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) -- Ignore unavailable to you. Want to Upgrade?


To: IanBruce who wrote (51047)3/24/1998 12:56:00 AM
From: Paul Engel  Read Replies (2) | Respond to of 186894
 
Ian - Re: Copper

Why don't you tell us about IBM's 0.20 micron all copper PowerPC that only runs at 480 MHz and Intel's 0.25 micron all aluminum Deschutes that runs at 450 MHz.

Even IBM admits that the copper resulted in lower capacitance - not lower resistance - for its interconnects.

You do have that information don't you?

Paul



To: IanBruce who wrote (51047)3/24/1998 3:14:00 PM
From: Time Traveler  Read Replies (1) | Respond to of 186894
 
Ian,

From your length article regarding copper interconnects, I do have a few comments:

...gigahertz parts...CMOS process using copper metallization, integrating six metal levels, a 0.2-micron drawn gate and a 0.12-micron effective channel length.

What is the big deal about this copper technology here? From 0.13um which is what IBM is describing, digital ICs can easily achieve 1 GHz, even with aluminum metallization.

ICs with inductors using copper technology for GPS system

These guys cannot be serious here, integrating inductors into an IC! For years, TRW has been using copper plating for their MMIC devices, highly integrated RF circuitry. That is probably a better solution for GPS.

"We are aggressively pursuing copper technology, but while copper solves a near-term problem, we think reducing the capacitance is even more of a critical issue because unless it is solved, chip performance, power and operating voltage ultimately will be limited by the interconnect," --- Robert Havemann of TI.

I agree with that comment. Notice he is hinting copper technology is not as beneficial as IBM has pointed out. He then changed the whole subject towards the dielectric material.

"We're aggressively investigating copper and dielectric materials, but we're still doing scaling of aluminum lines. We're trying to improve things by going to tall or high-aspect-ratio aluminum lines. That way, you gain performance advantages and get to keep your existing processes and production equipment." --- Mark Bohr of Intel.

If it ain't broke, why fix it! Remember, the name of the game is to make money.

"It's not actually the copper wires that are the main issue. It's the capacitance and that gets to the dielectric," --- Mark Pinto of Lucent.

Again, Mr. Pinto is emphasizing Mr. Havemann's comment.

"On most products, the most important capacitance is between metal lines on the same layers. If you use copper as the trace, it has about 30 percent less resistivity than aluminum, so you can make it 30 percent thinner. If the spacing is kept the same, the capacitance is reduced." --- Suresh Venkatesan of Motorola.

By making a trace thinner, you are risking design rule violation, besides most of the capacitance encountered by a gate is not in the metal trace but the semiconductor junction.

Over all, is copper really that good? Given that ICs today are limited by their power handling capacity, an increase in trace resistance does not really matter. Remember, it is not resistance of the trace but the capacitance that is governing how much power is dissipated in an IC.

John
East Greenwich, Rhode Island



To: IanBruce who wrote (51047)3/24/1998 10:41:00 PM
From: Time Traveler  Read Replies (1) | Respond to of 186894
 
Ian,

How come you are not replying to my post regarding this Copper technology?

Well, here is more I would like to add to:

Message 3820784.

The resistance which copper technology reduces is only the trace resistance! You still have the on-resistance of the junction to worry about. Guess which is overwhelmingly dominant? Right, the junction, not the metal trace. So that 30% reduction in trace resistance is really negligible.

When IBM made this information public, I believe it is not really hyping it. This company traditionally takes pride in its technology and make it very obvious known to the world. Remember a few years ago, IBM was bragging about how it can deposit single atom, one at a time.

It was the analysts who barely understood Ohm's Law (V = I R) really hyped it beyond all possible means.

John.