Who will win the next Nobel Prize in physics?
12/01/2000
 |
For the most part, the Nobel Prizes in physics have been awarded to contributors to our understanding of fundamental laws of nature (sometimes theorists, sometimes experimentalists) rather than inventors. The Prize will go to an Einstein or a Gell-Mann (whose eightfold way showed how the profusion of sub-atomic particles could be logically organized in a sort of Periodic Table) over an Edison or Rutan (the legendary exotic aircraft designer). The prize for the laser went to those, such as Townes, who worked out the theory rather than to a protégé who actually built the first working device. In electronics, De Forest did not get one for coming up with the three-terminal vacuum tube amplifier, but Shockley, Bardeen and Brattain won the Prize for adapting the concept to solid-state physics and coming up with the transistor.
Putting more than one device onto the same semiconductor chip seemed almost like an obvious step, yet the Nobel Committee chose to award a share of this year's Prize to Jack Kilby, retired from Texas Instruments, for his foresight in taking it (if Robert Noyce, a co-founder of Intel, were still alive, he would have deserved to share this honor with Kilby). Their first integrated circuits were crude forerunners of today's powerful devices with many millions of devices on them, but their innovation has dramatically transformed our world.
It may be that the progress in discovering basic laws of physics has become somewhat stalled until we build higher energy accelerators. While there is fascinating theoretical work going on in areas such as string theory, playing games with supersymmetry, and translating the laws of physics into complex variables, without experimental validation it is hard to know whether the models are correct or valuable in extending our knowledge. Sometime soon, someone may find the giant Higgs boson, create a grand unified theory that incorporates gravity, or find the source of a mysterious force that keeps the universe from collapsing on itself. Some remarkable advance like this might lead to an early Nobel Prize.
But it is quite possible that nothing of this magnitude will occur between now and the next round of prizes. If so, it might make sense to look at how the laws of physics have been bent to our advantage in recent times in the Information Age, allowing devices to be packed onto ICs with features even smaller than the wavelength of the light used to create them. A few years ago, responsible physicists would have declared sub-wavelength patterning to be impossible. It would be like trying to dig neat, three-inch square holes with a garden shovel. No matter how carefully you manage the shovel, it just cannot make neat holes that small.
But today, chipmakers use 248nm light to pattern 180nm devices (with some features even smaller than that) for mass production. They are moving toward 150nm and 130nm devices using the same light, and some fabs may extend it to 100nm or even smaller. Incredibly, working CMOS transistors have been made in the lab with 27nm features using a commercial 248nm stepper. These advances will help us make affordable gigabit memories and multi-GHz processors with several functions, if not a whole system, on the same chip. Since optical lithography involves a complex, delicately balanced interaction of hardware, chemicals, light sources, and maskmaking infrastructure, extending its life — and the usefulness of multimillion dollar steppers — for more generations of chips has tremendous economic benefits. The secret to these achievements lies in reticle enhancement techniques (RETs): optical proximity correction, phase shifting masks, and off-axis illumination (OAI).
This remarkable, even astounding, transition to sub-wavelength manufacturing was accomplished through the contributions of a number of innovative physics practitioners. Optical proximity correction (OPC) was first used in the 1970s, in the waning days of Rubylith maskmaking (remember those glass-topped drafting tables and the X-acto knives?). Bahaa Saleh of Boston University is credited with first suggesting the use of computers to predistort mask patterns in 1981, and, in 1982, Andy Neureuther and his students at UCal Berkeley reportedly put the idea into practice successfully. The phase shifting mask (PSM) is credited to Marc David Levenson, then of IBM (now a consultant and editor of Microlithography World) and Masato Shibuya of Nikon, also in the early 1980s. The first suggestion for adopting Lord Rayleigh's concepts of OAI for microlithography is credited to Chris Mack of Finle Technologies (now part of KLA-Tencor) in 1989.
There may be other creditworthy advances in the evolution of chipmaking, but it is clear that this group of physicists has contributed to a major breakthrough in optical lithography. Their work is allowing the industry even to push ahead of the progress projected by Moore's Law. In our view, their contributions should be worthy of the highest honor in the world of physics.
Robert Haavind
Editor in Chief