Foresight Update 53
A publication of the Foresight Institute
Progress with molecular devices, machinery, and building blocks
Continuing a tradition now 14 years old, several hundred researchers from more than a dozen countries converged on the 11th Foresight Conference on Molecular Nanotechnology, held October 10 to 12, 2003 at the San Francisco Airport Marriott in Burlingame, California, to survey progress in a variety of nanoscale sciences and technologies, and (in some cases) to ponder whether or not this progress was leading toward an ability to engineer molecular machine systems. This year's Conference was co-chaired by James T. Spencer, Department of Chemistry, Syracuse University, and Chris Gorman, Department of Chemistry, North Carolina State University. The Conference included 33 oral presentations and a poster session with 48 additional presentations—far too many to summarize here.
Tutorial on Molecular Nanotechnology precedes Conference
On Oct. 9, a Tutorial on Molecular Nanotechnology, chaired by Hicham Fenniri of the National Institute for Nanotechnology, National Research Council and the University of Alberta, afforded attendees overviews of key areas of nanoscale science and technology, with a particular focus on recent research in self assembly of nanostructured materials and application to catalysis, energy, environmental remediation, and molecular electronics.
Mark S. Lundstrom, Purdue University, provided "A Top-Down Look at Bottom Up Electronics," taking an in-depth look at CMOS electronic circuit technology and the limitations it will soon encounter, particularly heat dissipation as the ultimate limit on the density of devices. Molecular electronics does not offer a way to extent this heat dissipation limit, and so will complement rather than replace CMOS. Lundstrom nevertheless looks for progress in circuit and system design to advance from the current billion-transistor chips to trillion transistor chips (terascale integration).
Susannah Scott, University of California at Santa Barbara, covered "Nanostructured Catalysts," in which the use of various nanoparticles as catalysts brings both enormous economic impact and enormous environmental benefit by selectively speeding up specific chemical transformations. New methods of more precisely controlling nanoparticle structures, and making more complex nanoparticles, are leading to more effective catalysts.
Thomas E. Mallouk, The Pennsylvania State University, considered "Implications of Nanotechnology for Energy and Environmental Remediation," emphasizing the "massive" need for an inexpensive source of energy that does not increase carbon emissions and the role of nanotechnology in making solar power affordable through use of self-assembly techniques to fabricate nanomaterials with unique physical properties. In particular, nanoparticle structures (nanowires, core-shell particles) may lead to efficient multi-bandgap and polymer junction devices, and quantum dot structures could be important for more advanced devices. In terms of remediation, (zero valent) iron nanoparticles hold promise for removing chlorinated organics and metal ions from water and soil.
Steven C. Zimmerman of the University of Illinois at Urbana-Champaign presented "Self-assembly Approaches to Nanoscale Materials," covering the gamut from chemical synthesis of covalent nanostructures (adamantane diamondoids, conjugated organics for nanowires, dendrimers, carbon nanotubes) to various self assembly regimes (crystal engineering, supramolecular systems—including DNA base-pairing, metal coordination, mechanical bonding, hydrophobic assembly, and the role of molecular chaperones in reverting errors of assembly).
Fraser Stoddart of the University of California in Los Angeles, in a talk titled "An Integrated Systems-Oriented Approach to Molecular Electronics," discussed the transfer of concepts like molecular recognition and self assembly from the life sciences into materials science in order to control and harness the molecular motions in switchable molecules like rotaxanes and catenanes to produce solid state devices. Molecular switches can be built around mechanical movement in bistable rotaxane molecules, in which an applied voltage will cause one section of the molecule to move, serving as an on-off switch. Originally designed to work in solution, these switches also work in Langmuir monolayers and Langmuir-Blodgett films, and in self-assembled monolayers on gold surfaces. A 64-bit molecular RAM based on amphiphilic bistable  rotaxanes was demonstrated. Current efforts span further device design and synthesis, and computer architecture development for using molecular electronic circuits.
Molecular devices and molecular machinery
Fraser Stoddart opened the Conference with a keynote address titled "Meccano on the NanoScale: A Blueprint for Making Some of the World's Tiniest Machines," in which he proposed that the path to developing molecular machinery lies in using the same principles (self-assembly, self-organization, structure-activity relationships) that Nature uses for its molecular machinery, but without using the same chemical building blocks (nucleic acids and amino acids) that Nature uses.
One theme Stoddart explored is rotary motion in biological and artificial molecular machines. The natural molecular machine is the ATPase rotary motor, in which the passage of protons through a membrane causes a shaft to spin, resulting in the synthesis of ATP. For an artificial molecular level device, defined as an assembly of a distinct number of the molecular components designed to perform a specific function, Stoddart turned to catenane molecules, with their mechanical bonds formed by interlocking rings. The same principles of molecular recognition and self-assembly used by biology were adapted to the development of supramolecular chemistry, in which non-covalent interactions are used to guide the assembly of molecular entities held together by covalent and mechanical bonds. In the case of catenanes, non-covalent interactions guide the threading of a linear molecule through a circular molecule and pre-organize the components for ring-closing by covalent bond formation, thus forming interlocked rings, now held together by a mechanical bond. These methods have been pushed as far as making catenanes with seven interlocking rings. The interlocked rings of catenanes exhibit a back and forth rocking motion, a reversible threading of one ring through the cavity of the other ring, controlled by charge changes—oxidation and subsequent reduction. Such mechanical catenane-based switches are only about one cubic nanometer in volume.To make devices, the catenane was anchored with phospholipid counterions into a Langmuir-Blodgett film and sandwiched between two electrodes. A crossbar junction can be made with 5000 catenane switch molecules. A variant device currently being investigated has the catenane molecules stacked along a single wall carbon nanotube.
A second theme is linear motion in molecular machinery—the biological example is the muscle proteins actin and myosin, and the artificial example is rotaxane, interlocked molecules in which a ring-shaped component is threaded on a dumbbell-shaped component, and mechanically trapped by the bulky ends of the dumbbell-shaped component. If the middle of the dumbbell-shaped component has two different sites to which the ring can bind, then a molecular shuttle can be formed in which the ring shuttles back and forth between the two sites, thus forming another type of switch. By making one end of the dumbbell hydrophobic and one end hydrophilic, the molecular shuttle could be incorporated into a Langmuir-Blodgett film and sandwiched between electrodes, forming an 8x8 crossbar to constitute a 64-bit molecular RAM from about 5000 molecules (actually only 56 bits worked when tested).
Moving from molecular devices based on mechanically linked rings to devices based on carbon nanotubes, Mark Lundstrom presented "Carbon Nanotube Electronics: Device Physics, Technology & Applications," focusing on computational methods to understand device physics, to optimize transistor designs, to assess ultimate performance limits of such devices, and to identify which applications would be most appropriate. Carbon nanotube transistors (CNTFETs) are particularly promising subjects for detailed theoretical studies because they can be either metallic or semiconducting (depending on how the graphene sheet is rolled up to make the tube), near ballistic electron transport is possible, there are no dangling bonds to complicate depositing additional atomic layers, and electronic and optical components can be on the same substrate. Lundstrom's computational approach used an atomically detailed representation of the nanotube and a quantum mechanical treatment using the Schrödinger equation and Green's function. Results included the fact that the one-dimensional geometry of the nanotube produces electrostatics very different from those of conventional silicon transistors, making the details of the metal contacts with the nanotube very important. Most results so far show carbon nanotube transistors behaving like Schottky barrier transistors, which would be difficult to use in CMOS circuits. The carbon nanotube diameter is a critical parameter in this behavior. Although CNTFETs might have limited use in CMOS, Lundstrom suggested that use in MOSFET might be more profitable. At the very least, CNTFETs are a great model for understanding transistors at atomic detail.
Hicham Fenniri turned attention to a different type of nanotube, rosette nanotubes inspired by the DNA double helix, that can be engineered to have a wide variety of structures and properties: "Organic Nanotubes with Tunable Dimensions and Properties". These structures are based on modules patterned after the guanosine-cytosine base pair found in DNA, hierarchically self-assembled and self-organized, guided by hydrogen bonds and hydrophobic interactions. The rosette nanotubes can be designed to have different diameters, lengths, and physical, electronic, and optical properties. Fenniri noted that the pattern of hierarchical self-assembly and self-organization to go from simple to complex structure is borrowed from Nature, which uses similar principles to form entire chromosomes, with a diameter of 1400 nm, from nucleosomes 30 nm in diameter, which are formed from proteins and the DNA molecule, which has a diameter of 2 nm. The basic rosette nanotube structure can be designed to have a diameter of 3- 4 nm, and a length of 20-200 nm, or in some cases, of mm. Despite being held together by non-covalent interactions, the structures are stable enough that they do not fall apart when scanned with an atomic force microscope. The chemistry of the rosette nanotubes allows for synthesis in kilogram quantities, and for a wide range of chemical substituents, providing different properties. For example, the nanotubes can be metallized by coating with, for example, gold or titanium atoms. The titanium-coated nanotubes adhered well to human osteoblast cells, indicating they can also be biocompatible.
Seth R. Marder of the Georgia Institute of Technology explained how "Two-Photon Materials Chemistry" provides a new way to integrate nanostructures with MEMS by providing a way to fabricate three-dimensional structures with ~200nm feature sizes. The process uses focused femto-second laser beams to take advantage of very weak processes in which a molecule is excited by absorbing two photons simultaneously. Since excitation and thus chemical change occurs only in the very small volume element where the two lasers both focus, patterns can be produced in materials with pinpoint control in three-dimensions. This system of two-photon 3D lithography has been demonstrated in both polymers and metals. The smallest volume element addressed so far was 170 nm2 by 500 nm, compared with a minimum feature size of about 50 microns for commercial stereolithography.
Tobin Marks of Northwestern University opened the second day of the Conference with a keynote address titled "Self-Assembly of Nanophotonic Materials and Device Structures," addressing the question of how to assemble molecules in a precise organization to perform precise functions. Marks reported the ability to make pinhole-free self-assembling superlattices suitable for electro-optical applications (non-centrosymmetric chromophores). The monolayers were furthermore robust, withstanding heating to several hundred degrees. These molecules seem applicable to nanoscale OLED (organic light emitting diodes)—they can be made as small as 40 nm.
Looking to biology as a toolbox for building molecular machines, Jacob Schmidt of UCLA ("Development of Biomimetic Devices using Membrane Proteins") asked how components adapted from biological systems could be incorporated into artificial scaffolds. Many of the most interesting proteins, such as the F1-ATPase rotary molecular motor, porin proteins that open and close in response to electrical signals to allow certain molecules or ions to pass, and proteins that sense mechanical force (natural piezoelectric devices), function naturally inserted in lipid bilayer membranes. However, proteins in lipid membranes have a short functional life (a few days), so biocompatible polymers were investigated that would mimic the environment of the membrane while allowing longer lifetimes, in the hope of developing a generic platform for engineering membrane proteins. Schmidt reported work with an engineered voltage-gatable pore protein inserted into monolayer planar membranes of self-assembled amphiphilic block copolymers. The hope is to develop membranes suitable for use in micro-fluidic chips, in which engineered porin proteins could be electrically addressed to serve as valves.
Donald A. Tomalia, Dendritic NanoTechnologies, Inc. and Central Michigan University, "Synthetic Control of Dendritic Nanostructures Both Within and Beyond Poly(amidoamine) Dendrimers," explored the possibility of developing chemical synthetic strategies that parallel the strategies evolved over the past 3-4 billion years for controlling biological nanostructures as a function of size, shape and the placement of chemical groups in specific regions. Dendrimers that chemists have synthesized, in which successive shells of dendron groups are added to make tree-like structures surrounding a central core, are similar to globular proteins in several ways. They have precisely controlled masses and definite three-dimensional structures. Progress was reported in making dendrimers further mimic the complexity of biological structural hierarchies—designing structures that are not spherically symmetric (that is, more like the cusps and points of globular proteins) and that self-assemble into larger structures. For example, dendron spheroids of different sizes or different chemical functionalities can be joined together by disulfide bonds. The assembly of components into predetermined structures can be accomplished by the binding of complementary DNA strands attached to the components. Other variations can be introduced by only partially filling the shells of the dendrimer structure, thus producing empty spots that can be reactive.
Gavin D. Meredith of the University of California at Irvine, "Novel chemical strategy to link protein to DNA for directed molecular assembly," has the goal of designing and manufacturing molecular machine systems by assembling some of the tens of thousands of different molecular machines (proteins) that biology has evolved and provided as "off-the-shelf" components for nanotechnology. He looks at the possibility of using complementary DNA strands attached to proteins to link together protein molecules in a predetermined manner, in particular, to obtain precise orientation of proteins on a surface. It is important to attach the DNA to a specific part of the protein that does not interfere with the function of the protein. For this purpose, the DNA is linked to the protein via a 3-part molecule: nitrilo acetate-benzophenone-maleimide, which reacts with a series of 6 histidine residues engineered into the desired position of the protein sequence. This process can be used to convert a DNA chip, in which a specific array of DNA strands is attached to a substrate, into a protein chip, in which an array of protein molecules is attached to the chip in a pre-determined arrangement.
Stefan Diez of the Max-Planck-Institute of Molecular Cell Biology and Genetics reported using molecular motor proteins to move and stretch DNA molecules on a surface: "Manipulating DNA Molecules in Synthetic Environments by Motor Proteins and Microtubules." A number of methods exist for manipulating DNA molecules on a surface, but one advantage of using molecule motor proteins is parallelization—manipulating many DNA molecules concurrently. The molecular motor protein kinesin takes 100 steps per second of 8 nm each, fueled by one molecule of ATP for each step, along a track formed by microtubules, 25 nm-diameter hollow cylinders formed from two different types of protein subunits. Kinesin molecules can be attached to a substrate in specific patterns and microtubules moved along the pattern by controlled addition of ATP. DNA molecules attached to the microtubules can thus be moved and stretched, forming networks that could be coated with metal atoms to form circuits.
The Sunday morning session began with keynote presentations by the winners of the 2003 Foresight Institute Feynman Prizes. Marvin L. Cohen and Steven G. Louie of the University of California at Berkeley received the theoretical prize for their contributions to the understanding of the behavior of materials. They spoke about the theory and computation of properties of nanotubes (carbon nanotubes, or those made from boron and nitrogen), using their plane wave pseudopotential method, now recognized as the standard model of solids. In this model, inner shell electrons are treated with the nuclei of the atoms, and the valence electrons are allowed to move and interact with light. The model correctly predicts various properties of the nanotubes; for example, with BN nanotubes the properties do not depend on how the tube is rolled up but instead these nanotubes are always semiconductors, unlike the case with carbon nanotubes, which can be either semiconductor or metallic, depending on the detailed structure (chirality) of the tube. Another result is that electrical transport in (n,n) metallic carbon nanotubes is very robust against impurities and local defects. Current work includes studies of friction and how mechanical energy is dissipated at the nanoscale in order to determine whether nanomachines could operate efficiently.
The prize for experimental work was awarded to Carlo Montemagno of the University of California, Los Angeles for his pioneering research into methods of integrating single molecule biological motors with nano-scale silicon devices. He spoke on engineering and embedding intelligence into materials and devices using integrative technology. In living systems higher order functionality that is observable at a higher level emerges from stochastic, non-linear interactions at a lower level. Montemagno's current work is following a strategy he calls integrative technology, in which nanotechnology is integrated with biotechnology (as the source of blueprints for molecular machinery), and informatics (which deals with the ways in which information flows). For example, a salient feature of living cells is how they are compartmentalized by lipid membranes that control the flow of materials and information. To adapt this strategy to nanotechnology to make artificial organelles, fragile lipid membranes are replaced by cross-linked polymers embedded with bacterial proteins that harvest light energy to produce ATP molecules. Current research is also attempting to make biorobotic systems based upon the muscle protein actin and using ATP as fuel. Other work is inspired by the calcium and potassium channels in membranes that allow cardiac cells to communicate and the heart to beat in order to build nanoscale devices that self-excite, and MEMS devices are being moved by cultured cardiac cells.
Results with another molecular motor were presented by Richard T. Pomerantz of the SUNY Health Science Center at Brooklyn: "RNA Polymerase as an Information-Dependent Molecular Motor." This work exploits the fact that RNA polymerase is a powerful motor, exerting 15-20 pN of linear force as it moves along the DNA molecule. The progress of the RNA polymerase motor can be controlled to an accuracy of 0.34 nm, the length of one base pair, by limiting the supply of one of the four nucleotide triphosphates required to copy a DNA sequence (immobilize the polymerase to a substrate; wash repeatedly with different mixtures of triphosphates according to the sequence of the DNA). The polymerase can be engineered to permit attaching and releasing cargo molecules without interfering with polymerase function.
New building blocks and a new approach to engineering molecular structure
Introducing a novel type of molecular building block, Luc Jaeger of the University of California at Santa Barbara discussed "How to play LEGO with RNA: design of RNA cellular automata." Jaeger used knowledge of the folding and assembly rules governing the three-dimensional shape of complex natural RNA molecules, such as the 23sRNA of the bacterial large ribosomal subunit, to generate "tectoRNAs," named after tectonics, the science or art of building or constructing materials. These are self-assembling RNA building blocks that are designed and programmed to generate RNA super-architectures in a highly predictable manner. Jaeger reported RNA motifs that contain a perfect right angle between two helices, which were used to build RNA squares of 9 nm or 13 nm. The hope is that now that the folding rules governing RNA structure are understood, it will be possible to adapt these rules to other folding polymers to give highly predictable structures, and that it will also be possible to do "nanoscale sculpting with RNA."
TectoRNAs are especially interesting from the standpoint of molecular manufacturing because they may provide a path to engineering molecules to fold into predictable structures. In 1981 Eric Drexler proposed a path to general molecular manipulation based on engineering proteins to fold predictably. Progress in this direction has been slow, in part because the rules governing protein folding are very complex, in part because proteins are commonly built from 20 different subunits with very different chemical properties. The rules of nucleic acid structure are much simpler, based upon four subunits, and these have been exploited for DNA by Nadrian Seeman and others to form nanoscale structures and devices. However due to differences in the sugar-phosphate backbone, RNA folding, although based on rules of Watson-Crick base-pairing similar to DNA, leads to a wider variety of natural three-dimensional structures than are found with DNA, more like the huge variety of natural structures found with proteins. Further, some RNA molecules have catalytic activity, as do protein enzymes. Thus tectoRNAs provide a promising addition to the tool kit for molecular nanotechnology.
Venture Capital for Nanotechnology
Other features of the Conference included a panel discussion on "Venture Capital for Nanotechnology" chaired by Ed Niehaus and with panelists: Steve Jurvetson of Draper Fisher Jurvetson, Alan Marty of JP Morgan Partners, Alex Wong of Apax Partners, and Jim Von Ehr of Zyvex. One theme that emerged was that panelists had seen many interesting nanotechnology ideas put forward for investment, but very few of these were ready for institutional investment. Instead, most were early stage proposals more appropriate for government funding. As an entrepreneur who used his own money to found his company, Jim Von Ehr was initially uninterested in government funding, but more recently successfully turned to the NIST/ATP to fund investment in a top-down approach to nanotechnology. He especially recommended NIST/ATP because of their focus on the business aspects of the proposal, in contrast to other government agencies that are primarily guided by peer reviews of the science behind the proposal. When asked how best to fund MNT development that might lead to an assembler, Steve Jurvetson responded that venture capital firms are unlikely to make such an investment, and that funding was more likely to come from a government "moon shot" program to develop MNT. Alex Wong noted that many of the nanotechnology business models that he has seen propose to license to someone else a great technology that the entrepreneurs have developed. These are, he said, unattractive models because investors are looking for companies that propose to directly develop products. "We love product companies." Alan Marty seconded the importance of products compared to technology—he does not want to invest in a company that is developing building blocks, but rather in one that is using somebody else's building blocks to take a product to market.
11th Foresight Conference on Molecular Nanotechnology, October 9-12 , 2003
Platinum Level Sponsor
Founded in 1982 by four idealistic young pioneers with a shared vision of decentralized, heterogeneous computing systems, Sun Microsystems, Inc. has emerged as a global Fortune 500 leader in enterprise network computing, with over $8 billion in revenues and operations in 150 countries.
Gold Level Sponsors
Silver Level Sponsors
NanoSIG promotes the commercialization of nanotechnology by creating nano business networks and providing information, infrastructure, people, & funding services required to launch nano business ventures. In Silicon Valley, DC, and Southern California NanoSIG hosts forums & conferences, conducts surveys, facilitates networking, and works on information and business clusters.
From Foresight Update 53, originally published 15 January 2004.