Nanocluster Active Sites Formed on Heterogeneous Thermal Catalysts and Electrocatalysts by Operando Reactive Environments

Figure 1. (A) A simplified, atomic-scale illustration of the cluster formation process on a metal nanoparticle catalyst in the presence of reaction intermediates (e.g., NH₃). (23) The scissors schematically denote where the adsorbate-induced metal–metal bond cleavage events might easiest take place (i.e., along the edges, corners, and kink sites), especially upon heating (represented by the triangle below the arrow), leaving vacancies behind. (B) Illustration of the metal nanocluster formation process on a close-packed terrace due to ejection of kink or step-edge atoms. Dashed circles indicate kink and step-edge vacancies formed after metal atom ejection. Green, blue, yellow, and red spheres represent metal atoms on a (kinked) step edge, adatoms and clusters on the terrace, terrace metal atoms (the shaded area denotes the upper terrace atoms at the top-left corner); and adsorbate species (atomic or molecular), respectively. (C) Schematic representations of each energy term in the definition for the adatom formation energy due to ejection of a (874) kink atom, both under vacuum or in the presence of an adsorbate. Figure adapted with permission from refs (23,29). Copyright 2023 The American Association for the Advancement of Science. Figure 2. Heatmap of calculated adsorbate-induced adatom formation energies on close-packed terraces: (111), (0001), and (110) for FCC, HCP, and BCC, respectively. First row in the table provides entries under vacuum. “*” denotes cases where the close-packed terrace is the preferred ejection source; in all other cases, the defect sites are the preferred ejection sources (kink sites for FCC metals, regular step-edge sites for BCC and HCP metals). Gray-faded numbers indicate systems in which the adsorbate hinders adatom formation. All energy values were evaluated at the low-coverage limit of each adsorbate. Energies were calculated with GGA-PBE. The metals are listed from left to right in ascending order of experimental bulk cohesive energy. (47) The figure is replotted using data/adapted with permission from ref (23). Copyright 2023 The American Association for the Advancement of Science, and adapted from ref (28). Copyright 2023, The Authors, American Chemical Society under a Creative Commons license (CC-BY-NC-ND 4.0, https://creativecommons.org/licenses/by-nc-nd/4.0/). Where the estimated temperatures (for 1 ejection/sec) based on the calculated thermochemical energetics (within ∼ 0.2 eV typical of DFT error (32)) of metal atom ejection and subsequent nanocluster formation are lower than or close to typical applied temperatures (as cited from the corresponding references for each application). Revisiting the nature of the active site and the details of the reaction mechanism in these systems might be warranted. Figure 3. (A) Brønsted–Evans–Polanyi correlation between the activation energy barrier and the reaction energy for the CO-induced Cu atom ejection from the (874) kink site at various CO coverages. (B) Calculated cluster formation energies due to Cu(874) kink atom ejection as a function of the Cu cluster size in the final state on Cu(111) under vacuum as well as at intermediate CO coverages [i.e., 5/14 ML of CO coverage on Cu(874); half of the Cu cluster atoms formed on Cu(111) were covered by CO adsorbates) and high CO coverages (7/14 ML of CO adsorbed on Cu(874); all peripheral Cu atoms and the center Cu atom of each cluster on Cu(111) were occupied by one CO molecule each]. (C) Comparison of the calculated transition-state energy for CO–O* TS (E_TS) on Cu(111) plus adclusters (represented by the inset images) and pristine Cu slab models, arranged in ascending order of the coordination number (CN) of the surface site on which the CO oxidation reaction takes place. All energy values were evaluated at the low-coverage limit of each adsorbate. Zero energy is defined as the total energy of the respective clean surface plus CO and 1/2 O₂ in the gas phase. Green, cluster Cu atom or adatom; orange, Cu atom belonging to Cu(111). (D) Kinetic Monte Carlo (KMC) snapshots of Cu(111) surface exposed to 0.3-torr CO at room temperature at various simulation times (t_sim) over a total period of 105.0 min, with DFT-calculated ejection energies lowered by 0.15 eV to match experimental STM data. A fully CO-covered Cu(874) kink-containing step-edge, with 25% of the total number of Cu(111) terrace atoms, was assumed to be implicitly present and available for atom ejection. Red, blue, and black circles indicate catalytically active Cu₃ clusters, CO adsorption on hollow sites on Cu clusters (also indicated by the blue arrow), and CO adsorption atop Cu(111) atoms, respectively. Figure adapted with permission from ref (23). Copyright 2023 The American Association for the Advancement of Science. Figure 4. (A) A simplified, atomic-level illustration of the adatom formation process on a close-packed bimetallic alloy surface A₃B(111) with a kink-containing step-edge. Green and yellow spheres denote metal A and B, respectively. Dashed red and black circles indicate vacancies formed after ejections from the step-edge and kink sites, respectively. (B) (Left axis) Calculated CO-induced adatom formation energy (E_form; vertical bars) on Au(111), Cu(111), and Au_xCu_y(111) due to the ejection of a (874) kink atom (preferred adatom formation process) or of a (211) step-edge atom. (Right axis) Experimental relative turnover rate (TOR; open red circles) for CO₂ electroreduction to CO on Au–Cu bimetallic nanoparticles at −0.73 V vs reversible hydrogen electrode, normalized to pure Cu. The line segments drawn to connect circles were added to guide the eye. TOR data were taken from ref (49). Figure adapted with permission from ref (31). Copyright 2025 Elsevier. In the hypothesized mechanism that includes ejection, diffusion, and coalescence of metal atoms into nanoclusters, ejection is typically the rate-determining step. The ejection step often has an activation energy barrier that is minimally higher than the reaction energy of that step, allowing for the use of the ejection energy as a first-order descriptor for the entire nanocluster formation process. Adsorbate coverage plays an important role in modulating the metal atom ejection energy and its kinetics. Ejections, and thereby cluster formation, could happen at much lower temperatures than previously thought, affecting many common catalytic systems. If the operando formed clusters themselves show promising catalytic activity (e.g., Cu trimers for CO oxidation), the overall rate could be dictated by the formation of such clusters. The miniature nature of those clusters resembles the active metal-centers of homogeneous catalysts, thus providing a novel avenue for achieving the coveted bridge between homogeneous and heterogeneous catalysis. This work is supported by the US Department of Energy, Basic Energy Sciences (DOE-BES), Division of Chemical Sciences, Catalysis Science Program, grant number DE-FG02-05ER15731. We used resources at the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the US Department of Energy under contract no. DE-AC02-05CH11231 using NERSC award number BES-ERCAP0032205. We thank Dr. Lang Xu and Evangelos Smith for valuable feedback on the content of this manuscript. The authors dedicate this work to the late mother of A.O.E., Prof. Dr. Neveen Yosef, the professor of political theory in the Faculty of Economics and Political Science, Cairo University, Egypt. This article references 69 other publications. This article has not yet been cited by other publications.