Foresight Update 20

page 2

A publication of the Foresight Institute

Recent Progress: Steps Toward Nanotechnology

by Russell Mills

Accessing atoms and molecules

Today's feats are significant compared to the parlor tricks of a year or two ago.

Nanoprobe instruments--such as the scanning tunneling microscope (STM) and the atomic force microscope (AFM)--are misnamed. They are proving to be just as adept at manipulating atoms as at imaging them, and it always seems awkward to call them "microscopes" while describing their latest feats of construction. And today's feats are starting to be significant compared to the parlor tricks of a year or two ago: spelling out company logos with atoms, or gashing a surface with the probe tip. Let's take a look at some recent developments:

Researchers at the Aono Atomcraft Project in Japan, using an STM, are now able to extract a single silicon atom from the surface of a silicon crystal and rebond it to the surface at a different location. Atoms translocated in this manner can be re-removed without disarranging the underlying atomic layers. Atoms brought from afar can be used to repair holes in the silicon surface, or they can be used to build structures on top of the surface. [J. Vac. Sci. Technol. B 12(4): 2429-2433, Jul/Aug94] Folks, this looks an awful lot like real nanotechnology.

An STM equipped with a platinum-rhodium tip enabled workers at California's Lawrence Berkeley Laboratory to carry out chemical reactions at localized sites on a substrate. The experiments took place in a chamber containing hydrogen gas and a working surface coated with unsaturated hydrocarbons - i.e., molecules containing double bonds that can react with hydrogen. When the STM tip was activated by voltage pulses, it began to act as a catalyst, splitting nearby hydrogen molecules and transferring the hydrogen atoms to hydrocarbons on the working surface. Areas of the working surface that were scanned by this activated STM tip underwent a characteristic change in texture, indicating that the hydrocarbons in these areas had indeed been chemically hydrogenated. The conversion appeared to be 100%, judging by the pictures. [Science 265: 1415-1418, 2Sep94]

In this work no effort was made to restrict the chemical reaction to a single site; instead, the probe was scanned over a rectangular area containing many substrate molecules. Presumably the hydrogenation could just as easily have been confined to a much smaller area, and perhaps to a single substrate molecule. The basic methods used here may work with certain other reactions besides hydrogenation, and could provide a toolkit of useful transformations for the construction of nano-objects.

AFMs are normally used to map the topography of a surface. But a group of chemists at Harvard and M.I.T. have used one to map a surface according to the types of molecular groups that are attached to it. The key to this technique is to "functionalize" the AFM probe by coating it with a layer of molecules that will interact with the molecules on the surface being mapped. As the probe tip is dragged across this surface, the forces generated by the interaction will vary with the nature of the molecular groups the tip encounters. These forces, which can be measured, provide a chemical profile along the track of the tip. If the surface has a pattern of several different kinds of molecules, then the AFM should be able to form an image of this pattern. And indeed, the images these researchers obtained of their samples correctly showed the molecular patterns they had put there. [Science 265: 2071-2074, 30Sep94]

Functionalizing the tips of nanoprobe instruments is an essential step in turning them into useful molecular constructors. Of course, randomly coating a probe tip with a layer of one chemical, as was done here, is not the final answer. But even this primitive functionalization may prove useful for such applications as DNA sequencing, where the different nucleotides would exert differing forces on the tip.

Physicists at the University of California at Santa Barbara have made AFM images of the enzyme lysozyme that seem to show the enzyme flexing as it interacts with a molecule of its substrate. The motions probably correspond to conformational changes which the enzyme undergoes as the substrate binds to it, and these show up in an AFM image as height fluctuations. [Science 265: 1577-1579, 9Sep94]

At AT&T Bell Laboratories in New Jersey, researchers have recorded the optical spectrum of single molecules at room temperature. Molecules of a fluorescent compound were dispersed on a thin plastic film and then examined with a near-field scanning optical microscope. When the fluorescent molecules were illuminated by an appropriate frequency of green light, they absorbed and re-emitted the energy at a variety of frequencies characteristic of their molecular structure and of their immediate environment. From the spectra obtained from 28 individual molecules, the researchers concluded that the act of observing a molecule does not perturb it sufficiently to invalidate the spectral data obtained from it. They suggest that this technique may be used to study the binding of enzymes with their substrates. [Nature 369: 40-42, 5May94]

The ability to obtain meaningful spectral measurements from single molecules (or even parts of a molecule) may be needed to confirm certain kinds of structures that will someday be built with nanoprobe instruments.

A more interesting concept, however, is the inverse process: feeding a designed spectral sequence to a single molecule (or parts of a molecule) in order to control its behavior. This has been tried in bulk chemistry as a way of controlling chemical reactivity, but without much success--probably because the molecules are flopping around randomly in solution in an unsynchronized manner. But the technique might work better if it were applied to a single molecule in the solid phase.

One approach to the problem of communicating with nanodevices is to design them to receive, interpret, and generate messages encoded as spectral sequences. The work described above could be a small step in that direction. So, too, could the following:

Chemists at Stanford University and Indiana University have detected individual fluorescent molecules in liquids, using a confocal fluorescence microscope (CFM). The CFM differs from the near-field microscope (NFM) in that the detection apparatus is much further from the sample-- micrometers as opposed to 10 nanometers. Although this arrangement precludes high-resolution measurements of a molecule's position, it offers some advantages: a larger number of photons can be delivered to a sample molecule to stimulate fluorescence, and this improves the time-resolution of measurements; the CFM is less invasive than the NFM since the apparatus is further from the sample; it operates in a three-dimensional sample space rather than a two-dimensional one, making possible the mapping of an extended sample by focusing at a series of depths. [Science 266: 1018-1021, 11Nov94]

Calvin Quate, a physicist at Stanford University, recently made a transistor in which one component (the gate) was created by using the probe tip of an AFM to make scratches in appropriate places on the silicon substrate. The resulting transistor performed correctly. Although the gate was as large as those that can be made by photolithography (about 100 nanometers wide), the AFM is believed to be capable of making much smaller structures. Quate envisions arrays of thousands of AFM tips working in unison to construct the electronic circuits of the future. His team has already begun experimenting with five-tip arrays. [Science 266: 543, 28Oct94]

The scratches made by the AFM in this experiment were not atomically precise. But someday soon, they probably will be. (We'll then have to call them something more dignified than "scratches".) Quate's vision of arrays of busy AFM tips bears a remarkable resemblance to the vision of arrays of molecular robot arms in Eric Drexler's Engines of Creation.

Foresight Update 20 - Table of Contents

Movies of molecules?

Many biological macromolecules, such as enzymes, carry out their tasks by going through machine-like motions: grabbing, holding, pinching, twisting, pulling, releasing, etc. Biologists would love to be able to watch these actions in detail as they take place--or, at least, watch them in a movie in which the motions are slowed down to a human time-scale. Now this dream is becoming a reality.

Research groups at Los Alamos National Laboratory and at the University of Chicago have captured the motions of the protein myoglobin as it seized and released small molecules such as oxygen. A standard recording technique was used: x-ray crystallography, in which the protein being studied is crystallized and bombarded with x-rays; an analysis of the scattered x-rays yields a 3-dimensional map of the protein. In the myoglobin work, the researchers had to slow the activity of the molecule by cooling it to 77 kelvins in order to get adequate x-ray exposures. In the future they hope to have access to brighter x-ray sources, which should drastically reduce exposure times and eliminate the need for physiologically unrealistic low temperatures. [Science 266: 364-365, 21Oct94]

Foresight Update 20 - Table of Contents

Rotary engines in mitochondria

The structure of a remarkable biological machine has recently been resolved in atomic detail. English biochemist John Walker used x-ray crystallography to reveal the structure of the catalytic portion of the enzyme ATP synthase. This enzyme, found in the mitochondria of all cells, is a collection of proteins responsible for making ATP, the energy transport molecule.

To make a molecule of ATP, two precursor molecules (ADP and phosphate) must be brought together and chemically joined. Since the precursor molecules are floating in solution inside mitochondria, the ATP synthase machines have easy access to them. But exactly how does ATP synthase go about grabbing these precursors, bringing them together, and then releasing the product? The details are beginning to emerge from Walker's map of the enzyme.

ATP synthase resembles a lumpy orange spinning on a long axle. The "orange" has six segments - three alpha subunits, and three beta subunits. The beta subunits are the ones responsible for making ATP. The "axle" is actually a tube, the other end of which is embedded in the mitochondrial membrane; it contains the still-mysterious mechanisms responsible for enabling the enzyme to spin. As the "orange" turns on the axle, the subunits are pushed and pulled in a cyclical pattern by the axle's irregularities; the resulting deformation of the subunits manifests itself as the mechanical movements required for making ATP. [Science 265: 1176-1177, 26Aug94]

One of the intriguing facts about this sophisticated enzyme is that it is found in every kind of cell, from bacteria to plants and animals. That presumably means that it is a very ancient biological device that evolved with the early bacteria or even before. It seems surprising that the creatures of that era already had the use of such an elegant machine.

Foresight Update 20 - Table of Contents

Peptide nanotubes

Nanotubes are forerunners of self-assembling structures that will help us modify cells

As discussed in this column in Foresight Update 18, researchers at Scripps Research Institute in 1993 designed rings of amino acids that self-assemble into nanotubes. The same researchers have recently taken this work a step further by creating self-assembling nanotubes that form channels through lipid membranes similar to those found in cells.

These new nanotubes are based on two design principles: first, the amino acids in each ring must form hydrogen bonds with the amino acids in adjacent rings so as to hold the nanotube together; and second, the exterior of the nanotube must attract (rather than repel) the lipid molecules that make up the membrane--otherwise the membrane would eject the nanotube.

Membrane-spanning channels are standard features of living cells. They usually contain valves to regulate the flow of molecules across the membrane. The Scripps researchers showed experimentally that their nanotubes were very effective channels for ion flow across the artificial membranes they were using. They suggest that such nanotubes may be useful for delivering drugs into cells in the body. [Nature 369: 301-304 & 276-277, 26May94; C&EN 26May94: 4-5]

The idea of using nanotubes to deliver drugs into cells, without first developing a valve to close the channel after the drug has entered, sounds like a bad idea. Punching holes in cell membranes and letting material pass freely through the holes is a tactic used by some toxic organisms to kill cells.

On the brighter side, it seems safe to say that these nanotubes are the forerunners of a large class of self-assembling structures that will help us to modify cells and make them do our bidding.

Dr. Russell Mills is research director at KAH Sciences in California.

Foresight Update 20 - Table of Contents

Molecular Robotics at USC

by Ari Requicha

Prof. Ari Requicha is leading a new effort in molecular robotics at USC.

Prof. Requicha of the University of Southern California taught the first course using Nanosystems as a textbook (see Update 18). Here he describes his new inter-disciplinary laboratory in nanotechnology:

Molecular Robotics is an emerging and highly interdisciplinary field that seeks to produce new materials and devices at a nanometer scale by direct interaction with atomic structures. Whereas conventional chemistry relies on bulk phenomena such as diffusion to create self-assembling structures, Molecular Robotics manipulates structures by applying external forces and precisely positioning atoms and molecules.

It is a revolutionary technology, which attempts to provide fine control over the structure of matter, analogous to the fine control we can now exert upon the bits and bytes of information structures. Future applications range from very fast and small computers to nanorobots and self-replicating machines. Nanorobots might be programmed to recognize and repair specific kinds of cells, and to perform a large variety of other tasks that are now impossible to accomplish. Large structures might be built by using massive parallelism.

The technology has the potential for major scientific and practical breakthroughs. However, today we lack most of the tools necessary for realizing the dreams of Molecular Robotics. Research opportunities abound.

The Laboratory for Molecular Robotics was established at USC in late Fall 1994 with primary support from the Zohrab A. Kaprielian Technology Innovation Fund. The initial focus of the laboratory is on sensing and manipulation using Scanning Probe Microscope (SPM) technology. SPMs are capable of atomic-level imaging, and of manipulation with the precision required for positioning atoms and molecules. These instruments can be viewed as robots with sensory feedback, whose probe tip must be equipped with grippers--specially-designed molecules--so as to manipulate atomic-level structures in an environment that has significant spatial uncertainty, making and breaking chemical bonds, or providing chemical sensitivity, as in the lock-and-key model for enzyme action.

Our interdisciplinary team is tackling problems of tip design and construction for Scanning Tunneling Microscopes (STMs) and Atomic Force Microscopes (AFMs); computer control and programming of these devices; design and visualization aids; design and construction of substrates that serve as"nanoworkbenches"; methods for attaching specific molecules to a tip; designing probes and tips that distinguish between molecules by tactile affinity; and determining the structure of materials by tactile probing.

We are working in the context of a specific, intermediate-term goal: the construction of "nanoarrays" composed of nanoscale entities, such as organic or biological molecules of interest in optical computing and related applications, placed on a regular array of microscale semiconductor mesas. In the long run we expect to address even more challenging problems such as sensing and manipulation of biomaterials, and especially DNA.

Senior personnel associated with the new laboratory include Bruce Koel, Professor of Chemistry; Aristides A. G. Requicha, Professor of Computer Science and Electrical Engineering, Lab Director; Anupam Madhukar, Professor of Materials Science and Physics; and Peter Will, Division Leader, USC Information Sciences Institute.

Address inquiries to Professor Requicha at Laboratory for Molecular Robotics, Computer Science Department, University of Southern California, Los Angeles, CA 90089-0781. Internet: lmr@lipari.usc.edu; tel (213) 740-4502; fax (213) 740-7512.

From Foresight Update 20, originally published 1 February 1995.