Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: J_F_Shepard who wrote (126883)	2/8/2001 11:47:00 AM
From: fingolfen	Respond to of 186894

...any guesses when INTC introduces high-k gates? Clearly the 0.13 micron process doesn't have them... and the 30nm transistor Intel demonstrated didn't use them either (although it used a thinner oxide than 0.13)... The problem is that the minimum thickness for a silicon dioxide film (even if it's nitrided) is on the order of 5A (roughly the thickness of a native oxide)... That's getting proportional to the silicon-oxygen bond length (which should be ~2A give or take). Intel's 0.13 micron process uses 15A gates (though it's unclear if this is an electrical or physical thickness). All of my discussion above is based on a physical thickness. Presuming that the 15A thickness is physical, you may get two more generations out (10A and maybe 5A)... if it's electrical... well, I can't remember which way the calculation runs, is the electrical thickness higher or is the physical thickness higher??? (see: intel.com for details on Intel's 0.13 micron process). Intel's 30nm transistor (overview at: intel.com ... I don't remember where the full paper is stored), talks about features "three atomic layers thick." That's got to be the gate, and that means they used a gate oxide sub-15A thick... Wonder how they managed to stop the gate etch on so thin an oxide layer... I'm guessing that gate probably represents the minimum practical thickness for SiO2... Given the difficulty of integrating a High-K material into a volume manufacturing process, I think Intel will try to delay its introduction as long as possible. So 0.13 clearly doesn't have it... They'll probably try to use SiO2 for 0.10 micron... My guess is they'll have to switch at 0.07 micron, or they won't be able to scale the gate dielectric with the process... Possibly a more important question is when is the semiconductor industry going to have to move away from CMOS to another transistor design like finfet...