pairs of silicon atoms bonded together in a barbell configuration exist in up or down states whose configurations could be changed as shown in this supercomputer simulation by a thin metallic tip as used in Scanning Tunneling Microscopy. Source: K. Cho, J. Joannopoulos
A single oxygen atom is made visible as a blue sphere in this supercomputer-generated image of electronic charge inside a silicon crystal. The oxygen atom straddles one of the silicon bonds in the bonding lattice represented by a warm-colored honeycomb. Source: IBM Yorktown Heights and MIT
The computing revolution is dramatically transforming virtually every aspect of our society--our work, our play, even our national security. This revolution started with the discovery of the transistor, the result of fundamental research in solid state physics and the earlier development of quantum theory. The next stage, development of complex microchips incorporating many transistors, drew from fundamental work in physics, chemistry, and materials science. Now applications such as smart military weapons, delivery of consumer services such as movies on demand, or means of transferring electronic funds in a secure manner are incorporating new discoveries in mathematics, engineering, and computer science.
One important frontier of the computing revolution is found in today's powerful supercomputers, which have the ability to perform hundreds or thousands of calculations simultaneously (so-called massively parallel computers). Within years, this field is expected to cross an important threshold, when the fastest computers will be capable of a thousand billion floating-point operations per second (teraflops). Petaflop computers (capable of a billion billion operations per second) may follow only a few years later.
These advances in computing technology draw heavily on fundamental science. But science and technology are closely intertwined: the technology is also driving forward the frontiers of science--ushering in new fields of research and extending the limits of inquiry in virtually all fields--which will in turn enable new technology. For example, for the first time scientists can now begin to simulate such complex physical and biological systems as the earth's climate, the atomic structure of novel materials, and the molecular structure of living cells. Applications of this new computationally-driven science will include improved microelectronic devices and rational drug design. Computational studies of silicon-- the semiconductor material on which most modern computing is based--illustrate the trend. Researchers are now beginning to simulate silicon-based materials with supercomputers, allowing them to perform "theoretical experiments" and--using new techniques for visualizing atomic scale structure--to "see" the results. Physicists, for example, can now use supercomputers to understand how oxygen impurities influence and impede the electrical properties of silicon wafers--a problem that has plagued semiconductor manufacturers for years. In a simulation, a researcher can introduce oxygen molecules into a silicon lattice and watch how it throws the local electrons into a tizzy--something no microscope can observe. The same approach can be used to study another important problem--the migration and diffusion of impurities within a silicon crystal. Insights from such simulations could lead to improved manufacturing processes.
Exciting results are also emerging from studies of the surface of silicon crystals. The outermost atomic layer seems to consist of pairs of atoms, bonded together in a "barbell" configuration. Theoretical experiments indicate that each barbell can exist in one of two states--up or down. This suggests the possibility of storing bits of data on an atomic scale--many thousands of times more compactly than in present computer memories. Other simulations show that a thin metallic tip, similar to those used in Scanning Tunneling Microscopy, can in principle establish the required orientation of the surface atoms. Thus the supercomputer simulations may lead to the development of revolutionary new information storage technologies.
The synergy between science and technology is crucial for developing the next generation of new technologies. Present computer designs will reach limits dictated by the laws of physics. Can faster, smarter machines be built to model the human brain? Can biological components be built into computer chips? What about using individual molecules as switches a thousand times faster than microelectronic devices? These are the kinds of breakthrough technologies realizable only through fundamental research--research that is itself supported by advanced technology.