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 : Intel Corporation (INTC)
INTC 36.08-3.2%3:59 PM EST

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Len Roselli who wrote (66737)10/15/1998 12:06:00 PM
From: Paul Engel  Read Replies (1) of 186894
 
Len - re: " but IBM seems to be a generation ahead of Intel (0.12
micron), or are they just touting a demo project?"

I am going to assume that these numbers came from ICE - Integrated Circuit Engineering.

First, let me put in a plug for ICE - they have been doing superb analyses of semiconductor processes and devices for many years. Their technical expertise and attention to detail is terrific.

Now, I will also assume that the PowerPC 750 report reflects a production part. ICE generally sells reports ONLY if they are given the go ahead by whoever paid for the analysis - and "demo" chips generally are analyzed but the reports are kept very private.

Now, IBM stated that the PowerPC 750 copper process was a nominal 0.18 micron minimum feature size. Normally, this number applies to the length of the smallest polysilicon gate that can be "normallly" imaged in a wafer stepper/scanner. However, by tweaking exposure controls and etch conditions, a slightly narrower poly feature can be achieved - say 0.15 micron FINAL gate length.

When the source/drain regions are implanted and spacers are added and the various IMPLANT DRIVES are performed, Arsenic (the n -dopant) will diffuse under the gate perhaps 0.01 or 0.015 micron. This happens from both gate edges, so the TOTAL underdiffusion of Arsenic is twice that, or around 0.02 or 0.03 microns.

(I am ignoring p-channel devices which are always slightly larger.)

Subtracting this from the 0.15 micron PRINTED CD, you get what is known as an Leff - L effective - of 0.12 microns. Normally, this has to be determined by electrical measurements but a good cross section and selective "staining" of the implant regions below and adjacent to the gates will reveal the Leff when viewed with a good Scanning Electron Microscope (SEM).

Now - is IBM ahead of Intel?

Yes and no.

Yes, IBM is ahead in getting a 0.18 micron process into production.

No - Intel is AHEAD of IBM because Intel's 0.25 micron process produces FASTER CPUs that IBM's copper-based 0.18 micron process. The PowerPC 750 chip you refer to was introduced at only 400 MHz as I recall whereas Intel is SHIPPING 450 MHz Pentium II's and Xeons built on their "slightly tweaked" 0.25 micron process.

When Intel introduces their 0.18 micron process in mid-1999, Intel's CPUs will be quite a bit faster than their current 450 MHz speeds of today.

Paul
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext