"How Many Transistors Can Fit on the Head of a Pin (And All Work)?" may recall a similarly-posed theological question associated with St. Thomas Aquinas, but the intent of Don Monroe from Lucent Technologies' Bell Labs in Murray Hill, NJ, was anything but theological when he addressed students, parents, and teachers at the Princeton Plasma Physics Laboratory's "Science on Saturday" series on 21 February 1998. In the year following the 50th anniversary of the transistor, heralded as a revolution in electronics, Monroe observed that the real revolution came not from the single transistor but our present technologies of mounting millions of them onto a single chip. Although the replacement of each tube in the earliest computers meant greater reliability, less power consumption, and the need for less space, he showed that hand-soldering a Pentium chip would be very costly: 5 million transistors x 4 leads/transistor x 10 s/lead = 50,000 hours. Although a chip containing a million transistors typically entails 350 steps of depositing substrates and photoresists, exposing photoresists with light, then etching patterns and dissolving photoresist, the automation of these processes in batches has achieved considerable economies and capabilities in the developments of modern electronics. Mounting transistors on a chip is as revolutionary as the transistor
Monroe likened a functioning 2 cm x 2 cm chip to the houses in New Jersey, connected by roads, water and gas pipes, and electric lines, with no malfunctions. He also looked at Moore's Law (named after Intel cofounder Gordon Moore, which states that the number of transistors on computer chips doubles every 18 months) as a self-fulfilling prophecy. Not only are there more transistors per chip, he noted, but the transistors are also smaller. At the same time, speed has been increased, while power and cost per bit have been reduced. Moore's Law is an exponential relationship, and it has held up for 35 years. But exponentials always end, Monroe counseled. When the limit to Moore's Law will be reached is a matter of prediction (a field mined with pitfalls), but Monroe suggested the following possible causes: 1) physical limitation, 2) cost, 3) the speed of light, and 4) satiation (already 10,000 transistors are made per person per day -- do we need any more?).
Future reduction of size mandated by Moore's Law requires shorter wavelengths, which are achievable with electron beams instead of light. Other possibilities foreseen by Monroe include replacing our present reliance on silicon with other materials: 1) plastic transistors and circuits, 2) single electron transistors, 3) organic transistors, 4) molecular electronics, 5) DNA computing, and 6) quantum computing.
Home Spring 98 Full Screen
The TEACHERS CLEARINGHOUSE FOR SCIENCE AND SOCIETY EDUCATION