Gas-Mediated Dynamic Structure Evolution of Bimetallic Alloy Catalysts

Abstract

Bimetallic alloys are widely used as heterogeneous catalysts due to their unique physico-chemical properties for improving catalytic reactions. Typically, the structures of alloy catalysts are inherently dynamic under gas environments, which plays a crucial role in their catalytic activity, stability and selectivity. One method of enhancing the catalytic performance of bimetallic nanomaterials is, therefore, to tune or control the surface structure of the nanomaterials, and tremendous progress has been made in this area in the past decade. In this review, we primarily focus on the dynamic structure evolution of binary noble metal alloy catalysts influencing their catalytic performance during the thermal catalytic reaction. First, we summarize the advantage of binary noble metal alloy catalysts and their structure correlation with catalysis. Then, we examine how the structure of precious-metal-based alloy catalysts evolves in response to varying gas environments and the resulting structures impacts on heterogeneous catalytic activity. Further, the advanced characterizing techniques, i.e., in situ scanning/transmission electron microscopy (in situ S/TEM) and near-ambient pressure scanning tunneling microscopy (NAP-STM) are outlined for visualizing these structural evolutions. Finally, we summarize the remaining challenges and outlooks for the future in this research field and offer the potential direction of rational design catalysts with high energy-efficient and sustainable catalytic processes.

1. Introduction

Metal catalysts play a crucial role in heterogeneous reactions [1,2,3,4,5]. In particular, the past decade has witnessed the explosion of research on noble metal (Au, Pt, Pd, Rh, Ir) catalysts in the fields, such as environmental pollution control [6,7,8,9], electrocatalysis [10], and energy and chemical industry [11,12,13,14,15]. However, noble metals are extremely scarce in the earth’s crust. Therefore, reducing the use of expensive metals while improving catalyst performance is a key research project. One effective means is to partly replace them with cheap transition metals to fabricate bimetallic catalysts. This allows for reduced use of precious metals, offering huge significant potential in various catalytic applications.

It was determined that, in bimetallic alloy catalysts, the distinct physico-chemical properties of each metal can enhance catalytic performance beyond that of corresponding monometallic catalysts [5,16,17]. As bimetallic catalysts possess a shared energy band structure of two-metal components, their surface electronic and geometric structures can be continuously tuned by altering alloy composition. Specifically, by tuning the distance between the metal’s d-band center and the Fermi level, their adsorption properties towards reactants and products can be modified, and thus affecting the catalyst’s activity, selectivity, and stability [18,19,20,21]. For example, during a methanol oxidation reaction (MOR), CO is generated and strongly adsorbed on the Pt catalyst surface to reduce the active sites, and thus affecting the catalytic activity and stability for MOR [22]. In the studies of PdPt and PtRu alloy catalysts, PdPt and PtRu alloy catalysts show a bifunctional active site for MOR [23,24]. The introduction of Pd(Ru) into Pt nanoparticles enables Pd(Ru) to form Pd-OH(Ru-OH) species through dissociating H₂O, while Pt catalyzes MOR to produce Pt-CO. Then, Pd-OH(Ru-OH) species react with adjacent Pt-CO intermediates, producing CO₂. These bifunctional sites in PdPt and PtRu alloy catalysts can result in the continuous exposure of Pt active sites, enabling a continuous MOR catalysis. In another case, Yoo et al. used Pt₃M alloy catalysts to oxygen reduction reaction (ORR) [25], confirming the catalytic activity has the following relationship: Pt < Pt₃Zr < Pt₃Co < Pt₃Ni < Pt₃Y. This is attributed to the change in Pt d-band center by the addition of transition metals, affecting the adsorption and desorption of Pt sites for oxygen-containing intermediates.

It was previously demonstrated that the catalyst’s activity, selectivity, and stability vary depending on the structural types of alloy catalysts [16,26,27,28]. According to the structure of bimetallic nanocrystals, they can be classified into rare alloys including single atom alloys, solid solution alloys, intermetallic compound, and core–shell structures. In early studies, there were various mechanisms applied to reveal the structure–activity relationships, such as size effects, geometric effects, synergistic effects, ligand effects, stain effects, and combinations thereof [24,29,30,31,32,33,34,35,36,37,38,39].

Over the past decade, there has been growing recognition that catalyst structures are dynamic during reactions [40,41,42]. In particular, the gas-mediated dynamics of alloy structures have garnered significant attention [43,44,45,46]. During catalysis, reaction environments (temperature, gases, reactant and product molecules) typically result in the complex dynamic changes in catalyst structures. These changes include atomic migration and reconstruction on the catalyst surface, the generation, transformation, or disappearance of dynamic active sites, as well as oscillations in chemical states, sizes, and shapes. For characterizations of dynamic structures of catalysts, in situ techniques, particularly transmission electron microscopy (TEM) and scanning tunneling microscopy (STM) [47,48,49,50,51], have enabled a direct observation of microstructural changes in materials under reaction conditions; these structural variations further alter catalyst’s performances [41,52,53,54,55,56,57,58,59]. Therefore, investigating the structural evolution and performance responses of catalysts under reaction conditions are essential for identifying active sites, elucidating reaction pathways, and understanding the catalytic mechanism.

Since the structure of alloy catalysts is highly sensitive to the atmospheric environments of heterogeneous catalysis, thus this review will first provide a concise overview of how atmosphere environments (oxidizing atmosphere, reducing atmosphere and redox atmosphere) influence the structure of binary noble metal alloy catalysts. It will then focus on the progress of in situ imaging technologies and their thermocatalytic applications. Finally, it aims to offer cutting-edge perspectives on the in situ identification of dynamically evolving surface structures in bimetallic alloy catalysts during heterogeneous catalytic reactions.

2. Gas-Mediated Dynamic Structure Evolution of Alloy Catalysts

Recent advances in atmosphere-mediated structure evolutions of binary noble metal alloy catalysts have been made (Figure 1, and Table S1), such as surface reconstructing, surface segregation, structural phase transition, and compositional alloying. Atmosphere-induced structure evolution is a dynamic process governed by the interplay of thermodynamics and kinetics. This phenomenon is driven by several underlying mechanisms [41,60,61,62,63]. For example, adsorbate-induced structural evolution of alloy catalysts. The chemisorption of gas molecules on the catalyst surface alters the surface electronic structure and surface energy [41,60,63]. To maximize (or minimize) this adsorptive interaction, the system drives components that form stronger (or weaker) bonds with the adsorbate to the surface. Another crucial mechanism is oxidation-/reduction-induced structural evolution. During this process, intensive changes in chemical potential under oxidizing or reducing atmospheres are the primary cause of structural transformation [61,62]. From the early monometallic catalyst studies to the current bimetallic catalyst studies, the atmosphere-mediated dynamic structure evolution of alloy catalysts has completely overturned the traditional understanding of material surface stability and has propelled the development of catalytic science.

2.1. Gas-Mediated Surface Reconstruction

Surface reconstructing of metal catalysts refers to the spontaneous rearrangement of atomic arrangement structure and chemical composition due to difference in the adsorption binding of gas molecule to different metallic component in alloys, alloying surface is typically subjected to surface rearrangement inspired by oxidating, reducing or oxidating- reducing atmosphere at a given temperature.

2.1.1. Oxidation-Driven Surface Reconstruction of Alloy Catalysts

Oxidation-mediated atomic and valent rearrangements are the key driving forces for the dynamic reconstruction of alloy catalysts, which widely appeared in various alloy systems. For instance, in the case of Pt₃Co alloying nanoparticles annealing at a temperature of 993 K for 30 min under the constant pressure of oxygen conditions, surface restructuring of PtCo alloys was unveiled using in situ STEM [64]. Experimental findings showed that oxygen annealing not only induces the initial disordered Pt₃Co alloy to transform into ordered L1₂ intermetallic Pt₃Co alloys, but also promotes Pt surface segregation, forming a Pt-rich shell (Figure 2A). Density functional theory (DFT) calculations showed that this process follows an energetically favorable path. The ordered intermetallic structure of Pt₃Co nanoparticles effectively prevents surface oxidation of Co. The Pt shell on the ordered intermetallic core can thus hinder Co segregation and protect the NPs from surface oxidation. In addition, several previous reports on Ag-based alloys also noted the oxygen-mediated surface restructuring phenomenon [45,65,66]. In a study of Ni-doped Ag alloys by Sykes et al. [45], the combination of the surface science technique and surface molecule adsorption revealed the influence of the oxygen atmosphere on Ni-doped Ag(111) surface, showing that the Ni-doped Ag surface can adsorb O₂ and promote its dissociation (Figure 2B). Through temperature-programmed desorption (TPD) experiments, it was reported that when the Ni-doped Ag surface was exposed to O₂ at 350 K, O₂ dissociated and the oxygen atoms released could migrate to the Ag surface, forming adsorbed oxygen species. These oxygen species desorbed at around 520 K, which is approximately 70 K lower than the desorption temperature on the surface of pure Ag, indicating that Ni doping promotes the adsorption and release of O₂. During this process, the scanning tunneling microscopy (STM) image found that under oxidizing conditions, Ni atoms were located on the Ag(111) surface, but under vacuum conditions, Ni atoms would migrate to the subsurface. This migration behavior indicates the dynamic changes in Ni under oxidizing and reducing conditions, further confirming the high dispersion and migration of Ni on the Ag surface. Using atom-resolved STM, Sykes’s group also indicated the dissociation behavior of O₂ on the surface of AgCu alloy and the changes in surface structure at the atomic scale [65]. In the O₂ atmosphere, the strong interaction between O₂ and Cu atoms reduces the surface energy of Cu in the Ag layer. Therefore, the Cu atoms originally covered by the Ag layer are exposed. With the increase in O₂ exposure, the increase in oxygen coverage will further promote the exposure of Cu atoms, forming the Cu₂O structure. On the basis of a combination of ambient pressure X-ray photoelectron spectroscopy (XPS) and STM, an Ag-capped PdAg surface alloy was studied under the O₂ atmosphere, confirming that exposure to 1 Torr of O₂ at 400 K causes underlying Pd to resurface, and the resulting structure is stable at low pressures and low temperatures (300 K) [66]. Furthermore, in the case of PdAu alloys with dilute Pd, extended X-ray absorption fine structure (EXAFS) spectra proved that at high temperatures (673 K), O₂ induces Pd atoms to segregate from subsurface layers to the surface, forming thermodynamically stable Pd-O structures through combination with oxygen [67].

2.1.2. Reduction-Driven Surface Reconstruction of Alloy Catalysts

Reducing atmospheres, such as H₂ or CO, have been shown be an efficacious tactic for modifying surface structure of alloy catalysts. For instance, it has been reported that H₂ treatment significantly changes the structure of Pd₃Cu alloy catalysts [58]. Using high-resolution STEM imaging, two types of surface segregation modes of Pd₃Cu alloys were revealed for the first time. At 673 K, hydrogen reduction causes Cu enrichment on the catalyst surface. When the temperature rises to 873 K, Pd atoms move to the surface due to the stronger H₂ adsorption on Pd, creating a Pd-rich layer. These structural changes alter the catalyst’s surface composition and electronic properties. Specifically, Cu-rich surfaces make the catalyst’s surface electron-rich, while Pd-rich surfaces make it electron-deficient. These electronic changes greatly affect the catalyst’s activity and selectivity in follow-up reactions. When PdAu alloys with dilute Pd was exposed to a H₂ atmosphere, H₂ promotes the dissolution of Pd into the subsurface of the Au matrix at high temperature (673 K), due to a reliable reason that Pd is more thermodynamically stable in the subsurface, when there is a single adsorbed H present. After CO treatment, more Pd atoms migrate to the surface in PdAu alloys (Figure 3A) [67]. A different study reported the surface restructuring of Pd₂Ga alloys resulting from H₂ molecules [68]. In situ TEM experiments were used to confirm the structuring crystal faces and the Pd placeholder on surface (Figure 3B). The authors found the phase transformation process of Pd₂Ga nanoparticles from Pd to Pd₂Ga in a hydrogen atmosphere, and the dynamic evolution of surface reconstruction of Pd₂Ga alloys when the temperature rises from 673 K to 773 K (such as from (013)/(020) to (011)/(002) crystal planes). At a low temperature of 673 K, the surface of Pd₂Ga is mainly composed of (013)/(020) crystal planes, and Pd atoms on the surface are arranged in a continuous trimer (Pd₃) arrangement. At a high temperature of 773 K, the crystal plane is reconstructed into (011)/(002), and Pd atoms are isolated by the Ga atoms, forming isolated Pd₁ sites.

Another approach to induce surface reconstruction of alloys are techniques using CO treatment. For instance, the surface dynamic reconstruction process of Pd atoms in the PdAu bimetallic catalyst under CO adsorption induction was uncovered through time-resolved temperature-programmed infrared reflectance absorption spectroscopy [69]. Below 320 K, CO adsorption can stabilize Pd single atoms on the surface. Above 320 K, CO desorption and Pd dissolution become the dominant processes. Using CO diffuse reflectance infrared fourier transform spectroscopy (CO-DRIFTS), Sykes et al. pinpointed the active geometric sites (single atoms or clusters) of PdAu bimetallic catalysts in a CO atmosphere [70]. At 303 K, after CO introduction, the CO adsorption peak (2076 cm⁻¹) corresponding to isolated Pd atoms increased over time, while the peak at 2109 cm⁻¹ decreased, indicating that CO drove Pd atoms from the Au bulk to the surface. Higher temperatures (343 K) lead to Pd aggregation. In addition, other studies on PdAu alloys also verified that the strong binding CO to Pd compared to Au results in more Pd atoms migrating to the surface [67,71].

2.1.3. Redox Driven Surface Reconstruction of Alloy Catalysts

It has been reported that surface reconstructing could be modulated by using a cyclic oxidative/reductive process [72,73,74]. In situ DRIFTS spectra have shown that during cycles between CO₂ and H₂ flows, the redox process of surface Fe atoms was observed [72]. In situ DRIFTS showed that Fe⁰ on the surface of Pt-Fe catalyst was partially oxidized to Feⁿ⁺ in CO₂. Fe⁰ generates CO by reducing CO₂, and H₂ reduces Feⁿ⁺ back to Fe⁰. In another case of PdAu alloys, in situ XAS and TEM have shown that the alternating atmosphere of O₂/H₂ can produce a change in structure of surface Pd sites [73]. Specifically, O₂ treatment makes Pd more inclined to distribute on the catalyst surface. The local structure of Pd changed, the Pd-Pd coordination number increased slightly, and the Pd-Au coordination number decreased accordingly. The results show that O₂ treatment promotes Pd surface enrichment to form trimers. H₂ treatment causes Pd to dissolve into the Au matrix. After H₂ treatment, the Pd-Pd coordination number decreases, while the Pd-Au coordination number increases. This indicates that H₂ treatment makes Pd atoms more dispersed in the Au matrix, reducing the aggregation of Pd on the surface to form monomers or dimers. For a CuNi alloy studied using near pressure XPS [74], detailing surface segregation and structural reconstruction (O₂ and H₂) were disclosed under diverse temperatures and gas environments. The oxidative pretreatment at 973 K in O₂ enriches the surface with Ni³⁺ species (Ni₂O₃/NiOOH) and drives the formation of Cu²⁺-enriched cores (Cu(OH)₂ and CuO). Subsequent exposure to a H₂ atmosphere at 673 K further enhances the segregation of metallic Ni to the surface.

2.1.4. Dynamic Structure–Activity Relationship During Reaction

Under catalytic reaction conditions, the accurate identification of the real active structures of alloy catalysts is crucial for the dynamic understanding of structure–activity relationships and the precise design of highly active and stable alloy catalysts. It was previously demonstrated that surface structure of alloy catalysts is dynamic during reaction atmospheres [75,76]. In situ TEM and spectrum characterizations have shown the dynamic structural evolution mechanism of Ni@Au core–shell structured nanoparticles during CO₂ hydrogenation reaction [77]. Traditionally, core–shell catalyst performance was thought to stem from its unique shell structure. However, this study shows that during reactions, the core–shell structure reversibly transforms into an alloy surface (Figure 4). Using environmental TEM (ETEM), Ni@Au nanoparticles were observed in situ in a CO₂/H₂ atmosphere at 9 mbar. The core–shell structure was stable at 723–773 K (450–500 °C) but became a NiAu alloy at 873 K (600 °C), with the core–shell structure reappearing upon cooling. Calculations based on DFT demonstrate that the whole process can be divided into two stages. In the first stage, the CO₂ is hydrogenated at the Ni surface. The second stage is that the energetically favored adsorbed CO molecules diffuse from the Ni site to the Au sites. Meanwhile, CO drove Ni surface migration, achieving 95% of CO selectivity. But this high selectivity is not due to the initial core–shell structure. Instead, it is because the formed NiAu alloy surface, where Ni atoms enable easier CO₂ activation for CO production. Similar surface reconstruction was also observed in Pd₃Cu alloy catalysts during CO₂ hydrogenation reaction [58]. It was found that in the reaction temperature range of 517–673 K, the Cu atoms in the Pd₃Cu catalyst migrate from the metal particles to the SiO₂ shell under the strong adsorption of oxygen-containing reaction intermediates and form small Cu clusters. This structural change leads to the formation of a unique core@shell@satellite (CuPd@SiO₂@Cu) structure (denoted as Cu₃Pd-673). The formation of this structure has an important influence on the activity of the catalyst. The specific manifestation is that the catalyst with enriched Cu surface shows low activity. This might be because the Cu enrichment on the surface weakens the dissociation ability of H₂, and the enriched-electron Cu sites reduce the activation ability of CO₂. By contrast, H₂-produced Pd₃Cu alloys with a Pd-rich surface at 873 K (denoted as Cu₃Pd-873) show higher activity. This is because the presence of Pd atoms can effectively dissociate H₂ and transfer the activated H atoms to the Cu surface through the electron overflow effect, thereby enhancing the hydrogenation reaction performance of the catalyst. In detail, within the temperature range of 517–673 K, the CO₂ conversion of the Cu₃Pd-873 catalyst was significantly higher than that of the Cu₃Pd-673 catalyst. When the temperature rose from 517 K to 673 K, the CO₂ conversion of the Cu₃Pd-673 catalyst gradually increased from 0.3% to 8.6%, while the CO₂ conversion rate of the Cu₃Pd-873 catalyst significantly increased from 1.3% to 12.7%. This indicates that the Cu₃Pd-873 catalyst exhibits higher activity in the CO₂ hydrogenation reaction.

2.2. Gas-Mediated Compositional Segregation

Atmosphere-mediated phase segregation in alloys has become a cutting-edge research topic in materials science and surface chemistry. The segregation of elemental compositions occurs under non-equilibrium conditions caused by chemical potential differences in specific atmospheres, forming two-phase interfaces. This offers a new perspective for controlling catalyst dynamic behavior.

2.2.1. Alloy Surface Segregation Under Heating Conditions

High-temperature treatment can cause the diffusions between heterogeneous atoms in alloys and thus result in variation in partial composition in alloys. Combined with in situ STEM and energy dispersive X-ray spectroscopy (EDX), Mirsaidov et al. observed the phase separation of AuRu alloy nanoparticles during heating increased from 623 K to 1023 K [78]. At the initial stage of phase separation, metastable phase decomposition of the alloy leads to the formation of small face-centered cubic (fcc) Ru clusters. When the temperature is increased, as the fcc Ru clusters grow larger, they undergo a phase transition to the more stable hexagonal close-packed (hcp) Ru structure. This transformation is attributed to the size-dependent competition between the interfacial and bulk energies of Ru domains. Under heating conditions of 973 K, the atomic diffusion of Au and Ru produced four Ru-rich domains. This finding showed that fcc AuRu alloy nanoparticles (containing about 85% Au and 15% Ru) will undergo phase separation, forming an fcc Ru intermediate phase from the fcc AuRu alloy, and finally forming a Ru phase with an hcp crystal structure.

2.2.2. Oxidation-Driven Surface Segregation

Previously, the dependency of surface segregation of alloy catalysts on oxygen conditions that was used for the treatment of catalysts has been observed in AuCu and PtCo alloys [79,80]. Using ETEM, Park et al. revealed that in an O₂ atmosphere, as the temperature increased, Cu atoms migrated from the interior of the AuCu nanoparticles to the surface and oxidized to form CuO_x species [79]. In this process, after O₂ was introduced at room temperature, the amorphous coating on the surface of the AuCu nanoparticles was removed and Cu₂O clusters were formed; as the temperature increased to 373 K, the Cu₂O clusters transformed into CuO phases, and the Au-Cu nanoparticles became smaller and flatter. These results indicate that the oxygen atmosphere treatment induced atomic rearrangement and surface structural segregation of AuCu nanoparticles, forming a CuO_x/Au heterostructure. The AuCu/SiO₂ catalyst (O₂-AuCu) pretreated with oxygen exhibited a significantly enhanced catalytic performance in CO oxidation reactions. Likewise, under the oxygen atmosphere, through in situ X-ray absorption spectroscopy (XAS) and STEM techniques (Figure 5), the dynamic structural evolution mechanism of Co₂Pt₃ nanocatalysts under the O₂ atmosphere was systematically revealed [80]. STEM observations found that after O₂ treatment, CoPt nanoparticles formed a distinct core–shell structure, with the shell being Co oxide (presumably Co₃O₄) and the core still being a CoPt alloy. Electron energy loss spectroscopy (EELS) analysis further confirmed the presence of oxygen in the shell, indicating that Co migrated to the surface during oxidation and formed an oxide layer, causing Pt atoms to be more confined inside the particles, thereby reducing the sites of Pt exposed on the catalyst surface.

2.2.3. Reduction-Driven Surface Segregation

How reduction treatments affect the compositional segregation for alloy catalysts has also been studied. For instance, using CO TPD and in situ CO polarization modulation infrared reflection absorption spectroscopy (PM-IRRAS) techniques to monitor the changes in surface composition and structure of Cu-Ni/SiO₂ catalyst under changed CO pressure and temperature, Goodman et al. revealed that CO adsorption can induce Ni segregation on the surface of Cu-Ni/SiO₂ bimetallic catalysts under normal pressure [81]. During this process, Cu is enriched on the surface of the CuNi alloys under ultra-high vacuum conditions. When the CO pressure was increased to the normal pressure, CO adsorption induced Ni segregation on the surface of the CuNi alloys. In situ PM-IRRAS results showed that under normal pressure CO conditions, the structure changed from Cu surface enrichment under UHV conditions to Ni surface enrichment under normal pressure CO, forming a surface Ni coating layer. Direct capture of the real-time structural evolution of PtPb@Pt nanosheets in CO environment has also been observed in situ TEM for the first time [82]. In a CO atmosphere (0.1 mbar, 623 K), the sharp corners of the hexagonal nanoplates gradually became rounded over time, and new lattice constants and orientations began to appear in some areas. After 10 min, these areas accounted for about 8%; after 75 min, the proportion increased to about 30% (Figure 6). High-resolution transmission electron microscopy (HRTEM) showed that the lattice spacing in these areas corresponds to the (220) crystal plane of Pt, indicating that CO induced the surface enrichment of Pt. During this process, lattice fringes generated by Pb-CO species were observed at the edge of the nanoplates, consistent with the structure of Pb(CO)₄. This suggests that CO interacts strongly with Pb, pulling out single Pb atoms to form Pb(CO)₄ species, which eventually accumulate to form an ultrathin Pb(CO)₄ overlayer.

2.2.4. Redox-Driven Surface Segregation

As mentioned previously, surface segregation could be modulated by using a cyclic oxidative/reductive atmosphere [83,84]. In the study of AuPd alloy [83], the authors found that after treatment in the air atmosphere at 773 K, the size of AuPd nanoparticles increased from 2.5 to 3.5 nm to 5.5 nm. At the same time, Pd atoms migrated to the surface to form an Au-rich core and Pd shell structure. When the Pd content increases, the surface Pd is further oxidized to form a PdO shell. After H₂ treatment at 773 K, PdO is reduced to metallic Pd, and a Pd-rich surface is formed on the surface. Using in situ TEM, Stach et al. revealed the dynamic structural evolutions of Au₀.₇₅Pd₀.₂₅ alloy nanoparticles under different gas environments (O₂, H₂, CO and CO₂) and elevated temperature conditions. O₂ environment (673 K, 1 bar) induced the Pd-enriched surface to form PdO, followed by H₂ treatment leading to PdO being reduced to metallic Pd. However, replacing H₂ with a CO atmosphere and increasing the temperature from room temperature to 473 K causes the destruction of the surface crystal structure of nanoparticles, the particle surface becomes irregular, and the surface roughness increases significantly. This is because the CO molecules adsorbed on the particle surface can change the surface energy, and thus cause surface segregation. When employing a CO₂ atmosphere rather than CO, no obvious surface variation was observed for nanoparticles exposed to CO₂ at 523 K, indicating that the interaction between CO₂ and Pd or Au is weak, and it is difficult to induce changes in the surface structure of nanoparticles. Similar studies on the redox cycles of alloy catalysts have been performed under H₂, CO, NO, and O₂ conditions using a near-ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) [52]. Using this surface technique by Somorjai et al., the elemental composition and chemical state changes on the surface of Rh-Pd nanoparticles are monitored under different NO, O₂, CO and redox atmospheres (Figure 7A). Under oxidative conditions (100 mtorr NO or O₂), Rh is almost completely oxidized in the surface layer and segregates to form a RhO-rich surface layer. Under a reducing atmosphere, Pd migrates to the surface, while Rh migrates to the interior. In the catalytic reaction (CO + NO), the chemical state of Rh also changes significantly, and RhO_x is reduced to metallic Rh⁰. The experiment also found that this structural change is reversible. When CO is removed and NO is retained, Rh migrates back to the surface and is oxidized again. Compared with Rh_0.5Pd_0.5, the shell of Pt_0.5Pd_0.5 nanoparticles is mainly composed of Pd under oxidative conditions. Under catalytic and reducing conditions, PdO_x is significantly reduced, but no obvious metal segregation is observed. In the studies on CuNi/SiO₂ and CuCo/SiO₂ catalysts [85]. In H₂ or O₂ atmospheres, Ni is concentrated on the surface of CuNi nanoparticles; in the O₂ atmosphere, Cu is enriched on the surface. In the presence of O₂, the oxidation state of Cu increases significantly, while the oxidation state of Ni does not change much. NAP-XPS results show that this surface enrichment phenomenon is reversible when re-exposed to H₂, indicating that the surface structure of CuNi nanoparticles is sensitive to changes in atmosphere. In contrast, in an O₂ atmosphere, Cu in CuCo alloys is irreversibly enriched on the surface, and the surface remains Cu-enriched even when re-exposed to H₂. NAP-XPS results show that the proportion of Cu in the surface layer increases significantly after O₂ treatment, and this enrichment does not reverse when re-exposed to H₂, while in H₂, part of the oxidation state is reduced, but the Cu enrichment still exists.

In addition, Zhang et al. employed in situ TEM and in situ scanning tunneling microscopy (STM) to unveil the micro-segregation behavior of PdFe alloys at the atomic scale (Figure 7B) [86]. The results indicated that the PdFe alloy phase formed under a hydrogen atmosphere exhibits a surface alloy structure of PdFe, suggesting that Fe atoms preferentially migrate to the sub-surface layers of Pd nanoparticles, thereby forming a Pd@PdFe core–shell configuration. During oxidative segregation, embedded two-dimensional FeO nano-islands were initially formed. With increasing temperature and pressure, these FeO nano-islands migrated from the embedded regions to the surface while maintaining a two-dimensional (2D) layered morphology due to strong metal–oxide interactions (SMOIs). In contrast, under a CO atmosphere, the strong interaction between CO and Fe led to the preferential formation of three-dimensional Fe nano-islands. As temperature and pressure rose, these islands underwent a structural transition into three-dimensional (3D) aggregated pinnacle-like features, revealing the microscopic pathway of phase separation in PdFe systems. Furthermore, the stronger binding of gaseous CO with Fe atoms stimulates Fe segregation out of the PdFe dilute alloys, resulting in 3D growth of Fe islands, whereas the dissociative adsorption of O₂ results in 2D layer-by-layer growth of segregated FeO as encapsulation overlayers that bind strongly with the Pd surface underneath. By alternating between CO and O₂ atmospheres, the segregation structure of Fe species demonstrate reversible characteristics.

2.2.5. Dynamic Structure-Activity Relationship During Reaction

Changes in the surface segregation of alloy nanoparticles during the real reaction have also been reported [87,88]. During a CO oxidation reaction, the structure of the PdCu nanoparticles is strongly related to the crystal plane of the carrier (Figure 8A) [87]. On the PdCu nanoparticles supported by CeO₂ cubes (CeO₂-C), the phase separation occurs, leading to a two-phase Cu₂O/PdCu structure formation, while the nanoparticles supported on CeO₂ octahedrons (CeO₂-O), a uniform CuPdO_x phase is formed. For catalysts supported by CeO₂-C, the Cu₂O/PdCu interface formed can promote the efficient conversion of CO. While for PdCu catalysts supported by CeO₂-O, the CuPdO_x phase formed shows a low activity compared to PdCu/CeO₂-C catalyst. This was mainly attributed to the structural differences formed during reaction process. Using in situ TEM, similar study on PdCu alloy catalysts found that in an O₂-rich atmosphere (CO oxidation reaction), PdCu alloys undergo phase separation to form Pd-CuO structures [88].

With in situ DRIFTS, H₂-TPR, CO-TPR, EPR, and isotope labeling experiments, it has also been observed that PtCu alloys in PtCu/MgO catalysts were subjected to dynamic structural evolution in a CO oxidation reaction [89]. At low temperature (<423 K), CuO_x species are formed on the surface of PtCu alloys, which promotes CO oxidation. CuO_x species can provide abundant active oxygen species and significantly enhance the CO oxidation activity at low temperatures. At high temperatures (>423 K), PtCu alloys synergize with surface CuO_x species to promote CO oxidation. In follow-up studies, NAP-XPS was utilized to reveal surface segregation of PtCo alloys under the conditions of the CO oxidation reaction [90]. NAP-XPS results showed that as the temperature increases (Figure 8B), Co gradually oxidizes from a metallic state into CoO and finally forms CoO. Pt remains in a metallic state throughout the reaction, indicating the formation of a Pt-CoO interface. The as-formed Pt-CoO interfaces significantly enhanced their catalytic activity for CO oxidation compared to the single-metal Pt and Co catalysts. Similarly, dynamic structural changes in heterostructured Ni-Rh nanoparticles were also found in the CO oxidation reaction [91]. In situ TEM revealed that under oxygen-rich conditions (O₂-rich), the Ni core began to become empty at 673 K, and the overall size increased, as well as part of the Ni was converted to NiO. While under oxygen-poor conditions, the contrast of Rh nanoparticles gradually decreased above 523 K, and the overall contrast was more uniform at 673 K, indicating that the Rh nanoparticles alloyed with the Ni core to form a NiRh alloy surface. The Ni-Rh nanoparticles under oxygen-rich conditions indicate high catalytic activity at low temperature, which is much higher than that of pure Ni and Rh nanoparticles. Under oxygen-poor conditions, the activity of NiRh nanoparticles decreased. This is because the surface of the NiRh alloy is more susceptible to CO adsorption, resulting in CO poisoning, which reduces the catalytic activity.

Figure 8. Structure–activity relationship of (A) PdCu/CeO₂ catalysts [87] and (B) PtCo catalysts during CO oxidation [90].

Figure 8. Structure–activity relationship of (A) PdCu/CeO₂ catalysts [87] and (B) PtCo catalysts during CO oxidation [90].

2.3. Gas-Mediated Dynamic Phase Transition

Atmosphere-mediated structural phase transition in alloys holds great importance in condensed matter physics, materials chemistry, and catalytic science. In this process, alloy materials, when exposed to specific atmospheres, undergo a transformation in their crystal structure. This is triggered by differences in chemical potential, the diffusion barriers, and interfacial energy of different elements, resulting in the atomic rearrangement that is entirely different from the original crystal structure.

2.3.1. Alloy Phase Transition Under Vacuum or H₂ Atmosphere

It was previously demonstrated that alloy phase transformation can emerge under vacuum or an inert atmosphere [92]. Employing in situ heating TEM under vacuum conditions (Figure 9A), Utkur Mirsaidov et al. elucidated the dynamic mechanism of the fcc-to-bcc phase transition in single-crystalline PdCu alloys [92]. They observed that the bcc phase originates at the edges of fcc nanoparticles and propagates inward to form an exclusive bcc phase. At 773 K, metastable fcc nanoparticles transform into the ordered B₂ bcc phase. Similar fcc-to-bcc phase transition in PdCu alloys was also realized by the effective method reported by Shen et al. [93]. The authors subjected the initial PdCu colloid to hydrogen treatment at 673 K, inducing a phase transformation to the ordered B₂ bcc structure (Figure 9B). This process enabled the formation of an ordered arrangement of Pd and Cu atoms within the nanoparticle. Additionally, to transform the B₂-phase PdCu nanoparticles into the disordered fcc phase, the authors first treated them with oxygen at 673 K, followed by the reduction with hydrogen at 773 K. In the resulting fcc phase, Pd and Cu atoms are randomly mixed without forming ordered Pd-Cu bonds.

In a study using in situ XRD, it has been reported that the changes in the PdCu crystal structure during annealing under a N₂ atmosphere reveal the phase transformation process from fcc to bcc. At low temperature (<573 K), the heat annealing treatment mainly forms PdCu with an fcc structure [94]. When temperature increases from 573 K to 773 K, the heat annealing treatment causes the PdCu to transform from the fcc phase to the bcc phase. When the annealing temperature reaches 773 K, the bcc phase becomes the dominant structure. Further increasing the annealing temperature (>773 K) will lead to a reduction in the bcc phase and the reappearance of the fcc phase, accompanied by particle aggregation and size increase. In addition, it has been reported that the structural conversion between disordered and ordered alloys in different gases conditions [95]. Subjecting disordered PdCu alloys (D-PdCu/C) to high-temperature treatment in a hydrogen atmosphere promotes the formation of atomically ordered structures. Annealing the D-PdCu/C catalyst in hydrogen at 773 K yielded the ordered PdCu/C catalyst. However, in a follow-up study [96], under a N₂ atmosphere, the atomically ordered Pd₃Bi was subjected to thermal annealing at 773 K to transform it into disordered Pd₃Bi alloys.

2.3.2. Dynamic Structure–Activity Relationship During Reaction

It was generally accepted that there is a close relationship between structural phase changes and catalytic performance for alloy catalysts. Adjusting the alloy’s phase structures can effectively optimize catalytic performance. This offers new methods for developing high performance catalysts. To investigate the relationship between alloy structural ordering and catalytic performance, the work by Zheng et al. [97], provides compelling experimental evidence. They first synthesized disordered PdCu alloy nanoparticles via a chemical reduction method. Subsequently, annealing these particles under a H₂ atmosphere successfully induced a structural phase transition, resulting in the formation of an atomically ordered PdCu structure. This transition from disorder to order significantly enhanced the alloy’s catalytic activity. Within the ordered PdCu structure, Cu and Pd atoms show a specific arrangement, generating a high density of Cu-Pd pairs. These Cu-Pd pairs act as catalytic active sites, significantly boosting the surface coverage of the CO intermediate. In the electrocatalytic CO reduction reaction, experimental testing revealed that at an applied potential of −1.03 V (vs. the reversible hydrogen electrode, RHE), the ordered CuPd catalyst achieved a partial current density for acetate of 425 mA cm⁻². This value is substantially higher than the performance observed for pure Cu nanoparticles and other samples. These results clearly demonstrate that the ordered structure of CuPd significantly boosts the activity for acetate production from CO₂ reduction. This work strongly validates the critical role of structural ordering in enhancing the catalytic performance of alloy catalysts.

Through thermal annealing of fcc PdCu/C under an Ar/H₂ atmosphere, Huang et al. achieved a phase transformation from an fcc structure to the bcc structure. Complete conversion to the bcc phase occurred at 623 K [98]. This change in crystal phase significantly affected the electrocatalytic performance for the nitrogen reduction reaction (NRR). The bcc PdCu catalyst exhibited superior electrocatalytic activity, achieving a NH₃ production rate of 35.7 mg h⁻¹ mg⁻¹_cat with a Faradaic efficiency (FE) of 11.5% at −0.1 V versus the reversible hydrogen electrode (RHE). Importantly, it demonstrated high selectivity, with no hydrazine (N₂H₄) detected. In contrast, the fcc PdCu showed a substantially low NH₃ yield, highlighting that the alteration in crystal phase structure markedly enhances the catalyst’s activity. Furthermore, the bcc PdCu catalyst exhibited excellent long-term stability. After five consecutive electrolysis cycles, no significant decay in current density was observed, and the NH₃ production rate and Faradaic efficiency showed only marginal changes.

In another case [99], Zhang et al. employed in situ XRD, XPS, and ETEM techniques to investigate in detail the temperature-dependent phase evolution of the PdO/ZnO/Al₂O₃ catalyst under syngas (CO/H₂). At low temperatures (<393 K), PdO is initially reduced to β-PdH_x hydride. As the temperature increases to 393–623 K, β-PdH_x gradually decomposes into α-PdH_x, and a surface-alloyed intermediate begins to form at the grain boundaries. Upon exceeding 623 K, this intermediate state facilitates the continuous incorporation of zinc and carbon, ultimately leading to the formation of a well-defined Pd₃ZnC_x intermetallic carbide phase. This intermetallic carbide phase demonstrates exceptional catalytic performance in the selective hydrogenation of acetylene. It maintains high selectivity (>90%) towards the desired C₂H₄ product and exhibits superior long-term stability. These performance enhancements are primarily attributed to the optimized electronic structure and geometric advantages of the Pd₃ZnC_x phase (Figure 10).

2.4. Gas-Mediated Compositional Alloying

Atmosphere-mediated alloying in bimetallic catalysts refers to bimetallic nanoparticles undergoing atomic-scale mixing under specific reducing environments, forming an alloy phase with homogeneously distributed compositions. This atmosphere-mediated alloying process can significantly alter the catalyst’s electronic structure and geometric configuration. Consequently, it provides a novel approach for designing and optimizing new generations of highly efficient alloy catalysts.

2.4.1. Reduction-Driven Alloying of Bimetallic Nanoparticles

Changes in the composition distribution of Pd-based bimetallic nanoparticles upon hydrogen have been reported by Lu et al. [100]. In situ characterizations including XPS, EXAFS, XRD and STEM were used to study the changes in the surface structure of Pd based catalysts in a H₂ atmosphere. At 573 K, the hydrogen reduction of InO_x-coated PdCu catalysts prompts InO_x to move from the SiO₂ support to the PdCu nanoparticle surface, thus creating PdIn alloys with Pd surface enrichment. Using similar characterizations, the structural changes in Pd@Ga₂O₃/Al₂O₃ catalyst were also revealed under a H₂ atmosphere, showing that the Pd/Ga₂O₃ catalyst mainly forms the Pd₂Ga phase, while the Pd@Ga₂O₃/Al₂O₃ catalyst forms both Ga-rich PdGa₅ and PdGa alloy phases [101]. In addition, it was found that the structure of Pd@Au core–shell nanoparticles changes significantly under different CO pressures [102]. Under low CO pressure (below 10⁻³ mbar), CO adsorption absorbs predominantly on the Pd core. As CO pressure increases, CO adsorption triggers the diffusion of Pd atoms from the core towards the surface of the Au shell, leading to the formation of a Pd-Au alloy layer. When CO pressure exceeds 10⁻³ mbar, Pd atoms driven by CO adsorption migrate further to the shell surface, resulting in the formation of Pd dimer structures.

On the other hand, thermal annealing in a H₂ atmosphere promotes nanoparticle alloying, such as Pt-based and Au-based catalysts [103,104,105,106]. Wang et al. proposed a general strategy for synthesizing high-entropy alloys (HEAs) under mild conditions by leveraging strong metal–support interactions (SMSIs) [104]. This work revealed that SMSIs promote the formation of multi-component alloys (MCAs) at significantly lower reduction temperatures (673–873 K), such as the synthesis of compositionally complex PtPdCoFe HEAs supported on anatase TiO₂. Using S/TEM studies of Pt@Ru core–shell catalysts treated by H₂ atmospheres [103], Shao et al. revealed that at 373 K, the Pt precursor is preferentially reduced to Pt nanoparticles, while most of Ru precursors remain as atomically dispersed species. At 573 K, a well-defined Pt@Ru core–shell structure forms, with Ru encapsulating the Pt core. As the annealing temperature is further increased to 723 K and 873 K, Pt atoms diffuse outward to the particle surface, ultimately forming a disordered PtRu alloy. Shen et al. revealed distinct structural transformations in Pt/TiO₂ catalysts under H₂ atmosphere at 773 K, which depend on Pt particle size [105]. For 2 nm Pt particles, a bulk Pt−Ti alloy formed. In contrast, 4–6 nm Pt particles formed a Pt₃Ti intermetallic compound, forming a Pt₃Ti@Pt core–shell structure. When the reduction temperature was further increased to 873 K, Pt particles in the 2–6 nm size range were fully converted into well-crystallized Pt₃Ti intermetallic compound structures. In a follow-up study [106], Cu@Au core–shell nanowires undergo a disorder-to-order phase transformation upon thermal annealing in a H₂/Ar atmosphere, ultimately forming Cu₃Au intermetallic compound nanowires. The results demonstrate that within the temperature range from 413 K to 593 K, the nanowires progress through three distinct stages. Below 473 K, partial alloying of Cu and Au predominates; a mixture of tetragonal CuAu (L1₀ phase) and Cu phase emerges between 473 K and 493 K; above 493 K, the Cu₃Au phase (L1₂ structure) becomes predominant.

2.4.2. Dynamic Structure–Activity Relationship During Reaction

In a recent study on Pd/FeO_x catalysts [107], in the CO₂ hydrogenation reaction, Du et al. revealed the dynamic evolution process of the Pd-FeO_x catalyst, particularly highlighting the “domino effect” between the reaction network and the structural reorganization of the catalyst (Figure 11A). The 5% Pd/FeO_x catalyst was studied with in situ XRD, XPS, DRIFT, and STEM during reaction. The results indicated that Pd particles with larger sizes (4–5 nm) induce a reactive metal–support interaction (RMSI), leading to the formation of a Pd₃Fe alloy. This alloy formation significantly alters the catalyst’s surface composition, promoting a shift in the reaction pathway from the redox pathway involving direct CO₂ dissociation to the hydrogenation pathway proceeding through a HCOO⁻ intermediate, increasing the yield of CO. Ryoo et al. reported the dynamic evolution behavior of Cu-Ni catalysts during CO₂ hydrogenation reaction [108]. Utilizing characterization techniques such as in situ XPS and DRIFTS, researchers discovered that under reaction conditions, Cu atoms migrate from the core to the surface. This migration transforms the initial Cu@Ni core–shell structure into a Cu-Ni alloy structure. NAP-XPS experiments revealed that at room temperature, the surface layer of Cu_0.5Ni_0.5 nanoparticles is Ni-enriched, exhibiting a Cu/Ni atomic ratio of 0.25. As the reaction temperature progressively increases, the Cu/Ni atomic ratio rises to 0.91, indicating significant surface segregation of Cu from the core to the surface and the creation of new active sites (Figure 11B). Following this structural alloying, the catalyst’s surface adsorption capacity for reaction intermediates is altered. For instance, in CO₂ hydrogenation, the Cu-Ni alloy surface more effectively adsorbs and activates CO₂ molecules, facilitating their conversion into adsorbed CO₂ species. Concurrently, the dissociation and adsorption of H₂ are enhanced, enabling more hydrogen atoms to participate in the reaction, thereby boosting overall catalytic activity. A further example of Pd-Pt alloying has been reported [109]. In situ XRD results have shown that at a reduction temperature of 523 K (250 °C), the Pd-Pt/In₂O₃ catalyst exhibited significant structural changes, indicating the formation of a PdPt alloy. Under CO₂ hydrogenation reaction, in situ XRD results showed that the PdPt alloy phase remained stable at 573 K (300 °C), indicating that this alloy structure has good stability during the reaction process.

3. In Situ Characterizations for Dynamics of Bimetallic Alloys

In the fields of materials science, energy, and catalysis, the resolution of numerous critical questions remains constrained by the limitations in detection thresholds, and instrumental sensitivity. These persistent challenges, however, can be fundamentally addressed through advanced methodologies in electron microscopy and surface science. Recent breakthroughs in in situ aberration-corrected transmission electron microscopy (in situ TEM) and surface probe techniques such as in situ or near-ambient pressure scanning tunneling microscopy (in situ/NAP-STM) deliver the experimental foundation for rationally designing nanoscale materials that tune performance, and for revealing the dynamic mechanisms that govern catalysis.

3.1. In Situ TEM Characterization

The wave-like nature of electrons, proposed by French physicist Louis de Broglie in 1923, provided the theoretical foundation for using electrons rather than visible light to probe the microstructure of matter [110]. Building on this insight, German scientists Max Knoll and Ernst Ruska constructed the first TEM in 1932, achieving a resolution better than 2 nm [111]. This breakthrough opened an era of rapid TEM development. So far, high-resolution TEM routinely delivers 0.2 nm resolution, and the advent of spherical-aberration correctors has pushed the limit below 80 pm, enabling direct, atom-scale visualization of material structures.

Recent advances in environmental/in situ aberration-corrected transmission electron microscopy (in situ TEM) and dedicated in situ specimen holders allow researchers to visualize, at atomic or molecular resolution, how a material’s microstructure evolves under realistic and multi-physical stimuli such as gases, heat, liquids, light, electric fields, and mechanical stress. Catalysis has emerged as one of particular application areas [42,46]. By engineering “micro-reactors” inside the microscope column, the entire catalytic process can be monitored in real time, thus mimicking operating conditions and yielding comprehensive and unambiguous insights into complex reaction pathways.

Over the past decade, many research groups have exploited in situ TEM to track gas- mediated structural dynamics of supported metal catalysts. The dynamic structural evolution of nanocrystals has been captured in situ, such as sintering and agglomeration, redispersion, phase segregation, alloying, and reconstruction phenomena [75,100,112,113,114,115]. These observed unconventional behaviors advance the study of structure–performance relationships in catalysis. In a typical example [116], Wang et al. delivered the anatase TiO₂ featuring a distinctive (1 × 4) reconstruction on its (001) facet. By leveraging highly ordered rows of 4-coordinate Ti atoms as active sites, they successfully visualized the dissociation of water molecules and the water–gas shift reaction process. Using in situ heating atomic-resolution STEM [117], Yun et al. have unveiled the atomic behavior and structural evolution of Pt−Sn alloys during distinct phase transformation processes (Figure 12). Their observations reveal that during the phase transition from PtSn₄ to PtSn₂, an intermediate phase (PtSn₂) begins to form as the temperature increases to 473 K. When the temperature reaches 773 K, the transformation from PtSn₂ to Pt₂Sn₃ is initiated, with Pt₂Sn₃ nucleating on the surface of PtSn₂ and subsequently growing along specific crystallographic directions. Above 773 K, PtSn₂ undergoes an isotropic phase transformation into PtSn species.

3.2. NAP-STM Characterization

In 1981, Binning, Rohrer et al. at the Zurich Institute developed the scanning tunneling microscopy (STM) based on quantum tunneling principles [118]. This breakthrough not only heralded humanity’s entry into the atomic-scale era of material characterization but also enabled atomic-scale material manipulation and single-atom control. The technique exhibits three defining advantages, such as atomic-scale resolution (~0.01 nm), direct visualization of local surface microstructures, and operational versatility across diverse environments (vacuum, ambient air, room temperature). This achievement provided the first real-time observation of individual atoms’ spatial arrangements on material surfaces and their associated electronic behaviors. It has generated profound implications for fundamental research in surface science, materials engineering, and so on.

Breakthroughs in NAP-STM have driven substantial advances in catalysis research in recent years [118,119], such as alloy surface segregation phenomenon, distribution of active sites in single-atom catalysts, identification of oxide defect sites and adsorption/desorption dynamics of catalytic reactants [51,120,121,122]. With high resolution and in situ characterization capabilities, STM provides fundamental insights into surface structure and active sites of catalysts, and reaction mechanisms at the molecular level. Berndt et al. studied the adsorption of C₆₀ on Au(111), finding that 49 of C₆₀ molecules formed a Si(111)-7 × 7-like structure, as shown in Figure 13 [123]. They proposed a model that substrate-induced moiré-effect-like weeny orientation changes of C₆₀ molecules allow the √589 × √589 R14.5° superlattice to offer epitaxial sites. These sites optimize intermolecular interactions, which is key for the superlattice’s stability.

In a case of Cu₂O thin film [50], Baber et al. used NAP-STM to study the evolution process of Cu₂O/Cu(111) in a CO atmosphere over time. Figure 14A depicts a fully oxidized Cu₂O surface. Upon exposure to a CO atmosphere for 281 s, metallic Cu domains emerge at the step edges (Figure 14B). Progressive expansion of these metallic regions occurs with prolonged CO exposure, as shown in Figure 14C–G. The results reveal the continuous growth of metallic Cu domains along both step edges during reduction. This evolution indicates the directional migration of Cu atoms to step edges, driving phase segregation between oxide and metallic components to ultimately form metallic Cu steps.

4. Summary and Outlook

In summary, binary noble metal alloy catalysts have always been a hot topic in the field of heterogeneous catalysis. They hold the unique effects between two metals compared to corresponding single-metal catalysts, such as crystalline surface effects, synergistic effects, ligand effects, stain effects, and combinations thereof, which could tune the surface reactivity for enabling them to achieve superior catalytic activity, selectivity, and stability. The prospects of using coupled in situ imaging (TEM and STM) and in situ spectroscopy techniques to distinguish the dynamic correlation between structure and activity of alloy catalysts have been summarized above. The systematic framework for understanding the dynamic structural evolution of alloy catalysts under reactive gas environments was provided. We delve into fundamental phenomena such as surface segregation, phase separation, lattice transformation, and alloying processes. Despite there having been remarkable advances in the controlled syntheses, catalytic performances, and in situ characterizations of bimetallic alloy catalysts in recent years, following research efforts are still needed in the following aspects.

(1)

Improving the structural stability of alloy catalysts under reaction conditions. Alloy catalysts, when subjected to high temperatures or prolonged reaction conditions, are prone to structural degradations such as particle sintering, and surface coarsening, which inevitably leads to a decline in catalytic activity and selectivity. Consequently, enhancing the stability the active structure of alloy catalysts remains a critical research priority during reaction.

(2)

Overcoming in situ TEM resolution and sensitivity limitations caused by environments and electron beam effects. When conducting in situ imaging characterizations, the clear differences between the experimental environments and the actual working conditions of catalysts must be carefully considered. These differences may include the reactor dimensions, gas pressure, flow rate, and other parameters. In in situ TEM, gas molecule scattering and the window membrane in gas cell holders can adversely affect imaging quality. Moreover, electron beam irradiation can induce structural modifications in catalysts through the electron transfer process and localized heating effects. Taking these factors into account, in situ TEM techniques still require further advancements in both resolution and sensitivity, particularly in achieving higher temporal and spatial resolution, to more accurately capture the dynamic structural evolution of alloy catalysts under realistic operating conditions. Bridging the gap between the model in situ studies and industrial conditions now remains a key direction for the predictive design of active catalysts.

(3)

Developing the advanced and intelligentized methodology for automated tracking of atomic or nanoparticle trajectories. In situ TEM experiments of catalysts can generate massive volumes of image and video data, which is used to capture dynamic structural and compositional changes at high temporal and spatial resolution. The enormous quantity and diversity of these datasets often exceed the capacity of conventional manual analysis methods, which are typically time-consuming and prone to the existence of subjective bias. Therefore, developing fast, accurate, and efficient data processing methodologies are needed, such as machine-learning-assisted image recognition and advanced image recognition algorithms for the automatic tracking of atomic or nanoparticle trajectories.