Conference presenters Viola Vogel (left) and Steven Block (center) enjoy dialoging with attendees.
Kinesin's head is only 10 nm in diameter, making it the smallest linear motor, the "physicist's myosin". Its structure was solved in 1997. The backbone of kinesin proved to exactly match a subset of myosin's structure. This could only be seen from the 3D structure, because the genetic sequences are too different to show it in the linear amino acid sequence.
Understanding motors' mechanisms requires knowledge of dynamics as well as of structure. In vitro assays of motor function have helped. Today we can bind a motor to a bead, and watch the bead crawl along a microtubule. With optical tweezers, we can exert a force on the micron-scale bead, which transfers them to a nanometer-scale motor.
Optical traps typically exert forces of ~1.5 pN, comparable to noncovalent binding forces. A typical trap width is 200 nm. The position of a bead can be sensed to within 10-3 wavelengths, with bandwidths around 103 Hz. One can use Poisson statistics to get just one motor on a bead, binding them to a sufficiently dilute solution that most beads bind no motors, and only a minute fraction bind two or more.
A typical speed for an unloaded kinesin motor is 500-800 nm/sec. Kinesin doesn't stay on a microtubule permanently. Typically it moves ~100 steps, then lets go. With an optical trap one can see the individual steps (~8 nm in length - matching the repeat spacing of the microtubule), and one can measure the speed vs. load curve for the motor, seeing it stall under a 7 pN load. Knowing the step size and structure of the motor, and knowing that it stays on the microtubule for many steps, one can model plausible gaits for the motor, e.g. an arm-over-arm motion.
Finer examination of the statistics of the kinesin motion, the details of the fluctuations in speed, helps constrain models of the dynamics. Kinesin is a chemical motor, and must go through a series of intermediates during an engine cycle. The variance in the motion goes down as the number of balanced rate limiting steps rises. By lowering the ATP concentration enough, the ATP binding step(s) can be forced to become rate limiting. This experiment showed that 1 ATP molecule is used per kinesin step.
Optical traps permit other types of statistical measurements as well. By setting up a feedback loop to keep the force on a motor constant, one can look at fluctuations in position in a way that cancels out the compliance of all the components attached to the motor. Information from this experiment can be used to refine models of the motor. The models now include a free energy landscape for the coupled ATP hydrolysis/head motion with two load-dependent states in the model.
Dr. Block also briefly described work on RNA polymerase, which operates in a "lower gear" than kinesin, with a step size corresponding to base spacing rather than microtubule repeat lengths, and a stall force of 25 pN. Efficiency of both motors is ~50%.
Following the 1998 Feynman Prize Presentation (see above), conference attendees enjoyed lunch while listening to Dr. Eric Drexler take (upon the urging of close colleagues, he claimed) a "deliberately rude" approach to re-framing the major technology policy issues raised by the anticipated development of molecular nanotechnology.
Foresight Institute Chairman, Drexler gives a hair-raising lunch talk.
Establishing his credentials to address his topic, he noted that:
To demonstrate such an obvious, simple claim, he predicted (holding up a penny) that if he let go of the penny, it would fall and land someplace on the floor. As examples of the kinds of claims he has in fact made, he cited:
The trick, according to Dr. Drexler, is to say simple, obvious things that nevertheless surprise people. For example, among the products of molecular machines that use positional control of chemical reactions will be:
Accepting the feasibility of molecular manufacturing led Dr. Drexler to speculate what technology policy would be like if it were formulated "as if reality mattered." It seems that several important issues would be impacted.
Wealth vs. poverty. Self-replicating assemblers fed with sunlight, CO2, and dirt would produce pharmaceuticals, solar cells, supercomputers, etc. for about $1/kg.
Clean environment vs. pollution. Using CO2 as a feedstock for an inherently clean manufacturing process, combined with solar cells cheap enough to use as paving and roofing means there will be no greenhouse problem. Dr. Drexler estimated that it should be possible, using only waste CO2 and garbage dumps, to sustain a population larger than the current US population at a standard of living higher than the current one. Thus the future could see a massive rollback of the human impact on the environment.
Further, Dr. Drexler rudely asserted that, given the future capabilities of molecular manufacturing, the most effective strategy to preserve biodiversity was to use current resources to collect samples for future study, rather than wasting resources on scientific study of ecosystems with current technology.
Health vs. disease. Here again, Dr. Drexler asserted that current priorities are misplaced. Where, he asked, are the efforts to build new molecular machines, which would lead to cell repair machines able to target viruses and cancer cells and to proofread DNA? A crucial capability provided by molecular manufacturing systems will be that cells and tissue samples could be placed on a diamond slab at liquid helium temperature, essentially freezing all molecular motion immediately, and then disassembled one molecular layer at a time, yielding a very precise picture of the differences between healthy and diseased cells. [Such information could then be used to program cell repair machines to restore diseased tissue to a healthy state.]
Considering the capabilities of future cell repair machines led Dr. Drexler to query the audience as to how many had made arrangements to preserve their personal information databases against the possibility of a system crash. [That is, how many had made arrangements to have their brains cryogenically preserved after legal death. Somewhat more than a dozen hands were raised.] Because the alternative is not to preserve the information within individual brains, Dr. Drexler asserted that it is criminal not to talk about cryonics as an option.
Peace vs. war. Noting that conflict between those with molecular manufacturing technology and those without would be like war between those with machine guns and those with spears, Dr. Drexler noted that the crucial issue is that the guys with the machine guns be the ones "who don't want to shoot." Thus discussions on security and strategy issues "loom large".
Summarizing, Dr. Drexler concluded that our national debate about 21st century technology is totally out of touch with reality -- most policy makers are discussing a fantasy world where we are stuck forever with 1980's technology.
Two rude remarks were addressed to research funding priorities:
Tying his last remark into another technology that will strongly impact the coming decades, Dr. Drexler noted that the "failure" of the artificial intelligence efforts of the 1970's was the inability to get human-like intelligence from computers with 10-6 times the computing capacity of the human brain; in the next few decades molecular manufacturing will provide computers with 106 times the computing capacity of the human brain.
|Foresight Update 35 - Table of Contents|
The National Science Foundation sponsored a forum in conjunction with the Sixth Foresight Conference on Molecular Nanotechnology: "From Scientific Discovery to the Nanotechnology of Tomorrow".
Speaker James Murday (left), with NSF sessions chairs: Mike Roco (center) and Ilhan A. Aksay (right)
Dr. Herb Goronkin, director of physical laboratories at Motorola, described nanoelectronic devices and the status of their development.
He noted that the power dissipation of a circuit depends on the number of active electrons, so the drive towards circuits with a great deal of function yet reasonable power dissipation drives the industry directly towards single electron devices. Conventional CMOS continues to be scaled down, but the gate oxide thickness is already nearing the point where tunneling current becomes a problem, and the problem becomes exponentially worse with further scaling. For applications which need more than a terabit, yet need low power, CMOS will not suffice.
An alternative device, the single electron transistor (SET) is already known to work. SETs have a conducting island coupled to two electrodes through tunnel junctions. They conduct when the voltage between the electrodes is high enough to charge the capacitance of the island with a single electron. Capacitive coupling of the island to other nodes can modulate this conduction. SETs have a long history. Anomalies in the charging of thin films due to coulomb blockade effects had been seen in the 1940s-60s. In the mid-80s, the blockade effects were understood. The temperature of operation has to be low enough that thermal energy doesn't wash out the charging energy, e2/2C, of the island. The tunnel barriers must be more resistive than a quantum channel, Rtunnel > h/e2 ~ 26 kohm.
Today, SETs have been incorporated into devices as large as a 128Mbit memory. SETs have been implemented in a variety of materials. Much of the research work has been done in GaAs, in 2D electron gases confined near the surface and controlled by electrode potentials. This work has required temperatures as low as 50 millikelvin. Other materials have included polysilicon grains, InAs dots deposited on a ridge of a GaAs crystal, and AuPd islands. The AuPd islands have operated at 77K. Single organic molecules bound to gold electrodes through thiol linkers have also demonstrated SET activity.
One application for SETs is clearly viable. They have been used as highly sensitive electrometers, with noise levels of 10-4 electron/Hz1/2. One proposed application is a quantum flash memory (which would operate at room temperature), but retention times are still short compared to existing flash memories. One limitation of current SETs is the need for very complex circuits. One example requires 8 SETs, a 3 phase clock, and 2 amplifiers to form an AND.
An alternative approach to high density, low power computation is the use of electrostatically coupled quantum cellular automata. In this approach the logic devices are simple but the wires are complex.
Prof. Joseph Ballyntyne of Cornell Univ. and Director of the National Nanofabrication Users Network provided an overview of projects at the Cornell Nanofabrication Facility. He pointed out that CNF provides a national facility for both nanofabrication and microfabrication that is available to any qualified researcher from a United States academic, industrial or government laboratory who has a good idea, but neither the equipment nor the expertise for nanofabrication. Work done at CNF is heavily subsidized by the National Science Foundation.
Currently there are 300-400 projects at CNF. As examples, Prof. Ballyntyne showed results from a few of these, including a MOSFET with channel widths of 50 nm, atomically flat (single plane, no terraces) silicon surfaces, quantum dot cellular automata for future digital logic circuits, aligning DNA molecules to pass by two near-field optical probes for DNA sequencing, and neurons growing on textured silicon or on micro-printed patterns of protein. These examples demonstrated that nanofabrication is now of use to a wide variety of disciplines, not just to the electronics area, where it has historically been of most importance.
As an example of the unpredictable evolution of nanoscale technology, Prof. Ballyntyne pointed to the early development work at CNF of "gene-gun" technology, in which small metal spheres are coated with DNA and then shot through cell walls, permitting the DNA to become part of the cell genome. This technology has already revolutionized plant breeding.
Dr. Meyya Meyyappan gave a talk describing work on nanotechnology at several NASA sites. He noted that "The future of space is information technology." He cited many subsystems in NASA's spacecraft that could benefit from further miniaturization, amongst them the devices that provide:
He pointed out that miniaturization allowed the use of new architectures, such as collections of microspacecraft, to accomplish missions. The kinds of target systems that NASA is interested in range from 5 kg spacecraft to avionics systems-on-a-chip.
The first group Dr. Meyyappan described was the Center for Integrated Space Microsystems (CISM), a group within JPL in Pasadena. CISM has interests in low power electronics (workers in this group have, for instance, fabricated an imaging array that uses only 20 nW to monitor 250 x 250 pixels) and in revolutionary technologies such as quantum computing. Their long-term goal is to progress from today's fixed hardware to reconfigurable, and then to evolvable, hardware.
The second group Dr. Meyyappan described studies nanotubes at the Johnson Space Center. They study the nanotube growth mechanism, using spectroscopic diagnostics. Oddly enough, this does not seem to be being studied in academia. They also study nanotube applications, such as hydrogen storage and composite materials.
Dr. Meyyappan also described his own group, the Integrated Product Team on Devices and Nanotechnology at NASA/AMES. This group started in computational physics and chemistry, including projects such as integrated modeling of electronic, optical, and thermal effects in optoelectronics, modeling of electronic and mechanical properties of carbon nanotubes, and ab-initio modeling of fundamental processes in microelectronics and in nanotechnology. Their experimental work includes nanotube fabrication (at the lowest temperature demonstrated thus far, 500C) and lithography at the 10 nm scale using nanotubes.
Prof. Hiroshi Komiyama of the University of Tokyo described studies of dispersions of nanoparticles of Au, Ag, Cu, and GaAs that were formed after sputtering those materials on a SiO2 surface. The surprising result was that, instead of forming continuous layers via layer-by-layer growth, the sputtered materials form nanoparticles. Specifically, they formed small nanoparticles that moved readily about the surface, coalescing and growing until they reached a size where the physical properties of the mesoscopic particles approached those of the bulk material.
For example, initial Au nanoparticles were 2 nm in diameter and contained about 500 atoms, but as sputtering times were increased, the particles grew to a size of about 20 nm, at which point they were too large to move. The smaller particles have lower melting temperatures (500 °C versus 1064 °C for 20-nm particles and for bulk Au) because the larger surface area to volume ratio means fewer bonds among Au atoms in the particles.
Prof. Komiyama speculated that this method of dispersing nanoparticles on surfaces could be applied to making 2-D quantum dot arrays, and further, by alternating layers of SiO2 and sputtered metal, 3-D quantum dot arrays could be made.
Additional information about this project is available at Prof. Komiyama's web site: http://www.komiyama.t.u-tokyo.ac.jp/ab-growth/index.html
Dr. James S. Murday (from NRL) gave a talk on three case studies of DOD research on nanostructures.
The first case that he described was the study of hydrogen desorption from silicon surfaces. This process can be caused by electrons tunneling from an STM tip and exciting vibrations in the Si-H bonds due to energy transfer from the electrons. It has been used to remove individual hydrogen atoms from the silicon surface. In addition to the promise of this process for atomically precise fabrication, Dr. Murday described how the study of this desorption enhanced current semiconductor processing. One of the desorption studies looked at isotope effects from substituting deuterium for hydrogen. This substitution reduced the desorption rate by orders of magnitude. The Si-H bonds studied in the STM environment are also used to passivate the Si/SiO2 interface in conventional microelectronics. Substituting deuterium for hydrogen in semiconductor processing increased transistors' lifetimes, and allowed use of smaller devices than would otherwise have been possible.
The second case that Murday described was the use of the mechanical force from the binding of a single molecule to allow the detection of that molecule. He described AFM experiments where DNA binding from a known number of complementary sites could be seen in the force curves, and described an instrument (from R.J. Colton et. al) which used a magnetic "immunobead" to exert a magnetically-controlled force on an AFM lever if an analyte molecule was present to bind the bead and tip together. This technique has already been shown to be a factor of 10-102 more sensitive than conventional detectors and may allow a factor of 106-108.
The third case that Murday described was nanoparticle research - seeking better materials properties by using particles (which need not be atomically precise) with nanometer dimensions. WC/Co, for instance, has 50% better hardness when nanometer scale WC particles are used than when they are larger. It was also found that residual stress was low in these materials, permitting them to be used in thick layers.
Professor Ilhan A. Aksay of Princeton University discussed hierarchically structured materials, composite structures in which different design principles pertain at different length scales. The advantages of such composites are demonstrated by biological shells, nanometer-scale organic-inorganic composites in which CaCO3 "bricks" are assembled and held together by templates of organic "cement", resulting in a fracture strength for the shell that is 100-fold higher than that of pure CaCO3, even though the organic component is only 5% by weight of the composite.
Prof. Aksay reported on improved ceramic fabrication methods in which self-assembly at scales of less than 1 nm provides templates for patterning an inorganic phase on scales of a µm or greater. Since various techniques exist for patterning useful materials on the µm-scale, the limiting factor is the availability of chemistries for self-assembly on the nm-scale, he stated.
Dr. John Mendel and Dr. M. Carmody of Eastman Kodak, Rochester, N.Y. reported on using mechanical size reduction of organic substrates to produce nanoparticle pigment systems. Conventional pigment particles are in the range of 500 nm, overlapping the wavelength of visible light, thus producing less intense colors because the pigment particles scatter much of the light. They reported that nanoparticle pigments, containing particles mostly in the range of 10 to 20 nm produced superior results.
Dr. Robert Tampé of Max-Planck Institute for Biochemistry, Martinsried, Germany described the organization of biomolecules into arrays, or 2D crystals. These arrays are based upon artificial membranes of self-assembling lipid molecules. Metal ions are used to insert biomolecules into the membranes. To do this, the lipid molecules are linked to chelator molecules that bind metals. The biomolecules are proteins that have been genetically engineered to insert a sequence of 5 to 6 histidine residues into the protein, thus providing a site for a metal to bind. Proteins can therefore be inserted or removed from the membranes via the addition or removal (by using a competing chelator) of metal ions. The result is that proteins are not only immobilized in the membrane, but also uniformly oriented.
A wide variety of proteins can be modified by genetic engineering to insert into these membranes, which in turn, Dr. Tampé reported, can be transferred via the Langmuir-Blodgett technique to flat surfaces, such as mica, and then imaged by AFM. One current application is the study of antigen processing by 2D arrays of proteasome complexes (cylindrical arrays of proteins active in the degradation of protein antigens to produce peptide fragments recognized by the immune system).
|Foresight Update 35 - Table of Contents|
The evening of November 13 was devoted to the poster session, which was very competitive, and included contributions from this year's senior co-chairman and the tutorial chairman.
A list of all abstracts, for both talks and posters, can be found at http://www.foresight.org/Conferences/MNT6/Abstracts/index.html.
A list of all full papers, for both talks and posters, can be found at http://www.foresight.org/Conferences/MNT6/Papers/index.html.
Two of the poster presentations seemed especially promising from the perspective of the development of molecular manufacturing.
Dr. Kevin D. Ausman et al. showed that nanotubes bent across a v-shaped substrate were selectively oxidized by HNO3 at the site of the bend. They have further plans to attempt Diels-Alder addition to bent nanotubes. This technique may enable us to build up covalently bound structures on the walls of nanotubes, at sites selected by readily controllable mechanical stresses. If some technique can be used to move the location of the maximum stress, perhaps an array of modifications to the nanotube wall can be fabricated.
Dr. Ausman has a full paper on the conference site, so further details can be seen there.
Dr. James E. Hutchinson et al. are building structures of Au55 metal clusters on DNA and other biomolecule templates. They are using ligand exchange reactions to select how the Au clusters will bind. They hope to assemble electronically active structures (from the Coulomb blockade effects in the Au clusters) with interesting topologies (from the biomolecule backbone). Since all of the components of this system (the Au clusters, the bridging ligands, and the biomolecule backbone) are atomically precise, this technique might allow construction of self-assembled atomically precise electronic devices.
The abstract for this work is here.
From Foresight Update 35, originally published 30 January 1999.
Foresight materials on the Web are ©1986–2014 Foresight Institute. All rights reserved. Legal Notices.