In order to get insight into the interplay between metal ultradispersion and catalytic performance, atomically dispersed Pt/γ-Al2O3 catalysts with 0.1 or 0.5 wt% Pt loading, were prepared by impregnation-based methods and tested for CO oxidation. According to scanning transmission electron microscopy (STEM), only single Pt atoms (SAs) are present on the 0.1 wt% catalyst, while they coexist with a small fraction of subnanometric clusters in the 0.5 wt% one. The Pt SAs are stable during CO oxidation up to 300 °C in oxygen excess, but exhibit relatively low activity. In contrast, after H2 pretreatment, the SAs aggregate into subnanometric clusters, which exhibit high CO oxidation performance. Environmental STEM under 5 mbar of H2 shows that the SA-to-clusters aggregation already occurs at room temperature, and the resulting Pt clusters appear stable in size up to 800 °C in this reducing atmosphere. 1. Scope Noble metals are highly active in many catalytic reactions but they are rare and expensive, limiting their industrial use and requiring a maximization of their efficiency, or their replacement by cheaper metals. An “atom-economical” strategy consists in dispersing small metal amounts, ideally in the form of isolated atoms, on a suitable support. This induces a change in surface electronic structure, which can lead to different/superior performances (activity, selectivity, stability) as compared to conventional catalysts. However, the preparation of so-called “single-atom catalysts” (SACs)1 is challenging because noble metal atoms tend to aggregate in order to minimize their surface energy. 2. Results and discussion In this work, we have prepared Pt/γ-Al2O3 catalysts using conventional impregnation methods. To maximize the metal dispersion, relatively low metal loadings (0.1-0.5 wt%) and an oxidizing thermal treatment were used. Aberration-corrected scanning transmission electron microscopy (STEM) analyses of the as-prepared 0.1 wt% Pt/γ-Al2O3 catalysts exclusively show single Pt atoms (Figure 1a). The increase in the Pt loading (0.5 wt%) leads to the formation of some subnanometric clusters in addition to single atoms (Figure 1b). The 0.1 wt% Pt1/γ-Al2O3 SAC was tested for CO oxidation in a flow-fixed-bed reactor. Three CO oxidation heating/cooling cycles separated by calcination or reduction treatments were performed (Figure 2). The catalyst was analyzed at each step by STEM. As a result, the Pt SAs are retained after the first calcination/reaction step, then partially converted into subnanometric clusters after reduction, and finally replaced by 2 nm-NPs after the following steps (Figure 2a). Considering only the first (heating) reaction runs, the CO oxidation activity follows the order: SAs < NPs < clusters (Figure 2b). Moreover, a strong hysteresis is observed in all cases, the catalyst being always much more active during the cooling stage than during the heating one. The relatively low activity of the SAC is consistent with the results of Moses-DeBusk et al. for Pt/θ-Al2O3.2 3. Conclusions Many recent works praise the use of SACs in various catalytic reactions, however the stability and the homogeneity of the metal dispersion during the reactions are not always systematically assessed. In this work, we have combined CO oxidation tests of atomically dispersed Pt/γ-Al2O3 catalysts with STEM analysis performed before, after or during the reaction/adsorption processes. Two kinds of Pt entities were present depending on the synthesis conditions: SAs only (0.1 wt% Pt loading) or a mixture of SAs and subnanometric clusters (0.5 wt%). Despite their low activity, Pt SAs are stable during O2-rich CO oxidation cycles up to 300 °C. In contrast, SAs aggregate into stable clusters upon H2 pretreatment. The latter are much more active than SAs and NPs in CO oxidation. The interplay between the size, the structure, the oxidation state and the catalytic performance of the supported Pt entities, together with the strong hysteresis phenomena, will be discussed in the light of additional microscopic and spectroscopic analyses. References 1. J. Liu, ACS. Catal. 2017, 7, 34-59. 2. M. Moses-DeBusk, M. Yoon, L. F. Allard, D. R. Mullins, Z. Wu, X. Yang, G. Veith, G. M. Stocks and C. K. Narula, J. Am. Chem. Soc. 2013, 135, 12634-12645.