Using AI/ML To Find And Correlate IC Test Data

What causes low yield in wafers? Usually it's due to design or process changes, but sometimes yield issues occur even if there haven't been any changes from one manufacturing lot to the next. Finding the cause requires some sleuthing, and the best approach for pinpointing problems is to mine design, process, and manufacturing data, and to correlate that data by date and time, by which equipment was used for test and inspection, which chamber was used for manufacturing, and where else low yield cropped up. Aftkhar Aslam, CEO of yieldWerx, talks with Semiconductor Engineering about how to find the root cause in dozens of process and inspection steps, why AI/ML is required to make this work, and why multi-die assemblies and chiplets add even more complexity to this process.