Computing

The holy grail of nanoelectronics is an integrated computing device.  In theory, such a computer would mimic the computational methods of present-day microchip processors.  However, a nano-computer would take advantage of higher component density, low power requirements, and greater overall efficiency.

One obstacle stands in our way: we can't fabricate components with accuracy or precision.  These two fundamental problems are the result of our poor control over the sizes and diameters of nanostructures and our inability to move them into position with ease.

On the accuracy side, we have the problem of large size distributions.  With any given synthesis method, there is usually a non-uniform distribution of sizes that lead to differences in characteristics from structure to structure.  In order to move further with nanoelectronics, we'll need tighter distributions.  One way to guarantee this has been the traditional top-down approach, but lithography and other destructive techniques never provide the product density we need and also damages the final product.  A top down technique introduces problems with placement.  For integrated devices, each component must be in the right relation to its adjacent/connected components.  We can't have components overlapping each other since each component must have proper isolation.  Again, a top down approach isn't the best to achieve this.

It would appear then, that some bottom up method must be devised.  On one side, we have our traditional semiconductor nanowire approach, and on the other, the molecular electronics approach.  Both hold some promise in achieve the goal of a nano-computer. 

For nanowires, it's obvious to researchers that it is the physical crossings (p-n junctions) that provide all the functionality.  Their research has been focused on creating large numbers of crossings in a reproducible process.  One such approach utilizes a Langmuir-Blodgett thin-film technique to apply layers of axially aligned nanowires on top of each other.  Using this technique, it's possible to layer orthogonal sheets of nanowires over each other, thus creating thousands of crossings in a very small area.

Molecular electronics is focused right now on creating useful components.  While nanowires have already been proven as building blocks for junctions, transistors, and logic gates, molecular devices have yet to demonstrate good numbers.  Most of the problem lies in the donor-acceptor model that molecules must operate under.  They are not semiconductors and hence do not conduct electricity in the normal manner.  It will be some time before molecular electronics take off, but when they do it's quite possible that they'll be easier to manufacture do to the bio-mimicry that the science utilizes.