Automated Detection and Characterization of Surface Restructuring Events in Bimetallic Catalysts

Surface restructuring in bimetallic systems has recently been shown to play a crucial role in heterogeneous catalysis. In particular, the segregation in binary alloys can be reversed in the presence of strongly bound adsorbates. Mechanistic characterization of such restructuring phenomena at the atomic level remains scarce and challenging because of the large configurational space that must be explored. To this end, we propose an automated method to discover elementary surface restructuring processes in an unbiased fashion using Pd/Ag as an example. We employ high-temperature classical molecular dynamics to rapidly detect restructuring events, isolate them, and optimize using density functional theory. In addition to confirming the known exchange descent mechanism, our systematic approach has revealed three new predominant classes of events at step edges of close-packed surfaces that have not been considered before: (1) vacancy insertion; (2) direct exchange; and (3) interlayer exchange. The discovered events enable us to construct the complete set of mechanistic pathways by which Pd is incorporated into the Ag host in vacuum at the single-atom limit. These atomistic insights provide a step toward systematic understanding and engineering of surface segregation dynamics in bimetallic catalysts.