Computational Nanotechnology

Computational nanotechnology is an interdisciplinary research field within nanoscience, focused on designing and using mathematical models to simulate nanoscale phenomena and interactions. Information technology has made the nanoscale accessible in many important ways, facilitating sophisticated instruments like the scanning tunneling microscope while also making solvable systems of equations that describe nanoscale behavior. Involving work from chemistry, physics, molecular biology, materials science, engineering, computer science, and mathematics, simulations generate predictions and images of material properties at the nano- and atomic scale. Because many interesting nanoscale phenomena are not yet accessible even with the most sophisticated instruments, simulations provide information that is not experimentally available.

Nanotechnology came of age in the era of accessible, relatively cheap, high-speed computing. Therefore, nanotechnology relies on the computational modeling activities of older disciplines, like chemistry and physics. Computational nanotechnology researchers bring together methods to model new materials, showing attributes from the subatomic scale to the optically visible bulk scale. However, the constantly changing limits of computing create a dynamic trade-off between size of the system being depicted and level of detail possible.

[Page 114]While researchers aim to model materials on a size continuum, there can be discontinuities and incommensurabilities between the models used to show different scales. The challenge to nanotechnologists is to bridge length scales. Models are often said to be of the meso-scale, that is the size range in which phenomena with both classical and quantum mechanical features dominate. Four modeling techniques are common, and each functions best at a particular scale.

At the largest end, finite elements are used to describe bulk properties. At the micron scale classical molecular mechanics is used to depict molecular interactions. Semi-empirical methods allow atomic structure, especially chemical bonds, to be relatively accurately modeled.

At the smallest scale, quantum mechanical effects are modeled; however, such models are often constrained by their computational complexity to relatively small numbers of atoms. Multiscale models, which patch together techniques for more than one length scale, can be extremely computationally complex with parallel calculations being carried out to exchange values between semiautonomous submodels.

The term computational nanotechnology was coined in a 1991 Nanotechnology article by computer scientist Ralph Merkle, then based at Xerox PARC. Merkle's interest in simulations of molecular systems stemmed from his frustration with chemists' and material scientists' inability to build the kinds of molecular structures he was able to imagine. As a computational scientist, Merkle was comfortable with turning the computer screen into an experimental space. Merkle believed that his work on simulated systems would significantly advance the field by helping to determine which designs were feasible and therefore warranted attention by lab scientists to really build them. Simulations would drive and direct experimental work. Merkle's vision for computational nanotechnology was drawn directly and largely from computational chemistry—a field that dates back to the 1940s.

If Merkle is the originator of the term computational nanotechnology, then K. Eric Drexler has been the approach's most famous supporter. Drexler's doctoral dissertation research, subsequently published as Nano-systems: Molecular Machinery, Manufacturing, and Computation used computational modeling, drawn primarily from electrical engineering and applied physics, to demonstrate the feasibility of molecule-scale machines, particularly self-replicating molecular assemblers. There has been extensive criticism of Drexler's simulations, but criticism has been focused on ways in which he may have made questionable decisions about physical features in his models, rather than challenging more broadly whether a computational approach is legitimate. Still, one consequence of Drexler's championing of computational approaches for proofs of concept has been some skepticism of computational nanotechnology and the characterization of some kinds of models as science fiction.

...