Physics of Atomic-Scale Wires: A Layman's Description

The word ``nanotechnology'' is becoming increasingly common, and reflects the desire among the scientific and technological community to pursue miniaturisation to its ultimate limits on the atomic scale. This is most easily seen in the computer industry, where the size of the transistors in microprocessors such as the Intel Pentium 4 or AMD Athlon tend to halve every two to three years. There is a layer of insulating material (silicon dioxide) in these transistors which is currently only about 10 atomic layers thick, and will soon be so thin that atomic-height steps will have a major effect on it. This is the background to this project.

The scale which is becoming known as ``extreme'' nanotechnology (where at least one dimension is truly atomic-scale) is far from having immediate technological implications, but is fascinating both in terms of fundamental physics and for studying problems which may well beset the realisation of genuine nanotechnological devices. In this fellowship, I am aiming to understand the formation of atomic-scale wires (both inorganic, for instance fabricated from metal, and organic molecules) and their atomic and electronic structure on (in particular) semiconductor surfaces, such as the silicon surface commonly used in microprocessor fabrication. I will also study their behaviour in the presence of a current: how their structure is affected; and how well they conduct, among other problems.

The interest in these systems is extremely wide, and focusses on different end points. There are hopes that individual transistors may be able to be made out of single molecules, which can then be patterned onto a surface to build computer chips with a much higher density than is possible at present. There is also interest in using these types of device to build quantum computers, which would result in a completely new way of doing computation. There are also new and exciting physical problems, for instance quantised conduction (where the amount of current flowing increases in well-defined jumps, not smoothly), which is also related to the structure of these wires; and new physics which arises in extremely constricted wires (for instance interactions between electrons and other electrons, or lattice distortions become important).

Wires on an atomic scale are rather different to conventional wires. For instance, impurities can completely change the conduction behaviour, as can the structure of the wire. Mechanisms of conduction change with temperature and length of the wire, and the forces on the wire from the current flowing through it become important. There are many open questions in the field.

These wires will take many different forms, both inorganic and organic, and will be constructed in different ways, both deliberately and through more disordered self-assembly processes. Possible candidates include: carbon nanotubes; conjugated polymers (such as polyacetylene); semiconductor nanowires; metal deposited on templates (for instance bismuth nanolines or dangling-bond wires on silicon); and rare-earth wires (including erbium disilicide on silicon).

The study of these wires requires the solution of the quantum mechanical equations of motion for the electrons and ions in the system being considered. But these equations are immensely complicated, requiring sophisticated computer programs, and the traditional techniques require time and effort that increases with the cube of the system size (so if we wanted to study a system twice as large, it would require eight times as long). I will use a radically different computer code called Conquest (which I have been developing in collaboration with Professor Mike Gillan) whose scaling is linear: a system twice as large requires only twice the effort. In order to model the effects of current flow on these wires, I will have to extend the code. The problem here is known as ``open boundary conditions'' - in other words, we're allowing the system that we want to study to be open to an environment (where traditionally it is a closed system). This will require a significant effort to redevelop the code.

The systems that I will study include both organic and inorganic wires. The inorganic systems have been mentioned above: bismuth nanolines (which form spontaneously on the silicon surface) and ``dangling-bond'' wires (which have to be etched on a hydrogen-covered silicon surface), with and without metals deposited on top of them. The organic systems will involve conducting polymers (e.g. polyacetylene) both bare, and encased in an insulating coat (made from a molecule called cyclodextrin). The studies will start with structural calculations and an exploration of the electronic structure under normal (closed boundary) conditions. Once Conquest has been altered to work with open boundaries, I will study the effects of current on these structures, and investigate their suitability for use as nanowires.