Atomic-level engineering of single Ag¹⁺ site distribution on titanium–oxo cluster surfaces to boost CO₂ electroreduction

Abstract

Precise control over the distribution of active metal sites on catalyst surfaces is essential for maximizing catalytic efficiency. Addressing the limitations of traditional cluster catalysts with core-embedded catalytic sites, this work presents a strategy to position catalytic sites on the surfaces of oxide clusters. We utilize a calixarene-stabilized titanium–oxo cluster (Ti₁₂L₆) as a scaffold to anchor Ag¹⁺in situ, forming the unique nanocluster Ti₁₂Ag_4.5 with six surface-exposed Ag¹⁺ sites. The in situ transformation from Ti₁₂L₆ into Ti₁₂Ag_4.5 clusters was traced through mass spectrometry, revealing a solvent-mediated dynamic process of disintegration and reassembly of the Ti₁₂L₆ macrocycle. The unique Ti₁₂Ag_4.5 cluster, featuring a surface-exposed catalytic site configuration, efficiently catalyzes the electroreduction of CO₂ to CO over a broad potential window, achieving CO faradaic efficiencies exceeding 82.0% between −0.4 V and −1.8 V. Its catalytic performance surpasses that of bimetallic Ti₂Ag₂, which features a more conventional design with Ag¹⁺ sites embedded within the cluster. Theoretical calculations indicate that the synergy between the titanium–oxo support and the single Ag¹⁺ sites lowers the activation energy, facilitating the formation of the *COOH intermediate. This work reveals that engineered interactions between active surface metal and the oxide support could amplify catalytic activity, potentially defining a new paradigm in catalyst design.

---

Introduction

Within the intricate sphere of catalysis, the sophisticated interplay of size, morphology, and structural attributes of materials is recognized for its profound influence on catalytic efficacy.^1–4 Recently developed single-atom catalysts (SACs), with isolated metal atoms dispersed on solid supports, are garnering significant attention due to their exceptional catalytic efficiency and ability to maximize the utilization of metal atoms.^5–9 Oxide-supported SACs, in particular, are among the most thoroughly investigated systems.¹⁰ Advances in the fabrication of these materials are noteworthy, but achieving uniform dispersion of catalytic sites on oxide substrates remains a significant challenge.¹¹ This requires not only advanced synthesis techniques but also sophisticated characterization methods.¹² The complexity is further exacerbated by the absence of detailed atomic structures, which obscures the analysis of the metal coordination environment—a pivotal factor in determining the performance of these catalyst systems.^13–15

To effectively tackle these challenges, it is crucial to develop molecular analogues for oxide-supported single-atom materials. Recent advances in TiO₂ analogues,^16–19 particularly through the development of crystalline titanium oxide clusters (TOCs), have facilitated comprehensive investigations of TiO₂ structures and reactivities at the molecular level.^20–24 These TOCs, when modified with catalytically active single metals, hold promise as effective molecular mimics for oxide-supported SACs. However, heterometal-doped TOCs are typically created via a one-pot solvothermal method, with heterometal sites embedded deep within the cluster core, often in fully coordinated states.^25–29 Such configurations limit the interactions with reactants, exposing a significant shortfall in their capacity to act as true molecular proxies. This insight highlights the urgent need for a new class of crystalline, cluster-stabilized single-atom materials, designed to more precisely emulate the structural and reactive properties of oxide-supported SACs at a molecular level.^30,31

Here, we present an approach designed to strategically place catalytic Ag¹⁺ sites on the surfaces of oxide clusters (Scheme 1). Employing thiacalix[4]arene (TC4A) as a protective ligand,^32–38 we meticulously engineered a titanium–oxo macrocycle, Ti₁₂L₆, through a one-step solvothermal method, which is enriched with surface O and S sites. We utilized Ti₁₂L₆ as a scaffold for the in situ loading of Ag¹⁺ ions, revealing a solvent-mediated assembly process through mass spectrometry. In N,N-dimethylformamide (DMF) solution, the scaffold undergoes decomposition under the cleavage of Ag¹⁺ ions; however, in acetonitrile, the scaffold remains stable and coordinates six accessible single Ag¹⁺ sites on the cluster surface, ultimately transforming into a Ti₁₂Ag_4.5 cluster. Ti₁₂Ag_4.5 has proven to be an exceptional catalyst for the electrochemical reduction of CO₂, exhibiting superior reactivity and CO selectivity compared to its bimetallic counterpart, Ti₂Ag₂, which possesses a more conventional structure with Ag¹⁺ sites embedded within the cluster matrix. Notably, Ti₁₂Ag_4.5 exhibits high selectivity for CO across a wide voltage range, with CO faradaic efficiency (FE_CO) consistently exceeding 82.0% from −0.4 V to −1.8 V vs. RHE, reaching a peak FE_CO of 92.7%. At an overpotential of ∼−1.4 V, the system remained stable for continuous electrolysis over 11 h with a CO partial current density exceeding 100.0 mA cm⁻², while the FE_CO consistently remains above 85.0%. We elucidated the reaction path using in situ ATR-SEIRAS technology and comprehensively calculated the Gibbs free energy changes for each elementary step of CO₂ conversion to CO, highlighting the critical contribution of exposed Ag centers to the observed catalytic prowess.

Scheme 1 Schematic representation of the cluster assembly featuring a Ti–oxo core with surface-exposed Ag¹⁺ catalytic sites.

Results and discussion

One-step synthesis of titanium–oxo macrocyclic carriers

The synthetic route commenced with the reaction of Ti(OⁱPr)₄ with TC4A in a solvent blend of CH₃CN/CHCl₃, maintained at 80 °C for three days. This process yielded a turbid yellow liquid. Extending the reaction duration by introducing acetone and continuing incubation at 80 °C for an additional three days led to the crystallization of yellow rhomboid crystals of Ti₁₂L₆. Interestingly, this reaction exhibits a high sensitivity to temperature changes. Upon heating to 130 °C, certain sulfur sites in TC4A oxidize, leading to the emergence of two distinct TOCs, Ti₈L₆ and Ti₈L₅. SCXRD analysis revealed that the Ti₁₂L₆ cluster has the composition of {H₆Ti₁₂O₁₈(HTC4A)₆}, featuring a core–shell architecture where a central Ti–oxo macrocycle is encapsulated by six TC4A molecules. Ti₁₂L₆ consists of six {Ti₂O(TC4A)} units interconnected by twelve μ₂-O²⁻ (Fig. 1A). TC4A adopts a mono-cone configuration, coordinating with two Ti⁴⁺ through three phenoxide groups and two S atoms, thereby forming the {Ti₂O(TC4A)} units. These units are geometrically arranged in a triangular pattern, bridged by five μ₂-O linkages to form a trimeric {Ti₆(TC4A)₃}. This trimer further aggregates into the hexameric {Ti₁₂(TC4A)₆} structure via the integration of two additional μ₂-O²⁻ ions (Fig. 1B). This configuration yields a “donut” shape with an outer diameter of 3 nm, representing the largest TC4A-stabilized TOC reported to date (Fig. 1C). The structure of Ti₈L₆ closely resembles that of Ti₁₂L₆, with a central Ti₈O₈ ring encircled by six TC4A molecules (Fig. 1D). In contrast, the core of Ti₈L₅ consists of a Ti₈O₁₀ cluster, surrounded by five TC4A molecules arranged in a pentagonal bipyramidal configuration. In both structures, some of the TC4A molecules undergo oxidation, with one or two sulfur sites being converted to S=O groups. Additionally, each structure has two Na⁺ ions adsorbed on its surface.

Fig. 1 (A) The overall structure of Ti₁₂L₆; (B) the Ti₁₂O₁₈ core in Ti₁₂L₆; (C) the polyhedral arrangement of six calix moieties in Ti₁₂L₆; (D) the overall structure of Ti₈L₆; (E) the overall structure of Ti₈L₅.

Solvent-mediated assembly of Ti₁₂Ag_4.5 clusters

The surfaces of this series of calixarene-stabilized TOCs are endowed with abundant O/S sites. Specifically, in Ti₈L₆ and Ti₈L₅, these sites facilitate the adsorption of Na⁺ onto the cluster surfaces. This characteristic has inspired us to utilize these clusters as carriers for positioning Ag¹⁺ ions. In our preliminary studies, we attempted to directly introduce Ag(I) salt into the reaction mixture for these clusters, but this approach resulted in the formation of crystals of Ti₂Ag₂ (Fig. 2A). Ti₂Ag₂ is a typical bimetallic cluster, with two Ag¹⁺ and two Ti⁴⁺ ions arranged in a quadrilateral configuration, sandwiched between two oxidized-TC4A ligands (Fig. 2B). Each Ag¹⁺ site maintains a coordination number of six, with the Ag¹⁺ seamlessly integrated into the cluster core, effectively shielded from external exposure (Fig. 2C). The composition of Ti₂Ag₂ was determined using electrospray ionization mass spectrometry (ESI-MS) (Fig. S28†), which revealed a +1 signal at m/z = 1794.18, corresponding to the species [HTi₂Ag₂O(O-TC4A)₂]⁺.

Fig. 2 (A) The overall structure of Ti₂Ag₂; (B) coordination mode of oxidized TC4A in Ti₂Ag₂. (C) Comparison of the positions of the Ag and Ti sites in Ti₂Ag₂.

Interestingly, using Ti₁₂L₆ crystals in reactions with Ag(I) salts yielded distinctly different outcomes. Introducing Ag₂SO₄ into DMF solution along with Ti₁₂L₆ crystals and allowing the mixture to react at 80 °C for three days resulted in the formation of the known clusters Ti₂Ag₂-DMF and Ti₂Ag₄.³⁴ Subsequently, by switching the solvent to CH₃CN, rhombic Ti₁₂Ag_4.5 crystals were obtained, emphasizing the critical role of solvent selection in steering the chemical pathway. Ti₂Ag₂-DMF and Ti₂Ag₄ have typical bimetallic configurations, with two or four Ag¹⁺ ions embedded between two {Ti(TC4A)} units. In contrast, in Ti₁₂Ag_4.5, multiple Ag¹⁺ ions are effectively anchored onto the surface of the Ti₁₂ core without altering its intrinsic structure. SCXRD analysis revealed the structure of Ti₁₂Ag_4.5 to be [H_1.5Ti₁₂Ag_4.5O₁₈(HTC4A)₆(CH₃CN)₄], with the cluster containing three crystallographically distinct sites for Ag. Structural analysis demonstrated that the Ti–oxo core of Ti₁₂Ag_4.5 closely mirrors the macrocyclic structure of Ti₁₂L₆, albeit with minor deviations (Fig. 3A). The Ag¹⁺ sites are symmetrically divided and positioned on both the upper and lower facets of the Ti–oxo macrocycle (Fig. 4B). Ag1 and Ag2, coordinated through phenoxide, sulfur, and μ-O₂²⁻, are placed between two TC4A ligands, exhibiting coordination numbers of 3 and 4, respectively. The distances between Ag1/Ag2 and the Ti–oxo core range from 2.392 to 2.607 Å. Notably, the Ag3 site, with an O₃N₂ coordination environment, is defined by two phenoxide groups, one μ-O₂⁻, and two CH₃CN, with Ag–O bond lengths ranging from 2.649 to 2.774 Å and Ag–N distances of 2.056 and 2.215 Å. The spatial arrangement of the Ag sites, especially the Ag3 sites proximal to the cluster surface as depicted in Fig. 3C, highlights their potential as catalytically active sites. This is particularly significant given the labile nature of the CH₃CN ligand, which may facilitate dynamic catalytic processes. Sites Ag1 and Ag2 exhibit full occupancy, while Ag3 displays a partial occupancy of 0.25. As a result, the average number of Ag sites in the cluster is 4.5, which is confirmed by the mass spectrometry analysis of Ti₁₂Ag_4.5. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) analysis reveals six distinct peaks with a mass difference of 107.86, corresponding to the ion {H₄Ti₁₂Ag₆O₁₆(TC₄A)₆(CH₂Cl₂)}⁺ (Fig. 3D) (x = 1–6). This pattern suggests that the cluster contains six Ag sites, which can sequentially detach under ionization conditions. Additionally, the atomic ratio of Ag to Ti in the cluster, as determined by ICP analysis, is 2.63, which is in excellent agreement with the theoretical value of 2.66 (Table S2†).

Fig. 3 (A) The overall structure of Ti₁₂Ag_4.5; (B) space-filling model of the structure of Ti₁₂Ag_4.5, highlighting the exposed Ag atoms on the surface; (C) bonding mode of six Ag¹⁺ sites to the titanium oxide core; (D) positive-mode MALDI-TOF-MS of Ti₁₂L₆ and Ti₁₂Ag_4.5 in CH₂Cl₂ solution.

Fig. 4 Time-dependent ESI-MS for the reaction of the Ti₁₂L₆ crystals and Ag₂SO₄ in CH₃CN (A) or DMF (B) solution at 80 °C at 0 h, 1 h and 6 h.

From the above results, it is evident that the solvent choice significantly influences the reaction pathway of Ti₁₂L₆ with Ag(I) salts; CH₃CN primarily leads to Ti₁₂Ag_4.5, whereas solvents like DMF result in different structures of Ti₂Ag₂-DMF and Ti₂Ag₄. The significant impact of seemingly minor variations in solvent choice on the synthesis outcomes is fascinating. To elucidate the formation mechanism of Ti₁₂Ag_4.5, its evolutionary process was monitored using ESI-MS.³⁹ Time-resolved ESI-MS analysis of the reaction mixture, containing Ti₁₂L₆ crystals and Ag₂SO₄ in CH₃CN, captured data at various intervals (Fig. 4A). Initially, ESI-MS detected two principal signal sets corresponding to the +2 and +3 charge states of {H_xTi₁₂O_18−y(TC4A)₆}, indicating that Ti₁₂L₆ maintained its integrity in the early stages of the reaction. After heating at 80 °C for one hour, new signals emerged, with a strong peak corresponding to {HTi₂(TC4A)₂}⁺, resulting from the fragmentation of Ti₁₂L₆. Additionally, subtle signals in the m/z range of 2000–2500, likely representing {Ti₁₂Ag_xO_20−z(TC4A)₆}²⁺, were observed. By the six-hour mark, the spectrum displayed distinct signal sets aligned with the +2 and +3 charge states of the {H_xTi₁₂Ag_yO_20−z(TC4A)₆} species, indicating a complete transformation of Ti₁₂L₆ into Ti₁₂Ag_4.5.

Conversely, in a DMF environment, ESI-MS revealed a completely different reaction scenario (Fig. 4B). Initially, Ti₁₂L₆ also maintained its structure in DMF (Fig. S34†). One hour post Ag(I) salt introduction, ESI-MS revealed the breakdown of the macrocycle, detecting {H_xTi₆O_z(TC4A)₃}⁺ and {H_xTi₂O_z(TC4A)₂}⁺ species. Notably, the emergence of {H_xTi₂Ag_yO_z(TC4A)₂}⁺ (x = 2–4) clarified the formation of Ti₂Ag₄ and Ti₂Ag₂-DMF clusters. By the six-hour mark, these signals intensified significantly, highlighting the progressive formation of these clusters in the DMF solution.

These findings indicate that the synthesis of Ti₁₂Ag_4.5 is not merely a simple process of Ag¹⁺ adsorption onto the Ti₁₂L₆ carrier. Instead, it involves a complex sequence of fragmentation and subsequent reassembly (Fig. 5). Under solvothermal conditions, the Ti₁₂L₆ framework undergoes a fragmentation process, resulting in the formation of {Ti₂O(TC4A)} units. A dynamic equilibrium forms between the larger {Ti₁₂O₁₂(TC4A)₆} structure and these smaller {Ti₂O(TC4A)} units. In the presence of CH₃CN solvent, during the reassembly phase, these {Ti₂O(TC4A)} fragments coordinate with Ag¹⁺ ions, facilitating the formation of Ti₁₂Ag_4.5. Conversely, using DMF as the solvent leads to a transformation of the {Ti₂O(TC4A)} units into more complex {Ti₂O(TC4A)₂} structures. These structures then complex with Ag¹⁺ ions to form Ti₂Ag₂-DMF and Ti₂Ag₄, highlighting a distinct synthetic pathway that is significantly influenced by the choice of solvent.

Fig. 5 Schematic diagram illustrating the solvent-mediated self-assembly process transitioning from Ti₁₂L₆ to Ti₁₂Ag_4.5.

Electrochemical CO₂ reduction

Ag-based nanoclusters are optimal catalysts for electrocatalytic CO₂ reduction reactions.^40–42 Tuning the coordination environment of Ag sites within the clusters is a key strategy to boost their catalytic activity and tune the selectivity for CO production.^43–51 From the structural comparison provided, it is clear that the distribution of Ag sites in the two Ti–Ag clusters, Ti₁₂Ag_4.5 and Ti₂Ag₂, is significantly different. For bimetallic Ti₂Ag₂, Ag¹⁺ and Ti⁴⁺ sites are indistinguishably mixed within the core of the cluster. In contrast, Ti⁴⁺ sites of Ti₁₂Ag_4.5 are embedded in the Ti–oxo core, while Ag¹⁺ sites are strategically distributed across the cluster surface. These structural differences between Ti₁₂Ag_4.5 and Ti₂Ag₂ provide a unique platform for investigating how variations in the Ag environments affect their catalytic activity.^52,53

Powder X-ray diffraction (PXRD) analysis shows that the Ti₁₂Ag_4.5 crystal maintains its crystalline structure even after exposure to a highly alkaline 1 M KOH solution for 24 hours (Fig. S35†). Given the resilience of Ti₁₂Ag_4.5, its eCO₂RR activity was assessed in a three-compartment flow cell with 1 M KOH as the electrolyte. Linear sweep voltammetry (LSV) results indicated a significantly higher current density and more positive onset potential for both the Ti₁₂Ag_4.5 and Ti₂Ag₂ clusters in the CO₂ flow electrolyzer compared to the N₂-purged system, confirming their CO₂ reduction capability, as illustrated in Fig. 6A. Gas chromatography detected only CO and H₂ as products, with no other liquid products identified by ¹H NMR spectroscopy (Fig. S39†). Control experiments under N₂-saturated conditions yielded no carbon-reduction products. Isotopic tracing experiments with ¹³CO₂ through GC-MS confirmed the production of ¹³CO (m/z = 29) (Fig. S40†). Additionally, the primary electrocatalytic product of the Ti₁₂L₆ catalyst was identified as H₂, indicating that the Ag¹⁺ components of these clusters predominantly drive the catalytic activity (Fig. S37†).

Fig. 6 (A) LSV of samples in N₂ or CO₂ saturated 1 M KOH solution; (B) FE values of Ti₁₂Ag_4.5 at different voltages; (C) FE values of Ti₂Ag₂ at different voltages; (D) CO partial current density (j_CO); (E) I–t test and FE_CO values of Ti₁₂Ag_4.5 at −1.4 V vs. RHE in different time periods; (F) the ATR-SEIRAS results from 2500 to 1000 cm⁻¹ of Ti₂Ag₂ and Ti₁₂Ag_4.5.

Fig. 6B and C illustrate the trends in product distribution across varying potentials for two distinct clusters. In a CO₂-saturated environment, Ti₁₂Ag_4.5 consistently exhibits a high FE_CO production, maintaining over 82.0% across an extensive potential range from −0.4 V to −1.8 V vs. RHE, peaking at 92.7% at a higher potential of −1.4 V (Table S3†). However, for Ti₂Ag₂, the predominant reaction between −0.6 V and −1.0 V is CO₂ reduction, reaching its highest FE_CO of 85.1% at −0.8 V. Beyond this, the reaction is largely overtaken by H₂ evolution in the potential range from −1.2 to −1.8 V, leading to a progressive decline in FE_CO as potential increases. At −1.8 V, the FE_CO of Ti₂Ag₂ plummets to merely 21.2%, whereas Ti₁₂Ag_4.5 still manages to maintain an FE_CO of 83.25%. Additionally, the CO partial current density (J_CO) for both clusters was also analyzed (Fig. 6D). The J_CO for Ti₁₂Ag_4.5 reached an impressive 130.1 mA cm⁻² at −1.8 V, which is above 2.4 times greater than that of Ti₂Ag₂. This comparative analysis emphasizes the pronounced differences in catalytic efficiencies between two clusters, attributing Ti₁₂Ag_4.5's superior performance to its unique structural configuration and effective Ag¹⁺ site utilization.

Electrocatalytic stability is a key indicator for evaluating the performance of electrocatalysts in the eCO₂RR. We employed PXRD to analyze the reduction of Ag⁺ in Ti₁₂Ag_4.5 after reaction at different applied voltages (Fig. S44†). PXRD analysis reveals that the catalyst remains stable when the applied voltage is below −1.8 V. However, when the voltage exceeds −1.8 V, signals corresponding to Ag nanoparticles appear, indicating the reduction of Ag⁺ to metallic Ag. To further evaluate the durability of the catalyst, we conducted a rigorous 11-hour chronoamperometric test at −1.4 V. During this process, the current density remained above 100 mA cm⁻², and the FE for CO remained stable above 85.0% (Fig. 6E). Additionally, ESI-MS analysis of the catalysts after electrolysis showed a signal for the {H_xTi₁₂Ag₆O_18−z(TC4A)₆}²⁺ species, confirming that the structural integrity of the Ti₁₂Ag₄.₅ catalyst was maintained (Fig. S45†). Further characterization by transmission electron microscopy, X-ray photoelectron spectroscopy, and infrared spectroscopy showed that the catalyst maintained its chemical composition and structural stability during the electrolysis process (Fig. S47–S49†). Differential pulse voltammetry (DPV) measurements reveal that the electrochemical gap of Ti₁₂Ag_4.5 is 1.91 V (Fig. S50†), which is greater than the 1.32 V observed for Ti₂Ag₂. This larger electrochemical gap further suggests enhanced stability of Ti₁₂Ag_4.5 during the electrocatalytic process. The stability of Ag⁺ in the Ti₁₂Ag_4.5 clusters can be attributed to the strong electronic and structural stabilization provided by the Ti–oxo support. This unique characteristic not only preserves the active sites but also prevents the competitive reduction of Ag⁺, thereby enhancing the suitability of these clusters for CO₂ reduction.

The hypothesized mechanism for CO₂ reduction to CO using Ag-based catalysts follows the pathway: CO₂(g) → *COOH → *CO → CO(g).^54–57 To validate this mechanism, we utilized in situ electrochemical attenuated total reflection surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS). This technique allows for real-time monitoring of the absorption of evanescent waves by substances on the catalyst surface, providing direct insight into the reaction dynamics. We set the potential range from −0.5 V to −1.3 V and monitored changes in absorption peaks with Ti₁₂Ag_4.5 and Ti₂Ag₂ as the electrocatalysts (Fig. 6F). The spectra display similarities, including a pronounced peak at 1225 cm⁻¹, attributed to C–OH stretching in *COOH, which intensifies with increasing voltage. Another peak at 1711 cm⁻¹, corresponding to C=O stretching in *COOH, also increases in intensity from −0.5 V to −1.3 V, suggesting a rise in the surface coverage of *COOH species with increasing voltage.^58,59 Additionally, a weak signal at 2127 cm⁻¹, attributed to the Ag–*CO vibration mode, indicates the presence of *CO adsorbed on the catalyst surface, especially at lower potentials where the *CO band intensity shows a slight increase. Notably, the intermediate characteristic peak of Ti₂Ag₂ becomes distinctly observable only from an electrode potential of approximately −1.1 V, while the characteristic peaks of Ti₁₂Ag_4.5 are already evident at −0.5 V. This observation indicates that Ti₁₂Ag_4.5 demonstrates a higher reaction activity and stronger catalytic proficiency for the eCO₂RR, aligning with experimental findings.

Density Functional Theory (DFT) calculations were conducted to elucidate the reactivity of two specifically designed clusters, Ti₁₂Ag_4.5 and Ti₂Ag₂. These models were optimized to mirror their actual crystal structures, simplified by replacing the tert-butyl groups on the TC4A with H for faster computational convergence. The results of these optimizations are depicted in the Gibbs free energy diagrams for both the eCO₂RR and the HER, as presented in Fig. 7A. DFT calculations specifically focused on the energetics of each step, revealing that the formation of the *COOH intermediate is the rate-determining step in the CO₂RR process. A critical finding from our study is the calculated Gibbs free energy for the formation of *COOH on the Ag3 site in Ti₁₂Ag_4.5, which was notably low at 0.55 eV. This value contrasts with the corresponding energy of 0.88 eV for the same process within Ti₂Ag₂. This substantial difference highlights that the exposed Ag sites on Ti₁₂Ag_4.5 are much more energetically favorable for catalyzing the conversion of CO₂ to CO compared to those on bimetallic Ti₂Ag₂ (Fig. 7B). Additionally, the Gibbs free energies for hydrogen adsorption (*H) were calculated, showing high values of 2.46 eV for both clusters, indicating that neither Ti₂Ag₂ nor Ti₁₂Ag_4.5 is favorable for the formation of H₂. This finding is important as it underscores the higher selectivity of both clusters for the CO₂RR-to-CO pathway compared to the HER.

Fig. 7 (A) Free energy diagrams for the eCO₂RR and HER on Ti₁₂Ag_4.5 and Ti₂Ag₂; (B) schematic diagram of the eCO₂RR process on Ti₁₂Ag_4.5.

By comparing the Gibbs free energy diagrams of the eCO₂RR and HER, the exceptional catalytic efficiency of Ti₁₂Ag_4.5 is clearly linked to its unique Ag coordination. This coordination is supported by a titanium–oxo core, with Ag sites evenly distributed across the cluster surface. Specifically, the Ag active sites in Ti₁₂Ag_4.5 exhibit a d-band center (ε_d) that is closer to the Fermi level at −3.04 eV. As a comparison, the bimetallic cluster Ti₂Ag₂ exhibits a lower ε_d value of −3.47 eV, likely attributed to the absence of the unique support effects found in Ti₁₂Ag_4.5. Furthermore, the projected density of states (PDOS) for Ti₁₂Ag_4.5 shows higher and narrower peaks at energies near the Fermi level, suggesting a higher and more localized density of electronic states (Fig. S53†). This enhanced electronic configuration facilitates stronger interactions with adsorbate *COOH molecules, improving the activation and subsequent transformation of CO₂. These electronic characteristics are critical for reactions requiring complex electron interactions, thus positioning Ti₁₂Ag_4.5 as a more effective catalyst for CO₂ reduction.

Conclusions

In summary, this study has designed and synthesized a new type of crystalline oxide cluster-stabilized single-atom material that can more precisely emulate the structural and reactive characteristics of oxide-supported single-atom catalysts at the molecular level. We have discovered a solvent-mediated assembly process that transforms the calixarene-stabilized titanium oxide cluster, Ti₁₂L₆, into the oxide-supported cluster, Ti₁₂Ag_4.5. This transformed cluster prominently features six accessible single Ag¹⁺ sites on its surface. The distinctive structure of Ti₁₂Ag_4.5 has demonstrated superior catalytic performance in the eCO₂RR, outperforming the traditional bimetallic cluster Ti₂Ag₂ in terms of reactivity and CO selectivity. In situ ATR-SEIRAS and DFT calculations were employed to investigate the catalytic mechanism, complemented by a structural comparison between Ti₁₂Ag_4.5 and Ti₂Ag₂ to clarify the factors underlying their different performances. These insights emphasize the importance of controlling the distribution of active metal sites on the surface of the cluster. This study promotes the intentional design of interactions between active surface sites and their cluster support, potentially establishing a new paradigm in single-atom catalyst design by significantly enhancing catalytic activities.

Data availability

The data that support the findings of this study are available in the main text and the ESI.†

Author contributions

C. L. supervised the project and conceived the idea. R. X. M., L. C. Z and L. K. W carried out synthesis, characterization and catalytic experiments of clusters. Q. T. and L. P. L performed the calculations for this article. Y.-L. S. was responsible for the single crystal testing. C. L. wrote the manuscript. All authors discussed the experimental results.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was supported by the Natural Science Foundation of Hunan Province (2023JJ30650), the Central South University Innovation-Driven Research Programme (2023CXQD061), and the National Natural Science Foundation of China (21901256 and 21805170). We thank Professor Mingzhao Chen and Dr Ning Wang from Guangzhou University for the mass spectrometry measurements.

Notes and references

1. L. Liu and A. Corma, Chem. Rev., 2018, 118, 4981–5079.
2. Y. Guo, S. Mei, K. Yuan, D. J. Wang, H. C. Liu, C. H. Yan and Y. W. Zhang, ACS Catal., 2018, 8, 6203–6215.
3. M. Murdoch, G. I. N. Waterhouse, M. A. Nadeem, J. B. Metson, M. A. Keane, R. F. Howe, J. Llorca and H. Idriss, Nat. Chem., 2011, 3, 489–492.
4. A. vonWeber and S. L. Anderson, Acc. Chem. Res., 2016, 49, 2632–2639.
5. B. Qiao, A. Wang, X. Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J. Li and T. Zhang, Nat. Chem., 2011, 3, 634–641.
6. S. Yang, J. Kim, Y. J. Tak, A. Soon and H. Lee, Angew. Chem., Int. Ed., 2016, 55, 2058–2062.
7. J. Zhang, X. Wu, W. C. Cheong, W. Chen, R. Lin, J. Li, L. Zheng, W. Yan, L. Gu, C. Chen, Q. Peng, D. Wang and Y. D. Li, Nat. Commun., 2018, 9, 1002.
8. C. M. Zhao, X. Y. Dai, T. Yao, W. X. Chen, X. Q. Wang, J. Wang, J. Yang, S. Q. Wei, Y. Wu and Y. D. Li, J. Am. Chem. Soc., 2017, 139, 8078–8081.
9. J. D. Yi, X. P Gao, H. Zhou, W. Chen and Y. Wu, Angew. Chem., Int. Ed., 2022, e202212329.
10. R. Lang, X. Du, Y. Huang, X. Jiang, Q. Zhang, Y. Guo, K. Liu, B. Qiao, A. Wang and T. Zhang, Chem. Rev., 2020, 120, 11986–12043.
11. X. F. Yang, A. Wang, B. Qiao, J. Li, J. Liu and T. Zhang, Acc. Chem. Res., 2013, 46, 1740–1748.
12. Z. Huang, X. Gu, Q. Cao, P. Hu, J. Hao, J. Li and X. Tang, Angew. Chem., Int. Ed., 2012, 51, 4198–4203.
13. C. T. Campbell, Acc. Chem. Res., 2013, 46, 1712–1719.
14. R. Jin, C. Zeng, M. Zhou and Y. Chen, Chem. Rev., 2016, 116, 10346–10413.
15. P. Liu, Y. Zhao, R. Qin, L. Gu, P. Zhang, G. Fu and N. Zheng, Sci. Bull., 2018, 63, 675–682.
16. W. H. Fang, L. Zhang and J. Zhang, Chem. Soc. Rev., 2018, 47, 404–421.
17. P. Coppens, Y. Chen and E. Trzop, Chem. Rev., 2014, 114, 9645–9661.
18. L. Rozes and C. Sanchez, Chem. Soc. Rev., 2011, 40, 1006–1030.
19. Y. Lv, J. Cheng, A. Steiner, L. Gan and D. S. Wright, Angew. Chem., Int. Ed., 2014, 53, 1934–1938.
20. C. Zhao, Y. Z. Han, S. Dai, X. Chen, J. Yan, W. Zhang, H. Su, S. Lin, Z. Tang, B. K. Teo and N. Zheng, Angew. Chem., Int. Ed., 2017, 56, 16252–16256.
21. G. Zhang, C. Liu, D. L. Long, L. Cronin, C. H. Tung and Y. Wang, J. Am. Chem. Soc., 2016, 138, 11097–11100.
22. W. H. Fang, L. Zhang and J. Zhang, J. Am. Chem. Soc., 2016, 138, 7480–7483.
23. M. Y. Gao, F. Wang, Z. G. Gu, D. X. Zhang, L. Zhang and J. Zhang, J. Am. Chem. Soc., 2016, 138, 2556–2559.
24. J. L. Hou, N. H. Huang, D. Acharya, Y. X. Liu, J. Y. Zhu, J. X. Teng, Z. Wang, K. Q. Qu, X. X. Zhang and D. Sun, Chem. Sci., 2024, 15, 2655–2664.
25. M. Zhang, M. Lu, M.-Y. Yang, J.-P. Liao, Y.-F. Liu, H.-J. Yan, J.-N. Chang, T.-Y. Yu, S.-L. Li and L. Y. Qian, eScience, 2023, 3, 100116.
26. Y. J. Liu, P. Shao, M. Y. Gao, W. H. Fan and J. Zhang, Inorg. Chem., 2020, 59, 11442–11448.
27. X. Fan, F. Yuan, D. Li, S. Chen, Z. Cheng, Z. Zhang, S. Xiang, S. Q. Zang, J. Zhang and L. Zhang, Angew. Chem., Int. Ed., 2021, 60, 12949–12954.
28. M. Y. Gao, K. Wang, Y. Y. Sun, D. J. Li, B. Q. Song, Y. H. Andaloussi, M. J. Zaworotko, J. Zhang and L. Zhang, J. Am. Chem. Soc., 2020, 142, 12784–12790.
29. N. Li, J.-M. Lin, R.-H. Li, J.-W. Shi, L.-Z. Dong, J. Liu, J. He and Y. Q. Lan, J. Am. Chem. Soc., 2023, 145, 16098–16108.
30. E. M Han, R. X Meng, Y. Q Tian, J. Yan, K. Y Liu and C. Liu, Chem. Commun., 2023, 59, 11097–11100.
31. S. Chen, Z. N. Chen, W. H. Fang, W. Zhuang, L. Zhang and J. Zhang, Angew. Chem., Int. Ed., 2019, 58, 10932–10935.
32. D. T. Geng, X. Han, Y. F. Bi, Y. C. Qin, Q. Li, L. L. Huang, K. Zhou, L. J. Song and Z. P. Zheng, Chem. Sci., 2018, 9, 8535–8541.
33. S. T. Wang, X. H. Gao, X. X. Hang, X. F. Zhu, H. T. Han, W. P. Liao and W. Chen, J. Am. Chem. Soc., 2016, 138, 16236–16239.
34. Y. Q. Tian, W. L. Mu, L. L. Wu, X. Y. Yi, J. Yan and C. Liu, Chem. Sci., 2023, 14, 10212–10218.
35. Z. Wang, L. Li, L. Feng, Z. Y. Gao, C. H. Tung, L. S. Zheng and D. Sun, Angew. Chem., Int. Ed., 2022, 61, e202200823.
36. Z. Wang, H. F. Su, Y. W. Gong, Q. P. Qu, Y. F. Bi, C. H. Tung, D. Sun and L. S. Zheng, Nat. Commun., 2020, 11, 308.
37. Z. Wang, F. Alkan, C. M. Aikens, M. Kurmoo, Z. Y. Zhang, K. P. Song, C. H. Tung and D. Sun, Angew. Chem., Int. Ed., 2022, 61, e202206742.
38. Z. J. Guan, F. Hu, S. F. Yuan, Z. A. Nan, Y. M. Lin and Q. M. Wang, Chem. Sci., 2019, 10, 3360–3365.
39. X.-M. Luo, S. Huang, P. Luo, K. Ma, Z.-Y. Wang, X.-Y. Dong and S.-Q. Zang, Chem. Sci., 2022, 13, 11110–11118.
40. L. J. Li, Y. T. Luo, Y. Q. Tian, P. Wang, X. Y. Yi, J. Yan, Y. Pei and C. Liu, Inorg. Chem., 2023, 62, 14377–14384.
41. Z. Liu, J. Chen, B. Li, D.-e. Jiang, L. Wang, Q. Yao and J. Xie, J. Am. Chem. Soc., 2024, 146(17), 11773–11781.
42. S. L. Zhuang, D. Chen, L. W. Liao, Y. Zhao, N. Xia, W. Zhang, C. Wang, J. Yang and Z. Wu, Angew. Chem., Int. Ed., 2020, 59, 3073–3077.
43. J. Y. Xu, L. Xiong, X. Cai, S. S. Tang, A. C. Tang, X. Liu, Y. Pei and Y. Zhu, Chem. Sci., 2022, 13, 2778–2782.
44. X. Liu, E. D. Wang, M. Zhou, Y. Wan, Y. K. Zhang, H. Q. Liu, Y. Zhao, J. Li, Y. Gao and Y. Zhu, Angew. Chem., Int. Ed., 2022, 61, e202207685.
45. S. Li, A. V. Nagarajan, D. R. Alfonso, M. K. Sun, D. R. Kauffman, G. Mpourmpakis and R. Jin, Angew. Chem., Int. Ed., 2021, 60, 6351–6356.
46. S. Zhao, N. Austin, M. Li, Y. Song, S. D. House, S. Bernhard, J. C. Yang, G. Mpourmpakis and R. Jin, ACS Catal., 2018, 8, 4996–5001.
47. L. Tang, Y. Luo, X. Ma, B. Wang, M. Ding, R. Wang, P. Wang, Y. Pei and S. Wang, Angew. Chem., Int. Ed., 2023, e202300553.
48. L. J. Li, W. L. Mu, Y. Q. Tian, W. D. Yu, L. Y. Li, J. Yan and C. Liu, Chem. Sci., 2024, 15, 7643–7650.
49. G. C. Deng, H. Yun, M. S. Bootharaju, F. Sun, K. Lee, X. L. Liu, S. Yoo, Q. Tang, Y. J. Hwang and T. Hyeon, J. Am. Chem. Soc., 2023, 145(50), 27407–27414.
50. L. B. Qin, F. Sun, X. S. Ma, G. Y. Ma, Y. Tang, L. Wang, Q. Tang, R. Jin, Z. Tang and Z. H. Tang, Angew. Chem., Int. Ed., 2021, 60, 26136–26141.
51. X. S. Ma, F. Sun, L. B. Qin, Y. G. Liu, X. W. Kang, L. K. Wang, D. E. Jiang, Q. Tang and Z. H. Tang, Chem. Sci., 2022, 13, 10149–10158.
52. S. Yoo, S. Yoo, G. C. Deng, F. Sun, K. Lee, H. Jang, C. W. Lee, X. L. Liu, J. Jang, Q. Tang, Y. J. Hwang, T. Hyeon and M. S. Bootharaju, Adv. Mater., 2024, 36, 2313032.
53. X. L. Liu, T. Ki, G. C. Deng, S. Yoo, K. Lee, B.-H. Lee, T. Hyeon and M. S. Bootharaju, Nanoscale, 2024, 16, 12329–12344.
54. G. Deng, J. Kim, M. S. Bootharaju, F. Sun, K. Lee, Q. Tang, Y. J. Hwang and T. Hyeon, J. Am. Chem. Soc., 2023, 145, 3401–3407.
55. J. Wang, F. Xu, Z. Y. Wang, S. Q. Zang and T. C. W. Mak, Angew. Chem., Int. Ed., 2022, 61, e202207492.
56. L. J. Liu, Z. Y. Wang, Z. Y. Wang, R. Wang, S. Q. Zang and T. C. Mak, Angew. Chem., Int. Ed., 2022, 61, e202205626.
57. Y. F. Lu, L. Z. Dong, J. Liu, R. X. Yang, J. J. Liu, Y. Zhang, L. Zhang, Y. R. Wang, S. L. Li and Y. Q. Lan, Angew. Chem., Int. Ed., 2021, 60, 26210–26217.
58. D. Wang, J. J. Mao, C. C. Zhang, J. W. Zhang, J. S. Li, Y. Zhang and Y. F. Zhu, eScience, 2023, 3, 100119.
59. X. Wang, M. H. Yu and X. L. Feng, eScience, 2023, 3, 100141.

---

Footnotes

† Electronic supplementary information (ESI) available: X-ray crystallographic file in CIF format, and full experimental and computational details. CCDC 2349621–2349625. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4sc07186g

‡ These authors contributed equally.

---