(265e) Elucidating How Nanoparticle Stability Metrics Depend on the Choice of Exchange Correlation Functionals Using Analytical Frameworks

Recent efforts seek to efficiently compute catalyst stability metrics, thus complementing existing capabilities in obtaining reactivity trends from volcano plots. These efforts determine metal-metal interactions using density functional theory (DFT)-derived surrogate models. Abild-Pedersen’s group formulated the alloy stability model to estimate energies of metal atoms in nanoparticles with active site specificity. This model unites linear functions correlating energies of metal atoms with varying composition at a fixed coordination number, and quadratic functions correlating metal atom energies with varying coordination numbers at a fixed composition. The metal atom energies in turn yield surface energies, cohesive energies of nanoparticles, segregation energies, and Wulff nanoparticles. Herein, leverage the analytical form of the alloy stability model to quantify the sensitivity of catalyst stability metrics on the DFT method employed. Across a family of different GGA functionals (PBE, PBESol, RPBE, PBE-D3, BEEF-vdW), we consistently obtain the linear and quadratic correlations within the alloy stability model, thus validating our model ansatz. Using surrogate models parametrized with these functionals, we estimate the surface energies of (111), (100), (110), and (211) surfaces on transition metals. These surface energies are used to generate Wulff constructions. The model predicted cohesive energies of Wulff nanoparticles varies inversely with nanoparticle size, thus following the classical Gibbs Thomson relationship across all considered functionals. We now illustrate how the density of stepped and terrace sites on Wulff nanoparticles depends on the DFT functional. We find that the densities of stepped and terrace active sites are weakly sensitive (within ~ ±10%) to the choice of the exchange correlation functional. This weak sensitivity implies that obtaining active site densities from Wulff nanoparticles represents a complementary approach to compute turnover frequencies from mass-averaged reaction rates. In summary, we elucidate an accelerated approach to quantify the sensitivity of the DFT method on catalyst stability metrics.