(684b) Predicting the Morphology of Metal Nanoparticles Decorated with p-Block Promoters on-the-Fly

Central to the success of liquid organic hydrogen carriers is finding dehydrogenation catalysts that selectively dissociate C-H bonds while preserving C-C bonds. This selectivity can be steered using p-block promoters like sulphur that sterically block unselective low coordinated catalytic sites while electronically modifying other active sites such that the binding of the dehydrogenation product is weakened. Promoters like sulphur are employed in industrial catalysts and their identity/concentration is determined by trial-and-error. Due to trade-offs between selectivity and reactivity, there is a need to systematically screen different promoters, determine their optimal coverages, and understand if promoters alter the nanoparticle morphology.

Equilibrium morphologies of promoted nanoparticles are determined by Wulff constructions-which require surface energies of promoter-decorated crystal planes as inputs. Computing these surface energies with density functional theory (DFT) is challenging because of the large configurational space, especially at high promoter coverages. We present a physics-based surrogate model that estimates the surface energies (Ω) of arbitrary (hkl) planes decorated with promoters, on-the-fly, with accuracies of ~0.003 eV/A² compared to DFT. Using the surrogate model, we determine Ω as a function of sulphur coverage for all possible configurations on Pt (hkl) facets. The ab initio phase diagrams reveal that equilibrium sulphur coverages for the (111) and (100) facets are between 0.5-0.7 monolayer while the (211) edge-sites are covered by 1 monolayer of sulphur. These Ω values are inputted into Wulff constructions, yielding Pt nanoparticles having diameters between 2 and 10 nm. By tracking the densities of edge (unselective) and terrace (selective) sites; and the fractional area of each (hkl) facet, we observe that sulphur induces morphological changes to the Pt nanoparticles. If combined with a microkinetic model, our surrogate model can identify suitable promoters, their concentrations, and nanoparticle sizes that maximize the rate per mass of precious metal for dehydrogenation reactions.