IMM Report Number 32

New nanofabrication method for metal nanostructures

F Rosei, M Schunack, P Jiang, A Gourdon, E Laegsgaard, I Stensgaard, C Joachim, F Besenbacher “Organic Molecules Acting as Templates on Metal Surfaces” Science 296, pp. 328-331, 12 Apr 2002.

Scientists from the University of Aarhus in Denmark and from CEMES-CNRS in France report a novel method of fabricating nanostructures that uses a large organic molecule known as “the Lander” as a template to cause rearrangements of copper atoms on a surface. These molecules are deposited on a Cu(110) surface at room temperature, where they diffuse toward the edges of a one layer-thick terrace of copper atoms. The copper atoms at these edges are mobile at room temperature, but at low temperatures (e.g., at 150 K) these copper atoms are static.

After lowering the temperature, isolated Lander molecules can be seen using scanning tunneling microscopy (STM) to be bound to the now fixed edges. By very precisely controlling the STM tip position (at low temperature), the authors are able to nudge individual molecules away from the edge without disturbing the rest of the copper surface. Where the Lander molecule was bound to the edge, there is now a “tooth” two copper atoms wide and about seven copper atoms long (about 0.75 nm wide and 1.85 nm long) extending from the edge of the terrace.

The C₉₀H₉₈ Lander (1.4 nm long and 1.2 nm wide) consists of a central “board” composed of a long, polyaromatic array of ten cycles (a molecular “wire”) supported by four bulky 3,5-di-tert-butyl-phenyl “spacer” legs, which elevate the central board about 0.5 nm above the substrate. The authors use both theoretical calculations and STM to explore the interactions between the Lander and the copper surface. The favorable interaction between the pi electrons of the Lander central board and the copper atoms in the tooth provides the energy needed to rearrange the copper atoms at the edge to form the tooth. The restructuring of the copper atoms is thermally activated since it does not occur if the Lander adsorption to the surface is done at low temperature.

The authors speculate “Using appropriately designed molecules, this [result] points to a new self-fabrication process at the nanoscale for integrated nanoelectronics.” It is more difficult to see how such metal nanostructures, held together by metallic rather than by covalent bonds, could be strong enough for nanomechanical devices, particularly for moving parts, but perhaps they might serve to anchor other structures to a surface.

Designing a working monomolecular machine looks hard

C Joachim, H Tang, F Moresco, G Rapenne, and G Meyer “The design of a nanoscale molecular barrow” Nanotechnology 13, pp 330-335, June 2002.

Researchers from CEMES-CNRS in France and from the Freie Universität Berlin in Germany designed a wheel barrow synthesized as a single molecule 1.6 nm wide x 1.5 nm long in the hope that it could be pushed by an STM tip across a Cu(100) surface. They demonstrated the synthesis of a crucial part of the molecule and calculated how it would be expected to behave. Unfortunately the calculations showed that the simple, intuitive design is not likely to work as planned.

The approach is based on molecules used successfully in several earlier studies manipulating molecules on copper surfaces with an STM. The rear and center of the barrow bears some resemblance to the Lander molecule used by Rosei et al. above. In this case, the rear part is made from a phenyl group with two covalently attached tert-butyl phenyl legs. This rear part is attached to a central board made from a tetracene (four phenyl rings fused into a straight line). The front part is a phenyl group covalently bonded to two acetylenyl groups, each of which is covalently bonded to a triptycene (a three-bladed molecule) wheel. This arrangement produces two three-cogged wheels that should freely rotate about the acetylenyl units. The rear legs were chosen because in several earlier studies these structures were used as legs on molecules that were shoved around copper surfaces by pushing with an STM tip. The tetracene central board was chosen to permit good electron tunneling between the wheels in the front and the STM tip pushing at the rear, so that rotation of the wheels could be measured by the tunneling current.

The front part was synthesized, deposited on a copper surface, and imaged by an STM. However, molecular dynamics calculations indicated that pushing the barrow with an STM would cause the rear legs to collapse and the tip to ride up over the central board. If the rear legs are artificially frozen in the calculation, the STM tip pushes the barrow, but the front wheels do not turn, apparently because there is not enough friction between the wheels and the copper surface. The authors suggest that attaching to the central board two ratchets similar in structure to the rear legs will allow the front wheels to rotate, but as of the publication, this structure had been neither synthesized nor dynamically simulated.

Self-assembled nanotube scaffolding

H Fenniri, B-L Deng, AE Ribbe, K Hallenga, J Jacob, and P Thiyagarajan “Entropically driven self-assembly of multichannel rosette nanotubes”, Proc. Natl. Acad. Sci. USA, 99, pp. 6487-6492, April 30, 2002.

Chemically very versatile nanotubes have been designed and demonstrated by researchers at Purdue University and Argonne National Laboratory. They began with an aromatic organic base designed to possess the the same pattern of hydrogen bond donors and acceptors as the Watson-Crick base pair guanine-cytosine. Provision was made to attach other chemical functions to this basic structure, and in the example reported, the chosen attachment was a crown ether, a class of compounds that have been extensively studied for their ability to selectively bind various “guest” ions and molecules to form a variety of photonic and electronic devices.

The geometry of the intermolecular hydrogen bonds and the hydrophobic nature of the aromatic bases cause these molecules to bind together in aqueous solutions to form a six-member “supermacrocycle” held together by 18 hydrogen bonds. In the rosette that results, the six aromatic bases form a central hydrophobic ring, which is surrounded by the six separate, hydrophilic crown ether rings.

The formation of the rosette is the first of two levels of self-organization. The hydrophobic bases in the inner ring cause the rosettes to stack to minimize solvent contact, forming a long multi-channeled nanotube with a central channel 1.1 nm in diameter. The crown ethers on alternate rosettes stack on top of each other so that the six crown ether rings on each rosette form twelve smaller channels around the central channel. The outside diameter of this multi-channel nanotube is about 4 nm.

The authors found that not only are these nanotubes quite stable, surviving heating to 95 C, but that heating the nanotubes causes them to grow longer, reaching micrometers in length. The increase in length with increased temperature is characteristic of an entropy-driven lengthening. The hydrophobic ends of the nanotubes cause an entropically unfavorable ordering of the surrounding water molecules, so adding more rosettes to lengthen the nanotube releases ordered water molecules into the bulk water. Such entropy driven processes are common in biological systems, such as self-assembly of protein fibrils and tubules.

The principal advantage of this class of nanotubes is that the design and synthetic strategy permit modifications of both size and chemical functionality. Accordingly it is easy to imagine a large assortment of wires that provide predefined chemical and physical properties. To provide structures more complex than one-dimensional wires, it would be useful to explore whether different chemical functions could be introduced in the same rosette, and whether different types of rosettes could be assembled in one nanotube or into different classes of nanotubes that could be made to interact with each other in a predefined manner.

Another DNA nanomotor

JJ Li and W Tan “A Single DNA Molecule Nanomotor” Nano Letters 2, pp. 315-318, April 2002.

Joining the lengthening list of ways to make a nanomotor from DNA (see Update 48) is a single molecule design developed by scientists at the University of Florida. This design takes advantage of yet another property of DNA — the 4-way non-Watson-Crick interaction that stabilizes the ends of eukaryotic chromosomes.

The 17mer DNA oligonucleotide (5′-TGGTTGGTGTGGTTGGT-3′), on its own, folds into a compact, single-molecule tetraplex, driven by the tendency of G-rich DNA sequences, such as those found in telomere sequences on the ends of eukaryotic chromosomes, to form four-stranded intra-molecular tetraplex structures (in which G tetrads replace Watson-Crick pairs). Think of the 17-mer as folded into a hairpin that is folded again to form a second hairpin so that the G residues at positions 2, 7, 11, and 16 form a square stacked under a second square formed by G residues 3, 6, 12 and 15. These tetrads keep the two ends of the 17-mer close to each other. This tetraplex is the nanomotor’s “shrunken” state.

In the presence of a sequence “alpha,” a 27-mer in which residues 11 – 27 form the Watson-Crick complement of the 17-mer, the tetraplex unfolds so that it can form a W-C duplex with alpha. This bi-molecular duplex is the “extended” state of the nanomotor. The nanomotor returns to the shrunken state when the 17-mer is displaced from the duplex by binding of alpha with “beta,” a 27-mer that is the entire W-C complement of alpha. Adding more alpha would start a new cycle, and each cycle produces a waste molecule, the alpha-beta duplex. Because this nanomotor switches between duplex (DU) and tetraplex (TE), the authors call it a “DUTE nanomotor.”

The authors demonstrate that the DUTE nanomotor can do work by loading the two ends of the 17-mer with two small molecules, a fluorophore on one end and a quencher on the other. Thus the molecule has little fluorescence in the shrunken state, when the quencher is close to the fluorophore, but fluoresces strongly in the extended state. Measurement of fluorescence changes upon addition of alpha or beta demonstrated the expected conversion between TE and DU states.

The authors further demonstrated that their DUTE nanomotor functioned when immobilized on nanoparticles. They added a biotin molecule to one end of the 17-mer so that it would bind to nanoparticles coated with the protein streptavidin, to which biotin very tightly binds. The nanoparticles were in turn bound to the inner surface of a microchannel, permitting alternating solutions of alpha and beta to be flushed by the nanoparticles. Over 12 cycles the nanoparticle efficiently fluoresced in the presence of solution alpha and darkened in the presence of solution beta. The authors calculate an extending force of about 20 pN and a shrinking force of about 2 pN. The latter is close to the forces determined for protein nanomotors like kinesin and myosin, but the former is about ten times larger than protein nanomotor forces.

None of the DNA nanomotors described as been demonstrated to perform useful work in the context of a molecular machine system, but the variety of motors of different sizes working according to different principles under different conditions increases the possibilities for success.

IMM would appreciate learning your thoughts on the above article.