Insights of Ammonia Decomposition on W--B Nanoclusters by Computational Simulations
Tungsten-boride nanoclusters represent a promising class of materials for catalytic applications, yet their structural stability and reactivity remain poorly understood. The evolutionary algorithm combined with density functional theory (DFT) are used to systematically explore the ground-state structures and stability landscape of W$_m$B$_n$ nanoclusters with up to 43 atoms. The resulting stability maps reveal a highly non-monotonic landscape characterized by isolated "magic" compositions, including WB$_{16}$, W$_2$B$_8$, W$_7$B$_{24}$, and W$_{11}$B$_{22}$, which exhibit pronounced local stability maxima. We further investigate the adsorption and initial decomposition step of ammonia on these clusters as a probe of their catalytic potential. Molecular NH$_3$ adsorption occurs exclusively on tungsten sites with energies ranging from -0.54 to -1.78 eV (average -1.43 eV), comparable to Pt$_n$ and Fe$_n$ clusters. Atomic hydrogen adsorption spans a broader range from +0.49 to -1.46 eV, reflecting high site sensitivity. Nudged elastic band calculations for the first N--H bond cleavage reveal forward barriers of 1.1-1.4 eV, with the dissociated NH$_2^*$ + H$^*$ state lying below the molecular adsorption state for most compositions. Notably, the activation barrier depends critically on the local environment available for stabilizing the detached hydrogen atom. These findings establish W--B nanoclusters as tunable catalysts for ammonia decomposition and provide a structural foundation for their rational design.