SI
SI
discoversearch

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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?


To: Sarmad Y. Hermiz who wrote (228905)3/25/2007 7:20:40 PM
From: neolibRead Replies (1) | Respond to of 275872
 
Yes I did. You said 100 atoms. This link, also quoting an IBM person mentions 14 atoms. And I read that Intel's 45 nm has some structures that are 2 atoms thick.

The 100 atoms are dopant ions in the channel, the 14 atoms you quote are oxide thickness. As noted, 65nm gate oxide is around 5-6nm (IBM is a bit thinner than Intel IIRC).

Most of the talk in recent years has centered on gate oxide scaling problems, so I found the paper interesting because it looks like channel implant is claimed to be a significant problem, much more so if you are counting atoms. The total atom count variation in the gate oxide layer is much worse than 100 atoms I would guess. So from an atomic standpoint, it looks like implant is a more critical process.

Another source of intrinsic device variability arises from atomic-scale oxide thickness variations. Physical gate oxide thickness is currently down to 1 nm, equivalent to approximately five inter-atomic spacings.

Yes, as I had mentioned. However, the total atomic count in the gate oxide is on the order of 200 x 200 x 5 = 200,000 atoms (back of the envelope) with +/-1 layer (per the article) so perhaps +/- 40K atoms. So while the layer count is critical in gate oxide, if you are looking for process variation that is approaching single atom issues, it looks like channel implant is the critical one by a wide margin.