Typically, fabrication of good quality semiconductor wafers without defects is challenging for any semiconductor manufacturer. It consists of sequential process steps performing physical and chemical operations on wafers. Usually, wafers are aggregated into so-called lots of size 25 or 50, which always pass through some operations in the production chain. Therefore, the entire manufacturing process can involve lengthy repetitive steps, and takes place in a sterile clean fabrication room designed to prevent even the tiniest speck of dust from falling on the pristine wafers. The fragile wafers may get scratched or get particulates from the clean room that could cause the micro circuits to malfunction when tested after the manufacturing process. Often, these flaws are microscopic and completely invisible to the naked eye which leads to poor production quality. If these flaws are not addressed at contamination phase, then there might be decrease in test yield that results in the wafer manufacturing costs. After a detailed analysis and research, Ai Labs developed and implemented a Proof of Concept (POC) using its proprietary engine Minsky to highlight the defects in wafer fabrication process. The objective of this case study is to automate inspections process and identify the defects by categorizing them as (good, Bad) in the process using Deep learning classification model depending on the response value set (Threshold value) by the semiconductor company.
After comprehensively evaluating the client ‘s challenges/data, we used Minsky to accurately model the historical data which comprised of image data that were labelled & classified into (good, bad ) based on 8 categories such as Centre, Donut, Edge Loc ,Edge Ring, Loc, Near full, Random, Scratch.
Minsky modelled and analysed these features using deep learning algorithms that were modelled on labelled image Historical labelled image data was used against the real time image to predict and identify the defects in wafer fabrication process by categorising them as (good, bad ).