Samsung Advanced Institute of Technology

In this study, we identify the local structures of ex-solved nanoparticles using machine-learned potentials. We develop a method for training machine-learned potentials by sampling local structures of heterointerface configurations as a training set, with its efficacy tested on the Ni/MgO system, illustrating that the error in interface energy is only 0.004 eV/？². Using the developed scheme, we train an MLP for Ni/La_0.5Ca_0.5TiO₃ ex-solution system and identify the local structures for both exo- and endo-type particles. The established model aligns well with the experimental observation, accurately predicting nucleation size of 0.45 nm. Lastly, the density-functional theory calculations on the established atomistic model verify that the kinetic barriers for the dry reforming of methane are substantially reduced by 0.49 eV on the ex-solved catalysts compared to the impregnated catalysts. From our findings, we offer insights into the local structures, growth mechanisms, and the underlying origin for the catalytic properties of ex-solved nanoparticles.