Junction Temperature During Burn-in: How Variable is It and How Can We Control It?

Abstract

Electronic devices are some of the most reliable consumer products. It is taken for granted that they will work but a consideration of the number of millions of interconnects and the complexity of the manufacturing processing makes this remarkable. The semiconductor industry has embraced the use of Burn-in to force the failure of marginal devices before they reach final assembly into end products. The burn-in process uses a combination of temperature and voltage to stress the device. For economies of scale this is usually accomplished by placing devices in a burn-in oven. The time to ensure the early failure of marginal devices is dependent on the junction temperature and this is a specified by the reliability engineer. However, do we really know the junction temperature? Is it estimated or controlled or how is it measured? This paper will review some of the aspects of burn-in and the way the junction temperature can be measured or inferred. The impact of on junction temperature for devices which have a wide power range will be evaluated and discussed. Methods to understand and control the junction temperature, during burn-in will be reviewed and the impact on burn-in time discussed. The ability to accurately control the process is dependent on an understanding of junction temperature. Accurately controlling the junction temperature, not the oven ambient temperature offers the opportunity to reducing burn-in time by raising the temperature. The potential for higher burn-in temperatures and the impact on burn-in time will be discussed.