Reliability Prediction of Radiation Induced Failures in Semiconductor Packages

Abstract

The growth and usage of semiconductor components has risen greatly due to evolution of applications in industries such as aerospace, defense, and automotive, in both the commercial and government sectors. Among the different types of components that make their way into such applications, digital logic and memory-based devices are of prime importance due to the ever-demanding need to store and compute vast amounts of data. This brings to the forefront the reliability of such components, especially in demanding applications in aerospace and defense where components are exposed to extremely harsh environments where high performance is critical. This is where we find that reliability is largely dependent on the susceptibility of semiconductor components to radiation induced effects - total ionizing dose (TID) and single event effects (SEE). Optimization in design, process, and packaging could make microelectronics more reliable to the radiation effects. Packaging material and thickness affect the incident flux of radiation on microelectronics and can help mitigate both dose and SEE effects in electronics. Additionally, packaging materials should not generate a high flux of secondary particles when exposed to environmental radiation (an example is alpha emissions from impurities in package mold compound or underfill). In this paper, the behavior of semiconductors at the component level due to radiation effects is evaluated. A reduced order model approach is presented where the analysis is performed at the functional block level then aggregated into component. The behavioral model can be used to predict the behavior across different operating, design, and process conditions.