Constructing Intermediate Encapsulation on Ru/CeO₂ Nanocatalysts to Enhance Ammonia Synthesis

ABSTRACT

Engineering the metal‐support interfacial structure in supported heterogeneous catalysts opens wide opportunities to tailor catalytic performance, particularly by exploiting strong metal‐support interactions (SMSI) that modify active‐site electronic and geometric properties to tune adsorption energetics, activity and stability. However, designing SMSI that simultaneously meets the geometric requirement of sufficient active site exposure, and the electronic requirement remains challenging. Here, we achieved an intermediate encapsulation by forming crystallographically defined CeO_x overlayers on Ru/CeO₂ via a Ru‐precursor‐mediated route. Using Ru₃(CO)₁₂ as the precursors creates a locally reducing environment on CeO₂{100}, inducing localized surface modification and formation of truncated‐pyramid‐shaped SMSI overlayers on Ru with a height of ∼2–3 atomic layers (Ru/CeO₂─CO). By contrast, catalysts prepared from Ru(NO)(NO₃)₃ (Ru/CeO₂─NO₃) show no SMSI encapsulation. Ru/CeO₂─CO exhibits significantly higher oxygen‐vacancy concentrations both overall and within the SMSI layer, which promotes electron transfer from the interface and overlayer to the Ru nanoparticles. The formation of electron‐rich Ru species, without blocking a significant number of active sites, facilitates nitrogen activation and enhances catalytic ammonia synthesis. This work presents a novel strategy to construct intermediate SMSI encapsulation and also provides mechanistic insight into precursor‐dependent activity at the nanoscale.

Strong metal‐support interaction (SMSI) with intermediate encapsulation showing truncated‐pyramid shapes was constructed via a Ru‐precursor‐mediated route. Such intermediate encapsulation with a higher concentration of Ce³⁺, effectively donates electrons to Ru without blocking a significant number of active sites, boosting ammonia synthesis.

1. Introduction

Metal‐support interactions play a pivotal role in heterogeneous catalysis due to their profound ability to modulate catalyst performance [1, 2, 3, 4, 5, 6, 7]. Among these, the strong metal‐support interaction (SMSI) which is characterized by the encapsulation of metal nanoparticles with suboxide overlayers and interfacial electron transfer, is widely recognized as one of the most effective strategies for tuning catalyst stability, activity, and selectivity [8, 9, 10, 11]. It is generally accepted that the SMSI overlayer can suppress nanoparticle sintering, thereby enhancing catalytic stability, although it may undergo dynamic changes under reaction conditions [8, 12, 13]. In contrast, the contribution of SMSI to activity and selectivity is not always straightforward. At the atomic scale, the formation of SMSI overlayers inevitably alters the geometry and number of active sites, either improving performance (e.g., CO₂ hydrogenation, Fischer–Tropsch synthesis) [6, 14, 15] or, when excessive, blocking metal sites and suppressing adsorption of key reactants (e.g., H₂, CO), thereby degrading overall performance [16, 17, 18, 19]. From an electronic‐structure perspective, SMSI induces bidirectional charge transfer across the metal‐support interface, either from the support to the metal or vice versa, thereby changing the electron density of the metal nanoparticles [20, 21, 22]. Even for the same catalytic system, the influence of SMSI overlayers on performance can be inconsistent [19, 22]; this variability likely reflects different SMSI configurations produced under varying synthesis and pretreatment conditions. In some cases, geometric and electronic optimization under SMSI compete: improvements in electronic properties come at the cost of fewer active‐site ensembles. Therefore, controlled construction of SMSI configurations that simultaneously satisfy both geometric and electronic requirements is of paramount importance.

In recent years, considerable effort has gone into optimizing metal encapsulations. For example, Matsubu et al. reported that Rh/TiO₂ catalysts treated under a CO₂ + H₂ atmosphere form amorphous, porous SMSI overlayers via an adsorbate‐mediated mechanism. Unlike the conventional robust, impermeable SMSI overlayers produced by high‐temperature H₂ reduction, this porous overlayer exhibits good reactant permeability and significantly enhances CO selectivity in CO₂ hydrogenation [6]. Similar porous and amorphous carbonate overlayer was successfully established in Au/MgO catalysts by reversible reactions between the CO₂ atmosphere and oxide support, which markedly enhances CO oxidation activity [23]. Wet‐chemical synthesis is also an effective approach for fabricating porous encapsulation layers and has been shown to yield porous overlayers in Pd/CeO₂ and Au/TiO₂ nanocatalysts [11, 24, 25]. More recently, Zhang et al. proposed a laser‐induced strategy to construct porous CeO_x overlayers on Pt/CeO₂ catalysts via ultrafast photo‐excitation of water molecules, which not only allow free diffusion of CO and O₂ molecules but also generate abundant Pt‐CeO₂ interfacial sites, leading to greatly improved CO oxidation performance [26]. Beyond the extensively studied porous or amorphous overlayers, construction of intermediate encapsulation that partially covers metal nanoparticles may offer an alternative promising strategy to reconcile active‐site exposure with electronic tuning [27, 28]; however, facile experimental routes and direct atomic/electronic evidence remain limited.

In this context, we adopted Ru/CeO₂ for ammonia synthesis [29, 30, 31] as a model catalyst and developed a Ru‐precursor‐mediated strategy to create intermediate SMSI encapsulation. When Ru₃(CO)₁₂ was used as the Ru precursors, a truncated‐pyramid‐shaped SMSI overlayer forms at the Ru─CeO₂{100} interface. The localized surface modification of CeO₂{100} and crystallographically defined CeO_x overlayers exposing {111} facets, result in an intermediate encapsulation with approximately two atomic layers in height. For comparison, Ru/CeO₂{100} nanocatalysts synthesized using Ru(NO)(NO₃)₃ precursors, which did not show evidence of SMSI, were also studied. Catalyst evaluation experiments demonstrate that the Ru/CeO₂{100} nanocatalysts with intermediate encapsulation show superior catalytic activity. Further investigations into the valence states and local charge transfer at the interface were conducted using electron energy loss spectroscopy (EELS) and density functional theory (DFT) calculations to elucidate the structure‐performance relationship.

2. Results and Discussion

2.1. Constructing Intermediate Encapsulation on Ru/CeO₂

CeO₂ supports were synthesized via a hydrothermal method [32, 33]. The corresponding high‐angle annular dark field scanning transmission electron microscopy (HAADF‐STEM) images were provided in Figure S1. The structural features include predominantly exposed {100} facets that are atomically smooth and particle sizes ranging from 20 to 150 nm. When the CeO₂ nanocubes impregnated with Ru(NO)(NO₃)₃ (1 wt.% Ru loading) were reduced under a hydrogen flow of 50 mL/min at 773 K for 3 h (denoted as Ru/CeO₂─NO₃), the resulting Ru─CeO₂ interface keeps atomically smooth, as shown by the HAADF‐STEM image along the CeO₂ [001] zone axis in Figure 1a. No detectable overlayer on the Ru surface was observed. In contrast, when the CeO₂ nanocubes were impregnated with Ru₃(CO)₁₂ and subsequently reduced in hydrogen at 773 K for 3 h (denoted as Ru/CeO₂─CO), the supports surrounding the Ru nanoparticles migrate onto their surfaces, a hallmark feature of SMSI, as shown by the arrow in Figure 1b. This migration leads to the formation of an intermediate encapsulation on the Ru nanoparticles, with the overlayer extending approximately two atomic layers in height from the interface. Furthermore, the overlayer exhibits interplanar spacings of 2.75 Å, corresponding to that of the CeO₂ {100} facets. Unlike the conventional layered, polycrystalline, or amorphous SMSI encapsulations that wrap along the nanoparticle surface [26, 34, 35], such intermediate encapsulation exhibits the single‐crystalline CeO₂ structure and maintains an epitaxial relationship with the underlying CeO₂ support [27].

FIGURE 1.

Microstructures of the Ru/CeO₂─NO₃ and Ru/CeO₂─CO nanocatalysts. (a–d) Atom‐resolved HAADF‐STEM images of the interface in Ru/CeO₂─NO₃ (a,b) and Ru/CeO₂─CO (b,d), viewed along the CeO₂ [001] direction (a,b) and CeO₂ [011] direction (c,d). The SMSI overlayers are indicated by black arrows in (b) and (d). (e–h) Top‐view SE‐STEM images (e,f) and intensity profiles (g,h) along the white dash rectangles of the Ru/CeO₂─NO₃ (e,g) and Ru/CeO₂─CO (f,h) nanocatalysts. The yellow dashed lines denote the Ru boundaries. The red dashed lines indicate the edges of the extended bright‐contrast regions associated with the CeO_x overlayers surrounding the Ru cores. (i,j) Three‐dimensional models of the interface configurations in Ru/CeO₂─NO₃ and Ru/CeO₂─CO, respectively. Scale bars are 2 nm.

The interface configuration between Ru and CeO₂ nanocubes was also characterized along the CeO₂ [011] axis, as shown by the HAADF‐STEM images (Figure 1c,d). Similarly, compared to the clean Ru surface and flat Ru─CeO₂ interface observed in Ru/CeO₂─NO₃, the Ru nanoparticle in Ru/CeO₂─CO is decorated by the CeO₂ support (marked by the arrows in Figure 1d), forming an intermediate configuration of SMSI overlayer with two atomic layers in height. The CeO₂ lattice is dominantly visible at the corners of the Ru nanoparticles, rather than extending across the nanoparticles as seen along the CeO₂ [001] direction in Figure 1b, suggesting an anisotropic encapsulation. Broader views are provided in Figure S2, confirming that the Ru₃(CO)₁₂ precursors can induce the localized surface modification of CeO₂, leading to intermediate encapsulations exposing the CeO₂ {111} facets on the Ru nanoparticles.

The Ru/CeO₂─NO₃ and Ru/CeO₂─CO nanocatalysts were further analyzed using secondary electron STEM (SE‐STEM) imaging from a top‐down perspective (Figure 1e,f). The Ru nanoparticles appear with bright contrast, positioned centrally, and the CeO₂ supports exhibit a well‐resolved lattice structure. For the Ru/CeO₂─NO₃ nanocatalyst, the uniform contrast and corresponding intensity profiles of the CeO₂ support demonstrate a smooth surface, indicating negligible surface topological variations (Figure 1e,g). In contrast, for the Ru/CeO₂─CO nanocatalyst, brighter contrast with higher intensity in the profile (highlighted by the green‐filled rectangles) can be clearly observed within a few atomic layers surrounding the Ru nanoparticles (Figure 1f,h). The bright‐contrast region around the Ru nanoparticle exhibits a square shape (Figure 1f), with its edges aligned parallel to the CeO₂ {110} surface (Figure 1f and Figure S3), rather than parallel to the perimeter of the Ru nanoparticle. This observation supports the idea that compared to the flat interface in Ru/CeO₂─NO₃, an anisotropic intermediate encapsulation with a truncated‐pyramid shape exposing CeO₂ {111} facets forms on the Ru nanoparticles in Ru/CeO₂─CO, as shown by the three‐dimensional models in Figure 1i,j. Furthermore, in the areas surrounding the bright‐contrast region (outside the green‐filled rectangles), a reduction in contrast is observed, indicating that Ce and O atoms in the overlayer originate from the surrounding support.

Recently developed environmental TEM [36, 37] enables in situ imaging of the emerging Ru, the ceria support and their interface at the same location to track the formation of the truncated‐pyramid SMSI encapsulation on Ru with Ru₃(CO)₁₂ being the precursors (Figure S4). The CeO₂ surface initially exhibits an atomically flat {100} facet covered by an amorphous layer after Ru₃(CO)₁₂ impregnation. Upon heating under H₂, Ru nanoparticles appear on the surface as the temperature rises, accompanied by the formation of intermediate encapsulation. Notably, the SMSI structure at the Ru─CeO₂ interface retains the same epitaxial alignment and exposed facets as in the ex situ experiments.

Very recently, metal‐metal interactions have been proposed as a key thermodynamic factor in determining whether SMSI can be constructed in a supported nanocatalyst, and lower oxidant states in supports can promote the SMSI encapsulation [38]. Although Ce⁴⁺ can be partially reduced to Ce³⁺ species under high‐temperature hydrogen treatment, SMSI in CeO₂‐based systems typically forms at relatively high temperatures and strongly reducing conditions [34, 39, 40]. In our experiments, the Ru/CeO₂─NO₃ nanocatalysts maintain an atomically flat CeO₂ surface after treatment at 773 K in hydrogen, indicating that this temperature is insufficient to induce SMSI formation. Even when the reduction temperature was raised to 973 K, the Ru surface remains free of significant SMSI encapsulation (Figure S5). At elevated temperatures, the CeO₂ support may undergo surface reconstruction and other structural changes [41]. Moreover, treatment under a CO atmosphere leads to substantial carbon encapsulation around the Ru nanoparticles (Figure S6) as a result of CO disproportionation, which inhibits the formation of SMSI encapsulation. These observations indicate that conventional high‐temperature reduction treatments are not effective for forming SMSI on CeO₂, let alone for the controlled construction of an intermediate SMSI configuration.

In contrast, Ru₃(CO)₁₂ binds to the oxide support surface via its ligands and decomposes at 773 K, releasing CO and leaving Ru in close contact with the support [42, 43]. These create a locally reducing environment that promotes the reduction of Ce⁴⁺ and the formation of oxygen vacancies on the CeO₂ (100) surface. The reduced CeO_x species migrate to cover the Ru surface driven by minimizing of the surface energy [4]. The resulting overlayer preserves the crystallographic structure of CeO₂, preferentially exposing the low‐energy {111} facets [41, 44, 45]. Such crystallographically defined CeO_x overlayers account for the intermediate encapsulation on Ru, typically only a few atomic layers in height. Upon treatment in an O₂ atmosphere, the encapsulation disappears (Figure S7), further supporting that this SMSI encapsulation depends on a locally reducing environment. A complementary investigation was conducted using ruthenium acetylacetonate (Ru(acac)₃) and ruthenium chloride (RuCl₃) as precursors. SMSI encapsulation exists on Ru with the former (Figure S8), whereas no encapsulation occurs with the latter (Figure S9). The precursor‐mediated strategy can also be extended beyond the Ru‐based systems, as shown in Figure S10. Similar intermediate encapsulation is observed in Pt/CeO₂, which was prepared using platinum bis(acetylacetonate) (Pt(acac)₂) as the precursor.

2.2. Catalytic Performance Toward Ammonia Synthesis

Catalytic ammonia synthesis performance of these Ru/CeO₂ nanocatalysts was evaluated using a mixed gas of 75% H_2%–25% N₂ at 673 K, across a pressure range of 0.1–3.0 MPa, as shown in Figure 2. As shown in Figure 2a, the NH₃ synthesis rate of the Ru/CeO₂─CO nanocatalysts is as high as 332 mmol g_Ru ⁻¹ h⁻¹ and approximately 2.4 times greater than that of Ru/CeO₂─NO₃ (136 mmol g_Ru ⁻¹ h⁻¹) at 673 K and 0.1 MPa. The NH₃ synthesis rates increase with the total gas pressures and reach 770 mmol g_Ru ⁻¹ h⁻¹ and 346 mmol g_Ru ⁻¹ h⁻¹ for Ru/CeO₂─CO and Ru/CeO₂─NO₃ at 3.0 MPa, respectively. Over the pressure range, the rates of Ru/CeO₂─CO nanocatalysts are consistently 2–3 times higher than that of the Ru/CeO₂─NO₃ nanocatalysts, and rank among the high level of typical Ru/CeO₂‐based catalysts (as listed in Table S1) [46, 47, 48, 49, 50, 51, 52, 53, 54, 55]. The enhanced NH₃ synthesis rates of Ru/CeO₂─CO nanocatalysts demonstrate that the SMSI construction with intermediate encapsulation is beneficial for achieving efficient NH₃ synthesis under mild conditions. Moreover, the Ru/CeO₂─CO catalyst exhibits excellent long‐term stability over 130 h of continuous operation under the reaction conditions (Figure S11a). The STEM images of the post‐reaction catalyst (Figure S11b) further confirm the stability of its SMSI structure, as ammonia synthesis itself occurs under reducing conditions. The Ru/CeO₂ nanocatalysts prepared using ruthenium acetylacetonate as the Ru precursor also exhibit superior performance compared with the Ru/CeO₂─NO₃ nanocatalysts (Figure S12), further validating the effectiveness of the precursor‐mediated strategy in optimizing ammonia synthesis.

FIGURE 2.

Catalytic performance of the Ru/CeO₂─NO₃ and Ru/CeO₂─CO nanocatalysts toward ammonia synthesis. (a) NH₃ synthesis rates at 673 K (H₂/N₂ = 3) and different pressures from 0.1 to 3 MPa. All error bars represent the standard deviation calculated from at least three independent measurements. (b) N₂ reaction orders (α) at 673 K, and 0.1 and 1.0 MPa. (c) H₂ reaction orders (β) at 673 K, and 0.1 and 1.0 MPa.

Kinetic analysis was performed to analyze the reaction orders at 0.1 and 1 MPa. As shown in Figure 2b, the N₂ reaction orders for the Ru/CeO₂─NO₃ nanocatalysts are 0.67 and 0.81 at 0.1 and 1 MPa, respectively; while the orders decrease to 0.58 and 0.63 respectively for the Ru/CeO₂─CO nanocatalysts. The lower N₂ reaction orders of the Ru/CeO₂─CO nanocatalysts imply that N₂ exhibits higher coverage on Ru/CeO₂─CO under the reaction conditions, indicating a promoted nitrogen dissociation [2, 56]. This was confirmed by the measurements of apparent activation energies (E_a) for ammonia synthesis (Figure S13), which show that Ru/CeO₂─CO exhibits a lower E_a (83 kJ/mol) than Ru/CeO₂─NO₃ (111 kJ/mol). Furthermore, the reaction orders of H₂ are 0.18 at 0.1 MPa and −0.33 at 1 MPa for the Ru/CeO₂─NO₃ nanocatalysts, which increase to 0.36 and −0.24 respectively for Ru/CeO₂─CO (Figure 2c and Figure S14). The increase of the H₂ reaction orders suggests that the H₂ poisoning is slightly alleviated on Ru/CeO₂─CO.

2.3. Electronic Structure of Ru/CeO₂

Ammonia synthesis over Ru is a typical structure‐sensitive reaction, jointly influenced by the abundance and the electronic structure of the exposed Ru active sites. The average Ru particle size, Ru loading and surface area were characterized to evaluate the number of exposed metallic active sites. The average Ru particle size was measured to be 1.68 ± 0.56 nm for Ru/CeO₂─NO₃ with a Ru loading of 0.96% and 1.73 ± 0.79 nm for Ru/CeO₂─CO with a Ru loading of 0.91% (Figure S15 and Table S2). The Ru loading and average particle size exhibit no significant differences, suggesting that these factors are not the main reasons for the observed performance variation. The BET surface areas are 17 and 18 m²/g, respectively, indicating that both catalysts support possess similar textural properties. On the other hand, although classical SMSI often eliminates gas adsorption due to the formation of suboxide overlayer on metal nanoparticle, the SMSI encapsulation in Ru/CeO₂─CO is typically only two atomic layers in height and covers approximately 16% of the exposed Ru surface area. Consequently, the adsorption capacity is not significantly reduced or eliminated, in contrast to the full SMSI encapsulation [9, 19]. This is further supported by CO diffuse reflectance infrared Fourier transform spectroscopy (CO‐DRIFT) and CO chemisorption measurements. The CO‐DRIFTS show clear adsorption signals on both catalysts (Figure S16), and CO chemisorption reveals similar CO uptake with only ∼11% difference between the two samples (Table S2), in good agreement with the estimation from our microscopic observations.

The electronic structures of the Ru/CeO₂─CO nanocatalysts with SMSI at the nanoscale were carefully characterized using EELS. Reference spectra from bulk CeO₂ and CePO₄ (Figure 3) establish the spectral fingerprints for Ce⁴⁺ and Ce³⁺, respectively. As evidenced by the reference spectra, the Ce⁴⁺ state exhibits higher energy‐loss at the M_4,5 edges and a lower second‐derivative intensity ratio (M₅/M₄) compared to Ce³⁺. Quantitatively, bulk CeO₂ (Point 1) shows peaks at 884.3 eV (M₅) and 902.7 eV (M₄) with a second derivative intensity ratio (M₅/M₄) of 0.9 (characteristic of Ce⁴⁺), while CePO₄ exhibits shifted peaks at 883.1 and 901.1 eV with a ratio of 1.3, confirming its Ce³⁺ state [57]. Utilizing this calibrated M₅/M₄ ratio as a quantitative metric, we analyzed the Ce valence at the free CeO₂ surface (Point 2) and the Ru─CeO₂ interface (Point 3) in both catalysts (Figure 3a,b). In Ru/CeO₂─NO₃, the Ce³⁺ concentrations at the free surface are 39%, which is much lower than that in Ru/CeO₂─CO (55% at the surface). This indicates that Ru/CeO₂─CO possesses a higher overall concentration of oxygen vacancies, which further implies a greater overall reduction extent. Electron paramagnetic resonance (EPR) characterization was performed to confirm the O_V concentration in those samples (Figure S17). The signal occurring at g = 2.003 can be ascribed to the superoxide anions (•O²⁻) attached with the Ce⁴⁺ ions [58], and its intensity is mainly associated with the concentration of O_V. The acquired EPR spectra demonstrate that Ru/CeO₂─CO has a higher concentration of O_V, consistent with the EELS analyses.

FIGURE 3.

Valence states and electronic structures of the Ru/CeO₂ nanocatalysts. (a,b) HAADF‐STEM images and corresponding EELS spectra of Ce M_4,5 edges extracted from the red points in Ru/CeO₂─NO₃ (a) and Ru/CeO₂─CO (b). EELS spectra of Ce M_4,5 edges for CePO₄ (Ce³⁺) are illustrated for comparison. (c) Ru L EELS spectra of the Ru nanoparticles in Ru/CeO₂─NO₃ and Ru/CeO₂─CO. The peak shift is highlighted by the dashed line and arrows.

In addition to the differences in the overall Ce³⁺ concentration, the distribution of Ce³⁺ at the free surface and the Ru─CeO₂ interface exhibits opposite trends between the two nanocatalysts. In Ru/CeO₂─NO₃, the Ce³⁺ concentration is highest at the interface, followed by the free surface. In contrast, the Ce³⁺ concentration at the interface in Ru/CeO₂─CO is lower than that at the free surface. This is because, in addition to the O_V concentration, the amount of Ce³⁺ also depends on charge transfer. As shown in Figure 4a,b, two models of a Ru₄ cluster on the CeO₂ (100) surface, one with and one without an O_V, were constructed using DFT calculations to investigate the effect of surface reduction on charge transfer between Ru and CeO₂. Figure 4a shows the CeO₂ (100) surface without an oxygen vacancy, where the Bader charge on the Ru atom bonded to the surface lattice oxygen is calculated to be +0.39, indicating electron transfer from Ru to the support and thus the Ce³⁺ concentration at the interface increases. In contrast, Figure 4b shows the CeO₂ (100) surface with an oxygen vacancy. Notably, the Ru atom directly above the vacancy exhibits a Bader charge of ‐0.14, indicating electron transfer from the reduced CeO₂ (100) surface to Ru, which results in a decreased Ce³⁺ concentration at the interface. Spin charge density analysis further elucidates the relationship between Ce³⁺ concentration and oxygen vacancies through direct visualization of the charge distribution (Figure S18). The presence of an oxygen vacancy reverses the charge transfer direction: while electrons flow from the Ru cluster to stoichiometric CeO₂ (reducing Ce⁴⁺ to Ce³⁺), the introduction of an oxygen vacancy causes one electron to transfer back from a Ce³⁺ site to the Ru cluster, thereby reducing the net number of Ce³⁺ ions. Combined with the EELS and Bader charge analyses, these results indicate that as the degree of surface reduction increases, the direction of electron transfer shifts from Ru‐to‐support to support‐to‐Ru.

FIGURE 4.

Bader charge of Ru atom on (a) CeO₂ surface and (b) CeO₂ surface with one oxygen vacancy, where white balls represent Ce atoms, red balls represent O atoms, green balls represent Ru atoms. pDOS analyses for (c) a three‐layer CeO₂ (100) surface, (d) a three‐layer CeO₂ (100) surface with one oxygen vacancy, (e) a thirteen‐layer CeO₂ (100) surface, and (f) a thirteen‐layer CeO₂ (100) surface with one oxygen vacancy.

EELS spectrum of the SMSI overlayer (Point 4) in Ru/CeO₂‐CO was also extracted, as shown in Figure 3b. The ratio of Ce³⁺/(Ce⁴⁺+Ce³⁺) in the SMSI overlayer was estimated to be 65%, higher than that at both the free surface and the interface. In order to assess the effectiveness of the SMSI overlayer with a thinner CeO₂ layer in donating electrons to Ru (i.e., electron transfer from Ce³⁺ to Ru), we constructed CeO₂ (100) surface models with different thicknesses, three‐layer and thirteen‐layer configurations, with and without O_V (Figure S19). Figure 4c–f shows the corresponding projected density of states (pDOS) plots for each structure. In the absence of O_V, the system contains no occupied 4f states (Figure 4c,e). However, when O_V is introduced into CeO₂, as shown in Figure 4d,f, the Ce 4f orbitals become partially occupied, indicating the presence of Ce³⁺ ions. This is a direct result of oxygen vacancy formation, which induces localized states in the Ce 4f orbitals. A comparison between Figure 4d,f reveals that the energy of the 4f orbitals is higher in the thinner CeO₂ layer, with a 4f band center of −0.30 eV, compared to −0.42 eV for the thicker CeO₂ layer. This suggests that electron localization in the thin‐layer CeO₂ is more active, thereby facilitating electron transfer. This finding highlights the critical role of the thin SMSI overlayer in effectively donating electrons to Ru nanoparticles and enhancing catalytic performance.

To confirm the electron transfer within the Ru/CeO₂ nanocatalysts, the Ru L edges were analyzed using EELS [59], revealing a shift in energy loss spectrum (Figure 3c). In Ru/CeO₂─NO₃, the Ru L edge is located at 2848.9 eV, while it shifts to 2845.5 eV in Ru/CeO₂─CO. This redshift indicates that the Ru nanoparticles in Ru/CeO₂─CO are more electron‐rich, because increased electron cloud density enhances the shielding effect on core electrons [59, 60]. X‐ray photoelectron spectroscopy characterizations further show Ru/CeO₂─CO exhibits approximately 0.3 eV redshifts in the Ru 3d_5/2 binding energy compared to Ru/CeO₂─NO₃ (Figure S20), supporting that Ru in Ru/CeO₂─CO is more electron‐rich overall. These conclusions are in agreement with the electron transfer predicted by DFT calculations.

The dissociative adsorption and activation of N₂ were explored on the Ru₄/CeO₂ system, both with and without oxygen vacancies. The DFT results show that Ru atoms adjacent to an oxygen vacancy possess higher electron density, leading to stronger dissociative adsorption of N₂ (−1.43 eV for system with an oxygen vacancy vs. −0.21 eV without) and a lower N‐N dissociation barrier (0.52 eV with an oxygen vacancy vs. 1.20 eV without). A similar trend is also found in the Ru/CeO₂ systems with larger Ru clusters (Ru₈, Figure S21) and Ce‐encapsulated Ru/CeO₂ system (Figure S22). As discussed, the increased oxygen vacancies and the formation of a thin intermediate SMSI encapsulation render Ru more electron‐rich at exposed sites, facilitating N₂ adsorption and activation on Ru/CeO₂, which is consistent with the conclusions from our catalytic performance tests and reaction kinetic analysis. Additionally, the electron‐rich Ru facilitates hydrogen spillover and alleviates hydrogen poisoning [61, 62, 63]. Both contribute to the superior performance of Ru/CeO₂─CO in ammonia synthesis.

3. Conclusion

Our work reveals that intermediate SMSI encapsulation can be rationally constructed by forming crystallographically defined CeO_x overlayers on Ru/CeO₂ via a Ru‐precursor‐mediated route. The crucial role of Ru precursors in modulating the Ru‐CeO₂ interfacial structures and ammonia synthesis was revealed at the microscopic level. The Ru₃(CO)₁₂ precursor enhances the reduction extent of the CeO₂ support, thereby facilitating the localized surface modification and the formation of SMSI overlayers with truncated‐pyramid shapes on the Ru nanoparticles. Such intermediate encapsulation with a higher concentration of Ce³⁺, effectively donates electrons to Ru without blocking a significant number of active sites, thereby promoting nitrogen dissociation and boosting ammonia synthesis. Unlike the conventional methods that form SMSI overlayers shaped by the contours of metal nanoparticles, we propose a Ru‐precursor‐mediated approach for constructing SMSI overlayers that preserve the crystalline morphology of CeO₂. This generates intermediate configurations of SMSI that meet both the geometric and electronic requirements essential for efficient ammonia synthesis. Furthermore, direct evidence of the distinct atomic and electronic structures at the interface of Ru/CeO₂ nanocatalysts with and without SMSI can provide deeper insight into SMSI and precursor‐mediated performance in supported catalytic systems.

Author Contributions

Z.H., Y.J. and Y.W. conceived and supervised this study. B.C. performed the STEM observations, data analysis, and drafted the manuscript. J.H. helped with the experiments. W.Y., Y.G. and Z.Z. participated in discussion. All the authors commented on the manuscript and have given approval to the final version of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll73911‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.