Search for: All records

Recently, we found that the atomic ensemble effect is the dominant effect influencing catalysis on surfaces alloyed with strong- and weak-binding elements, determining the activity and selectivity of many reactions on the alloy surface. In this study we design single-atom alloys that possess unique dehydrogenation selectivity towards ethanol (EtOH) partial oxidation, using knowledge of the alloying effects from density functional theory calculations. We found that doping of a strong-binding single-atom element ( e.g. , Ir, Pd, Pt, and Rh) into weak-binding inert close-packed substrates ( e.g. , Au, Ag, and Cu) leads to a highly active and selective initial dehydrogenation at the α-C–H site of adsorbed EtOH. We show that many of these stable single-atom alloy surfaces not only have tunable hydrogen binding, which allows for facile hydrogen desorption, but are also resistant to carbon coking. More importantly, we show that a rational design of the ensemble geometry can tune the selectivity of a catalytic reaction. 

Surface segregation in bimetallic nanoparticles (NPs) is critically important for their catalytic activity because the activity is largely determined by the surface composition. Little, however, is known about the atomic scale mechanisms and kinetics of surface segregation. One reason is that it is hard to resolve atomic rearrangements experimentally. It is also difficult to model surface segregation at the atomic scale because the atomic rearrangements can take place on time scales of seconds or minutes – much longer than can be modeled with molecular dynamics. Here we use the adaptive kinetic Monte Carlo (AKMC) method to model the segregation dynamics in PdAu NPs over experimentally relevant time scales, and reveal the origin of kinetic stability of the core@shell and random alloy NPs at the atomic level. Our focus on PdAu NPs is motivated by experimental work showing that both core@shell and random alloy PdAu NPs with diameters of less than 2 nm are stable, indicating that one of these structures must be metastable and kinetically trapped. Our simulations show that both the Au@Pd and the PdAu random alloy NPs are metastable and kinetically trapped below 400 K over time scales of hours. These AKMC simulations provide insight into the energy landscape of the two NP structures, and the diffusion mechanisms that lead to segregation. In the core–shell NP, surface segregation occurs primarily on the (100) facet through both a vacancy-mediated and a concerted mechanism. The system becomes kinetically trapped when all corner sites in the core of the NP are occupied by Pd atoms. Higher energy barriers are required for further segregation, so that the metastable NP has a partially alloyed shell. In contrast, surface segregation in the random alloy PdAu NP is suppressed because the random alloy NP has reduced strain as compared to the Au@Pd NP, and the segregation mechanisms in the alloy require more elastic energy for exchange of Pd and Au and between the surface and subsurface.