Artificial Intelligence Model for Predicting Power Consumption in Semiconductor Coating Process †

Abstract

We developed an artificial intelligence (AI) model to optimize the time efficiency, yield, and energy efficiency of the semiconductor coating process. A random forest-based model was developed for rapid modeling and analysis of the semiconductor coating process, thus allowing designers and operation managers to conduct an efficient and effective process. The developed AI model offers an objective and accurate basis for decision-making, thereby ensuring that each unit is operated energy-efficiently, stably, and reliably in the minimized operation time. The developed model assists Taiwan’s semiconductor industry in transitioning from engineer experience to data-driven approaches, thus accelerating the technological optimization of semiconductor factories and adding value to customers. This model considerably reduces the material, energy, resource, time, labor, and costs of thin film deposition. The model allows the semiconductor industry of Taiwan to consolidate its competitive advantage by achieving net-zero carbon emissions and sustainability.