Chen, Z. Fan, Z. Luo, et al., Adv. Mater. 29 (2017) 1701331. doi: 10.1002/adma.201701331

  8. [8]

     P. Li, J. Bi, J. Liu, et al., J. Am. Chem. Soc. 145 (2023) 4675–4682. doi: 10.1021/jacs.2c12743

  9. [9]

     Y. Wang, M. Zheng, Y. Li, et al., Angew. Chem. Int. Ed. 61 (2022) e202115735. doi: 10.1002/anie.202115735

  10. [10]

      Y. Lv, P. Liu, R. Xue, et al., Adv. Sci. 11 (2024) 2309813. doi: 10.1002/advs.202309813

  11. [11]

      G. Qin, Y. Hao, H. Ma, et al., Adv. Funct. Mater. 34 (2024) 2312744. doi: 10.1002/adfm.202312744

  12. [12]

      C. Ma, H. Zhang, J. Xia, et al., J. Am. Chem. Soc. 146 (2024) 20069–20079. doi: 10.1021/jacs.4c04023

  13. [13]

      L. Gao, X. Li, Z. Yao, et al., J. Am. Chem. Soc. 141 (2019) 18083–18090. doi: 10.1021/jacs.9b07238

  14. [14]

      G. Wu, W. Zhang, R. Yu, et al., Angew. Chem. Int. Ed. 63 (2024) e202410251. doi: 10.1002/anie.202410251

  15. [15]

      X. Cao, Y. Tian, J. Ma, et al., Adv. Mater. 36 (2024) 2309648. doi: 10.1002/adma.202309648

  16. [16]

      M. Armbrüster, K. Kovnir, M. Behrens, et al., J. Am. Chem. Soc. 132 (2010) 14745–14747. doi: 10.1021/ja106568t

  17. [17]

      Y. Zhang, X. Liu, T. Liu, et al., Adv. Mater. 34 (2022) 2202333. doi: 10.1002/adma.202202333

  18. [18]

      M. Song, Q. Zhang, T. Shen, G. Luo, D. Wang, Chin. Chem. Lett. 35 (2024) 109083. doi: 10.1016/j.cclet.2023.109083

  19. [19]

      B. Xu, T. Liu, X. Liang, et al., Adv. Mater. 34 (2022) 2206528. doi: 10.1002/adma.202206528

  20. [20]

      W. Tong, B. Huang, P. Wang, et al., Angew. Chem. Int. Ed. 59 (2020) 2649–2653. doi: 10.1002/anie.201913122

  21. [21]

      W. Xiao, M.A.L. Cordeiro, G. Gao, et al., Nano Energy 50 (2018) 70–78. doi: 10.1016/j.nanoen.2018.05.032

  22. [22]

      B. Chen, P. Duan, S. Wang, et al., J. Mater. Chem. A 13 (2025) 17242–17260. doi: 10.1039/d5ta02448j

  23. [23]

      J.A. Dean, Lange’s Handbook of Chemistry, McGraw-Hill, Inc., New York, 1999.

  24. [24]

      M. Zhou, J. Liu, C. Ling, et al., Adv. Mater. 34 (2022) 2106115. doi: 10.1002/adma.202106115

  25. [25]

      Y. Zhang, G. Li, Z. Zhao, et al., Adv. Mater. 33 (2021) 2105049. doi: 10.1002/adma.202105049

  26. [26]

      J.N. Hausmann, M. Ashton, S. Mebs, et al., Small 20 (2024) 2309749. doi: 10.1002/smll.202309749

  27. [27]

      C. Bianchini, P.K. Shen, Chem. Rev. 109 (2009) 4183–4206. doi: 10.1021/cr9000995

  28. [28]

      Y. Xiong, J. Dong, Z.Q. Huang, et al., Nat. Nanotechnol. 15 (2020) 390–397. doi: 10.1038/s41565-020-0665-x

  29. [29]

      H. Lv, L. Sun, Y. Wang, S. Liu, B. Liu, Adv. Mater. 34 (2022) 2203612. doi: 10.1002/adma.202203612

  30. [30]

      Q.H. Wu, S.G. Sun, X.Y. Xiao, Y.Y. Yang, Z.Y. Zhou, Electrochim. Acta 45 (2000) 3683–3690. doi: 10.1016/S0013-4686(00)00462-X

  31. [31]

      Z.X. Liang, T.S. Zhao, J.B. Xu, L.D. Zhu, Electrochim. Acta 54 (2009) 2203–2208. doi: 10.1016/j.electacta.2008.10.034

  32. [32]

      S. Han, H. Sun, C. Ma, et al., Adv. Funct. Mater. 34 (2024) 2403023. doi: 10.1002/adfm.202403023

  33. [33]

      C. He, C. Ma, J. Xia, et al., Adv. Funct. Mater. 34 (2024) 2311683. doi: 10.1002/adfm.202311683

  34. [34]

      B. Fei, Z. Chen, J. Liu, et al., Adv. Energy Mater. 10 (2020) 2001963. doi: 10.1002/aenm.202001963

  35. [35]

      L. Ji, H. Che, N. Qian, et al., Appl. Catal. B: Environ. 328 (2023) 122521. doi: 10.1016/j.apcatb.2023.122521

• 

  Figure 1  (a) XRD pattern of Pd₈Sb₃ NDs. The inset illustrates the corresponding crystal structure model. (b) TEM image and (c) diameter distribution of Pd₈Sb₃ NDs. (d) HRTEM image, (e) EDS mapping and (f) EDS spectrum of Pd₈Sb₃ NDs.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) HAADF-STEM image of a single Pd₈Sb₃ ND. (b) The corresponding FFT pattern of (a). (c) The simulated FFT pattern of Pd₈Sb₃ intermetallic along [001] direction. (d, e) The high-resolution HAADF-STEM image taken from areas marked in (a). (f) Atomic-resolution elemental mappings of a Pd₈Sb₃ ND. Crystal structure model viewed along the (g) [001] and (h) [100] directions. Red and green spheres denote Pd and Sb atoms, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Normalized XANES spectra at the Pd K-edge of Pd₈Sb₃ NDs, Pd foil, and PdO. (b) Normalized XANES spectra at the Sb K-edge of Pd₈Sb₃ NDs, Sb foil, Sb₂O₃, and Sb₂O₅. (c) Pd K-edge FT-EXAFS spectra of Pd₈Sb₃ NDs, Pd foil, and PdO. (d) EXAFS fitting curves at the Pd K-edge of Pd₈Sb₃ NDs. (e) Sb K-edge FT-EXAFS spectra of Pd₈Sb₃ NDs, Sb foil, Sb₂O₃, and Sb₂O₅. (f) EXAFS fitting curve at the Sb K-edge of Pd₈Sb₃ NDs. (g) WT-EXAFS contour plots of Pd₈Sb₃ NDs and the reference samples.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) CV curves of Pd₈Sb₃ NDs and commercial Pd/C tested in a 1.0 mol/L KOH aqueous solution. CV curves normalized to (b) the mass of Pd and (c) ECSAs tested in 1.0 mol/L KOH + 1.0 mol/L ethanol. (d) Histogram of mass activities (MAs) and specific activities (SAs) for EOR. (e) i-t-test of Pd₈Sb₃ NDs and Pd/C recorded at −0.3 V (vs. SCE). CV curves of (f) Pd₈Sb₃ NDs and (g) Pd/C before and after the i-t-test. (h) CO-stripping measurements of different catalysts in 1.0 mol/L KOH (scan rate: 50 mV/s). (i) ¹H NMR spectra of the electrolytes before and after i-t-test.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) The scheme of reaction pathways on Pd₈Sb₃ (001). Free energy profiles of EOR on Pd₈Sb₃ (001) and Pd (111) through (b) Pathway Ⅰ, (c) Pathway Ⅱ and (d) Pathway Ⅲ. (e) The PDOS of Pd₈Sb₃ (001). (f) The interaction between CO molecule and Pd 3d orbital in Pd (111) and Pd₈Sb₃ (001).

  下载: 全尺寸图片 幻灯片
•