ABSTRACT
In an ever more connected world, semiconductor devices are at the heart of every technically sophisticated system. Safety and security in operation, on which many times vital personal or business data or our lives depend on, is critical. The market for semiconductors is tremendous, and rogues also to get their share by selling counterfeit products which potentially jeopardize that very safety and security. Trust into semiconductor devices can be created by securing the supply chain or by verifying the electrical characteristics, the physical layout and the manufacturing technology against the design and specifications. The objective of this work is to propose a verification pipeline for semiconductor devices utilizing their technological features computed by the means of an automated device cross-section analysis. The emphasis lies on the confluence of an established industrial analytic process with novel possibilities provided by the advances in data processing and machine learning. This framework, its technical implementations, and exemplary results of our proposed autonomous technology analytics approach are presented in this work. Furthermore, the results are compared against a manual expert’s measurement which underline the high performance of the framework and its effective multi-stage realisation.
