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: wanna_bmw who wrote (142651)	9/3/2001 3:19:15 PM
From: Saturn V	Read Replies (1) \| Respond to of 186894

Ref < Bipolar vs MOS > At an individual transistor level, I will not dispute that bipolar may be faster than MOS. However the power dissipation is so much higher, that completely bipolar processor chips have seen their twilight days. BiCMOS, which allows a mostly CMOS chip with bipolar SiGe transistors for a few critical locations, still remains a viable technology .

To: wanna_bmw who wrote (142651)	9/3/2001 10:14:42 PM
From: J_F_Shepard	Respond to of 186894

" Recent advances have brought BJT transistor sizes down close to that of CMOS, and they are a lot faster." 'Bipolar device densities will never be the same as MOS, at the same lithography ground rules, since BiP's require a separate isolation region for each device whereas MOS devices are self-isolating. "The only problem is that yield was abysmal" Bipolar yield will probably always be poorer, at the same device densities, than MOS because the primary yield detractor of a bipolar device, at the silicon level, is emitter -collector shorts. In an MOS device this is analogous to gate shorts but, historically and inherently, gate short defect densities have always been about a factor of 10 lower than E-C shorts. This is because the base region of a bipolar device resides in single cyrstal silicon whereas the the gate insulator of an MOS device is an amorphous material, ie SiO2. Single cyrstal silicon inherently has more defects than amorphous SiO2. If you accept this premise you can see that compared to CMOS, BiCMOS device yield will be limited by the number of bipolar devices designed into it. SiGe does not help bipolars yield-wise since a heterojunction material can not be expected to be more defect free than pure silicon alone. MOS devices have largely caught up to bipolars in speed because the channel lengths of MOS are close to or equal to the dimension of the base width of bipolars. But bear in mind that the bipolar still needs to include an isolation area and hence will occupy more real estate.