Physics-of-Failure Methods and Prognostic and Health Management of Electronic Components

Abstract

This tutorial will discuss the role of reliability physics methods and artificial intelligence algorithms, in developing reliable electronic systems for the era of “more than Moore” and heterogeneous integration. Electronic systems are recognized to be highly complex multi-physics and multi-scale systems, extending from mm length scale well down into the nanometer length scale. These systems have to perform reliably under complex combinations of life cycle environmental stresses and operational stresses. Assuring reliable operation and high availability requires systematic co-design that combines electrical, mechanical, thermal, and chemical analyses to design for performance, design for manufacturability, design for testability design for reliability, design for supportability/availability, and design for affordability. Systematic approaches to achieve these co-design goals will have to use judicious combinations of reliability physics and artificial intelligence (based on data-analytic machine learning algorithms). This tutorial will present the underlying principles and a few simple illustrative examples.