CompoundSemi Online - The Original Compound Semiconductor Industry Newspaper

July 4, 2007...Famed semiconductor industry futurist, MIT Professor Eugene Fitzgerald, kicked off our annual CS Vision Executive Business Forum in Austin, Texas USA, last month in grand style. If he had actually placed a crystal ball on the podium in front of him, nobody would have been surprised. But his personal charisma makes up for any lack of actual mystical abilities. His talk, titled "Innovation in IC’s: New Materials, New IC’s?" was an apt keynote for this unique gathering of CEOs, their lieutenants, consultants, investors, and market prognosticators.

Gene Fitzgerald is the Merton C. Flemings-SMA Chair Professor of Materials Engineering at MIT. He's kind of a rock star of the semiconductor industry, on tour seemingly constantly, brightening every arena he enters. He's also a really nice guy and we felt privileged to have him wedge CS Vision into his busy agenda. I'm not implying that I understood everything he said because much of it was over this lowly journalist's head. (I'm not a technologist, but I absolutely recognize a rock star when I see one). But after 35 years hanging around and down inside advanced semiconductor wells of expertise, correct me if I'm wrong, but I think I got the gist of his message. That being... that there's definitely a place for compound semi materials in next gen ICs, provided that they recognize their place in the established semiconductor industry pecking order.

Gene started by reminding us that the building of the overall semiconductor industry (which, as we know, is comprised of many industries) can be categorized in three phases: Phase I: Materials Selection and Vertical Integration; Phase II: Age of Scaling; and Phase III: Age of New Materials. And the latter, for Gene, is focused on what he calls sustained silicon and CS-Si CMOS. The catch phrase that caught my ears and eyes was: Creating a common platform at the lowest level (materials) produces innovation, strategic advantage, economic growth at all the before mentioned levels.

The ICs and systems of the future are definitely bounded by reality, he pointed out, citing that good old Moore’s Law (i.e. transistor density) is actually not the main overarching paradigm and that calculations/$ at the system level is a long term metric. Gene's generalization is that economical miniaturization is still the goal, and that Moore’s Law paradigm has created integrated systems based on MOSFETs that calculate at low V,I very well... but, the bottleneck is now at the digital/communication interface, which is actually outside the silicon IC world. So he poses the following to the shakers and movers of the CS industry, i.e. CS Vision attendees: "We will use silicon manufacturing because it is the lowest-cost way to make any complex patterned object." That one's no surprise. Any sensible CS company bases its business model on the premise that: If you can do your design in silicon, by all means do. It's the cheapest most sensible route to go. But if you can't... you have no choice but to look to the CS world for new materials. We all should know that mantra by now. Based on that, Gene's point is that wireless and optical interconnects/functionality, embodied within their optimal materials (CS), need to be integrated on silicon. OK. We can work with the other team. Why not? Once they recognize the compounds exist and bring something to the party, great! His questions then become: How do we do this? and... Who will do this?

Gene's answers lies in what he calls the SOLES integration scheme. SOLES stands for Silicon on Lattice-Engineered Substrate. Ah... sounds like SiGe to me. But it's more sophisticated than that. I'll not outline it here because it's been detailed in the extended trade press news and at other conferences, but suffice to say, the key appears to be that it involves coplanar integration using standard semiconductor processing, a template layer matched to GaAs buried to allow standard SOI wafer processing, and after CMOS processing, etch device Si and SiO2 is used to access a GaAs template, and then you integrate the III-V and Si devices with standard Si back-end processing. Piece of cake or "POC" (my term). It's evidently already being employed to make monolithic visible LED arrays on Si using SOLES and CMOS-compatible processing, which of course he detailed, which is when I started to see some clouds in his crystal ball, but then again, as I said, I'm not a technologist. I'm just a technology journalist. (There's a big difference.)

I did understand when he pointed out that a disaggregated industry removes the ability to innovate across the supply chain because I can spot a cool new phrase when it's throw out and "disaggregated industry" pretty well nails what we have at this point and time within the compounds. And he also noted that a huge, long term investment is required if this is to be implemented to the level of commercial profitability (like we don't always hear that) and that vertical knowledge (yet another way cool term) is required for novel new products. He also noted that commercial sustainability in small market size is possible, and that virtual vertical (man, can this guy come up with new totally cool terms) brings with it high security compared to a mature horizontal, high strategic advantage. Then he again challenged the audience with: Who Will Do this?

Given this was a heavily CEO audience, he set the tone for the rest of the two days by asking if The Who should be silicon IC companies? Heck, they have a terrible reputation for starting new markets that are too small for their now giant appetites. The decade timescales needed for implementation are beyond their patience. So how about the CS industry's IC companies? Maybe... but they might be wedded to their own infrastructure (hey... we can change... heck, we might even learn to embrace the word standards!) Gene posed that maybe here's a place for new entrants. If you're a newbie company, what Gene is proposing poses what looks to the newbie as a large market. What it will take, he challenged, is an openness to establish a relationship between the CS and the silicon industry. What a concept!

Maybe that's why we sometimes see people from the likes of Intel and AMD sneaking into our companies or executive business forums for a quiet look see. Remind me to (quietly) invite some of the leading silicon IC company strategic planners to CS Vision '08, which will be held in late January or early February (the timeframe it and its predecessor event, CS Outlook, use to be held) in some nice warm spot where you don't have to worry about flying into a snowstorm.

Avanex Expands Transmission Product Portfolio with Asset Acquisition
CompoundSemi News Staff

SiGe Semiconductor Taps Former Motorola Semiconductor President for Chairman Position
CompoundSemi News Staff

Skyworks Front-End Modules in iPhone

Aixtron Announces New MOCVD Tool Order from Arima for LED Production
CompoundSemi News Staff

Avago Completes Acquisition of Infineon’s Polymer Optical Fiber Business
CompoundSemi News Staff

Philips Sells Stake in GaAs Developer, Ommic S.A.S.
CompoundSemi News Staff

MRV Communications Completes Acquisition of Fiberxon
CompoundSemi News Staff

Cree Makes XLamp with Minimum of 100 lumens at 350mA Available in Volume
SSLDesign News Staff

Photovoltaic Startup, Stion Raises $15 Million in Series B Financing
CompoundSemi News Staff

Anadigics PA Shipping for Samsung Smartphone

RFMD Advances in Rankings for 2006
CompoundSemi News Staff

Gene Fitzgerald's Crystal Ball Virtually Glowed at CS Vision
Jo Ann McDonald, founding editor

Avanex Expands Transmission Product Portfolio with Asset AcquisitionCompoundSemi News Staff

SiGe Semiconductor Taps Former Motorola Semiconductor President for Chairman PositionCompoundSemi News Staff

Skyworks Front-End Modules in iPhone

Aixtron Announces New MOCVD Tool Order from Arima for LED ProductionCompoundSemi News Staff

Avago Completes Acquisition of Infineon’s Polymer Optical Fiber BusinessCompoundSemi News Staff

Philips Sells Stake in GaAs Developer, Ommic S.A.S.CompoundSemi News Staff

MRV Communications Completes Acquisition of FiberxonCompoundSemi News Staff

Cree Makes XLamp with Minimum of 100 lumens at 350mA Available in VolumeSSLDesign News Staff

Photovoltaic Startup, Stion Raises $15 Million in Series B FinancingCompoundSemi News Staff