Optimization of dual-fluid wafer cleaning system through CFD simulation and dimensional analysis of dynamic liquid film behavior

As semiconductor devices continue to scale down, wafer cleaning is a critical step for manufacturing yield and device reliability. This study presents a framework integrating Computational Fluid Dynamics with experimental validation to investigate dual-fluid wafer cleaning by examining liquid film dynamics on rotating wafers. CFD simulations, coupling the Volume of Fluid and Discrete Phase Model approaches, modeled droplet impact, liquid film evolution, and key hydrodynamic parameters. These CFD predictions were validated through physical cleaning experiments using SiO₂ nanoparticles. The validation demonstrated a correlation between simulated wall shear stress and experimentally measured wafer cleaning efficiency, and also between simulated film uniformity and experimental cleaning uniformity. Key findings indicate that while an increasing Reynolds number enhances turbulence and shear stress, it may compromise liquid film uniformity. Increasing the Weber number not only contributes to higher cleaning efficiency but also promotes more uniform liquid film coverage. An empirical equation derived from Re and We offers a predictive tool for liquid film thickness, intended for scalable cleaning process optimization. This work provides an experimentally validated theoretical and practical foundation for optimizing semiconductor wafer cleaning, contributing to considerations of efficiency and process scalability.