a. Similarities: small scale, electronic quantum effects. Microtechnology has enabled the fabrication of micron-scale mechanical devices. These share basic scaling properties with nanomechanical devices — and so, for example, electrostatic motors are preferred over electromagnetic motors in both micro- and nanotechnology (Section 2.4.3). Further, quantum electronic devices of kinds now being explored with microfabrication technologies may become targets for molecular manufacturing.
b. Differences: fabrication technology, scale, molecular phenomena. Microfabrication relies on technologies (photolithographic pattern definition, etching, deposition, diffusion) essentially unrelated to those of molecular manufacturing. In a sense these two fields are moving in opposite directions: microfabrication attempts to make bulk-material structures smaller despite fabrication irregularities; molecular manufacturing will emerge from attempts to make molecular structures larger without losing the atomic precision characteristic of stereospecific chemical synthesis. Making structures consisting of a few dozen precisely-arranged atoms seems unachievable using microfabrication, but is routine in chemical synthesis. The gears, bearings, and motors described in Chapters 10 and 11 differ in volume from their closest microfabricated counterparts by a factor of ~109, and rely on molecular structures and phenomena in their operation.*
a. Similarities: molecular structure, processes, fabrication. Chemical principles describe the basic steps of molecular manufacturing, since each consists of a chemical transformation. Chemical knowledge can help in evaluating the stability of products, and chemical research has produced the most useful models of the mechanical behavior of molecular objects. Organic chemistry is particularly relevant owing to the superiority of carbon-based structures for most mechanical applications. Fundamental chemical concepts such as °bonding, °strain, °reaction rates, °transition states, °orbital symmetry, and °steric hindrance are all applicable; familiar chemical entities such as °alkanes, °alkenes, °aromatic rings, °radicals, and °carbenes are all useful.** Solution-based organic synthesis can make precisely structured molecular objects; it has even been used to make molecular gears (Mislow, 1989), although of a sort having no obvious utility for nanomechanical engineering.
b. Differences: machine-phase systems, mechanosynthesis. The chief differences between the present subject and conventional chemistry stem from the properties of machine-phase systems and of mechanosynthesis. These have been summarized in Section 1.1.2 and are discussed at length in Chapter 8.
a. Similarities: molecular machines, molecular systems. Molecular biology, like molecular nanotechnology, embraces the study of molecular machines and molecular machine systems. Ribosomes — like mechanisms in flexible molecular manufacturing systems — can be viewed as numerically controlled machine tools following a series of instructions to produce a complex product. Molecular biology and biochemistry stimulated the train of thought that led to the concept of molecular manufacturing (Drexler, 1981), and their techniques offer paths to the development of molecular manufacturing systems (Section 15.2).
b. Differences: materials, machine phase, general mechanosynthesis. Biology is a product of evolution rather than design, and molecular biologists study systems that differ greatly from the eutactic systems described here. Unlike molecular manufacturing systems, the molecular machines found in cells can synthesize only relatively small molecules and a stereotyped set of polymers; they cannot synthesize a broad class of °diamondoid structures. Larger biological structures typically acquire their shapes through the action of weak forces (°hydrogen bonds, °salt bridges, °van der Waals attraction, °hydrophobic forces). As a consequence of stronger bonding, the strength and °modulus of diamondoid components can exceed those of biological structures by orders of magnitude. The bearings, gears, motors, and computers discussed in Part II are accordingly quite different from the bacterial flagellar motor, the actin-myosin system, systems of neurons, and so forth. Biological and nanomechanical systems are organized in fundamentally different ways. For example, cells rely on diffusion in a liquid phase — although they contain molecular machines, they are not machine-phase systems.***
As Section 1.2 indicates, the study of molecular nanotechnology spans multiple disciplines. This circumstance has hampered both evaluation of the existing concepts and research aimed at extending and superseding them. One purpose of the present volume is to assemble a large portion of the necessary core knowledge in a form that requires no specialized knowledge of the component disciplines. An effort has been made (and a glossary provided) to make key chemical concepts accessible to nonchemists, solid-state physics concepts accessible to nonphysicists, and so forth, assuming only a basic background in both chemistry and physics (and a willingness to skip past the occasional obscure observation aimed at a reader in a different discipline). The intended contribution of this work consists not in extending the frontiers of existing fields, but in combining their basic results to lay the foundations of a new field.
To facilitate understanding, several mathematical results in Part I are derived from fundamental principles. Many of these results appear in existing textbooks; others (so far as is known) are novel, being motivated by new questions. The exposition of these mathematical models includes an unusually large number of graphs that illustrate equations in the text; these are provided to facilitate design, which is a synthetic as well as an analytic process. In the analysis of a given system, a calculation based on an equation with a single set of parameter values frequently suffices. In synthesis, however, the designer usually wishes to understand how system properties will vary as controllable parameters are changed; for this, a graph can be more useful than a bare equation.
Different fields have applied different energy units to molecular-scale phenomena, including the kilocalorie per gram-mole of items (≈ 6.95×10–21 J per single item) and the kilojoule per gram-mole of items (≈ 1.66×10–21 J per single item) of chemistry, and the electron-volt (= eV ≈ 160×10–21 J) of physics. The standard SI unit of energy, of course, is the joule itself. To avoid allusions to hypothetical moles of identical systems or to electrons not involved in the problem, and (more important) to enable mechanical calculations involving force, work, kinetic energy, and so forth to proceed without frequent unit conversions, this volume adheres to the joule as the unit of energy. The attojoule (= aJ = 10–18 J) and milli-attojoule (= maJ = 10–21 J) are convenient fractional units.
* The term nanotechnology, first widely used to refer to what is here termed molecular nanotechnology, has since been applied to many small-scale technologies, including conventional microfabrication techniques working in the submicron size range. Accordingly, discussions of the history, status, and prospects of so-called nanotechnology often confuse essentially dissimilar concepts, as if the term ornithology were used to describe the study of flying things, thereby stirring together birds, bats, spacecraft, balloons, and bombers into a single conceptual muddle.
** Discussions phrased in terms of controlling and building with individual atoms (Drexler, 1986a) have fostered a perception that molecular manufacturing would employ individual, unbonded, and hence highly reactive carbon atoms. This rightly strikes chemists as implausible. Indeed, unbound atoms would be difficult to produce and control; more conventional reactants seemed appropriate from the start. The same volume speaks of using reactive molecules as tools to bond atoms together a few at a time, and the first paper on the subject (Drexler, 1981) speaks of positioning reactants and reactive groups. Controlling the motions and reactions of individual molecules, of course, implies controlling the motions and destinations of their individual constituent atoms.
*** It is sometimes suggested that artificial molecular machine systems cannot improve on biological systems because the latter have been shaped by billions of years of evolution. In specific engineering parameters, however, the products of evolution have already been surpassed: graphite composites are stronger than bone, copper is more conductive than axonal cytoplasm, and so forth. Biological systems do, however, excel in their capacity to evolve, and it can be shown that several of their characteristic differences from eutactic nanomechanical systems (including the use of diffusive transport rather than mechanical conveyance) are important to this capacity (Drexler, 1989a).
Copyright © 1998 by John Wiley & Sons, Inc.