It looks like some fab folks are again poking around in the silicon to get more out of power MOSFETs. That's an interesting turn of events, because less than two years ago the industry proclaimed packaging as the last roadblock. But the red flag's already out: Even some of today's best-of-class solutions are already seen as not good enough to secure higher efficiency and longer battery life in tomorrow's portables, and packaging for the moment has taken the power chips about as far as they can go.
Just as there appears to be no way out, it's a return to the silicon for better or worse, and round and round it's destined to go. It rather reminds me of those endless repeating dream sequences in a few episodes of The Twilight Zone, or that Monty Python skit in which some chap in jail facing a firing squad apparently awakens from his nightmare, only to have his mum tell him, "No, son, this is the dream, you're still back in the cell!"
As luck would have it, going "back into the cell" is again important for MOSFETs, with vendors taking another look at how to achieve higher cell densities to help lower on-resistance in particular. While device packaging seems as important as ever, the popular sentiment two years ago decrying the silicon's "diminishing law of returns" has changed. Now, the focus has returned to such activities as cleverly pruning some internal length, width or geometry of the semiconductor device.
The most recent gain comes from ON Semiconductor, which has found a new way to minimize cell pitch, ultimately boosting channel length and equivalent channel density in its trench devices. This technology, says the company, reduces on-resistance an average of 40 percent over competing products, which seems significant, given that there's been so much success to date in reducing it.
That's just one example. Other challenges on the road map include improving the speed of trench devices as well as the figure-of-merit in planar devices in the still-contentious planar vs. trench debate. Both issues seem destined to incorporate new materials technologies.
Ultimately, success will boil down to making the new manufacturing processes cost-effective. Fortunately, it's likely many of the evolving processes will be applicable to the new materials.