NANO: EMT: Revolutionary 3D Nanoarchitectures to Organize the Assembly of Computing Elements

New York University, New York NY

Investigators

Abstract

Improvements in computational capability have entailed making the components smaller and smaller, to the point where we are reaching the limits of the current top-down fabrication methodologies. Furthermore, fabrication methods currently in place are inherently limited to two dimensional arrangements of components. With the research to be pursued under this award, we plan to address both of those problems: We will develop methods to assemble computing components from the bottom-up, using chemical methods. In addition, we will extend our ability to assemble computational elements from two dimensions to three dimensions. How do we plan to accomplish these ambitious goals? We will use the most effective and selective self-assembling system that is found in nature -- DNA. We are all aware that DNA serves as the genetic material of living organisms, from bacteria to humans. One of the key features of DNA that enables it to fulfill this role is complementarity between the two strands of the double helix: If one side of the helix contains a unit known as an A, the other side will always have another unit know as a T; likewise, two further complementary units known as G and C work in the same way. We now live in an era when it is very easy to synthesize DNA strands with particular sequences on them, so we can program a given strand on one side, and then make its complementary strand on the other side. Furthermore, DNA is a nanoscale object, with a width of about 2 nm, and a helical repeat of about 3.5 nm. In addition, DNA is the molecule whose intermolecular interactions are the most readily programmed, from the perspectives both of affinity (which molecules will bind to which others), and structure (what will they look like when they combine). This intermolecular programming is accomplished by using short single-stranded segments on the ends of each molecule, which are called 'sticky ends'. But wait. Nature already makes linear DNA molecules in profusion. Synthetic molecules would just be relatively short 'lines' of matter, so what use would they be? The advantage of synthetic DNA molecules is that they can be programmed to associate into branched, rather than linear molecules. Thus, joining a bunch of linear DNA molecules together can lead to longer lines, but joining branched molecules together can lead to connected networks of DNA. In the past, we have built a variety of two dimensional crystalline arrays with specific patterns; to do this we have designed DNA strands to form two dimensional motifs that could they self-assembled using sticky ends. In the research to be pursued under this award, we will design and self-assemble motifs that will form motifs that self-assemble to produce three dimensional arrays, related to conventional crystals, such as sugar crystals, except that their repeating units will be much larger. We will characterize these molecules using the standard method for examining three-dimensional matter, x-ray crystallography. How does the assembly of three dimensional DNA crystals relate to revolutionary computing? We have just discussed that DNA has outstanding architectural properties. However, it does not seem well-suited to serve as a computational component. By contrast, there are numerous newly discovered nanoscale-sized systems, such as carbon nanotubes or quantum dots that would seem to be ideal for computational purposes, if only we could organize them into circuitry for our purposes. Once we are able to assemble DNA into three-dimensional arrangements, we will undertake to use its architectural features to act as scaffolding in three dimensions for electronic components such as these. Thus, we will combine the outstanding architectural properties of DNA, which we know how to control, with demonstrated outstanding electronic components to form circuitry in three dimensions. This achievement ultimately will lead to extremely dense memory units and extremely rapid computation.