Stable Strain State of Single‐Twinned AgPdF Nanoalloys under Formate Oxidation Reaction

Surface reconstruction as common phenomenon during catalysis complicates the prediction and modeling on catalytic activity of the nanoalloy, hence developing a stable structure to be resistant to surface restructuring would provide an ideal prototype for substantial and reliable mechanism analysis. Herein, the single‐twinned structure in inverse AgPdF catalyst is constructed to enhance the catalytic activity and stability for the formate oxidation reaction (FOR). The single‐twinned AgPdF nanoalloy (t‐AgPdF) catalyst exhibits an enhanced peak current density of 4.6 A mgPd−1, a reduced onset potential of 0.44 V, a higher activity retention of 55.7% after 600 cycles, and a longer activity retention time of 55.9 h. Additionally, the t‐AgPdF catalyst presents a higher hydrogen generation rate of 1.11 mL mgPd−1 than that of single‐crystalline AgPd nanoalloy (AgPd) catalyst, and density functional theory calculations reveal that t‐AgPdF(111) surface exhibits a reduced activation energy of 0.59 eV for formate decomposition reaction. Impressively, the t‐AgPdF maintains compressive and tensile strain state along the Σ3 twin boundaries before and after the FOR, in contrast to AgPd. This is the first time to reveal that the nanotwinned structures contribute inverse t‐AgPdF catalysts the catalytic active sites with stable strain state since starting reaction for the FOR.


Introduction
Direct formate fuel cells have emerged as a promising solution to meet the practical requirements of future fuel cell technology, primarily owing to the secure storage-transportation and easier operation characteristics of formate fuel over other candidates. [1]owever, the slow kinetics of the formate oxidation reaction (FOR) has hindered its practical applications in electric vehicles and electronic devices. [2]Numerous efforts have been made to improve the FOR catalytic properties of Pd-based catalysts, such as through alloying with other metals, miniaturizing to nanoscale, and controlling morphologies. [3]Nevertheless, the synergistic modulation to physicochemical properties of intrinsic active surface is crucial for promoting Pd-based nanoalloys to touch the ceiling of catalytic activity for the FOR. [4]In this regard, the strained nanostructure of catalysts can be constructed through compressing or expanding the atomic geometry of active sites, where a minor strain introduced into the catalyst can significantly modulate the adsorption properties and ultimately impacts the catalytic activity. [5]enerally, Pd-based nanoalloys undergo the structural evolution during catalytic reaction, known as surface reconstruction; the reconstructed structures may be more dynamic due to the higher catalytic active. [6]owever, the complicated evolution information of surface reconstruction makes it challenging to identify the structural model of actual active sites and, in turn, to conduct the substantial and reliable mechanism analysis. [7]To eliminate the impact of reconstructing, an effective strategy involves constructing the steady configuration during catalysis to directly trigger the catalytic cycle without structural transformation.For example, Sun et al. reported  the atomically dispersed Ru on Ni-V layered double hydroxide scaffold with the bifunctional reactive sites, which can preserve the surface configuration and suppress the occurrence of surface restructuring through a strongly atomic metal-support interaction. [8]Huo et al. designed the Re-doped IrO 2 as highly active catalyst for oxygen evolution reaction, and the strong confinement of Re within the IrO 2 lattice can suppress Ir dissolution and enhance the catalytic stability. [9]Liu et al. explored the dislocation-strained IrNi nanoparticles as a highly efficient hydrogen evolution reaction catalyst, and the strain-induced structures arising from dislocations could be resistant to surface restructuring during catalysis. [10]Therefore, developing a stable structural configuration to be resistant to surface restructuring is essential for optimizing the catalytic properties of designed catalysts.DOI: 10.1002/sstr.202300110Surface reconstruction as common phenomenon during catalysis complicates the prediction and modeling on catalytic activity of the nanoalloy, hence developing a stable structure to be resistant to surface restructuring would provide an ideal prototype for substantial and reliable mechanism analysis.Herein, the single-twinned structure in inverse AgPdF catalyst is constructed to enhance the catalytic activity and stability for the formate oxidation reaction (FOR).The single-twinned AgPdF nanoalloy (t-AgPdF) catalyst exhibits an enhanced peak current density of 4.6 A mg Pd À1 , a reduced onset potential of 0.44 V, a higher activity retention of 55.7% after 600 cycles, and a longer activity retention time of 55.9 h.Additionally, the t-AgPdF catalyst presents a higher hydrogen generation rate of 1.11 mL mg Pd À1 than that of single-crystalline AgPd nanoalloy (AgPd) catalyst, and density functional theory calculations reveal that t-AgPdF(111) surface exhibits a reduced activation energy of 0.59 eV for formate decomposition reaction.Impressively, the t-AgPdF maintains compressive and tensile strain state along the Σ3 twin boundaries before and after the FOR, in contrast to AgPd.This is the first time to reveal that the nanotwinned structures contribute inverse t-AgPdF catalysts the catalytic active sites with stable strain state since starting reaction for the FOR.
material, has also been identified in nanomaterials, which are formed by splitting the crystal lattice into two mirror-symmetric planes. [11]The stress and strain distribution in nanoalloys can lead to the formation of Shockley partial dislocations, which can move to the surface due to the nanoscale size of a single nanoparticle.The movement of Shockley partial dislocations leaves stacking faults behind and twin boundaries can result and be bounded by two stacking faults. [12]Liu et al. utilized metastable Pd nanoparticles with twin boundaries to catalyze the ethanol oxidation reaction, minimizing poisoned active sites and accelerating the catalytic process. [13]Huang et al. introduced the increasing twin boundaries in Pd catalyst to enhance the methane combustion activity, and the defective structure with strain can lower the activation and desorption barriers. [14]The twin boundaries can not only directly function as active sites, but also indirectly enhance catalytic activity through strain effect. [15]Furthermore, constructing the correlation of twinned structure with structural reconstruction is critical for the research of future FOR catalysts.Surface treatment of nanoalloys can induce the formation of inverse catalysts, where metal oxides or other compounds are decorated on metal surfaces. [16]Chen et al. reported a PdNW@CuO x nanostructure with CuO x nanolayer deposited onto Pd surface, and the Pd/CuO x interface showed higher activity and stability for the alcohol oxidation reaction. [17]Su et al. constructed a Pd-Pd 4 S/C heterostructure to enhance the catalytic performance toward the hydrogen oxidation reaction, leveraging strong electronic interaction at the interface to strengthen OH adsorption. [18]Our previous work developed an inverse AgPdF nanoalloy to replace the surface self-adapting structure in AgPdO true catalyst, and the stable interface maintained the enhanced catalytic properties during the long-term FOR. [19]The inverse catalysts offer abundant diverse active sites, which can against further oxidation to improve catalytic stability. [20]Accordingly, combining twinned structures with inverse catalysts has the potential to yield insight into the possible reconstruction behavior in nanoalloy catalysts, and provides an opportunity to synergistically enhance catalytic performance for the FOR.
Herein, we reported the single-twinned AgPdF nanoalloy (t-AgPdF) with stable strain state during the FOR.Unlike the single-crystalline AgPd nanoalloy (AgPd) with a distribution of compressive strain at the edges, the t-AgPdF has a distribution of both compressive and tensile strains along the twin boundary.The t-AgPdF catalyst maintained its strain distribution along the twin boundary, and exhibited the lower onset potential, the higher peak current density, and activity retention.During long-term potentiostatic polarization tests, the t-AgPdF catalyst with a stable strain state showed a higher current density and a longer activity retention time.Additionally, the t-AgPdF catalyst demonstrated enhanced catalytic performance for hydrogen generation, with a higher volume of generated H 2 gas.Theoretical calculations revealed that the t-AgPdF(111) surface showed a reduced activation energy for the formate decomposition reaction.To the best of our knowledge, this study first reported that the twin deformation in the AgPdF nanoalloy has been realized to improve the catalytic properties for the FOR.

Results and Discussion
Figure 1 shows the micromorphology, strain, and composition analysis of the single-crystalline AgPd nanoalloy (AgPd), whose low-magnification transmission electron microscopy (TEM) image is shown in Figure S1a, Supporting Information.Figure 1a shows the high-resolution transmission electron microscopy (HRTEM) image of a typical nanoparticle with clear lattice fringes and a particle size of approximately 6 nm.The inset image presents the atomic model of AgPd.In Figure 1b, the strain distribution image of an individual nanoparticle is displayed.The significant compressive strain in e yy direction is represented by the color of dark blue distributed at the edges and surface of the nanoparticles.3a,21] Figure 1e,f depicts the highangle annular dark field-scanning transmission electron microscopy (HAADF-STEM) image, energy-dispersive X-ray (EDX) elemental mapping, and line profile of the single nanoparticle, illustrating that both Ag and Pd elements are uniformly distributed throughout the nanoparticle, while the content of O element is almost negligible.
The micromorphology, strain, and composition analysis of the t-AgPdF are presented in Figure 2, with a low-magnification TEM image shown in Figure S1b, Supporting Information.Figure 2a depicts the HRTEM image of a single nanoparticle with a size of 8 nm in the field of view, and surprisingly, a twin boundary can be marked out at the junction of lattice fringes with the mirrorsymmetrically extension directions.The inset image shows the atomic model of t-AgPdF.Figure 2b further shows the strain distribution image of an individual nanoparticle.The dark red and blue regions are concentrated at the edge or twin boundary of the nanoparticle; obviously, the alternate distribution of dark blue and red regions exhibits the coexistence of compressive and tensile strains in the e yy direction of the t-AgPdF.The fluorination treatment in nanoalloys typically involves three processes: 1) the chemical reaction between F and Ag atoms; 2) the diffusion of F ions driven by a concentration gradient; and 3) the resulting volume expansion of nanoalloys from the embedded F ions.The fluorination reaction is inevitably accompanied by the breaking/reforming chemical bonds between the F, Ag, and Pd atoms.The released reaction energy and the concentration gradient of F within the nanoalloys drive the diffusion of F ions inward, as the driving force is balanced by the accumulated internal stress induced by the embedding of F ions.The diffusion of F ions ceases at some position, resulting in volume expansion and high internal strain energy, leading to the stress distribution in the fluorinated nanoalloys. [22]Figure 2c presents the FFT pattern of the corresponding region in the HRTEM image.Different from AgPd, two sets of bright spots are marked by red and cyan circles, respectively.Combined with the HRTEM image, four pairs of the diffraction spot belonging to the ( 111) and ( 220) characteristic crystal planes are offset by a symmetrical distribution from each other.And the diffraction spots belonging to different sets coincide into a pair of diffraction spots, belonging to the (200) characteristic crystal plane, which is also the crystal plane information of twin boundary.Figure 2d displays the IFFT pattern corresponding to the FFT pattern, which also highlights the twin boundary at the edges of the lattice fringes extending in various directions.We express the results using the coincidence site lattice method designated by the symbol ΣN with the twin boundary corresponding to Σ3. [12a] In the compressive strained region of the t-AgPdF, the interplanar spacing of the lattice fringe for the (111) crystal plane is 0.225 nm, while it is 0.233 nm in the tensile strained region.Figure 2e shows the HAADF-STEM image and EDX elemental mapping of the t-AgPdF, which indicates that the F element is evenly distributed throughout the nanoparticle.Figure 2f presents the EDX line profile extracted from the typical line in the HAADF-STEM image, and the t-AgPdF contains 50.8 wt% Ag, 48.5 wt% Pd, and 0.7 wt% F. Different from the decrease trend of Pd element content, unexpectedly, the contents of Ag and F elements are enriched at the same position.22b] With the increase of the chemical potential of F element, the adlayer on the surface gradually transforms into the surface fluoride. [23]herefore, the enriched Ag and F elements is associated with the formation of surface fluoride.The structure of surface fluoride covering the metal surface is a typical inverse catalyst, which indicates the synergy between single-twinned structure and inverse catalyst in the as-prepared t-AgPdF.Figure S2, Supporting Information, illustrates the X-ray diffraction (XRD) patterns of the t-AgPdF and AgPd.The diffraction peak of the characteristic (111) crystal plane is located between standard diffraction paeks of Pd(111) and Ag(111) planes, indicating the existence of alloying structures in the t-AgPdF and AgPd.The similarity of diffraction peaks between t-AgPdF and AgPd suggests that the surface fluoride on the surface of t-AgPdF is amorphous structure.
To gain a better understanding of the aforementioned characterization results, the slab models of the AgPd(111) and t-AgPd(111) surfaces were constructed by cleaving the crystal face of the conventional cell.The molecular dynamics (MD) simulation results of the strain fields for the AgPd(111) and t-AgPd(111) surfaces are presented in Figure 3, with top and left-side views displayed in Figure 3a,b, respectively.To conduct the MD simulations of slab models with the LAMMPS package, the corresponding embedded-atom-method (EAM) interatomic potential is required. [24]However, to our knowledge, there is no available EAM interatomic potential for AgPdF, which prevents the MD simulation of the t-AgPdF slab model.Previous theoretical calculations about the mechanistic origin of strain fields demonstrate a strong dependence of strain on the morphology of the nanocrystal, mostly independent of size. [25]Thus, the strain distributions of AgPd with twinned boundary is calculated to justify the strain distributions observed in the t-AgPdF during the experiment, and a reasonable agreement is expected between the internal stain states of t-AgPd and t-AgPdF.The t-AgPd(111) surface exhibits more blue areas around the twin boundary in comparison to the AgPd(111) surface, indicating that compressive strain can be concentrated around the twin boundaries.The surface compositions of t-AgPdF and AgPd are compared in Figure S3, Supporting Information, where intense peaks for each contained element are visible after calibration with C 1s at 284.8 eV. Figure S3a,b, Supporting Information, presents the high-resolution X-ray photoelectron spectroscopy (XPS) spectra of t-AgPdF and AgPd, respectively.The binding energies of Ag 3d 5/2 and Pd 3d 5/2 peaks in the t-AgPdF are positively shifted to 367.5 and 334.6 eV, respectively, when compared to those of the AgPd.The same trend can be observed by comparing the Ag 3d 3/2 and Pd 3d 3/2 peaks in the high-resolution XPS spectra of both nanoalloys.3a,6b] The positive shift of the binding energies corresponding to the Ag 3d peaks could be ascribed to the slight electron transfer from Ag to Pd/F.Similarly, the positive shift of the binding energies corresponding to the Pd 3d peaks indicates the modification of the electronic structure of Pd by a slight electron transfer from Pd to Ag/F. Figure S3c, Supporting Information, shows the appearance of an F 1s peak in the t-AgPdF, which is located at the binding energy of 688.8 eV.Table S1 and S2, Supporting Information, list the XPS spectral peak locations and surface compositions of Ag and Pd in the various nanoparticles.In addition, Figure S3d, Supporting Information, displays the XPS valence band spectra (VBS) for t-AgPdF and AgPd, where the d-band center of the -AgPdF is located at À4.99 eV, a negative shift compared to the AgPd located at À4.79 eV.The downshift of the d-band  center in the t-AgPdF is commonly attributed to electron transfer. [26]igure 4 demonstrates the catalytic performance of t-AgPdF catalyst in comparison to that of AgPd catalyst for the FOR.The correlation between catalytic activity and Na 2 H 2 PO 2 concentration for different AgPd catalysts is summarized in Figure S4, Supporting Information.As the Na 2 H 2 PO 2 concentration increases, the d-band centers of the catalysts shift downward from À4.53 to À4.86 eV, while the peak current densities in the cyclic voltammetry (CV) curves initially increase and then decrease.3a,28] However, according to the Sabatier principle, the best catalyst should bind atoms and molecules with an intermediate strength: not too weakly in order to be able to activate the reactants, and not too strongly to be able to desorb the intermediates and products. [29]This leads to a volcano-type relationship between activity and bond strength.Therefore, modulating the electronic structure can promote the activity to approach to the very peak of the volcano curve, thus achieving the highest catalytic activity.Figure 4a displays the CV curves of t-AgPdF and AgPd catalysts with red arrows indicating the scan direction.The t-AgPdF catalyst exhibits the highest oxidation peak current density of 4.6 A mg Pd À1 , which is 1.8 times higher than that of AgPd catalyst.Figure 4b shows the CV curves in the positive direction only, in the early potential range of 0-0.6 V, derived from Figure 4a.The t-AgPdF catalyst has the lowest onset potential at 0.44 V, which is significantly better than that of AgPd catalyst at 0.52 V. Figure S5, Supporting Information, presents the CV curves normalized by the electrochemically active surface areas (ECSA) and the ECSA of the t-AgPdF and AgPd catalysts.Compared with AgPd catalyst, the current density of the t-AgPdF catalyst increases to 30.7 mA cm À2 , and the ECSA of the t-AgPdF catalyst increases to 15 m 2 g Pd À1 .Figure 4c presents the activity retentions of t-AgPdF and AgPd catalysts over 600 cycles of cyclic voltammetry, while Figure S6, Supporting Information, shows the CV curves in the 1st, 300th and 600th cycles of these catalysts.The activity retention of t-AgPdF catalyst is 55.7%, while AgPd catalyst exhibits comparable activity retention only during the initial 100 cycles, and it drops significantly in the subsequent cycles with a final activity retention of only 35.5%. Figure S4d, Supporting Information, compares the peak current densities in the 1st and 600th CV curves of these catalysts, and the t-AgPdF catalyst exhibits much higher peak current density than AgPd and AgPdF catalysts after 600 CV cycles.Table S3 and S4, Supporting Information, provide the catalytic properties of the t-AgPdF catalyst and previous FOR catalysts, showing that the t-AgPdF catalyst has the highest mass activity and activity retention among the reported FOR catalysts.
The chronoamperometry (CA) curves in Figure 4d,e reveal important insights into the catalytic stability of t-AgPdF and AgPd catalysts.The t-AgPdF catalyst presents an approximate plateau shape with a higher current density than that of AgPd catalyst, which decreases significantly over time.However, the t-AgPdF catalyst maintains a current density of 0.49 A mg Pd À1 after the CA test, which is twice that of AgPd catalyst.In the long-term CA curves shown in Figure 4e, the current density of AgPd catalyst decreases rapidly and drops to 0.1 A mg Pd À1 after 17.7 h, while the t-AgPdF catalyst exhibits an activity plateau for up to 2.5 h in the initial stage, then slowly decayed and continued to maintain catalytic activity for 55.9 h, which is 3.2-fold that of AgPd catalyst.3a,3e,28] Hence, the increase of activity retention time fully accounts for the enhanced catalytic stability of the t-AgPdF catalyst.Remarkably, the t-AgPdF catalyst boasts the longest reported catalytic activity retention time to decay to 0.1 A mg Pd À1 among the FOR catalysts.Figure 5 presents the catalytic performance of hydrogen generation for the t-AgPdF catalyst compared with that of AgPd catalyst.Figure 5a displays the hydrogen production volume and reaction time curves of both catalysts during the initial 20 min reaction period.The almost straight curves reflect the hydrogen production activity of each catalyst.After 20 min, the volume of generated H 2 gas of the AgPd catalyst is 0.19 mL mg Pd À1 , while the t-AgPdF catalyst generates 1.7-fold more H 2 gas, with a volume of 0.32 mL mg Pd À1 .Figure 5b shows the volume of generated H 2 gas of the above catalysts during a long-term hydrogen generation reaction.However, the volume of generated H 2 gas of the t-AgPdF catalyst continues to increase during 120 min reaction, and the volume of generated H 2 gas of the t-AgPdF catalyst is 1.11 mL mg Pd À1 after 120 min reaction.Figure 5c,d presents the density functional theory (DFT) calculation for formate oxidation pathway on the t-AgPdF(111) and AgPd(111) surfaces, with additional data for the t-AgPd(111) surface shown in Figure S8, Supporting Information.Figure S7, Supporting Information, shows the optimized structures of the reactant, intermediate, and product in FOR.The FOR steps are as follows where H Ã , CO 2 Ã , and HCOO Ã indicate that H, CO 2 , and HCOO are adsorbed on the catalyst surface, respectively.Figure 5c illustrates the free energy diagram for each reaction step on both surfaces.The reaction step (3) with a more positive free energy change ΔG is considered the potential limiting step, with values of 0.2 and 0.34 eV for the t-AgPdF(111) and AgPd(111) surfaces, respectively.Figure 5d compares the activation energies for the formate decomposition reaction on both surfaces.The t-AgPdF(111) surface shows a reduced activation energy for the rate-limiting step, with an activation energy of 0.59 eV compared to 0.69 eV for the AgPd(111) surface.Together with the dynamic simulation, these results provide a theoretical basis for the catalytic improvement of the twinned structure in the t-AgPdF nanoalloy through the reduction of potential limiting and activation energy.
The catalytic stability of AgPd and t-AgPdF catalysts is related to the dynamic reconstruction under working condition, and the compressive and tensile strain states of nanoalloys are further investigated after the FOR. Figure 6 showcases the micromorphology and strain analysis of the AgPd catalyst after FOR.On the left side, the HRTEM image of a typical nanoparticle with a size of 8 nm is presented in Figure 6a, which shows the AgPd catalyst after the CV test.The strain distribution image of the nanoparticle in Figure 6b  image, which shows that the content of O element is almost negligible.Figure S11, Supporting Information, shows the morphology and strain characterizations of the AgPdF nanoalloy, and the presence of red area at the edge of the AgPdF nanoalloys indicates the existence of tensile strain due to the formation of surface fluoride.After CV and CA tests, the strain state remained stable, which can be attributed to the stable interface structure in AgPdF nanoalloy, consistent with the experimental results in our previous report. [19]igure 7 illustrates the micromorphology and strain analysis of the t-AgPdF catalyst after FOR.On the left side, the HRTEM image in Figure 7a shows a nanoparticle with a size of 9 nm, with twin boundaries marked.The strain distribution image in Figure 7b reveals that the tensile and compressive strains concentrate along the twin boundary, especially along the red and blue regions, indicating that these strains are more pronounced in the t-AgPdF catalyst after CV test.The corresponding FFT and IFFT patterns in Figure 7c,d display two sets of bright spots, circled in red and cyan colors, representing lattice fringes of the (111) crystal plane with interplanar spacing of 0.225 and 0.234 nm in compressive and tensile strained regions, respectively.Notably, the Σ3 twin boundary, represented by coincident diffraction spots in the FFT pattern, can be observed between two lattice fringe regions with different interplanar spacing in the IFFT pattern.In comparison to the AgPd catalyst, the compressive and tensile strains concentrate along the twin boundary in the t-AgPdF catalyst after CV test.The EDX elemental mapping and line profile of the HAADF-STEM image in Figure S9e,f, Supporting Information, reveal the enrichment of Ag and F content in the same position, and the t-AgPdF catalyst after CV test contains 51.5 wt% Ag, 47.8 wt% Pd, and 0.7 wt% F. It is evident that the distribution of Ag, Pd, and F elements in the t-AgPdF catalysts remains uniform before and after CV test, and the changes of their respective contents are almost negligible.
On the right side of Figure 7, the HRTEM image of t-AgPdF catalyst after CA test shows a stable twin boundary in a nanoparticle with a size of 6 nm.The strain distribution image in Figure 7f indicates that both dark and red regions are preserved along the twin boundary, indicating that the retained compressive and tensile strains are stabilized along the twin boundary in the t-AgPdF catalyst after CA test.The corresponding FFT and IFFT patterns in Figure 7g,h display two sets of bright spots, circled in red and cyan colors, representing lattice fringes of the (111) crystal plane with interplanar spacing of 0.226 and 0.233 nm, similar to those in the t-AgPdF nanoalloys in Figure 2. The Σ3 twin boundary can also be highlighted between two lattice fringe regions with different extending directions in the IFFT pattern.Figure S12, Supporting Information, shows the full-range XRD patterns of the t-AgPdF catalysts after CV and CA tests, the diffraction peak of the t-AgPdF catalysts after CV and CA test remains unchanged, which indicates the crystalline structures of the t-AgPdF catalysts remain stable and do not undergo reconstruction during the reaction process.The EDX elemental mapping and line profile of the HAADF-STEM image in Figure S10e,f, Supporting Information, show the uniform distribution of three elements and the simultaneous enrichment of Ag and F contents.In a word, the single-twinned structure in the t-AgPdF remains unreconstructed under working condition, and the strain distribution of the t-AgPdF maintains stable before and after the catalysis, indicating that the t-AgPdF with enhanced catalytic properties is resistant to surface restructuring during the FOR.

Conclusion
In summary, the t-AgPdF presents an improved catalytic activity and stability toward the FOR.The t-AgPdF catalyst demonstrated a higher peak current density of 4.6 A mg Pd

À1
, a lower onset potential of 0.44 V, and a higher activity retention of 55.7% after 600 cycles.During long-term potentiostatic polarization tests, the t-AgPdF catalyst maintained a current density of 0.49 A mg Pd

À1
after 3600 s and a longer activity retention time of 55.9 h.Additionally, the t-AgPdF catalyst also exhibited enhanced catalytic performance of hydrogen generation with a volume of generated H 2 gas up to 1.11 mL mg Pd À1 after 120 min of reaction.Furthermore, DFT calculations showed that the t-AgPdF(111) surface had reduced potential limiting of 0.2 eV and activation energy of 0.59 eV for the formate decomposition reaction.Unlike the relaxation of compressive strain in single-crystalline AgPd nanoalloy after FOR, the t-AgPdF maintained the compressive and tensile strain distribution along the Σ3 twin boundaries.These findings suggest that constructing the twinned structure in inverse catalyst can remain the stable strain state after FOR and maintain improved catalytic properties during the formate oxidation.

Figure 1 .
Figure 1.a) HRTEM image of the single-crystalline AgPd nanoalloy (AgPd); the inset image is the atomic model of AgPd.b) Strain distribution of e yy direction for the typical nanoparticle.The blue color represents the significant compressive strain in the AgPd.c) The FFT and d) IFFT patterns of the corresponding region in HRTEM image.e) HAADF-STEM and EDX elemental mapping images of the AgPd nanoalloy.f ) EDX line profiles extracted from the typical lines in HAADF-STEM image.

Figure 2 .
Figure 2. a) HRTEM image of the t-AgPdF, the inset image is the atomic model of t-AgPdF.b) Strain distribution of e yy direction for the typical nanoparticle.The distribution of blue and red regions exhibits the compressive and tensile strains in the t-AgPdF.c) The FFT and d) IFFT patterns of the Σ3 twin boundary structure, showing (111) orientation with d = 0.225 nm is in compressive strained region and that with d = 0.233 nm is in tensile strained region.e) HAADF-STEM and EDX elemental mapping images of the t-AgPdF.f ) EDX line profiles extracted from the typical lines in HAADF-STEM image.

Figure 3 .
Figure 3. Molecular dynamics simulations of AgPd(111) and t-AgPd(111) surfaces.a) Top views and b) left-side views of strain fields of AgPd(111) and t-AgPd(111) surfaces, respectively.The color indicates the strain labeled in the color map.

Figure 4 .
Figure 4. a) CV curves of the FOR for t-AgPdF catalyst compared with that for AgPd catalyst.b) Enlarged CV curves only in positive direction from (a). c) Activity retentions within 600 CV cycles of t-AgPdF and AgPd catalysts.d) CA curves at 0.4 V of the t-AgPdF and AgPd catalysts for 3600 s. e) Long-term CA curves of the t-AgPdF catalyst compared with that of AgPd catalysts.The inset presents a comparison of the duration of activity decay to 0.1 A mg Pd À1 , with t-AgPdF catalyst showing the highest activity retention time up to 55.9 h.

Figure 5 .
Figure 5.The catalytic performance of hydrogen generation for the t-AgPdF and AgPd catalysts.Volume of generated H 2 gas normalized by the mass of Pd versus reaction time of a) 20 min and b) 120 min for the t-AgPdF catalyst compared with that of the AgPd catalyst.c) Free energy diagram of the formate (HCOO) oxidation reaction (FOR) pathway and d) the potential energy profiles for the formate decomposition reaction on t-AgPdF(111) and AgPd(111) surfaces.The t-AgPdF(111) surface exhibits the reduced potential limiting and activation energy.
Figure 6.(left) The AgPd catalyst after the CV test under the FOR: a) HRTEM image, b) strain distribution, c) the FFT, and d) IFFT patterns of the corresponding region in HRTEM image; (right) the AgPd catalyst after the CA test under the FOR: e) HRTEM image, f ) strain distribution, g) the FFT, and h) IFFT patterns of the corresponding region in HRTEM image.

Figure 7 .
Figure 7. (left) The t-AgPdF catalyst after the CV test under the FOR: a) HRTEM image, b) strain distribution, c) the FFT, and d) IFFT patterns of the Σ3 twin boundary structure; (right) the t-AgPdF catalyst after the CA test under the FOR: e) HRTEM image, f ) strain distribution, g) the FFT and h) IFFT patterns of the Σ3 twin boundary structure.