merkle.pa@xerox.com (11/10/89)
The First Foresight Conference on Nanotechnology The First Foresight Conference on Nanotechnology, hosted by the Stanford Department of Computer Science and sponsored by the Foresight Institute and Global Business Network, was the first major conference to examine molecular systems engineering as a path to nanotechnology. Held on October 27-29 in the wake of the Bay Area earthquake, the conference in Palo Alto drew about 150 invited participants from three continents and many disciplines. It was a success by any measure. A proceedings volume is planned, availability will be announced in this news group (sci.nanotech). The sessions were audio and video taped, and the Foresight Institute plans to make these available at some point. Again, the availability of this material will be announced in this newsgroup. The Saturday sessions featured scientists defining the state of the art in various enabling technologies leading to nanotechnology. By Saturday afternoon, participants had a good overview of where work stands in these fields: further along than conference chairman Eric Drexler predicted in 1986, but still an undefined number of decades away from nanotechnology, which was defined as "thorough control of the structure of matter." Researchers in protein design, chemistry, biochemistry, biology, scanning tunnelling microscopy, quantum electronics, computer science, micromachines, physics, molecular modeling, and molecular electronics were all drawn together to discuss a common theme: understanding and building structures, devices, and systems on the scale of molecules. The excitement was palpable. Asked to rate the conference on a scale of one to ten, one conference attendee said "Eleven!" Nanotechnology has been described as the manufacturing technology of the 21st century, which some argue will be able to manufacture almost any chemically stable structure at low cost. If realized, such precise fabrication abilities could be used both to improve existing products and to build products that are impossible with present technology. Based on estimates of parts count and power dissipation, components of molecular size could make a single desk-top computer of the future more powerful than all the computers in existence today combined. Devices smaller than a red blood cell might circulate through the body and attack and remove both fat deposits and infectious organisms. These are potential long-term applications of nanotechnology, but the conference started with an examination of where we stand today in efforts to engineer molecular systems. Michael Ward of Du Pont described the design of self-assembling systems by controlling the charge on individual molecules. If the pattern of electrostatic charge on individual molecules is properly controlled, then it is possible to control many properties of molecular aggregates. Federico Capasso, head of Quantum Phenomena and Device Research at AT&T Bell Labs, discussed current work on exploiting quantum effects in devices built with controlled bandgap variations on a nanometer scale. A major limit in building and commercializing smaller devices is fabrication. Tracy Handel of Du Pont discussed the de novo design and construction of a protein by William F. DeGrado's group. This work provides a dramatic illustration that protein engineering is possible, and thus that objects of multi-nanometer scale can be designed and built to precise molecular specifications. Jay Ponder, of the Department of Molecular Biophysics and Biochemistry at Yale, described systems for molecular modeling and for the computer-aided design of proteins. He reports that an algorithm developed in collaboration with Frederic Richards has been quite successful in generating sequences of hydrophobic amino acids which will successfully pack to form the core of a protein with a specified backbone geometry. Molecular modeling is of general importance in molecular systems engineering because the proposed structures are at present often expensive to synthesize and characterize; longer-term proposals (under examination for exploratory purposes) may involve structures that are entirely beyond today's synthetic capabilties. In either case, molecular modeling can frequently distinguish between workable and unworkable proposals. Robert Birge, Director of the Center for Molecular Electronics at Syracuse University, reported on attempts to build a large optical memory with access times below 2 nanoseconds, using bacteriorhodopsin as an optically activated molecular switching element. They currently can achieve 20 nanosecond access times, the major limitation being the speed at which the optical beam can be positioned to "read" or "write" single bits. A later talk by Hiroyuki Sasabe of Japan's Institute for Physical and Chemical Research reported on the current state of molecular engineering research in Japan. He described a broad range of interdisciplinary projects in "intelligent materials" and molecular electronics. John Foster, manager of Molecular Studies for Manufacturing at IBM's Almaden Research Center, presented work with STM (scanning tunneling microscopy) technology, describing advances in both surface imaging and surface modifications. The latter could in theory be used to construct a memory device with storage densities on the order of 100,000 million bits per square millimeter, through a demonstrated mechanism which involves pinning individual molecules to a surface. Joe Mallon, Co-president of Nova Sensor, described the wide ranging abilities of current micro machines. These devices, typically measured in tens of microns, are made primarily of silicon using semiconductor fabrication technology, but are mechanical in nature. Electrostatic motors, gears, levers, joints, sensors, turbines, pumps, and a wide variety of other mechanical devices have been made in this size range and shown to work. Norman Margolus, of MIT's Laboratory for Computer Science, explained the known theoretical limits to computation, perhaps more properly termed the lack of known limits. Quantum uncertainty, thermal noise, and other factors commonly thought to limit computation are, instead, merely constraints. By designing computers in an appropriate way (for example, by building reversible computers) these constraints can at least in principle be satisfied without loss of speed and without requiring any fixed energy dissipation per logic operation. Even with practical constraints, quantum computers that dissipate much less energy than thermal noise per gate operation seem possible, and gate speeds in the femtosecond range seem plausible. Eric Drexler presented recent work that clarifies technical issues in the design of an "assembler," a device capable of guiding the synthesis of virtually any specified chemically stable structure via positional control of chemical reaction sites. Both in his talk and in an accompanying inch-thick preliminary draft, he outlined the design of a sub-micron scale articulated mechanism capable of positioning its tip with a standard deviation in position of less than 0.04 nanometers, despite both thermal and quantum effects. He also presented design sketches for proto-assemblers: cruder devices that might be made in the next decade which could be used both to experiment with positional control of chemical reactions and to build more sophisticated devices. His proposal that AFM (atomic force microscope) tips might be capped by engineered molecular structures, thus providing precise atomic control of the structure at the tip of the AFM (something that is notably lacking at the present time), was met with particular interest. On Sunday afternoon several talks explored the future implications and policy issues raised by this new technology. This process was perhaps the other major achievement of the meeting: consideration of the consequences of a powerful new technology decades before development is completed. Bill Joy, Vice President of Research and Development at Sun Microsystems, discussed what might be done with a trillion processors. He said truly large amounts of computational power would provide us with a new tool which would let us model and understand both physical phenomena and our environment better, and so let our society make better decisions. Lester Milbrath, Director of the Reseach Program in Environment and Society at the State University of New York at Buffalo, expressed his concern that the anticipation of nanotechnology development and its proposed use in environmental cleanup would make policymakers overly optimistic. He doubts that nanotechnology can be developed in time to head off the environmental problems now facing us. Ralph Merkle, a computer science researcher at Xerox PARC, discussed techiques for controlling artificial self-replicating systems. While attractive from an economic point of view, such systems must be designed to avoid any opportunity for unchecked replication and mutation. While "Star Trek" has popularized the idea that "nanites" could rapidly evolve into intelligent social beings capable of negotiating for their own planet, this popular vision appears highly implausible. The simplest and most practical artificial self-replicating systems will have inflexible designs and special raw-material requirements, making them unlike anything able to survive in nature and unable to change. Nonetheless, regulation of the design and use of such systems seems essential to ensure that dangerous new capabilities are not added by irresponsible or malicious parties. Greg Fahy, a researcher with the American Red Cross, discussed the medical implications of progress toward nanotechnology. Aging is a consequence of molecular changes that take place within the body, including changes in genes and their expression. Experimenters have successfully slowed aging in experimental animals; if this work can be extended to humans it should result in increased decades of healthy life. Progress in molecular design on the path to nanotechnology is likely to continue and strengthen this trend, eventually allowing the retention of good health for a prolonged period. The conference closed with two presentations on the broader impacts of technological advance. Economist Gordon Tullock of the University of Arizona cited historical trends showing that, although individuals can be hurt economically by technological advances, the overall effects have been positive. Arthur Kantrowitz of Dartmouth argued for keeping research programs open rather than classified, suggesting that if classified programs must exist, they will benefit from parallel research programs which are open. While it is too early to tell the ultimate impact of this first international conference on nanotechnology, it has clearly raised the level of interest and focused greater attention on both the technology and its consequences. It may well prove to have been the seminal event in the coalescence of a new field and in the emergence of a new and powerful technology.