The ESD Handbook

15ESD Power Electronics Design

15.1 Reliability Technology Scaling and the Reliability Bathtub Curve

In this chapter, the focus will be on electrostatic discharge (ESD) in power devices and analog components [1–85]. As technology is scaled, the reliability of semiconductor devices is affected. This can be observed from the reliability “bathtub” curve. The reliability bathtub curve has three regimes to predict failure rate on a logarithm–logarithm plot of FITs versus time. The FIT rate is the number of fails in one billion hours. The first region is known as the infant mortality regime, followed by a second time regime, known as the use or useful life regime, followed by the end-of-life (EOL) regime. The infant mortality is a decreasing linear regime on a log FIT vs log time plot. The second useful life regime is time independent and a low flat FIT rate. As one approaches the EOL regime, reliability “wear-out” begins leading to a linear increase in the FIT rate (Figure 15.1).

As technologies are scaled, pre-mature wear-out is occurring with a continued decrease in the length of the useful life regime. As technologies are scaled from 180 nm to below 65 nm, the length of useful life is decreasing, and wear-out will be a larger issue.

This indicates that the fundamental devices within a semiconductor chip are becoming weaker with technology scaling; it will be important to improve the reliability of components by improving electrical overstress (EOS) robust circuits through layout, ...