SiO2 thin film creation in Diffusion furnace - Process Functional Modeling

This project applies Functional Modeling to analyze the single-wafer wet etch process in semiconductor manufacturing. The model shows that wet etch defects are not caused only by chemistry, but also by interactions between the wafer, liquid flow, air, rotation, rinsing, drying, drain, and chamber environment. Key challenges include incomplete wetting, trapped air or bubbles, residue and particle redeposition, evaporation-driven watermarks, and contamination during rinse or drying. The project identifies improvement directions such as better wafer pre-wetting, controlled airflow, reduced evaporation, optimized rinse and drying conditions, and alternative chamber concepts such as face-down processing in a shallow liquid bath.

Wet cleaning is widely used in microchip manufacturing. Single wafer equipment is working as follows. A wafer rotates, and chemistry is poured from a movable nozzle. Water rinsing is performed at the end of the process. Loading of a new batch of the chemistry resulted in excursion - a strongly increased amount of defects was observed on the wafer after the processing. The project is dedicated to the failure analysis and creation of innovative solutions.

This project investigates particle-defect formation during SiO₂ dry etch. Functional modeling revealed that defects can be generated by two coupled mechanisms: chemical formation of SiO₂/SiOFₓ particles from SiF₄ interaction with moisture, and mechanical generation of particles by excessive ion bombardment and sputtering. The proposed improvement direction is to balance chemical etching and ion-assisted etching by properly adjusting the source and bias generators, reducing moisture, improving by-product evacuation, minimizing sputtering, and using pulsed or multi-step process recipes. The goal is to reduce particles and micromasking while maintaining etch rate, profile control, and process stability.