Abstract
The electrodeposition of polycrystalline nickel coatings with tailored microstructures is critical for enhancing mechanical performance and surface quality. However, while experimental studies show that nanoparticles can refine grains and prevent columnar growth, it is still not clear what nanoparticle types, concentrations and shapes minimize this phenomenon. In this work, we combine phase-field modelling and a machine learning approach to study the impact of material and process parameters on the resulting microstructures during electrodeposition. We simulated polycrystalline nickel electrodeposition in the presence of nanoparticles with varying geometries (spherical and 2D) and concentrations, and used Random Forest classifier, interpreted via Shapley additive explanation analysis, to identify the key parameters controlling microstructural evolution. Our results quantitatively demonstrate that the gradient coefficient and overpotential are the most influential factors, while nanoparticle concentration and shape are critical in suppressing columnar growth. Specifically, we observed that 2D nanomaterials like MoS2 or graphene are of high interest when their volume fraction is larger than 3% since they reduce drag and promote smooth interface movement, thus preventing columnar growth.