Current computer chip technology is based on lithographic methods that limit components to ~0.3 microns in size, due to the wavelength of light, and the photoresist/coating/etching processes. The size directly determines computer speed, complexity and cost, and advances in computers over the years have mostly been due to reduction in component size. It is here proposed to construct nanowires that are approximately 2 nm in diameter, or 150 times smaller than currently available. For 2 dimensions, this translates into a 1502 = 22,500-fold computational advantage. Additionally, 3 dimensional construction is proposed, bringing the potential improvement factor to 3,375,000. While it is probably unrealistic that this factor of packing density can be fully achieved, even several orders of magnitude improvement over current technology would be significant.
A wire width 2 nm may be achieved by placing gold quantum dots along a DNA template. Ends of the DNA-nanowire may be designed with sequences to attach by hybridization to complementary sequences on target connection pads, so that the two ends will seek and automatically wire correctly in solution. This strategy is easily adaptable to 3-dimensional wiring. Conduction between gold quantum dots may be studied as a function of spacing, size and coatings. In addition, the gold dots catalyze additional metal deposition leading to continuous metal nanowires. Preliminary results have demonstrated high loading of DNA strands with 1.4 nm Nanogold clusters spaced ~2 nm apart, showing basic feasibility.
DNA is close to an ideal wire template since it is of narrow diameter (2 nm), is flexible, can be targeted uniquely at both ends, and also has the property that its length can be exactly controlled by its production by DNA synthesizer or enzymes following a template. This is in contrast to other polymers and nanotubes whose lengths are poorly controlled and with nanotubes, have restricted flexibility.
Close packing of nanowires may require insulation. This may be achieved by several methods: 1) surface oxidation; 2) reaction with silanes to form a glass; or 3) coating with alkane thiols. All of these form insulating coatings and have been demonstrated for metal surfaces and are applicable to the DNA nanowires.
Double stranded DNA (from T7 bacteriophage) was labeled with 1.4 nm Nanogold clusters, and examined in the Brookhaven STEM (Scanning Transmission Electron Microscope); see Fig. 1. The average spacing between gold clusters was ~2 nm. This is of interest since the distance for tunneling of electrons between metal particles is also about 2 nm. Although a variety of methods are possible to link gold clusters to nucleic acids, including covalent attachment, photoreaction, intercalation, the method chosen here was to use positively charged Nanogold binding to the negatively charged DNA.
Metallographic methods have been developed to specifically deposit additional metal on the small gold clusters, and their enlargement can be followed by electron microscopy. Confluency between neighboring clusters could be achieved (Fig. 2). Here, gold was deposited catalytically, but silver and other metals have also been used in a similar process. In this way, continuous solid metal wires may be constructed.
Fig. 1 (Left): 1.4 nm gold clusters (bright spots) bound to double stranded bacteriophage T7 DNA (rope-like strands). Dark field, unstained STEM image on a thin carbon substrate. Full width 128 nm.
Fig. 2 (Right): Nanogold clusters nucleating further gold deposition so that they become contiguous. Metal deposition was then stopped at various times to demonstrate growth of cluster size. Top image after 5 min, bottom after 10 min. BNL STEM micrograph, darkfield, elastically scattered signal; full width of each image 230 nm.
- The authors wish to thank Dr. Joseph Wall, Dr. Martha Simon, Ms. Beth Lin, and Mr. Frank Kito for STEM operation. Research supported by the Office of Biological and Environmental Research of the U.S. Department of Energy under Prime Contract No. DE-AC02-98CH10886 with Brookhaven National Laboratory, and by National Institutes of Health Grant 2 P41 RR01777 and SBIR grants.
Thanks to Springer-Verlag for allowing us to reproduce this online.