Mechanism of Adhesion Particle Removal Using Supercritical CO₂ and Simulation of Deep Groove Cleaning in Semiconductor

The removal of adhered particles is critical in semiconductor manufacturing. This study develops a kinetic model for particle removal in supercritical carbon dioxide (scCO₂), analyzing the dynamic variations of forces and revealing key mechanisms. Results show that particle removal typically occurs via rolling motion. Under laminar flow, removal is influenced by both driving and lift forces, while lift force dominates in turbulent flow. For Cu and Au particles, van der Waals forces dominate at smaller sizes, whereas deformation forces are significant for larger particles. For polystyrene latex (PSL) particles, deformation forces consistently dominate. A Stokes number-based evaluation method is proposed to assess scCO₂ cleaning performance in deep trenches. Mixing methanol or ethanol with scCO₂ enhances cleaning, enabling the removal of particles smaller than 400 and 10 nm in trenches 25 μm and 50 nm wide, respectively. These findings provide theoretical guidance for optimizing scCO₂ cleaning in semiconductor manufacturing.