Breakthroughs in technology are often born of persistence, ingenuity and sometimes serendipity. The invention of the semiconductor silicon germanium involved all of these, plus the most humbling of uncontrollable variables ... accident.
The emergence of silicon germanium—often abbreviated by its chemical symbols, SiGe, and pronounced “sig-ē”—is especially noteworthy, having resulted from not one, but two missteps. In correcting these errors, IBM’s Dr. Bernard Meyerson laid the groundwork for an explosive advancement in wireless products, such as cell phones, global positioning system (GPS) devices, wireless broadband Internet (WiFi) and mobile TV.
In 1979, Meyerson was a doctoral student at the City College of New York, pursuing his interest in semiconductor technology. In the lab one day, he inadvertently dropped a one-inch piece of silicon he’d just cleaned in hydrofluoric acid. Without delay, he retrieved the sample from the floor and, rinsing it under a faucet, noticed that the silicon wafer was water-repellant. Meyerson knew well that exposure to air coats silicon’s surface with a thin layer of water-retaining oxide—or so he had been led to believe by 30 years of published literature. Most likely, something other than oxide was protecting the sample, and it wasn’t dirt from the lab floor. At that moment Meyerson could have set out along an important quest for answers, but he had research to complete, so he chose to mentally file the incident away.
In 1980, Meyerson accepted a research position at IBM. At that time, the company’s semiconductor experts realized they could not continue to shrink microprocessors without running into performance problems, and so began looking into improving chips by employing alloys. Meyerson was focused on combining silicon with germanium. The main obstacle was the need to reach 1000 degrees Celsius to rid silicon of contaminating oxide, thus preparing its surface for the growth of crystalline silicon and silicon germanium layers—or so the scientific community thought. The problem was that added SiGe simply would not withstand the extreme heat.
In time, Meyerson brought his own experience to bear on the problem by returning to the mysterious dropped silicon incident of three years before. He once again cleansed a sample in hydrofluoric acid and then subjected it to precise measurement. Meyerson realized that the hydrofluoric acid had covered the silicon in a thin, protective sheath of hydrogen. At 600°C the unseen layer of hydrogen would effectively blow away, allowing oxide to form. Only then was it necessary to reach 1000°C—to cleanse the silicon of the newly acquired oxide contaminant.
“... We’ve gone from sort of a rather routine working premise of ‘let’s make it smaller; it will then run faster; and here’s the next technology generation’ to ‘I’ve scheduled four inventions between now and the next generation of technology.’ The sobering thought is that someone then asks, ‘Uh, what are they?’ I reply, ‘I’m not really sure; that’s why they’re inventions.’”
"Inventions: A Result of Risk-Taking, Diversity, and Holistic Thinking—An interview with Bernard S. Meyerson,” IEEE Design and Test of ComputersSep./Oct. 2005
“This was originally done for a big mainframe program. But along the way what happened was I discovered that these transistors had absolutely outstanding properties for use in communication; unbelievably advanced ‘comms.’”
Telephone interviewMarch 4, 2011
“This new transistor shows there is a real road map in place with silicon germanium.”
“IBM Creates 350 GHz Transistor,” Internetnews.comNovember 4, 2002
“That was the epiphany,” explains Meyerson. “The layer that everybody thought you had to remove didn’t exist until you actually formed it on your way up to 1000°C. Simply growing materials below 600°C avoided the whole problem. It was the most bizarre finding I've ever had, but it also gave us a 10-year head start on the rest of the world because nobody understood the effect, so away we went.”
“Away” meant growing silicon germanium at 550°C, producing literally flawless, uncontaminated films and tremendously fast transistors. SiGe was far more efficient than silicon alone. Equally important, its costs were a tiny fraction of those for gallium arsenide, a leading alloy of the day used in communication chips. With SiGe, IBM alone had the manufacturing know-how along with patent protection.
Meyerson soon found himself in charge of 100 technologists focused on SiGe research and development for mainframe computers. However, within months, a decision was made to reduce investment in SiGe while focusing on an alternative technology in its mainframes—complementary metal-oxide semiconductor (CMOS) transistors. Meyerson, convinced of SiGe’s potential for new markets, retained a small SiGe team and focused on ways to build demand. And while SiGe’s potential for telecommunications and wireless communication was clear to Meyerson, IBM was not a participant in these newer markets. For these reasons the company did not invest further in the alloy.
In 1992, the initial SiGe team was scaled down to two: Meyerson and electrical engineer, David Harame. Yet, the duo remained confident in SiGe and decided to seek funding outside of IBM. Meyerson became a one-man sales force, forging alliances with several pioneering communications firms. He struck financial arrangements whereby the companies would pay IBM to first develop and then manufacture SiGe chips. The funds came in; IBM raced through development, into production and, virtually overnight, brought new fields of wireless technology to life.
For more than a decade, beginning in the mid 1990s, IBM reigned as SiGe manufacturer to the world, setting virtually every device performance record in silicon technology. The reliability, speed and low cost of SiGe enabled rapid growth in various wired and wireless networks, shrinking the size and power needs of WiFi, cellular phones, GPS systems and many other products. It also brought the prestigious designation of IBM Fellow to Bernard Meyerson and David Harame.
Selected team members who contributed to this Icon of Progress:
- Dr. Bernard Meyerson Vice President, Innovation, and IBM Fellow
- Dr. David Harame Senior Manager of the RF Analog Modeling and Design Kit Department, Director of the RF/Analog and Mixed Signal Process Development, IBM Fellow