LG Innotek, led by CEO Moon Hyuksoo, has announced the development and implementation of the industry’s first AI-powered inspection system for incoming raw materials. This system is designed to identify defects at the point of receipt, preventing substandard materials from entering the production process.
The AI-based technology combines material data with advanced image processing to ensure high-quality control in the production of semiconductor substrates, particularly in the Radio Frequency System-in-Package (RF-SiP) process. By utilizing this solution, LG Innotek aims to enhance production efficiency and maintain the highest standards in raw material selection.
LG Innotek has recently introduced its AI-based inspection system to the production process of Flip Chip Ball Grid Array (FC-BGA) semiconductor substrates. Previously, raw materials were subjected only to visual inspections, which were insufficient as semiconductor substrate technology advanced. Despite efforts to eliminate in-process defects, failures in reliability evaluations persisted, pointing to the quality of incoming materials as a crucial factor.
The primary raw materials for semiconductor substrates, including Prepreg (PPG), Ajinomoto Build-up Film (ABF), and Copper-Clad Laminate (CCL), consist of intricate blends of glass fibers and inorganic compounds. In the past, imperfections like air pockets or foreign particles generated during the mixing process had minimal impact on product functionality. However, as the requirements for substrates have become more precise, these imperfections have started to lead to defects, rendering traditional visual inspections inadequate.
LG Innotek developed an AI-powered system trained on tens of thousands of data points to address this challenge. This system can detect defective materials with over 90% accuracy in under a minute. Analyzing and visualizing quality deviations in raw materials allows LG Innotek to prevent faulty materials from entering production and ensures uniform quality in each lot of raw materials.
The AI system’s ability to visualize and quantify material configurations enables LG Innotek to optimize the design of raw materials. This innovation has reduced costs associated with identifying and resolving defect causes.
LG Innotek intends to further improve its AI inspection system by collaborating with customers and suppliers in the substrate sector to share data related to raw materials. This digital partnership aims to enhance the system’s detection capabilities, ensuring higher accuracy in identifying material defects. The company also plans to extend the application of this technology to optical solutions, including camera modules, where detecting material defects through image processing is essential.
According to LG Innotek’s CTO, S. David Roh, the AI-based inspection system will complete the company’s unique AI ecosystem, providing exceptional value to customers by identifying product defect causes early in production.
