Atomic-level Insights Into Atomically Dispersed Metal Catalysts

Abstract

Atomically dispersed metal catalysts have demonstrated promise in heterogeneous catalysis with both nearly 100% utilization efficiency of precious metals and high intrinsic activity of metal sites. As an emergent type of atomically dispersed metal catalysts, single-atom catalysts (SACs) have been extensively investigated with broad applications. However, rational design of improved SACs for targeted chemistry faces tremendous challenges, largely due to lack of atomic-level understanding into the site structures and reaction mechanisms under operating conditions. Herein, we report the efforts that we have been put in to gain atomic-level insights into the nature of SACs, and thus to design atomically dispersed metal catalysts with improved catalytic performance. By combining density functional theory (DFT) calculations with experiments, the complex nature of the Ir single atoms supported on anatase TiO2 (Ir1/TiO2) for CO oxidation is studied systematically (Chapter 2). The local configuration of the Ir1/TiO2 catalysts during the reduction in CO(g) and the reaction in CO(g) and O2(g) is elucidated. Based on the elucidated structures, we systematically investigate the reaction mechanisms and thus identify the two rate-determining elementary steps (RDSs) for CO oxidation on the Ir1/TiO2 catalysts, including CO adsorption/oxidation and O2 dissociation. This is further demonstrated by three Ir1 complexes along the reaction cycle that were isolated and identified using in-situ spectroscopy. To further elucidate the CO oxidation kinetics of the Ir1/TiO2, detailed MKM is performed (Chapter 3). However, even though DFT and DFT-based MKM show promise in revealing active sites and mechanisms, it falls short to capture detailed reaction kinetics, largely due to the uncertainties of DFT energetics. To bridge the gap between the theory and experiments, we assessed the uncertainties of DFT energetics and implemented such uncertainties into the MKM. With the catalytic nature of Ir1/TiO2 now better understood, we next investigate the fundamental factors that govern the catalytic activity trends in M1/TiO2 systems(Chapter 4). By analyzing the electronic structures of the initial states and transition states of the two RDSs, we found that the electronic properties of the Ir1 determine the CO oxidation kinetics of the Ir1/TiO2, while the TiO2 surface mainly provides anchoring sites for the Ir1 intermediates. Furthermore, a linear scaling between O binding energy (ΔEO, the property of M1 to bind *O) and O2 dissociation energy is found, suggesting a tradeoff between the removal and regeneration of the active *O on the M1/TiO2 catalysts. Thus, a simple rule between the catalytic activity and ΔEO is constructed, dictating the activity trends in M1/TiO2 for CO oxidation. However, even if the better SACs can be designed for targeted chemistry, their performances are not comparable to the state-of-art metal cluster catalysts. This can be arguably attributed to their overly rigid local structures and the isolation of active sites which lack synergistic effects from neighboring metal atoms. Thereafter, we develop atomically dispersed metal ensembles to overcome the limitation of SACs. In contrast to Mo and Fe monomers, and Fe2 dimers in the channel of ZSM-5, homonuclear Mo2 and heteronuclear MoFe dimer catalysts possess high thermodynamic stability under the He pretreatment. This enhanced stability may explain the superior performance of the Fe2(MoO4)3/ZSM-5 catalyst compared to the MoO3+Fe2O3/ZSM-5 catalyst (Chapter 5). To expand this scope, we used Pt single atoms embedded into the CeO2 lattice as "seeds" to construct embedded Pt atomic single layer (PtASL) catalyst (as a type of atomically dispersed metal ensembles) which exhibits the highest turnover frequency for CO oxidation among Pt catalysts (Chapter 6). Such high performance of the embedded PtASL originated from the appropriate Pt-CO bonding and improved activation/reactivity of lattice oxygens within CeO2.