Overview of the Application of Artificial Intelligence in Semiconductor Manufacturing Process Optimization and Defect Detection

Abstract

Semiconductor manufacturing involves highly complex and tightly coupled process flows, stringent quality requirements, and the generation of vast amounts of heterogeneous data throughout the entire manufacturing lifecycle. As device dimensions shrink and process integration becomes increasingly complex, traditional statistical and physics-based methods face growing limitations in handling nonlinear interactions, process variability, and real-time decision-making needs. Therefore, artificial intelligence (AI) has emerged as an effective data-driven approach to enhance manufacturing performance by enabling advanced process modeling, optimization, and fault analysis. This paper reviews the latest advancements in the application of artificial intelligence in semiconductor manufacturing, with a particular focus on process optimization and defect detection. It explores the applications of AI in process parameter optimization, yield improvement, and virtual metrology, highlighting its role in reducing variability and improving manufacturing efficiency.