What’s The Rush?
By Vipin Tiwari
Most of the device manufacturers are moving to high k/metal gate for the 32nm and 28nm process nodes.
On paper, this seems like a clear-cut decision. The technology reduces static leakage or boosts performance. In some cases, it can accomplish both. In others, it can provide significant gains in one or the other. It’s an extra knob to adjust for chipmakers, depending upon what they’re making and which market they’re making it for.
But from the manufacturing side, it’s going to be a race against time to be ready with process qualifications for next-gen products using high k/metal gate technology. And if all the pieces aren’t in place, that could cause needless delays.
It’s clear that the current Poly-SiON technology with strained silicon doesn’t make sense for 22nm. At that process node, device manufacturers will have no choice but to move to high k/metal gate or some other combination of technology that could even include 3D structures such as finFETS.
But what do companies do at 32nm and 28nm? Is the extra cost of high k/metal gate really warranted and necessary? And what won’t be ready in time?
These are important questions, and so far there aren’t many answers.
–Vipin Tiwari is senior manager of strategic product planning at Virage Logic
