As the technology feature size is reduced, the thermal management of high-performance very large scale integrations (VLSIs) becomes an important design issue. The self-heating effect and nonuniform power distribution in VLSIs lead to performance and long-term reliability degradation. In this paper, we analyze the self-heating effect in high-performance sub-0.18-µm bulk and silicon-on-insulator (SOI) CMOS circuits using fast transient quasi-dc thermal simulations. The impact of the self-heating effect and technology scaling on the metallization lifetime and the gate oxide time-to-breakdown (TBD) reduction are also investigated. Based on simulation results, an optimized clock-driver design is proposed. The proposed layout reduces the hot-spot temperature by 15 • C and by 7 • C in 0.09-µm SOI and bulk CMOS technologies, respectively.

No abstract

Burn-in is a quality improvement procedure challenged by the high leakage currents that are rapidly increasing with IC technology scaling. These currents are expected to increase even more under the new burn-in environments leading to higher junction temperatures, possible thermal runaway, and yield loss during burn-in. The authors estimate the increase in junction temperature with technology scaling. Their research shows that under normal operating conditions, the junction temperature is increasing 1.45 /generation. The increase in junction temperature under the burn-in condition was found to be exponential. The range of optimal burn-in voltage and temperature is reduced significantly with technology scaling.

This article describes how CMOS IC technology scaling impacts semiconductor burn-in and burn-in procedures.Burn-in is a quality improvement procedure challenged by the high leakage currents that are rapidly increasing with IC technology scaling. These currents are expected to increase even more under the new burn-in environments leading to higher junction temperatures, possible thermal runaway, and yield loss of good parts during burn-in. The paper discusses the effect of junction temperature on device reliability, aging, and burn-in procedure optimization. The effect of device thermal runaway and the requirements it forces on commercial burn-in ovens, device package, and device cooling are also described.