Advanced Atomic-scale Insights into the Chemical Mechanical Polishing Process with Ceria Abrasives using Molecular Dynamics and Neural Network Potential

We have advanced the development of a Chemical Mechanical Polishing (CMP) slurry for polishing silica substrates with ceria abrasives, crucial for semiconductor manufacturing. By integrating Molecular Dynamics simulations with Neural Network Potential and using the Open Catalyst 2022 Dataset, we have explored atomic-scale tribochemical interactions, with a special focus on silica surfaces of varying silanol group densities and geometries. This method enables cost-effective and accurate simulations, offering insights into the reactivity of ceria and silica, particularly regarding the impact of hydroxyl group concentration and mechanical pressure on Ce-O-Si bond formation and silica hydrolysis. The incorporation of Virtual Reality technology enhances the visualization of these simulations, encouraging collaborative progress. Our research enriches the understanding of CMP, contributes to the refinement of CMP slurries, and marks a significant leap in semiconductor fabrication technology.