Early product learning (EPL) and failure analysis flow and strategy

• Main Author: Veeraraghavan Anurathi
• Other Authors: Zhou Xing
• Format: Theses and Dissertations
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/73113
• Institution: Nanyang Technological University

The Early Product Learning (EPL) project was started to improve yield and efficiency of ICs and minimize customer return spills. There are four main phases of every IC failure analysis project. The first phase is non-destructive testing, where tools like acoustic microscopy and x-ray imaging are applied to learn as much about the sample without permanently altering it in any way. Next comes fault verification, where the analyst attempts to replicate the failing conditions reported by the customer. Without confirming the failure, the project cannot continue, since there is no evidence to show that a defect even exists. With the fault confirmed, the next phase is fault isolation, where the analyst identifies a site for an in-depth destructive physical analysis and examination. Finally, destructive analysis and documentation serve as the culmination of all the data collection from which an analyst will perform deprocessing, a cross-section, or other destructive technique to reveal the defect in its entirety and identify its most likely cause. Various advanced semiconductor failure analysis techniques were used, including Scanning Electron Microscopy (SEM), Energy-Dispersive X-ray spectroscopy (EDX), Curve Tracer, Scanning Acoustic Topography (SAT) and X-ray inspection among others to isolate the root cause of electrical failure in multiple types of devices based on varied process technologies. Failure Analysis flow and strategy: The failure analysis team had managed to create a FA flow and strategy iCommunity page where people could share their analysis examples, summarize helpful hints related to the analysis, point out common pitfalls and collect their inputs from. This would ultimately result in the transparency of the analysis flow and improve their flexibility. Its helps connect analysts from all over the world which leads to an effective knowledge sharing association.