Morris Chang attended his first IEDM in 1956---the same year I was born: another way of putting it is that he's been in the chipmaking biz practically since its inception and before mine.
Fifty-one years later, the TSMC chairman returned to offer his thoughts in an old-school, Powerpoint-less luncheon speech called "A Life In Semiconductors." As one of the last still-active participants from the industry's early days, Chang offered many colorful, entertaining and insightful anecdotes and observations during his talk to the soldout crowd.
The central theme of his presentation focused on what he identified as the six major disruptive innovations, breakthroughs without which "there would be no IEDM," he noted. The half-dozen landmarks he cited were not exactly controversial, for the most part. Starting off the list were the invention of the solid-state transistor at Bell Labs in 1947 (which celebrated its official birthday yesterday, Dec. 16, btw) and the codiscovery of the integrated circuit by Jack Kilby of Texas Instruments and Robert Noyce, then at Fairchild, in 1958-1959.
Chang, who worked with Kilby at TI, reminisced about how Jack, coffee cup in hand, would wander into fellow newbie non-Texan Morris's office and tell him about his work on that first kludgey IC. "The fact that it happened under my nose didn't clue me in to its significance," he recalled. "'This will be highly impractical,' I thought."
The third innovation was Gordon Moore's mid-1960s concept that became the industry law. "The genius of Moore's Law was really its boldness," Chang opined, noting that at the time of Moore's pronouncement, the only practical devices were bipolar-junction chips. "The MOS transistor was still primarily a lab curiosity." He also characterized Moore's Law "as the taskmaster for scaling over the past 50 years." As if on cue, the next disruptor on Chang's list was more like a series of breakthroughs: the demos and proliferation of the MOSFET, silicon gate, and CMOS design/process, respectively.
Number five was the invention of the microprocessor in the 1970s, what he called "the last truly major disruptive innovation...because it elevated semiconductor devices from the component level to the subsystem level." Some of Chang's colleagues who he consulted while preparing his IEDM remarks took issue with this choice, saying that maybe the invention of memory-chip technology deserved a place on the list, but he disagreed. The septegenarian also noted that the invention of the microprocessor was the "first example where the first movers' advantage became insurmountable," pointing to Intel's still-dominant industry position.
Chang's final disruptive innovation also marked the first commercial seepage into his speech, as he chose the development of the dedicated, pure-play foundry business model---pioneered by TSMC and its long-time chairman. But the audience forgave Chang this minor indulgence, in part out of respect for the venerable industry icon and also because he had a point. The emergence, maturation, and tremendous growth of the foundries and their symbiotic fabless partners has profoundly changed the semiconductor design and manufacturing ecosystem.
As he wrapped up his speech, Chang continued this slightly self-serving though still-prescient theme. "Any foundry that fails to provide differentiatible technology service to its customers does so at its own peril. And any customer that fails to take advantage of a differentiatible technology that his foundry vendor can provide to him does so at his peril....In the future, the foundry and customer will work hand hand as partners in collaboration."
But an article in the current December 2007 issue of IEEE Spectrum, which I picked up at IEDM, reminded me of one very important innovation left off Chang's list, its progenitor not even mentioned in his speech: the invention of the planar transistor and subsequent planar manufacturing process by Jean Hoerni (who passed away in 1997), one of the original Fairchild Eight (or the Traitorous Eight, if you prefer William Shockley's angry label for the group who founded Fairchild Semi in 1957).
Michael Riordan's "The Silicon Dioxide Solution" offers a fascinating overview of the irascible Hoerni and his collaborators' (such as Moore, Robert Noyce, and Jay Last) essential contribution to the advancement of semiconductor technology. "Fifty years ago, sitting alone in his office, he elaborated a radically new kind of transistor: a more compact, flatter device whose sensitive parts were protected beneath a thin layer of silicon dioxide," writes Riordan.
"Hoerni's brilliant idea, more than any other single factor, allowed the fledgling firm to begin printing transistors on silicon. Planar transistors would prove to be much more reliable and perform far better than other designs [such as the mesa--ed.], in effect rendering the competition's offerings obsolete....Hoerni's elegant idea helped to establish Silicon Valley as the microelectronics epicenter of the world."
To read Riordan's piece in IEEE Spectrum, click here.
|
Blogs 
