Nanotechnology and the future of information technology
Thomas N. Theis
IBM T.J. Watson Research Center, Yorktown Heights

March 10, 2004, 8:30 a.m.


Information technology has prospered as scientists and engineers have learned to make "bits" ever smaller. Current manufacturing processes cannot build much structure into an object at a length scale less than the minimum lithographic dimension of 90 nm, but there is no physical reason we cannot learn to design and build objects with complex structure defined on all length scales down to the atomic scale. Opinions on how this might be done are often divided into apparently conflicting camps - "top down" versus "bottom up", lithography versus chemical synthesis, "nanobots" and "molecular assemblers" versus "self assembly". None of these distinctions is very meaningful from the point of view of physics. What is meaningful is the error rate with which structural information can be imparted to an object by some dynamical process. "Digital" processes, with a few energetically accessible states, can have very low error rates, but are energetically costly. ünalog" processes, with many accessible energy states, have higher error rates, but can be very energy efficient. Both types of information transfer are discernible in every manufacturing process. Above the scale of the minimum lithographic dimension, lithography is inherently digital. Chemical synthesis tends to be analog. The trick is to combine the two modes of imparting structural information so that the complex, hierarchically-organized systems of information technology can be manufactured at minimum cost. Conventional optical lithography won't allow us to impart structural information at the atomic and molecular scale, but scanning-probe and other emerging lithographic processes will. Combined with increasingly sophisticated chemical synthetic processes, it should become possible over the next few decades to design and control the structure of an object on all length scales, from the atomic to the macroscopic, and to do so cheaply and reliably in manufacturing. The emergence of such a mature nanotechnology would ensure continued exponential reductions in the cost of information technology hardware, and generate yet unimagined new products and industries.



Share
Nanotechnology and the future of information technology
Thomas N. Theis
IBM T.J. Watson Research Center, Yorktown Heights

March 10, 2004, 8:30 a.m.


Information technology has prospered as scientists and engineers have learned to make "bits" ever smaller. Current manufacturing processes cannot build much structure into an object at a length scale less than the minimum lithographic dimension of 90 nm, but there is no physical reason we cannot learn to design and build objects with complex structure defined on all length scales down to the atomic scale. Opinions on how this might be done are often divided into apparently conflicting camps - "top down" versus "bottom up", lithography versus chemical synthesis, "nanobots" and "molecular assemblers" versus "self assembly". None of these distinctions is very meaningful from the point of view of physics. What is meaningful is the error rate with which structural information can be imparted to an object by some dynamical process. "Digital" processes, with a few energetically accessible states, can have very low error rates, but are energetically costly. ünalog" processes, with many accessible energy states, have higher error rates, but can be very energy efficient. Both types of information transfer are discernible in every manufacturing process. Above the scale of the minimum lithographic dimension, lithography is inherently digital. Chemical synthesis tends to be analog. The trick is to combine the two modes of imparting structural information so that the complex, hierarchically-organized systems of information technology can be manufactured at minimum cost. Conventional optical lithography won't allow us to impart structural information at the atomic and molecular scale, but scanning-probe and other emerging lithographic processes will. Combined with increasingly sophisticated chemical synthetic processes, it should become possible over the next few decades to design and control the structure of an object on all length scales, from the atomic to the macroscopic, and to do so cheaply and reliably in manufacturing. The emergence of such a mature nanotechnology would ensure continued exponential reductions in the cost of information technology hardware, and generate yet unimagined new products and industries.



Share